Alternately using the accelerator and brake doesn't make sense under conditions where you can see down the road and can know, can have some look-ahead to approaching conditions, and your preference is to conserve energy. Arriving at a particular time with some precision (no sooner, no later) is of reduced importance. In other uses of an automobile, the accelerator is frequently alternately operated with the brake, and is often operated at the same time as the brake. This is done in racing, where performance is given more priority than economy of fuel consumption. In control system theory, you can think of accelerating with the brake on or rapidly switching between accelerator and brake as a means of increasing the control system gain. On your weed whacker, if the motor's braking energy is recovered and fed back to the battery, the losses are somewhat reduced. Regardless, the motor controller can't see the thick grass approaching the cutting head and know to up the output in advance of the additional load, like a car driver can see an approaching hill and press the accelerator farther to compensate, and not lose speed. So being immediately ready to respond to a variance in load is critical, otherwise the weed whacker will seem weak in some circumstances and too powerful in others. Or, under certain steady loads, the motor speed may oscillate. Consider that in high performance motion control systems (not steppers, and typically systems that move faster than CNC machine tools), the motor drive amplifier typically drives the motor forward and backward simultaneously in order for the system to position a load stationary under varying demand. A DC servo motor may go one direction with one polarity and another direction with the opposite polarity, but to brake it, or keep it still, a judiciously chosen AC signal is used. It is essentially being rapidly alternately driven forward and backward. If you put your fingers on the motor shaft and try to make the motor shaft turn, the drive amplifier mixes in a DC bias with the AC to counter your efforts. The mixture of AC and varying DC bias keeps the motor excited and ready to respond immediately to any variance in command, which would be in response to any variance in effort from your fingers. AC servo systems do exactly the same thing, but the motors are typically very similar to BLDC motors, with a permanent magnet armature and 3 phase stator. For a great example of high performance servos, look to industrial electronics assembly machinery. Or, automated military fire control systems. On your weed whacker, I imagine the motor control uses field oriented control, where the voltage of windings that aren't energized is measured as a means of judging speed. Perhaps there's too much delay in processing that and applying the result to the motor drive output isn't fast enough to prevent a lag in speed correction, or some other aspect of the system presents a lag that is relatively expensive to improve. This usually means the "system" needs more gain. (If you're driving a 4-banger, how much earlier do you have to press the accelerator to have some chance of keeping up when the hill arrives?) They can't simply increase the amplification of the speed correction, depending on the circumstance, this causes oscillation. They can't leave the gain of the speed correction low, this causes poor performance. To compensate, system gain needs to increase. Increasing battery voltage would help, but that makes the product more expensive and or heavier, requires more expensive power transistors, etc. Increasing the gear reduction would help, but at the cost of slower maximum cutting head speed. They could add a sensor to provide the control system continuous true tachometer feedback, but that would be more expensive and perhaps less rugged. At the cost of some battery capacity, they use this alternating acceleration/brake technique to up the system gain to provide higher, or perhaps just acceptable performance. The user has to go back to the shed and swap battery packs a little earlier in exchange for that additional performance. Anyway, that's what I'm thinking they're doing.
@@lennyf1957 Not available in the Matthias Editition. You need M.A.D. (Matthias After Dark). It's a little more wild and the experiments with Legos are much harder.
I did not see any hall effect sensors, so they are probably running open loop until the motor spins fast enough to read the back EMF. They are most likely running a control with a current regulator as well. Generally, you accelerate the motor at a fixed rate and fixed current setpoint then switch over to sensorless control after a certain period of time. The acceleration rate, current setpoint value, and time are normally determined by application. In this case, they almost have to assume the motor will be loaded. This all gets screwed up when you run it unloaded! The best way to tell what is really happening is to measure the current, I bet you will see it being a constant value. Also, watch for the way the motor is aligned to start up. A more sophisticated control will use induction to determine the rotor position at start. A simpler control will force the rotor into a known fixed position by applying current through certain motor phases for a certain period of time. If the control is using induction to determine position, you can be sure the designers know how to PWM properly!
@@rpavlik1 That is the general overview of how BLDC sensorless drives work, but there are lots of little details I skipped over. The braking portion is probably a byproduct of their control logic and regulators that are tuned for operating under a higher load. For example, they may have a speed PI regulator that tells the system to brake when it overshoots. This would be an easy way to quickly stop the motor when the trigger is released (assuming the trigger is an input to this PI).
Being a brushless motor though, could/would the rotor will ever spin faster than the phases are being clocked around it? If not, there shouldn't be any reason other than to determine locked rotor (no back EMF) or how much load (how fast the frequency of the back EMF decays). Otherwise, would they able to determine the load by monitoring the phase angle difference between the voltage and current on the windings to see how far behind the rotor is from the generated field (slip)?
Seems like the controller was designed for something else and it was just drafted into this machine....maybe was cheaper to modify existing stock or design?
@dothemathright 1111 You are right about the coasting. I missed that part. Regardless, the fact it is braking the motor rather than coasting tells me that the control is in some strange state. Looking at the current and/or operating it under normal load would give us some more clues. Sensorless control can be difficult when you are not controlling something easy like a fan or pump. Since the controller is so far away from the motor, they probably saved a decent amount of money by removing the sensors and wiring.
This is just speculation... but... I'm willing to bet that it is optimized for operation under load... cutting through weeds could put a somewhat significant load on it, which would prevent the trimmer head from getting into the higher RPM's where it hits the brakes... and perhaps due to the nature of the trimmer head (string flying around with large centrifugal forces) the braking you are seeing is an overspeed protection to prevent the forces from becoming too high... they didn't bother to do it via the more complex PWM because in 99% of cases the trimmer will be under load while using it which would prevent the overspeed condition in the first place.
cryptk That would be my guess too, but hard to say for sure without looking at some actual technical specs for the trimmer and the motor assembly and driver.
And the people who trim early in the morning while the grass is at its wettest - you can't tell if the motor is burning out or if it's just a mist that is rising.
When I use a petrol strimmer on heavy growth near sharp fence posts there is a big variety of load. Often the nylon is cut off by square fencing at high speed or can slow down with tough weeds.
As primitive as that controller is, they set it up like that because there is a lot more torque required when using the tool with string and cutting grass, I think you will find it running almost full out without braking to slow it down.There is another story with ecm motors that this controller does not make use of and that is back emf from each coil, the good controllers actually feed the voltage spikes into the next coil pairs to smooth out operation and add torque.There was a company that used a solid state disk for the rotor about 10 years ago but I don't recall the name.
I think they designed that way because it is the cheapest way of implementing speed control. It looks like bang-bang control, like a heating system with a thermostat. There is no need to modulate the combustion because the inertia in the system smooths out the response (the room warms slowly). In the same way, DeWalt can use a cheap voltage amplifier and just switch it off when the reference speed is achieved and the inertia in the system will maintain the angular velocity. It would be interesting to see if you could make it more efficient with modulation of the voltage or frequency.
there is no "cheap" control in BLDC motors, you have to have your controller estimating the position of the rotor (basically it has to do what brushes would be doing) and keeping a tab on it. The only thing might be the absence of PWM, but 1) it's cheap, the "costly" part is the back-EMF sensing because it's noisy and you can't filter too much 2) they could also commute late to limit the torque.
Thinking of how a trimmer is used, I can maybe rationalize a reason. Trimmers tend to go from full load to coasting very quickly and often. When loaded (trimming) the motor will almost always need full power. Then you have to make sure it doesn’t spin too fast, or the string will break too easily. I’m betting they built the prototype with a different control system and had problems, likely with it going too fast and breaking the string, so their solution was this braking mechanism. It would be interesting to test it in use.
I believe the Sopwith Camel (?) had no throttle, i.e., it was wide open all the time. There was a kill button on the stick and the pilot controlled motor speed by pressing and releasing the kill button. You can hear this in many of the WWI movies made in the 1960s thru the 1980s. Most were done in B&W.
You need to test it with the wire going against grass. That should cause quite a bit of resistance, and maybe, just maybe, we can say the trimmer is not engaging break mode for most of its operation time.
I was thinking it was using maximum voltage so that it would either cut the grass or not cut, rather than half cutting and stalling. The foot on the gas comparison Matthias made was pretty accurate, except imagine the car is electric. Full power when your foot is on the gas, no power when it's not. Now whether this behavior is beneficial or not, I have no idea. Only the guy who made it knows what the thought process behind it was
I grew up learning to pulse the throttle of my weed whacker to keep the rpm low to avoid sending rocks through my windows; It can't be a coincidence that it's the only tool that performs the same thing electrically is the weed wackier.
Around 6:40: this is done to maintain both speed and maximum torque. You could regulate voltage accordingly to current, but that would be a waste of battery power to heat up another bunch of transistors, and probably spin up a fan to cool them.
It seems that in the null part of the duty cycle, when the coils arn't being driven, they aren't disconnecting the coils. They are driving the mosfets to short the coils instead. This doesn't allow the motor to coast, but would act as an electromagnetic brake. The question I (and I think, Matthias) am asking is why they do not drive the mosfets to open-circuit, so the motor doesn't generate current and is not slowed down. And when an electromagnetic brake would be useful - when the trigger is released - then the mosfets are driven to open-circuit, and the motor coasts. My thought is that they could be using a tuned circuit between the coils and some capacitors, perhaps, When they are not driving the coils, they resonate instead. Measuring the voltage across the coils may not show this, as you are really only measuring the voltage across the shorted mosfets.
That's bizarre. I'd love to see a schematic of the driving circuit. My shitty best guess is that either it was designed by a battery company or its somehow capturing some of the wasted energy and reapplying it to the next full-load. It would be interesting to see a current reading from the battery in sync with the coil signal.
But in actual practice, most of the time a weedwacker is running, it's not under load (the load comes in quick bursts.) So it seems like a poorly designed weedwacker.
I have some experience with motors and inverters. Shorting a synchronous motor that is rotating isn't as bad as you might think: contrary to rotating it at low speed and shorting it, where the resistive component dominates the reactive circuit of the shorted motor (and it really acts like a brake), at higher rates of rotation the inductive component dominates the circuit (the motor is decelerated for "half a pole" an then gets accelerated by the induced current again for the next half pole). So the only losses you get in that mode are the current losses in the windings. This cancels out most of the hysteresis losses in the stator, as the magnetic field is kept out of it by the shorted windings. Contrary to that, if you leave the motor open while it is truning you get the hysteresis losses in the stator instead of the resistance losses in the coils. Maybe(I'm not sure) there might be motors where the hysteresis losses when open would be higher than the resistive losses when shorted.
