If the new servo has the same PCB inside you might be able to exploit the ST-Link firmware-dump bug to extract the firmware from the new servo and flash it to a fresh microcontroller for the old servo. That's assuming the microcontroller's read protection is actually enabled; if not it's even easier to dump the firmware.
How can one typically extract firmware from an existing STM32? What kind of interface or dongle is needed to connect the the chip's pins to the computer?
@@pitot1988 I haven't specifically looked into this board but one thing I'd suggest researching is search for UART. It seems to be a serial connection and programming standard interface. The other you could try was searching for JTAG.
@@pitot1988 you could likely do it right through the programming header on the board, if there's no protection enabled, dumping the ROM is pretty straightforward, I'm sure someone has a video about the process here somewhere.
Do you have a lot of electronics background? As an EE of many years, I think you did a great job of explaining that whole process and identifying the operation of the various circuitry. Awesome job! And I think you made the right move buying the same motor again, at that price point it would be silly to re-engineer it.
"buy cheap, buy twice" ... in this case let's wait for part three ... Jokes beside, nobody can tell if the expensive branded option would also have failed. Because of the circumstances that lead to this failure ... like over-voltage, shocks or the usual "produced on a Monday-product" ( hehe ). But all this is speculation and only one kind of person can exactly know what is the right thing to do: UA-cam commentators ... :P But what you CAN say is that the brand-manufacturer provides guarantees. F.E. exchange in-between 2 or 5 years, or lifetime; Customer support and possibly help with the design of your contraption if some hurdles occur. You just have to RTFM!:) I don't see any reason to hold court about Clough42's decision to go for the cheaper exchange part. One way may be as good as the other. But one thing he said in the video made me think: that it is practical now, that he can just exchange the motor and has nothing new to adjust. Branded and certified products (this is why we in the industry use them ... and also to not land behind bars ... **g** ) guarantee a time-span when they are produced and can be supplied[1]. This is also guaranteed for spare parts and software for such systems, which are often modular and extensible. My opinion to the question if that was a good choice: It depends ... and you do you!:) [1] Edit and Example: For example and out of the field, the not so cheap motor and motor driver parts including all connection systems and cabinet parts from Siemens come with extensive personal advice and guarantees on availability and service. Some parts will be in stock for 20 years, some are even guaranteed to be in stock and with maintenance support for 50 years. You will certainly look in vain for this at the "Alibaba" dealer.
Mtbf is 2 years. I can tell you from experience that the clear path would have lasted significantly longer as long as it was sized and installed correctly. Just my opinion, but decisions like that add up, and you can find yourself “going broke saving money” quickly. I watch people do it every stinking day.
@@VoteForBukele normally i would agree...but its obvious he does know a bit in electronics and is likely to solve many possible cenarios... guys like this are very valuble for any company...
I think you did the right thing, hopefully this servo will work just fine. I'm saying it because I recently finished building your ELS for my lathe and I used the same servo. Hope you are reading this comment, I cannot thank you enough for your brilliant design and fantastic videos that guided me through the entire build process.
Hi James - I’m very interested in knowing more about the dental glasses. Source? Cost? Would you recommend them? Pros? Cons? Optical clarity? FWIW, please do a video on the glasses. There are a bunch of us old guys out here, and not much info in YT machinist land for vision aids. OptoVisors work only up to a point.
+1. Looked for this comment to see if there was anyone else who is also old and blind. I tried one cheap rando dental loupe from Amazon and it was useless rubbish. I know some must be better than others, but I aborted my search.
James I’d say it was a good idea to go with an even exchange. Now with that said, if this one craps out in a year or two, it might be worth stepping up (pun intended! 😂) to a more reliable long term solution. Thanks for the video. They are always a pleasure to watch.
There are usually some internal regulators inside the microcontroller, outputting internal core voltages onto lowish value ceramic bypass caps near the micro. Those multilayer ceramics are some of the most common things to fail - they can crack when the board flexes or vibrates. If one of those caps is shorted, I'd expect the regulator in the micro to get hot. You might be able to identify the culprit with a multimeter, or by just pulling the bypass caps off one by one.
That sounds like a reasonable possibility. A servo is certainly something that can have some vibration, especially mounted on a large vibrating machine. I don't know how much the boards would flex, they looked reasonably well mounted. But it is certainly a possibility. The interesting question is whether the microcontroller would have survived after the internal heating.