Do you suppose the hysteresis losses are greater than the copper losses? Also, copper losses plus some amount of volage drop across the electronics as its not a perfect short.
It's a possibility - But I don't know of any other motor inverter that does it this way. The main point is: Shorting a synchronous motor while it's spinning fast does not waste as much energy as one might think, opposed to shorting a DC motor, where all of the energy would be dissipated in the winding resistance quickly.
Yes, try testing with the inductive pickup while cutting grass to put it under a true load. The motor is freewheeling once it reaches max rpm and the controller senses overspeed and applys the brakes
I don't think it's doing the PWM strictly for the motor, but rather as a feature to drive the head in a different way. When cutting grass and weeds, there's a lot to cut through. The spool cuts most effectively when it's whipping, not just rotating. It's sort of like an impact driver's effect on a bolt. Try measuring the output RPM. It's probably more or less the same RPM throughout the trigger range if there's no load.
Simple answer what’s your definition of efficiency?... for automobiles adding aggressive air intake and a high flow exhaust creates a lot of efficiency focused on driving the car faster ... where as dampening the motors exhaust and slowing down its air intake will make the motor run more fuel efficient in the way it saves you money in gas... so I’ve used trimmers to cut jungles of grass from 5’ to 4” and a gas trimmer running full out throttle when cutting long grass will generate a ton of wind which blows the grass or momentarily bends the grass and it get missed by the trimmers cutters. So our old dads would tell us do the job proper and pulse the throttle and not to go full out. Or we will have to trim the whole yard again tomorrow! Watching the scope when you had the trimmer hooked up I immediately thought about being in a grass cutting situation in which to cut efficiently you have to operate the throttle non efficiently...
Here in the UK, our loan mowers are added with a brake to stop the blades turning once the power trigger is released. Maybe its same for strimmers? The EU implemented this safety device on our new lawn mowers some years back.
The Slot in the Motor Back Cover is for Hall Sensors. Its the same Motor as the one used in their brushless Drills just without the Feedback. The Feedback gives you more torque at low rpm.
What I noticed was the motor lacking a feedback loop via a hall effect sensor. The controller only has a vague idea what the motor is doing. I've tinkered with brushless motors used in computer equipment where a fine grain control is needed and they all have hall effect sensors mounted by the rotor. The controllers are fairly low voltage and use PWM for control but they are still brushless motors.
Sensorless brushless motors are very common. Usually only two of the coil (pairs) are driven and the third coil (pair) is used as the "sensor". The only drawback is that the motor has to turn at a minimum speed to generate enough voltage in the third coil for the controller to "read". So for a tool it's totally fine to not have a sensored motor/controller.
If the controller only had a vague idea of what the motor was doing, it wouldn't work at all. You can't just wing it when you are commutating a motor. You have to know where the rotor is with reasonable accuracy to turn your transistors on and off at the right time.
Hi Matthias, little late to the party here. But one observation is that shorting the windings doesn't automatically incur high resistive losses in a brushless motor, especially if it has high leakage inductance. If you're operating the motor electrically much faster than the L/R time constant of the leakage inductance, then the rotor flux is mostly excluded from the stator by the opposing currents induced in the shorted windings, and the torque is low (equivalent to exceeding the critical slip speed of an induction motor). Yes there’s still induced current circulating, but depending on the magnitude of the winding resistance, this might not actually embody a large power loss. Shorting the windings will potentially accomplish two positive things. First, it could reduce the cogging torque. Cogging torque arises due to the variation in the reluctance which the rotor flux sees versus angle, as it tries to find a path through the stator. Shorting the windings excludes the flux from the stator entirely and so could reduce the cogging torque. Second and probably more importantly, it will reduce the iron losses in the stator. Is that thing even laminated? At up to 500Hz, the eddy current losses in a solid core will be pretty darn severe. Even in laminations it would be bad. The hysteresis loss is also always there too, but eddy current losses scale as frequency squared. So by shorting the windings and excluding the rotor flux from the stator, you may actually get _lower_ drag torque than if you left the windings open circuit. This might in fact be literally the most efficient way to drive such a cheap, high RPM motor. When you want to apply torque, you need some of the rotor flux to penetrate the windings. So you allow all the flux through the windings and apply as much torque (current) as possible for a time, so the ratio of injected power to power loss is high (high efficiency). Then you exclude the flux from the stator completely to stop that iron from absorbing tons of eddy current loss and creating a huge drag torque.
Yes, core losses will be reduced by shorting it. But - if I completely let go of the trigger, then it's open circuit, and it takes a long time to spin down. So the non shorted torque from core and cogging losses has to be quite low.
That's a fair point @@matthiasrandomstuff2221. The system might have quite a lot of inertia compared to how much power is flowing though it, so a slow ramp-down might still represent quite a lot of internal electromagnetic friction loss. But I would have expect a pretty rapid slow-down if the core loss was really huge. Was it laminated by the way? Couldn't really see.
This is not the dumb way to drive an BLDC motor, this is the cheap way. Almost every ESC uses this method except high end controllers like the vesc which has the option to use FOC (field orientated control) which is what you think would be the right choice, but if you think about it, it would greatly increase the cost of that plastic contraption. Brushless servo motors can benefit from FOC because they sometimes need to turn slow and that way they don't have oscillating tourqe. Motors with high rpms and flywheels can get away quite happily whith this simple driving method you examined in this video. It may not be as efficient as foc but the difference would be minor unlike the cost difference. In conclusion I think it is the smart way to drive a brushless weed wacker. :) If you are interested in this kind of topics check out Great Scott's channel.
Perhaps try it with the string out? That may put more load on the motor and it would get out of what seems to be an over speed protection mode. Another thought: is the driver pulsing its output so that it can ‘read’ the motor speed in between applications of power?
I wonder if the driving in this manner doesn't work better when the string trimmer is loaded. A single string trimmer would only be loaded for maybe 120 degrees of the rotation. The rest of the cycle, would be building up inertia for the impact cutting during loading. So the motor runs free with the expectation that the loading will average it out, but on the test bench it has to brake.
I just bought the same exact thing lol. Now at least I know what that sound is (the sound it makes when its done spinning down ). In any case I was wondering if once the spindle is spun up if it will charge the battery when I release the trigger, since it doesn't stop right away. Will the big gear or small gear wear out first with all the chatter? Maybe depends on how its operated. I'm pretty happy with it so far, but it's a new purchase. Even my 1.5 Ah batteries are enough for my yard but its nifty you can use the heavier batteries with it too. My only issue with it is the vent holes on top don't stay clear without a little care, much unlike the drive shaft powered gas trimmers of yore. PSA: Stay away from cable as a drive shaft trimmers, if the shaft has a curve you know its a cable drive shaft... the cable will wear out, especially if pulsing the throttle.
I can't speak to brushless motors, but I have some observations on that weed trimmer. I have the same model. When I run it continuously on "High," the battery only lasts a few minutes. Continuously on "Low," I was able to run it continuously and take care of my whole yard with power to spare (> 30 minutes, maybe an hour). Without taking any power measurements, I supposed this was due to the lesser current draw on the LiIon batteries. Based on your waveform there, I'm now wondering if the "brain box"/potted electronics module is doing some kind of regeneration, rather than simply wasting the power.
I believe there is nothing stupid, this waveform is representing only one phase out of three, so when it is turned off (suppressed) another coil is actually turned on in order to complete the cycle, it's just how 3ph BLDC motors work. Note: opposite coil are usually connected in series and generate both S and N poles when energized. Nothing wrong with your basics, your basics are really good.
That was my thought, without probing all three phases you aren't getting a complete picture of what's going on. It's obviously a very basic method of control, with no PWM, but it would be interesting to see all phases. Perhaps it's capturing the phase current from the off phases to feed to the active phase? My theory is a little weak on how difficult it would be, but it's worth a thought.
However if it's doing that very quickly and switches of it's cycles it doesn't become much less efficient if any less efficient at all. Because PWM is also just a way of modulating the signal. And it decreases or extends the off time of the motor.
maybe, because it is more efficient to cut grass with that pulse freqency, than with constant frequency. the grass will be blown down with constant frequency, but will still rise with the pulse frequency. my other tought, the string will having a whip effect with the pulse frequency. just my tought, never try that kind of thing.
My thoughts as well, it's like the difference between chopping vegetables and slicing through a cake, the former needs a repetitive force and the later a continuous force. Could even be designed that way to more closely mimic the force of a gas trimmer. I'd be interested to see Mattias test a cheaper (perhaps a black & decker, as I had one like that and hated it, the battery died very fast) trimmer and see if it does the same thing.
Matthias, your scope is aliasing the pwm waveforms. Common brushless switching frequency is 20 khz. Your scope is only sampling at 10 ksps, which is waaaay too low to see this. You need a better scope to see the fast pwm switching. At 20khz switching, this is 50 microseconds per switch, and you want scope resolution at least 10x faster to get a clean trace of that.
Yes, saw the comment. but it's not aliasing (What are the odds that for all speeds its running at the same frequency as the scope samples at). It's using block communtation. Just because you know of something that can happen doesn't mean that it did.
@@matthiasrandomstuff2221 hi Matthias, the switching frequency of the mosfet is fixed. The duty cycle is modulated to create variable DC voltages. The switching frequency doesn't change. It is always at a constant 20 khz. The commutation rate does change with the motor rpm, however. I have been working on brushess DC motor controls for 6 years now. The scope is aliasing the pwm signal, I guarantee it.
@@superxpro12 You have very good "comment posting etiquette", you kept your cool and focused on your comment without any hard feelings towards his comment previous to you.
3:05 a bit late now but next time instead of voiding warranty needlessly, YSK the parts diagrams are available online for all dewalt products (and pretty much for all consumer tool manufacturers.)
Looking at the scope screen at 7:08, it seems that the measured phase is powered two thirds of the time. This makes perfect sense, since running the motor on open loop requires one phase to take current in, the second phase to pass that current through and out, with the third phase being used to measure the back-emf signal to accurately time the next stage (i.e. power A-B, read C, then power B-C, read A, etc.). Sinusoidial control (FOC) would be cool to see in a consumer product, but that is still more widely seen in larger/more expensive BLDC applications... Not here :)
Except that you can't drive the phases independently. Only three wires going to the motor. And even if they were independewntly wired, it would be stupid to short circuit the winding you are trying to sense on.
@@matthiasrandomstuff2221 My answer probably wasn't the most understandable, so here's try number two: normally, a BLDC controller drives the motor in such a way, that there is current going into/through one phase and out/through another, while the third one is being monitored. The phases are interconnected either in delta or star (Y) and as such are "shorted" if looking at things resistively. I'll try and find a link to better illustrate what I'm trying to say. (Also, the reason you're not seeing a constant current in a phase before the "dead" zone is because the current would rise too much and thus is pulsed in a way that the measurement result looks like a sinusoid).
images.elektroda.net/88_1274660220.jpg has the three power "signals" underneath the not-so-relevant hall sensor signals. This is how power is applied to each phase/phase pair 2/3 of the time. As far as control tech goes with sensorless BLDC control, there's some decent info here: www.roboteq.com/index.php/applications/applications-blog/entry/sensorless?format=amp
You are mixed up. For several full cycles of the back EMF, the observed voltage is zero (or very low). When the motor is spinning fast, the only way the back EMF can be that low is if it is shorted out by the drive circuitry. This leads to high current in the phase wires, strong braking force, and rapid temperature rise in the motor windings, but not to battery charging.
Yes, I had an older version like that and a tab holding the ground cover broke, I ran it for a short while cutting heavier grass without that cover and very soon smoked the motor - dead once stopped - so I think that design just ran out to maximum and wanted friction of the plastic cord on the cover as one passive speed governing control.
Love these video Matthias! Keep Making them! I started watching your videos years ago for the woodworking concepts and content. In a way, your transition into this style of video for me is convenient because I'm now an electrical engineering student and I find these videos as a source of inspiration, and a break from all the textbooks and studying.
Remember the coil is measuring the derivative of the field. So those spikes would indicate a square wave. Down load the waveform into a spreadsheet and integrate the waveform. Looks like fun.
Nothing strange happening with the weed trimmer, you're just measuring it wrong :P The motor has three sets of coils in opposite pairs, you need to scope all three coils. Have a look at the wiring, if the motor in delta or star config? I'm guessing delta, because I can only see three sets of wires? EDIT: Also the controller likely isn't braking the motor (what would be the benefit of that?) but rather it's using the coil's off period to measure back EMF, which the controller uses to measure speed as well as angular position. I don't think that motor has any hall effect sensors, you need to have another look :)
that's not how you measure back-EMF in a synchronous motor controller. You measure it during the cycle, detect the zero crossing, wait a bit for timing, and then shift your state machine one state. here is a paper on the topic: www.nxp.com/files-static/product/doc/AN1914.pdf Mathias thinks the motor has been put in short circuit because the back-EMF is extremely weak, in free wheeling it would be about the same size as when driven, so it probably has been driven low by the H-bridge-> that's a brake.
The voltage from phase to phase was zeroed out for several cycles. If the motor was coasting as you say, then you would see the sinusoidal back EMF from phase to phase. So, no, not coasting to measure back EMF. The controller is definitely braking by shorting from phase to phase.
Matthias, it would be useful for you to measure the maximum power output of your brushless tools before disassembly and after disassembly and see if there is any degradation of torque etc. The reason I mention this is that I have a number of expensive stepper motors made in the 80s/90s that expressively warn against disassembly because the permanent magnetic field of the rotor is damaged (removing the rotor from the stator changes the magnetic path from a low-reluctance path to high-reluctance, and that can push the field on the B-H curve to a lower intensity stable B field).
Just a couple comments about screws in plastic; 1) don't over tighten, use a hand screwdriver and 2) start the screws with your fingers turning the screw gently counterclockwise, you should be able to feel the screw "drop" into the original thread. This preserves the most plastic strength.
Silly question - does the drag of the string make a difference? spin it up with the string at full length and see if you still get less than full drive cycle. I know mine, brushless EGO brand sounds a lot smoother as if sine and the board get along.
It needs a load, the gas ones I use have a large RPM drop with the standard load from thick weeds. I imagine once its in the lower RPM range it will stop the breaking like that.
I didn't understand much of what you said, so please forgive me for being puzzled over your remark about braking. How do you know this is not just an intermittent PWM pulse of energy to save battery life? Once these get to full speed, the torque requirements are minimal, in contrast with what you would want for an electric drill.
A BLDC is like an alternator. If it is spinning, it is generating voltage as the magnets move past the coils. If it was "coasting" and not being driven by the controller, you would still see that sinusoidal voltage across the wires where he put his oscilloscope. You may be thinking that when drive voltage is removed from the motor, it will just rest at zero volts, but that is not the case.
I think the conclusion here is that Matthias doesn't understand why they're doing this. Weed whackers need this pulsed motion to cut weeds efficiently.
almost seems like for the "Brake", they are dumping the energy back into the battery. They probably drive the the motor hard so their driver transistor is either on or off and has a minimal power loss across it.
So, all that waveform is normal without load. Load the motor, then take measurements. brushless controllers are a closed loop system, and it cuts out due to overspeed sense in the circuitry. When you provide a 20-30% load, the frequency should smooth out because the motor is driven purely under load but at speed. Make sense? I hope?
In this closed loop, the feedback from inductive kick during off time generates an error signal to regulate speed. Unless you have an inductive pickup for your scope, what your seeing isn't real due to the very high impedance mismatch. I hope this helps. But, this thing is supposed to be as cheap as possible so I guess they don't care about any of this and just make it turn. Idk
Mathias I find your videos intriguing. I don't know what the heck you are saying but I am fascinated non the less. What I understood from this video was Dewalt trimmers are stupid.
Very different presentation to AvE but just as insightful and it's also a case of... I bought the tool with my own money so I can be honest in my opinions about it. A breath of fresh air compared to some channels. Only thing I don't like about Matthias is how much smarter he is than me..
I'm so old, I keep to the scythe. It sings quietly in the grass and I can let the thoughts fly. It doesn't even hurt- says the grass- falling to the scythe. ^.^
Pure speculation of course: My guess is that they decided to use a very cheap, dedicated BLDC motor controller IC, that only supports running at max. speed. Doing high-speed PWM (>10x faster than RPM) requires fast MOSFET drivers and fast transistors to avoid losses during the switching, and that costs more money. As to why they're actively braking: maybe they need the motor brake to stop quickly when you turn it off, and they didn't invest time/electronics to switch between coasting while on, and motor brake when stopping, and just enabled the brake by default? And then they added the speed control by doing slow PWM on the motor driver enable line, which results in the "drive"-"brake"-"drive"-"brake" pattern. Altogether, it does look like a very suboptimal solution to save a few bucks (or someone who just cobbled together a circuit drawing without really understanding what they're doing). All just guess work, impossible to say without having a look at the circuit. Maybe they have another, better reason that actually makes sense, but I can't think of anything.
But when Matthias drops the trigger, instead of then shorting the coils and braking the motor, they instead drive it to open circuit and the motor coasts...?
Most likely and perhaps some line extended to vary the load. A proper test would record the current draw under true load and then simulate that load in the water. I wonder about cavitation or wheater that would be a factor.
I would say it allows them to have a smaller motor so saves weight and money for them. In the end a 3:1 torque increase isn't all that much when slashing through the tall grass.
Put some trimmer line in the head and test it at the length used for cutting grass. I don’t know if the head is auto feed or not but possibly as the cutting line wears shorter it reduces the load on the motor thus the motor revs higher to a point we’re the engine brakes and this triggers a mechanism in the head to release more trimmer line.
It looks like a PWM signal is controlling the sine wave instead of approximating one. Which, as you say, is a weird or stupid approach. Do you think it has to do with the output current available? Do you have the capability to make a test circuit to see how much current it draws with an always on sine wave at the same values as what it is showing here? ( just speculating here now, since I used to be a tech for the prototyping at Rigid for their vacuums, and engineers are notorious for head scratching reasons to do some things) Perhaps it is trying to decrease the strain on the battery, or an attempt at staying within a spec that really should have been met differently?
Isn't that what you name as stupidity by chance a current chopping technique quite often used in bldc drivers to avoid excessive current in bridge driving the motor?
Signal patern is probably like that because transistors are more efficient at lower frequencies, they heat up more with pwm modulating the phase current, which makes electronics much more durable as it doesnt need to dissipate that much heat.
I have already had this problem using quite beefy mosfets but they were overheating, for sample e-bike controllers for 1-1.5kW usually have 18 mosfets and a whole aluminum case to cool them.
Just as a wild thought from someone with no background in electronics: could it recuperate power during the braking? Like charging a capacitor and then using it to drive motor next cycle? If this is possible, it may be simpler, than implementing PWM.
is it regenerative braking.i have discover if a bldc esc motor controller is in a motor break possition and i turn the motor with me hand it wil give a charge back to the battery.
I think under load it might not do that. Maybe it shorts the coils to allow for some kind of budget speed sensing circuit? Maybe when it is under load it can regulate speed without shorting the coils as the back-emf would be greater and a more elegant speed sensing method can be used.
I think this is all about saving money on the micro and half bridges. A proper sinusoidal processor requires 6 reasonable speed pwms. With this implementation known as "bang bang" modulation, you can do the control with a really cheap micro and half bridges.
You need 6 pwms if you are going to control the half bridge dead time. And if you dont control the dead time then your driver trebles in price. I have designed a few to find this out. These were high performance though
you can drive them so you only need to pwm one phase at a time, www.infineon.com/dgdl/Infineon-AP32359_BLDC_Motor_Control_Software-AN-v01_00-EN.pdf?fileId=5546d46258fc0bc101596988325e3f87
would it not have to do with the the third set of coils that are not attached, to the scope (not registering a signal, the third set would always keep a magnetic inershia thus prolonging the battery load, tri-lobal magnetic field limits the load requirement to start and maintain the rotational field,basically they are never truly north south , and why the stator clings to the magnetic field when stationary 33.33% of a pole rather than 50%opposition , similar to S.E.G.? lol ,got me thinking of of a AC MOTOR AND ELECTRICAL MAGNETISM course i took in moncton nb in the late 80's
makes me want to go out and tear apart my Makita 36V ($400) weed wacker that is smooth and quiet and runs for 1hr+ continuous with 2 4Ah batteries. I have a scope so could rig up a coil... but sound like too much work. What I want to know is why my Dewalt job site table saw 7480 motor (110V AC) screams so loudly.