I decided on mine to go with a 36 volt power supply because the technical book that came with the servo says to leave room for back feed EMF during deceleration. I wonder if your pushing it past what its little buck convertor can handle over time. Just thinking out loud.
You might be onto something...wonder what voltage the buck converter is supplying to VCC at the micro when you add reverse EMF to the 42 volts on his input? There is a 3 terminal regulator before the micro, but the manual does specifically mention leaving headroom for back EMF
I had that same thought. The motor may be rated for 48V but that drive circuitry is rated for much less. It looks like they depended on the body diode on the FETs to prevent back EMF instead of a robust design that uses an external diode.
@@PNWPrototyping Buck converter (LM5007MM) is rated up to 75v. Back-efm from the deceleration will simply charge the cap on the 42V rail to higher voltage, no matter what diode will be used. Body diode is enough in bldc motor drivers, you can reuse mosfets shown on the pcb as a ideal diodes.
Put a 3.6 zener on 3.3v rail and it might save this servo. Often over voltage on VCC kills processor. For example turning motor by hand making generator might do it.
I watched this with keen interest (and was "rooting for you" through the diagnosis, hoping you'd find a simple fix) because I have the exact same servo driving your ELS on my MX-210V. Mine has been working perfectly well for about a year now; hoping for many more. Still, for $99, it's a terrific bargain -- unless these failures become a pattern. Thanks for the excellent walk through of the circuitry; very informative and entertaining.
I built the ELS about 1 year ago and love it , no more gear changing , I was wondering if you had any plans to add any features to the firmware, such as automatic stop and start and return to the tap cycle Just a thought Thanks for the great projects looking forward to the future
Leadshine was the company that ripped off the stepper drive design from Gecko years ago. They sold them so cheap that Gecko developed an improved design that used programmable chips to slow down the Pirates. I wouldn't be surprised if your servo drive was also stolen. That's the world we live in and the Chinese have different ethics than we're used to. You may try and contact them to see if you can buy the replacement board without the motor to save a few pennies on a backup motor.
A gremlin let the magic smoke out of that servo. I can honestly say that I have done the same, bought the same thing that betrayed me and sometimes it betrayed me again. Thanks for the video keep on keeping on.
If it's not broken, don't fix it. If this one drops you can upgrade. With the way tech goes it'll probably be smaller and have more power by the time that one farts and let's the smoke out.
100$ for 24 Month of work. Comes down to about 0.14$ per day. I could certainly live with this, too. And maybe this STM32 will last a little longer. We'll know in 2 years time.
It's rated voltage is 36 V. You can see at 2:36 that you're operating the device outside of it's rated operating range. This is most probably why the device failed. An avalanche breakdown of the buck regulator would lead to avalanche breakdown on the linear regulator which would expose the STM32 to the full supply voltage, causing internal breakdown and latchup. This would explain why the microcontroller is drawing high currents at roughly two diode drops. You should get a different power supply if you want to avoid this $99 mistake from repeating.
HolUP HOLUP the thermal pad is connected to the insutator card and connected to the back cover with thermal past and the back cover looks to be made of PLASTIC? I think that maybe the problem. cut out a 40/40mm hole in the cover and add a ALLOY sink to it directly to the thermal pad and it maybe a 40mm fan as well. Those mosfets must be cooking ?
Purely anecdotal at this point but I swear there's a design flaw in those STM chips. I repair arcade machines, and one manufacturer *loves* those things, there's one on practically every single board... and every time a board fails in any of the games from them, it's *that.* EVERY time.
Hi James, I thought it was a good idea to replace the servo with the same type as you will have an opportunity to extract the firmware from STM32 on the new servo. Could be the making of another great video for the future. Never a bad idea to replace a part with a known good one. I know lots of mechanics that do that very thing each and every day. Thanks again for all your efforts in producing great videos.
All my years of buying inexpensive chinese whatever that fails all too often, it's cheaper in the long run to buy the more expensive option. You don't know the quality of the parts going in to those boards , as they might be factory seconds or thirds which will fail over time. Buying the same part again would be the faster - easier option but if the servo fails again then not the cheaper option over time. The servo did have voltage limits and maybe for a reason.
If the new servo's microcontroller doesn't have an internal security fuse (or if it does and the fuse hasn't been burned) you could extract the firmware and possibly repair your dud. You just have to decide whether it's worth going down that rabbit hole for a $100 item. At those prices a one for one swap seems like a wise move. If you find yourself faced with having decide whether to buy a third one at some point, then go for a higher quality one. My $0.02.