I had a DeWalt trimmer just like this. Had it for 2 seasons, in my tiny yard, and never fully uncharged the battery. Then after this past winter I went to use it and the battery was dead. Wouldn't charge up. I call customer support and they tell me it's out of warranty, and that they don't make that battery anymore. Although I could buy a replacement battery at Lowes... for $265. The trimmer / battery came together for $150. A brand new trimmer / battery was around the same price. Meanwhile I've had a Ryobi mower, trimmer, and leaf blower and 3 batteries that are absolutely _packed_ with juice and I can take care of my entire yard on less than 2.5 batteries. Never buying DeWalt battery-powered lawn equipment again. My experience is that they make solid cordless hand tools but I don't trust the brand anymore after they abandoned this whole battery pack after 2 measly years. The kicker was that when I gave the serial number for the battery, the lady at customer support told me it was manufactured in 2016. So I bought a brand new trimmer that came with a year and a half old battery. Great.
You may be able to take the battery apart and replace the cells. I don't have one of these machines, but I have seen videos of people doing that for various battery-powered tools.
It would be awesome to pry open the siliconed brainbox and probe directly on the drive transistors to verify brake mode enganging. The suppression of the back emf looks very odd to me. A good dead short should not let +-3v on the coils...
Not trying to be a DeWalt fan boy but I really like my weed eater regardless of how inefficient the motor is it cuts grass pretty well and is easily handled by my wife and kids. I don't like the spool mechanism very much though. I have a half acre in one battery will do the whole thing pretty easily
It's because DC motors are more efficient at high voltages/currents as the inefficiencies are static and/or diminishing. So tl;dr if running 10W through it gives 1 unit of power out, 100W would give not 10 but 12. So they use this method to get the motor working in it's sweet spot and then it's supposed to idle, which you're probably seeing as back EMF and thinking is lower power input. Great video though. Too bad the loop wasnt very sensitive.
For safety reasons, that trimmer needs to stop fast. The PWM is just on-off control: drive full speed or brake. It would be better if they braked only when the trigger is released but that would probably require a micro controller and software instead of a cheap driver IC. DeWalt probably has no idea what is going on, they just picked a Chinese factory to make it and drove a hard bargain, giving the factory little financial room for improvement.
You should have it connected while wheat whacking. With the amount of load wheat whacking generates it should be fine. Or at least with a wire so it has some resistance going through air.
you notice there is a craftsman version of that weed trimmer? i have a feeling that stanley black and decker are using something they aquired from the craftsman buy out
Very good video, the problem is called chineeze engineering. Make it as cheap as possible, sell it to the public as a brushless DC motor which most would think its going to last. What will go before the hardware is the battery, that is THEE most EXPENSIVE item on this device. The same goes for their Brushless DC Motor (BDCM) Drills.
When the motor is braked I wonder if they store the energy generated somehow in the caps to use in the next power cycle? We used to do something similar driving solenoids at high frequency to reduce power consumption.
It might just be because it has to be able to push harder under load than at idle, so when it encounters resistance it can push hard without going too fast. I assume the limit is there because they tried it without it and something broke... and instead of redesigning it, they just "fixed" it.
yes, but you could just drive it less of the time and not brake in between, and still save power. And a weed trimmer is mostly not running at full load when in use.
It‘s just a 3 Phase Motor like we have it in germany, they just take the DC Input signal and convert it to a 3 Phase AC Output with transistors, this thing is called an ESC (Electronic Speed Controller).
I dont have a large lot so I got a small black n decker lithium model, best weedeater I have ever used (cause I can just do it and be done instead of mixing gas and fussing around with those tiny devil 2 strokes)
Kinda like some OLED technology, whereby to show a dim picture, it just ups the refresh rate of the pixels so much so that the brightness is never at max.
2:00 It doesn't put one magnetic pole on one set of coils and second pole on the other two sets of coils, at least not fully like that. If you look at a signal from all 3 sets of coils it should be 3 sinusoidal lines, each moved 120 degrees from the others( upload.wikimedia.org/wikipedia/commons/c/cc/3_phase_AC_waveform.svg ), so you should have said "when one set has full north, at the same time the second set is in transition to full south and the third one is in transition from the full south". You said it like at the same moment one set has full north and the other two sets have full south. I dont mean to criticise too much, I just wanted to make the information more precise for people that may be learning for the first time about 3 phase from this video. I love your videos, Matthias :)
The drive voltage for the motor would not be a smooth sinusoid (assuming you use a scope with decent bandwidth). That would imply linear control and very high power dissipation in the MOSFET's or IGBT's. The drive would be more like a class D audio amplifier. That is, PWM. Also, sine-wave drive is not universal for BLDC's. Often they are driven with six-step commuation. You can look that up on your own time (six step commutation of bldc).
As with batteries, it might be best to test under load: Without load, batteries exhibit near-to-nominal voltage, even if they are practically empty. With a 50 Ohm resistor as load, things turn out a little differently. I suppose the same goes with electronically controlled motors: No serious load, no usable measurements.
Maybe they are trying to account for a variety of operating styles. I had a neighbor who used to pulse the trigger on weed eater twice a second while trimming and edging. It was very annoying to listen to. He can't be the only idiot who runs them this way, but maybe this control scheme allows for that type of nonsensical operation.
To really understand what that weed trimmer is doing is to take the electronics apart compare it to the electronics of other brushless tools. I would say bread board a circuit using an arduino as the PWM and drive it through some FETs and see if the motor runs the same. I see that various BDCM run differently as to how they are driven by different circuits and the frequency you subject them to. The quickest way to drive 3 phase DC Motor is to use an RC ECS module. That uses PWM.
Maybe it's because the H-bridge would clamp the induced voltage to supply, causing even more drag if they don't short the winding? I would also not have expected this, but probably best not to call a man stupid until you have walked a day in his shoes, likely there are good reasons. Interesting that Dewalt is again betting on a geared system while the Makita trimmer uses direct drive, just like the chainsaw.
It's inefficient, but it's cheap. I think they're doing it this way, because now they can use a far slower and simpler controller, with some cheap dedicated chips to drive the motor
That was my thought, DW is considering total cost of BOM, not ownership. They'er going to make 10's of K's of these and want to keep production costs low, so what if the owner has to recharge more often.
@@SlaveToMyStomach And if the owner has to recharge more often, the battery wears out faster, so they have to get a new one, and hopefully by that time, "Oh, look, we don't make that battery anymore, I guess you have to get a whole new tool!"
Actually you can drive the coils sinusoidal and minimize the cogging. The sine waves can be synthesized with modern power electronics. Of course that motor may have a simplier drive.
@@superdau less torque ripple, smoother current waveform, and whatever you lose in top speed you can potentially regain if you do phase advance (field weakening).
Alternately using the accelerator and brake doesn't make sense under conditions where you can see down the road and can know, can have some look-ahead to approaching conditions, and your preference is to conserve energy. Arriving at a particular time with some precision (no sooner, no later) is of reduced importance. In other uses of an automobile, the accelerator is frequently alternately operated with the brake, and is often operated at the same time as the brake. This is done in racing, where performance is given more priority than economy of fuel consumption. In control system theory, you can think of accelerating with the brake on or rapidly switching between accelerator and brake as a means of increasing the control system gain. On your weed whacker, if the motor's braking energy is recovered and fed back to the battery, the losses are somewhat reduced. Regardless, the motor controller can't see the thick grass approaching the cutting head and know to up the output in advance of the additional load, like a car driver can see an approaching hill and press the accelerator farther to compensate, and not lose speed. So being immediately ready to respond to a variance in load is critical, otherwise the weed whacker will seem weak in some circumstances and too powerful in others. Or, under certain steady loads, the motor speed may oscillate.
Consider that in high performance motion control systems (not steppers, and typically systems that move faster than CNC machine tools), the motor drive amplifier typically drives the motor forward and backward simultaneously in order for the system to position a load stationary under varying demand. A DC servo motor may go one direction with one polarity and another direction with the opposite polarity, but to brake it, or keep it still, a judiciously chosen AC signal is used. It is essentially being rapidly alternately driven forward and backward. If you put your fingers on the motor shaft and try to make the motor shaft turn, the drive amplifier mixes in a DC bias with the AC to counter your efforts. The mixture of AC and varying DC bias keeps the motor excited and ready to respond immediately to any variance in command, which would be in response to any variance in effort from your fingers. AC servo systems do exactly the same thing, but the motors are typically very similar to BLDC motors, with a permanent magnet armature and 3 phase stator. For a great example of high performance servos, look to industrial electronics assembly machinery. Or, automated military fire control systems.
On your weed whacker, I imagine the motor control uses field oriented control, where the voltage of windings that aren't energized is measured as a means of judging speed. Perhaps there's too much delay in processing that and applying the result to the motor drive output isn't fast enough to prevent a lag in speed correction, or some other aspect of the system presents a lag that is relatively expensive to improve. This usually means the "system" needs more gain. (If you're driving a 4-banger, how much earlier do you have to press the accelerator to have some chance of keeping up when the hill arrives?) They can't simply increase the amplification of the speed correction, depending on the circumstance, this causes oscillation. They can't leave the gain of the speed correction low, this causes poor performance. To compensate, system gain needs to increase. Increasing battery voltage would help, but that makes the product more expensive and or heavier, requires more expensive power transistors, etc. Increasing the gear reduction would help, but at the cost of slower maximum cutting head speed. They could add a sensor to provide the control system continuous true tachometer feedback, but that would be more expensive and perhaps less rugged. At the cost of some battery capacity, they use this alternating acceleration/brake technique to up the system gain to provide higher, or perhaps just acceptable performance. The user has to go back to the shed and swap battery packs a little earlier in exchange for that additional performance.
Anyway, that's what I'm thinking they're doing.
Thank you sir.
Ditto
I feel like this is BOLTR Matthias edition.
I feel like BOLTRs are WoodenGears enginerding videos autopsy, I mean AvE style.
Scott Cress except matthias actually knows what hes talking about.
Matty Ice they’re both pretty clearly well trained. It’s a flavor preference more than anything else.
OK, so where are all the f***ing profanities?
@@lennyf1957 Not available in the Matthias Editition. You need M.A.D. (Matthias After Dark). It's a little more wild and the experiments with Legos are much harder.
I did not see any hall effect sensors, so they are probably running open loop until the motor spins fast enough to read the back EMF. They are most likely running a control with a current regulator as well. Generally, you accelerate the motor at a fixed rate and fixed current setpoint then switch over to sensorless control after a certain period of time. The acceleration rate, current setpoint value, and time are normally determined by application. In this case, they almost have to assume the motor will be loaded. This all gets screwed up when you run it unloaded! The best way to tell what is really happening is to measure the current, I bet you will see it being a constant value. Also, watch for the way the motor is aligned to start up. A more sophisticated control will use induction to determine the rotor position at start. A simpler control will force the rotor into a known fixed position by applying current through certain motor phases for a certain period of time. If the control is using induction to determine position, you can be sure the designers know how to PWM properly!
I was wondering if that was just how a sensor-less ESC worked, though the evidence of braking/shorting rather than just "not driving" was unexpected.
@@rpavlik1 That is the general overview of how BLDC sensorless drives work, but there are lots of little details I skipped over. The braking portion is probably a byproduct of their control logic and regulators that are tuned for operating under a higher load. For example, they may have a speed PI regulator that tells the system to brake when it overshoots. This would be an easy way to quickly stop the motor when the trigger is released (assuming the trigger is an input to this PI).
Being a brushless motor though, could/would the rotor will ever spin faster than the phases are being clocked around it? If not, there shouldn't be any reason other than to determine locked rotor (no back EMF) or how much load (how fast the frequency of the back EMF decays). Otherwise, would they able to determine the load by monitoring the phase angle difference between the voltage and current on the windings to see how far behind the rotor is from the generated field (slip)?
Seems like the controller was designed for something else and it was just drafted into this machine....maybe was cheaper to modify existing stock or design?
@dothemathright 1111 You are right about the coasting. I missed that part. Regardless, the fact it is braking the motor rather than coasting tells me that the control is in some strange state. Looking at the current and/or operating it under normal load would give us some more clues. Sensorless control can be difficult when you are not controlling something easy like a fan or pump. Since the controller is so far away from the motor, they probably saved a decent amount of money by removing the sensors and wiring.
"Good haven't broken it yet". A saying we have all said a lot.
With emphasis on YET! :o)
oh, parts left over...
Isn't it: "It hasn't failed, yet" ?
@@satxsatxsatx "OK, good, I haven't busted it yet"
This is just speculation... but...
I'm willing to bet that it is optimized for operation under load... cutting through weeds could put a somewhat significant load on it, which would prevent the trimmer head from getting into the higher RPM's where it hits the brakes... and perhaps due to the nature of the trimmer head (string flying around with large centrifugal forces) the braking you are seeing is an overspeed protection to prevent the forces from becoming too high... they didn't bother to do it via the more complex PWM because in 99% of cases the trimmer will be under load while using it which would prevent the overspeed condition in the first place.
cryptk That would be my guess too, but hard to say for sure without looking at some actual technical specs for the trimmer and the motor assembly and driver.
And the people who trim early in the morning while the grass is at its wettest - you can't tell if the motor is burning out or if it's just a mist that is rising.
When I use a petrol strimmer on heavy growth near sharp fence posts there is a big variety of load. Often the nylon is cut off by square fencing at high speed or can slow down with tough weeds.
As primitive as that controller is, they set it up like that because there is a lot more torque required when using the tool with string and cutting grass, I think you will find it running almost full out without braking to slow it down.There is another story with ecm motors that this controller does not make use of and that is back emf from each coil, the good controllers actually feed the voltage spikes into the next coil pairs to smooth out operation and add torque.There was a company that used a solid state disk for the rotor about 10 years ago but I don't recall the name.
I think they designed that way because it is the cheapest way of implementing speed control. It looks like bang-bang control, like a heating system with a thermostat. There is no need to modulate the combustion because the inertia in the system smooths out the response (the room warms slowly). In the same way, DeWalt can use a cheap voltage amplifier and just switch it off when the reference speed is achieved and the inertia in the system will maintain the angular velocity. It would be interesting to see if you could make it more efficient with modulation of the voltage or frequency.
there is no "cheap" control in BLDC motors, you have to have your controller estimating the position of the rotor (basically it has to do what brushes would be doing) and keeping a tab on it. The only thing might be the absence of PWM, but 1) it's cheap, the "costly" part is the back-EMF sensing because it's noisy and you can't filter too much 2) they could also commute late to limit the torque.
Thinking of how a trimmer is used, I can maybe rationalize a reason. Trimmers tend to go from full load to coasting very quickly and often. When loaded (trimming) the motor will almost always need full power. Then you have to make sure it doesn’t spin too fast, or the string will break too easily. I’m betting they built the prototype with a different control system and had problems, likely with it going too fast and breaking the string, so their solution was this braking mechanism. It would be interesting to test it in use.
I believe the Sopwith Camel (?) had no throttle, i.e., it was wide open all the time. There was a kill button on the stick and the pilot controlled motor speed by pressing and releasing the kill button. You can hear this in many of the WWI movies made in the 1960s thru the 1980s. Most were done in B&W.
You need to test it with the wire going against grass. That should cause quite a bit of resistance, and maybe, just maybe, we can say the trimmer is not engaging break mode for most of its operation time.
I was thinking it was using maximum voltage so that it would either cut the grass or not cut, rather than half cutting and stalling.
The foot on the gas comparison Matthias made was pretty accurate, except imagine the car is electric. Full power when your foot is on the gas, no power when it's not.
Now whether this behavior is beneficial or not, I have no idea. Only the guy who made it knows what the thought process behind it was
At least retest it with string in the spool.
Should be fairly substantial difference with cord in trimmer head
or someones head.... wwe weed wacker match.
Ah, so that explains that twitter man's Tesla design.
I grew up learning to pulse the throttle of my weed whacker to keep the rpm low to avoid sending rocks through my windows; It can't be a coincidence that it's the only tool that performs the same thing electrically is the weed wackier.
Around 6:40: this is done to maintain both speed and maximum torque. You could regulate voltage accordingly to current, but that would be a waste of battery power to heat up another bunch of transistors, and probably spin up a fan to cool them.
It seems that in the null part of the duty cycle, when the coils arn't being driven, they aren't disconnecting the coils. They are driving the mosfets to short the coils instead. This doesn't allow the motor to coast, but would act as an electromagnetic brake.
The question I (and I think, Matthias) am asking is why they do not drive the mosfets to open-circuit, so the motor doesn't generate current and is not slowed down.
And when an electromagnetic brake would be useful - when the trigger is released - then the mosfets are driven to open-circuit, and the motor coasts.
My thought is that they could be using a tuned circuit between the coils and some capacitors, perhaps, When they are not driving the coils, they resonate instead. Measuring the voltage across the coils may not show this, as you are really only measuring the voltage across the shorted mosfets.
That's bizarre. I'd love to see a schematic of the driving circuit. My shitty best guess is that either it was designed by a battery company or its somehow capturing some of the wasted energy and reapplying it to the next full-load. It would be interesting to see a current reading from the battery in sync with the coil signal.
measurement under load would be more predicative then measurement in idle
'cept all tool specs (by Mfg.) are done with no load lol
But in actual practice, most of the time a weedwacker is running, it's not under load (the load comes in quick bursts.) So it seems like a poorly designed weedwacker.
I have some experience with motors and inverters. Shorting a synchronous motor that is rotating isn't as bad as you might think: contrary to rotating it at low speed and shorting it, where the resistive component dominates the reactive circuit of the shorted motor (and it really acts like a brake), at higher rates of rotation the inductive component dominates the circuit (the motor is decelerated for "half a pole" an then gets accelerated by the induced current again for the next half pole). So the only losses you get in that mode are the current losses in the windings. This cancels out most of the hysteresis losses in the stator, as the magnetic field is kept out of it by the shorted windings. Contrary to that, if you leave the motor open while it is truning you get the hysteresis losses in the stator instead of the resistance losses in the coils. Maybe(I'm not sure) there might be motors where the hysteresis losses when open would be higher than the resistive losses when shorted.
Do you suppose the hysteresis losses are greater than the copper losses? Also, copper losses plus some amount of volage drop across the electronics as its not a perfect short.
It's a possibility - But I don't know of any other motor inverter that does it this way. The main point is: Shorting a synchronous motor while it's spinning fast does not waste as much energy as one might think, opposed to shorting a DC motor, where all of the energy would be dissipated in the winding resistance quickly.
Its just overspeed, needs a load.
Yes, try testing with the inductive pickup while cutting grass to put it under a true load. The motor is freewheeling once it reaches max rpm and the controller senses overspeed and applys the brakes
I don't think it's doing the PWM strictly for the motor, but rather as a feature to drive the head in a different way.
When cutting grass and weeds, there's a lot to cut through. The spool cuts most effectively when it's whipping, not just rotating. It's sort of like an impact driver's effect on a bolt.
Try measuring the output RPM. It's probably more or less the same RPM throughout the trigger range if there's no load.
Simple answer what’s your definition of efficiency?... for automobiles adding aggressive air intake and a high flow exhaust creates a lot of efficiency focused on driving the car faster ... where as dampening the motors exhaust and slowing down its air intake will make the motor run more fuel efficient in the way it saves you money in gas...
so I’ve used trimmers to cut jungles of grass from 5’ to 4” and a gas trimmer running full out throttle when cutting long grass will generate a ton of wind which blows the grass or momentarily bends the grass and it get missed by the trimmers cutters. So our old dads would tell us do the job proper and pulse the throttle and not to go full out. Or we will have to trim the whole yard again tomorrow! Watching the scope when you had the trimmer hooked up I immediately thought about being in a grass cutting situation in which to cut efficiently you have to operate the throttle non efficiently...
That's brave, posting DeWalt on this channel again ;)
I really liked his video about "how to fix the Dewalt cordless tablesaw you just bought"
Something tells me he's not caring about DeWilt anymore as a brand.
which one is that?
I don't see any problems with Dewalt as a brand... Unless he is being a shill for them then yeah it's a problem.
Gentlemen! Welcome back to the shop!
*today* a treat une spéciale...
Why did I read that in AVE's voice?
gentlemanzzzz xD
Focus you f&$@!!!
Here in the UK, our loan mowers are added with a brake to stop the blades turning once the power trigger is released. Maybe its same for strimmers? The EU implemented this safety device on our new lawn mowers some years back.
I'm loving these explorations. Keep up the good work. Cheers.
The Slot in the Motor Back Cover is for Hall Sensors. Its the same Motor as the one used in their brushless Drills just without the Feedback. The Feedback gives you more torque at low rpm.
What I noticed was the motor lacking a feedback loop via a hall effect sensor. The controller only has a vague idea what the motor is doing. I've tinkered with brushless motors used in computer equipment where a fine grain control is needed and they all have hall effect sensors mounted by the rotor. The controllers are fairly low voltage and use PWM for control but they are still brushless motors.
Sensorless brushless motors are very common. Usually only two of the coil (pairs) are driven and the third coil (pair) is used as the "sensor". The only drawback is that the motor has to turn at a minimum speed to generate enough voltage in the third coil for the controller to "read". So for a tool it's totally fine to not have a sensored motor/controller.
If the controller only had a vague idea of what the motor was doing, it wouldn't work at all. You can't just wing it when you are commutating a motor. You have to know where the rotor is with reasonable accuracy to turn your transistors on and off at the right time.
Hi Matthias, little late to the party here. But one observation is that shorting the windings doesn't automatically incur high resistive losses in a brushless motor, especially if it has high leakage inductance. If you're operating the motor electrically much faster than the L/R time constant of the leakage inductance, then the rotor flux is mostly excluded from the stator by the opposing currents induced in the shorted windings, and the torque is low (equivalent to exceeding the critical slip speed of an induction motor). Yes there’s still induced current circulating, but depending on the magnitude of the winding resistance, this might not actually embody a large power loss.
Shorting the windings will potentially accomplish two positive things.
First, it could reduce the cogging torque. Cogging torque arises due to the variation in the reluctance which the rotor flux sees versus angle, as it tries to find a path through the stator. Shorting the windings excludes the flux from the stator entirely and so could reduce the cogging torque.
Second and probably more importantly, it will reduce the iron losses in the stator. Is that thing even laminated? At up to 500Hz, the eddy current losses in a solid core will be pretty darn severe. Even in laminations it would be bad. The hysteresis loss is also always there too, but eddy current losses scale as frequency squared. So by shorting the windings and excluding the rotor flux from the stator, you may actually get _lower_ drag torque than if you left the windings open circuit.
This might in fact be literally the most efficient way to drive such a cheap, high RPM motor. When you want to apply torque, you need some of the rotor flux to penetrate the windings. So you allow all the flux through the windings and apply as much torque (current) as possible for a time, so the ratio of injected power to power loss is high (high efficiency). Then you exclude the flux from the stator completely to stop that iron from absorbing tons of eddy current loss and creating a huge drag torque.
Yes, core losses will be reduced by shorting it. But - if I completely let go of the trigger, then it's open circuit, and it takes a long time to spin down. So the non shorted torque from core and cogging losses has to be quite low.
That's a fair point @@matthiasrandomstuff2221. The system might have quite a lot of inertia compared to how much power is flowing though it, so a slow ramp-down might still represent quite a lot of internal electromagnetic friction loss. But I would have expect a pretty rapid slow-down if the core loss was really huge. Was it laminated by the way? Couldn't really see.
what does it do under load? Maybe the design assumption is that it'll be under load most of the time and so there won't be any braking cycle
This is not the dumb way to drive an BLDC motor, this is the cheap way. Almost every ESC uses this method except high end controllers like the vesc which has the option to use FOC (field orientated control) which is what you think would be the right choice, but if you think about it, it would greatly increase the cost of that plastic contraption. Brushless servo motors can benefit from FOC because they sometimes need to turn slow and that way they don't have oscillating tourqe. Motors with high rpms and flywheels can get away quite happily whith this simple driving method you examined in this video. It may not be as efficient as foc but the difference would be minor unlike the cost difference. In conclusion I think it is the smart way to drive a brushless weed wacker. :) If you are interested in this kind of topics check out Great Scott's channel.
but is it better to short the windings while not driving it?
Perhaps try it with the string out? That may put more load on the motor and it would get out of what seems to be an over speed protection mode.
Another thought: is the driver pulsing its output so that it can ‘read’ the motor speed in between applications of power?
"Back EMF" (electromotive force) is definitely a thing. Could be a means of speed regulation when they don't want it running flat-out.
I wonder if the driving in this manner doesn't work better when the string trimmer is loaded. A single string trimmer would only be loaded for maybe 120 degrees of the rotation. The rest of the cycle, would be building up inertia for the impact cutting during loading. So the motor runs free with the expectation that the loading will average it out, but on the test bench it has to brake.
I just bought the same exact thing lol. Now at least I know what that sound is (the sound it makes when its done spinning down ). In any case I was wondering if once the spindle is spun up if it will charge the battery when I release the trigger, since it doesn't stop right away. Will the big gear or small gear wear out first with all the chatter? Maybe depends on how its operated. I'm pretty happy with it so far, but it's a new purchase. Even my 1.5 Ah batteries are enough for my yard but its nifty you can use the heavier batteries with it too. My only issue with it is the vent holes on top don't stay clear without a little care, much unlike the drive shaft powered gas trimmers of yore. PSA: Stay away from cable as a drive shaft trimmers, if the shaft has a curve you know its a cable drive shaft... the cable will wear out, especially if pulsing the throttle.
Try switching the controller for an ESC to see if it improves the way the trimmer works
Replace the electronic speed controller with an electronic speed controller?
I can't speak to brushless motors, but I have some observations on that weed trimmer. I have the same model. When I run it continuously on "High," the battery only lasts a few minutes. Continuously on "Low," I was able to run it continuously and take care of my whole yard with power to spare (> 30 minutes, maybe an hour). Without taking any power measurements, I supposed this was due to the lesser current draw on the LiIon batteries. Based on your waveform there, I'm now wondering if the "brain box"/potted electronics module is doing some kind of regeneration, rather than simply wasting the power.
I believe there is nothing stupid, this waveform is representing only one phase out of three, so when it is turned off (suppressed) another coil is actually turned on in order to complete the cycle, it's just how 3ph BLDC motors work.
Note: opposite coil are usually connected in series and generate both S and N poles when energized.
Nothing wrong with your basics, your basics are really good.
That was my thought, without probing all three phases you aren't getting a complete picture of what's going on. It's obviously a very basic method of control, with no PWM, but it would be interesting to see all phases. Perhaps it's capturing the phase current from the off phases to feed to the active phase? My theory is a little weak on how difficult it would be, but it's worth a thought.
Hilarious auto-captions:
Let’s monitor the waveforms this way [Music]
However if it's doing that very quickly and switches of it's cycles it doesn't become much less efficient if any less efficient at all. Because PWM is also just a way of modulating the signal. And it decreases or extends the off time of the motor.
great analysis, would love to see more of these.
maybe, because it is more efficient to cut grass with that pulse freqency, than with constant frequency. the grass will be blown down with constant frequency, but will still rise with the pulse frequency. my other tought, the string will having a whip effect with the pulse frequency. just my tought, never try that kind of thing.
My thoughts as well, it's like the difference between chopping vegetables and slicing through a cake, the former needs a repetitive force and the later a continuous force. Could even be designed that way to more closely mimic the force of a gas trimmer. I'd be interested to see Mattias test a cheaper (perhaps a black & decker, as I had one like that and hated it, the battery died very fast) trimmer and see if it does the same thing.
Matthias, your scope is aliasing the pwm waveforms. Common brushless switching frequency is 20 khz. Your scope is only sampling at 10 ksps, which is waaaay too low to see this. You need a better scope to see the fast pwm switching. At 20khz switching, this is 50 microseconds per switch, and you want scope resolution at least 10x faster to get a clean trace of that.
that's spot on superxpro12 - I hope Matthias sees your comment!
Yes, saw the comment. but it's not aliasing (What are the odds that for all speeds its running at the same frequency as the scope samples at). It's using block communtation. Just because you know of something that can happen doesn't mean that it did.
@@matthiasrandomstuff2221 hi Matthias, the switching frequency of the mosfet is fixed. The duty cycle is modulated to create variable DC voltages. The switching frequency doesn't change. It is always at a constant 20 khz. The commutation rate does change with the motor rpm, however. I have been working on brushess DC motor controls for 6 years now. The scope is aliasing the pwm signal, I guarantee it.
@@superxpro12
You have very good "comment posting etiquette", you kept your
cool and focused on your comment without any hard feelings
towards his comment previous to you.
3:05 a bit late now but next time instead of voiding warranty needlessly, YSK the parts diagrams are available online for all dewalt products (and pretty much for all consumer tool manufacturers.)
Can you please link me dewalt DCB105 schematic?
had to look up PWM: Pulse width modulation! nice video, thanks.
Looking at the scope screen at 7:08, it seems that the measured phase is powered two thirds of the time. This makes perfect sense, since running the motor on open loop requires one phase to take current in, the second phase to pass that current through and out, with the third phase being used to measure the back-emf signal to accurately time the next stage (i.e. power A-B, read C, then power B-C, read A, etc.). Sinusoidial control (FOC) would be cool to see in a consumer product, but that is still more widely seen in larger/more expensive BLDC applications... Not here :)
Except that you can't drive the phases independently. Only three wires going to the motor. And even if they were independewntly wired, it would be stupid to short circuit the winding you are trying to sense on.
@@matthiasrandomstuff2221 My answer probably wasn't the most understandable, so here's try number two: normally, a BLDC controller drives the motor in such a way, that there is current going into/through one phase and out/through another, while the third one is being monitored. The phases are interconnected either in delta or star (Y) and as such are "shorted" if looking at things resistively. I'll try and find a link to better illustrate what I'm trying to say. (Also, the reason you're not seeing a constant current in a phase before the "dead" zone is because the current would rise too much and thus is pulsed in a way that the measurement result looks like a sinusoid).
images.elektroda.net/88_1274660220.jpg has the three power "signals" underneath the not-so-relevant hall sensor signals. This is how power is applied to each phase/phase pair 2/3 of the time.
As far as control tech goes with sensorless BLDC control, there's some decent info here: www.roboteq.com/index.php/applications/applications-blog/entry/sensorless?format=amp
@@SaerX
The 2/3 of the time is valid for *within* one of the peaks, not over multiple peaks.
You are mixed up. For several full cycles of the back EMF, the observed voltage is zero (or very low). When the motor is spinning fast, the only way the back EMF can be that low is if it is shorted out by the drive circuitry. This leads to high current in the phase wires, strong braking force, and rapid temperature rise in the motor windings, but not to battery charging.
When using the trimmer i would imagene it is not braking that much because it has resistance from the thing you are cutting.
fl0r1svdk that’s exactly what I was thinking. Matthias can you make a “probe” and attach it to the scope and test it under a load?
It doesn't matter whether it is cutting. It just needs a piece of string in it. The string has a lot of air resistance at high speed.
Yes, I had an older version like that and a tab holding the ground cover broke, I ran it for a short while cutting heavier grass without that cover and very soon smoked the motor - dead once stopped - so I think that design just ran out to maximum and wanted friction of the plastic cord on the cover as one passive speed governing control.
Love these video Matthias! Keep Making them! I started watching your videos years ago for the woodworking concepts and content. In a way, your transition into this style of video for me is convenient because I'm now an electrical engineering student and I find these videos as a source of inspiration, and a break from all the textbooks and studying.
Remember the coil is measuring the derivative of the field. So those spikes would indicate a square wave. Down load the waveform into a spreadsheet and integrate the waveform. Looks like fun.
Right, so it's driving it with a square wave at a much lower frequency than the motor is rotating at? That makes a shitload of nonsense.
Nothing strange happening with the weed trimmer, you're just measuring it wrong :P
The motor has three sets of coils in opposite pairs, you need to scope all three coils. Have a look at the wiring, if the motor in delta or star config? I'm guessing delta, because I can only see three sets of wires?
EDIT:
Also the controller likely isn't braking the motor (what would be the benefit of that?) but rather it's using the coil's off period to measure back EMF, which the controller uses to measure speed as well as angular position. I don't think that motor has any hall effect sensors, you need to have another look :)
He measured the same effect without probing any specific coils.
Also you can hear it while it's running doing exactly what he's saying it's doing.
that's not how you measure back-EMF in a synchronous motor controller. You measure it during the cycle, detect the zero crossing, wait a bit for timing, and then shift your state machine one state. here is a paper on the topic: www.nxp.com/files-static/product/doc/AN1914.pdf Mathias thinks the motor has been put in short circuit because the back-EMF is extremely weak, in free wheeling it would be about the same size as when driven, so it probably has been driven low by the H-bridge-> that's a brake.
The voltage from phase to phase was zeroed out for several cycles. If the motor was coasting as you say, then you would see the sinusoidal back EMF from phase to phase. So, no, not coasting to measure back EMF. The controller is definitely braking by shorting from phase to phase.
Matthias, it would be useful for you to measure the maximum power output of your brushless tools before disassembly and after disassembly and see if there is any degradation of torque etc. The reason I mention this is that I have a number of expensive stepper motors made in the 80s/90s that expressively warn against disassembly because the permanent magnetic field of the rotor is damaged (removing the rotor from the stator changes the magnetic path from a low-reluctance path to high-reluctance, and that can push the field on the B-H curve to a lower intensity stable B field).
Just a couple comments about screws in plastic; 1) don't over tighten, use a hand screwdriver and 2) start the screws with your fingers turning the screw gently counterclockwise, you should be able to feel the screw "drop" into the original thread. This preserves the most plastic strength.
Silly question - does the drag of the string make a difference? spin it up with the string at full length and see if you still get less than full drive cycle. I know mine, brushless EGO brand sounds a lot smoother as if sine and the board get along.
Definitely. My gas trimmer rev limits without string. But runs very smoothly with string.
It needs a load, the gas ones I use have a large RPM drop with the standard load from thick weeds. I imagine once its in the lower RPM range it will stop the breaking like that.
I didn't understand much of what you said, so please forgive me for being puzzled over your remark about braking. How do you know this is not just an intermittent PWM pulse of energy to save battery life? Once these get to full speed, the torque requirements are minimal, in contrast with what you would want for an electric drill.
A BLDC is like an alternator. If it is spinning, it is generating voltage as the magnets move past the coils. If it was "coasting" and not being driven by the controller, you would still see that sinusoidal voltage across the wires where he put his oscilloscope. You may be thinking that when drive voltage is removed from the motor, it will just rest at zero volts, but that is not the case.
in the continued series Matthias applies for a job : Matthias tries to get hired by saying you are doing a bad job.
Yeah maybe the said company DeWalt NEEDS engineers with even HALF a decent brain not to make crappy tools and expect people to just keep on buying.
Matthias walk in to DeWalt and says :"Everything your doing here is wrong...i just want you to know that".
I think the conclusion here is that Matthias doesn't understand why they're doing this. Weed whackers need this pulsed motion to cut weeds efficiently.
almost seems like for the "Brake", they are dumping the energy back into the battery. They probably drive the the motor hard so their driver transistor is either on or off and has a minimal power loss across it.
So, all that waveform is normal without load. Load the motor, then take measurements. brushless controllers are a closed loop system, and it cuts out due to overspeed sense in the circuitry. When you provide a 20-30% load, the frequency should smooth out because the motor is driven purely under load but at speed. Make sense? I hope?
Motor spends most of its time under very light load in use, so that case is more important
In this closed loop, the feedback from inductive kick during off time generates an error signal to regulate speed. Unless you have an inductive pickup for your scope, what your seeing isn't real due to the very high impedance mismatch. I hope this helps. But, this thing is supposed to be as cheap as possible so I guess they don't care about any of this and just make it turn. Idk
@@matthiasrandomstuff2221 just adding string will load it quite a bit. Doesn't need to be cutting stuff to get resistance.
Mathias I find your videos intriguing. I don't know what the heck you are saying but I am fascinated non the less. What I understood from this video was Dewalt trimmers are stupid.
Amen, brother!
Very different presentation to AvE but just as insightful and it's also a case of... I bought the tool with my own money so I can be honest in my opinions about it. A breath of fresh air compared to some channels. Only thing I don't like about Matthias is how much smarter he is than me..
And he's less bullshit than Ave....
I'm so old, I keep to the scythe.
It sings quietly in the grass and I can let the thoughts fly.
It doesn't even hurt- says the grass- falling to the scythe.
^.^
I agree, but there are places you just can't get to with a scythe.
@@randomguy7253 Like onto an aeroplane.
Pure speculation of course: My guess is that they decided to use a very cheap, dedicated BLDC motor controller IC, that only supports running at max. speed. Doing high-speed PWM (>10x faster than RPM) requires fast MOSFET drivers and fast transistors to avoid losses during the switching, and that costs more money. As to why they're actively braking: maybe they need the motor brake to stop quickly when you turn it off, and they didn't invest time/electronics to switch between coasting while on, and motor brake when stopping, and just enabled the brake by default? And then they added the speed control by doing slow PWM on the motor driver enable line, which results in the "drive"-"brake"-"drive"-"brake" pattern. Altogether, it does look like a very suboptimal solution to save a few bucks (or someone who just cobbled together a circuit drawing without really understanding what they're doing).
All just guess work, impossible to say without having a look at the circuit. Maybe they have another, better reason that actually makes sense, but I can't think of anything.
But when Matthias drops the trigger, instead of then shorting the coils and braking the motor, they instead drive it to open circuit and the motor coasts...?
@@robertbackhaus8911 Mmh, yes, good point. I guess that invalidates most of my hypothesis...
Test it under load! you could simulate with paddles and a tub of water.
tbh the regular weed cutting head placed underwater would probably provide more than enough resistance, without needing paddles.
Most likely and perhaps some line extended to vary the load. A proper test would record the current draw under true load and then simulate that load in the water. I wonder about cavitation or wheater that would be a factor.
@@SouthernEngineering I suspect the cutting line would be quite draggy under water. I would expect a lot of cavitation off each line.
hard to say, but I think the line would most likely just wrap around the drum.
A small fan would be a good load, good enough for just air.
When zoomed out it looks like an AM radio signal. Is there correlation between how long it brakes the motor and how hard you squeeze the trigger?
Does that setup have any torque benefits in the way the trimmer is used? To prevent it from stalling in heavy grass?
I would say it allows them to have a smaller motor so saves weight and money for them. In the end a 3:1 torque increase isn't all that much when slashing through the tall grass.
Put some trimmer line in the head and test it at the length used for cutting grass. I don’t know if the head is auto feed or not but possibly as the cutting line wears shorter it reduces the load on the motor thus the motor revs higher to a point we’re the engine brakes and this triggers a mechanism in the head to release more trimmer line.
It looks like a PWM signal is controlling the sine wave instead of approximating one. Which, as you say, is a weird or stupid approach. Do you think it has to do with the output current available? Do you have the capability to make a test circuit to see how much current it draws with an always on sine wave at the same values as what it is showing here? ( just speculating here now, since I used to be a tech for the prototyping at Rigid for their vacuums, and engineers are notorious for head scratching reasons to do some things) Perhaps it is trying to decrease the strain on the battery, or an attempt at staying within a spec that really should have been met differently?
Isn't that what you name as stupidity by chance a current chopping technique quite often used in bldc drivers to avoid excessive current in bridge driving the motor?
My hunch is that they want to build it as cheaply as possible and nothing else matters.
Signal patern is probably like that because transistors are more efficient at lower frequencies, they heat up more with pwm modulating the phase current, which makes electronics much more durable as it doesnt need to dissipate that much heat.
But tens of kHz is not a problem. Besides, they already switch several times each cycle.
I have already had this problem using quite beefy mosfets but they were overheating, for sample e-bike controllers for 1-1.5kW usually have 18 mosfets and a whole aluminum case to cool them.
@@jansaljaj3042 To be fair, kHz, even a few MHz, is still low frequency.
Just as a wild thought from someone with no background in electronics: could it recuperate power during the braking? Like charging a capacitor and then using it to drive motor next cycle? If this is possible, it may be simpler, than implementing PWM.
Your channel ticks all my boxes. Fantastic.
Nice Metal Gear
Metal Gear? ..... no, it can't be! - GRAYFOX!
Is the next video going to be you building your own circuitry to drive this motor?
is it regenerative braking.i have discover if a bldc esc motor controller is in a motor break possition and i turn the motor with me hand it wil give a charge back to the battery.
I think under load it might not do that. Maybe it shorts the coils to allow for some kind of budget speed sensing circuit? Maybe when it is under load it can regulate speed without shorting the coils as the back-emf would be greater and a more elegant speed sensing method can be used.
I think this is all about saving money on the micro and half bridges. A proper sinusoidal processor requires 6 reasonable speed pwms. With this implementation known as "bang bang" modulation, you can do the control with a really cheap micro and half bridges.
plenty of BLDC drivers based on AVR, so it doesn't take much and one pwm shared between the phases is enough
There isn't really any advantage in trying to drive the coils with sinusoidal voltage, except for noise.
yeh, afaiu the principal different between BLDC and PMSM is one is wound for trapezoidal EMF the other sinusoidal
You need 6 pwms if you are going to control the half bridge dead time. And if you dont control the dead time then your driver trebles in price. I have designed a few to find this out. These were high performance though
you can drive them so you only need to pwm one phase at a time, www.infineon.com/dgdl/Infineon-AP32359_BLDC_Motor_Control_Software-AN-v01_00-EN.pdf?fileId=5546d46258fc0bc101596988325e3f87
would it not have to do with the the third set of coils that are not attached, to the scope (not registering a signal, the third set would always keep a magnetic inershia thus prolonging the battery load, tri-lobal magnetic field limits the load requirement to start and maintain the rotational field,basically they are never truly north south , and why the stator clings to the magnetic field when stationary 33.33% of a pole rather than 50%opposition , similar to S.E.G.? lol ,got me thinking of of a AC MOTOR AND ELECTRICAL MAGNETISM course i took in moncton nb in the late 80's
makes me want to go out and tear apart my Makita 36V ($400) weed wacker that is smooth and quiet and runs for 1hr+ continuous with 2 4Ah batteries. I have a scope so could rig up a coil... but sound like too much work. What I want to know is why my Dewalt job site table saw 7480 motor (110V AC) screams so loudly.
I heared somewhere that brushless motors tend to be louder than traditional ones, also with Makita.
I had a DeWalt trimmer just like this. Had it for 2 seasons, in my tiny yard, and never fully uncharged the battery. Then after this past winter I went to use it and the battery was dead. Wouldn't charge up. I call customer support and they tell me it's out of warranty, and that they don't make that battery anymore. Although I could buy a replacement battery at Lowes... for $265.
The trimmer / battery came together for $150. A brand new trimmer / battery was around the same price.
Meanwhile I've had a Ryobi mower, trimmer, and leaf blower and 3 batteries that are absolutely _packed_ with juice and I can take care of my entire yard on less than 2.5 batteries.
Never buying DeWalt battery-powered lawn equipment again. My experience is that they make solid cordless hand tools but I don't trust the brand anymore after they abandoned this whole battery pack after 2 measly years. The kicker was that when I gave the serial number for the battery, the lady at customer support told me it was manufactured in 2016. So I bought a brand new trimmer that came with a year and a half old battery.
Great.
You may be able to take the battery apart and replace the cells. I don't have one of these machines, but I have seen videos of people doing that for various battery-powered tools.
It would be awesome to pry open the siliconed brainbox and probe directly on the drive transistors to verify brake mode enganging. The suppression of the back emf looks very odd to me. A good dead short should not let +-3v on the coils...
Not trying to be a DeWalt fan boy but I really like my weed eater regardless of how inefficient the motor is it cuts grass pretty well and is easily handled by my wife and kids. I don't like the spool mechanism very much though. I have a half acre in one battery will do the whole thing pretty easily
It does work fine. Though I think the battery would last longer if the electronics were smarter.
It's because DC motors are more efficient at high voltages/currents as the inefficiencies are static and/or diminishing. So tl;dr if running 10W through it gives 1 unit of power out, 100W would give not 10 but 12. So they use this method to get the motor working in it's sweet spot and then it's supposed to idle, which you're probably seeing as back EMF and thinking is lower power input. Great video though. Too bad the loop wasnt very sensitive.
Test it under load. I would like to see the diagramm then.
For safety reasons, that trimmer needs to stop fast. The PWM is just on-off control: drive full speed or brake. It would be better if they braked only when the trigger is released but that would probably require a micro controller and software instead of a cheap driver IC. DeWalt probably has no idea what is going on, they just picked a Chinese factory to make it and drove a hard bargain, giving the factory little financial room for improvement.
Ma reasonable explanation I read here so far.
You should have it connected while wheat whacking. With the amount of load wheat whacking generates it should be fine. Or at least with a wire so it has some resistance going through air.
Wheat wacking..... bwhahahahah
@@JimNichols You are laughing at me for my mistake while you making one yourself. Bravo! It's WEED wHacking and I'm sorry for my terrible mistake.
*waiting for DeWalt feedback*
you notice there is a craftsman version of that weed trimmer? i have a feeling that stanley black and decker are using something they aquired from the craftsman buy out
www.lowes.com/pd/CRAFTSMAN-V20-20-volt-Max-13-in-Straight-Cordless-String-Trimmer-with-Edger-Capability-Battery-Included/1000730288
Very good video, the problem is called chineeze engineering. Make it as cheap as possible, sell it to the public as a brushless DC motor which most would think its going to last. What will go before the hardware is the battery, that is THEE most EXPENSIVE item on this device. The same goes for their Brushless DC Motor (BDCM) Drills.
When the motor is braked I wonder if they store the energy generated somehow in the caps to use in the next power cycle? We used to do something similar driving solenoids at high frequency to reduce power consumption.
shorting it doesn't produce recoverable energy
It might just be because it has to be able to push harder under load than at idle, so when it encounters resistance it can push hard without going too fast. I assume the limit is there because they tried it without it and something broke... and instead of redesigning it, they just "fixed" it.
yes, but you could just drive it less of the time and not brake in between, and still save power. And a weed trimmer is mostly not running at full load when in use.
It‘s just a 3 Phase Motor like we have it in germany, they just take the DC Input signal and convert it to a 3 Phase AC Output with transistors, this thing is called an ESC (Electronic Speed Controller).
I dont have a large lot so I got a small black n decker lithium model, best weedeater I have ever used (cause I can just do it and be done instead of mixing gas and fussing around with those tiny devil 2 strokes)
I agree. I still don't mind fiddling with the cord. An electric mower is easy once you get used to minding the cord.
Kinda like some OLED technology, whereby to show a dim picture, it just ups the refresh rate of the pixels so much so that the brightness is never at max.
Excellent video... I shall be watching some more of these.
2:00 It doesn't put one magnetic pole on one set of coils and second pole on the other two sets of coils, at least not fully like that. If you look at a signal from all 3 sets of coils it should be 3 sinusoidal lines, each moved 120 degrees from the others( upload.wikimedia.org/wikipedia/commons/c/cc/3_phase_AC_waveform.svg ), so you should have said "when one set has full north, at the same time the second set is in transition to full south and the third one is in transition from the full south". You said it like at the same moment one set has full north and the other two sets have full south. I dont mean to criticise too much, I just wanted to make the information more precise for people that may be learning for the first time about 3 phase from this video. I love your videos, Matthias :)
True. Brushless motor drivers are more complex than a single measurement point can reveal, and it's also more about current than voltage.
The drive voltage for the motor would not be a smooth sinusoid (assuming you use a scope with decent bandwidth). That would imply linear control and very high power dissipation in the MOSFET's or IGBT's. The drive would be more like a class D audio amplifier. That is, PWM. Also, sine-wave drive is not universal for BLDC's. Often they are driven with six-step commuation. You can look that up on your own time (six step commutation of bldc).
Also they could've put the driving circuit next to the motor, resulting in much less interference with other electronics
As with batteries, it might be best to test under load: Without load, batteries exhibit near-to-nominal voltage, even if they are practically empty. With a 50 Ohm resistor as load, things turn out a little differently. I suppose the same goes with electronically controlled motors: No serious load, no usable measurements.
Maybe they are trying to account for a variety of operating styles. I had a neighbor who used to pulse the trigger on weed eater twice a second while trimming and edging. It was very annoying to listen to. He can't be the only idiot who runs them this way, but maybe this control scheme allows for that type of nonsensical operation.
Keith Stephens maybe he’s trying to save battery lol
To really understand what that weed trimmer is doing is to take the electronics apart compare it to the electronics of other brushless tools. I would say bread board a circuit using an arduino as the PWM and drive it through some FETs and see if the motor runs the same. I see that various BDCM run differently as to how they are driven by different circuits and the frequency you subject them to. The quickest way to drive 3 phase DC Motor is to use an RC ECS module. That uses PWM.
Maybe it's because the H-bridge would clamp the induced voltage to supply, causing even more drag if they don't short the winding?
I would also not have expected this, but probably best not to call a man stupid until you have walked a day in his shoes, likely there are good reasons. Interesting that Dewalt is again betting on a geared system while the Makita trimmer uses direct drive, just like the chainsaw.
It's inefficient, but it's cheap. I think they're doing it this way, because now they can use a far slower and simpler controller, with some cheap dedicated chips to drive the motor
That was my thought, DW is considering total cost of BOM, not ownership. They'er going to make 10's of K's of these and want to keep production costs low, so what if the owner has to recharge more often.
@@SlaveToMyStomach And if the owner has to recharge more often, the battery wears out faster, so they have to get a new one, and hopefully by that time, "Oh, look, we don't make that battery anymore, I guess you have to get a whole new tool!"
Is it autofeed? Maybe it continually lets out tiny bits of line as it bumps along.
Actually you can drive the coils sinusoidal and minimize the cogging. The sine waves can be synthesized with modern power electronics. Of course that motor may have a simplier drive.
well, yes, that was my point.
Well.. unless it need to be smooth motion or silent, don't expect sinusoidal drive. It takes a bit of extra stuff to realise.
Except for noise there isn't much of an advantage to drive the coils with sinusoidal voltage.
@@superdau less torque ripple, smoother current waveform, and whatever you lose in top speed you can potentially regain if you do phase advance (field weakening).
Maybe it's a feature that helps cutting of the nylon wire in some way?
avoid the wire tangling or something... just a thought.
If it's overpowered before the suppression, would that be for when it's cutting thicker brush to maintain power through the cut?
I would imagine it doesn't do that under heavy load, but that would be hard to test