I have 3 of "those" servos but made by jmc (IHSV57) The PCb looks nearly identical. only thing is that the LDO is on the bottom and the Buck is on Top. The Failure stayed the same. There seems to be some sort of problem with this kind of Servos.
Time is money bro. Also no need to use other motor if this one works good. Accidents can happen with any motor. You can shift to other motor if this happens more frequently.
You could read the flash to extract the binary file from the new motor's STM32and then replace the uC on the old motor, to then program the new uC.Will cost you a second servo anyway, but you'll have a replacement, at least.
STM32C031C6T6 actually possible to buy one of these, ST Micro's parts have been out of stock for ages due to supply chain and stupid management issues.
So how do you find those glasses, I have been looking at the same thing that is popping up on my FB all the time. Im in New Zealand so by the time I get it here its not as cheap as I suspect it would be in the US so Im a little hesitant having been bitten by some previous adjustable glass's I purchased that were useless. I do like the look of these and I assume I can fit them onto my prescription glass's and they dont fall off etc. Any suggestions/recomendations
Get what you pay for I guess, but if you can replace it 4 times for the price of a "premium" device... I can't judge. If it occurs again - I'd be inclined to lower that supply voltage some more, to under 40V. You're current 20% over the rated voltage, so there might be some undue stress there, who knows - may even be a firmware bug that allowed some I/Os to trigger and cause this failure that you're not going to see with the new hardware.
One failure does not establish a pattern. If the second suffers a similar failure, it will call for an upgrade. BTW, having lived with your solid tool post base for a while now, do you like it? Is it worth the expense, effort and loss of versatility for a general purpose hobby lathe? I know Stefan and Robin like theirs, but they kinda work on a different level from most of us. Me, at least.
FYI, some Leadshine resellers or volume users can obtain new spare parts for the motors including the PCBs. I have replaced some of their encoder/hall sensors before now on their regular servos.
This was the comment I was looking for. Contact Leadshine and send a link to your video. Tell them you will do a video of the repair when they sell you the spare PCB or Microcontroller.
These STM32 are extremely sensitive to EOS and EDD stress. If for the smallest reason your voltage supply went too high, it could burn up the MCU. Typically I give STM32 good isolation from peripherals and power supply for this reason.
^ This right here. I've also had designs where the input protections were marginal, and had the uC fail in the same manner as what happened here in the video. Hot chip, low voltage on the VCC line, presumably due to an internal short. They seem to be a little more sensitive than some other manufacturer's chips I've worked with.
You were correct to replace with the same, stm32 blowing seems a random failure, unless there is back emf appearing, very unlikely with what looks like optical isolators before the power electronics… would have been interesting just to break the power supply to the microcontroller and see if the psu and regulator recovered to 3.3v more out of interest then anything else.
The only issue I see here is you didn't say you bought 2. Put one in and keep one as a backup. For low cost mission critical parts like this I always try to keep at least 1 extra on hand so I can get back up and running ASAP then order the next spare with a little bit less pressure to have it right now.
Genius! Thank you for posting this video. I’m configuring these servos with a masso. Got the db9-usb cable from A..zon and finally have a connection. I’ll try 500 steps too as a starting point. I greatly appreciate your channel and hopefully your getting rewarded for your hard work!
Hi James, useful walk through to show how to diagnose the problem. Thank you for spending your hard earned money to let us know if it is a generic fault! I've still got your wonderful ELS on a breadboard ...... one day, one day.
Is it possible to cut rifling using the electronic lead screw? I collect and shoot antique firearms. Every now and then i have to put a barrel liner in a gun. Colts originally had Progressive rifling. so it would start off straight and then progressively add more and more twist. I would love to be able to replicate this rifling. How difficult do you think this would be using electronic lead screw?
Hi James, i would have done the same as you and bought the cheaper servo. You were just unlucky as STM micro processors are used in cheap as well as expensive equipment and could have failed in either. The other point i have is the cost versus benefit of only $1.00 per week based on $100 for 2 years use, i think that is a win in my book. Cheers
I mean there are definitely poor design choices that can be made to increase the likelihood of killing a microcontroller, but otherwise, I agree that as long as it isn't a recurring fault it's worth getting the same motor.
I too have this same servo and all the components to build an ELS, I simply have not had the time to put it all together yet. I hope this new motor lasts longer than a couple years as well. Have you considered adding stops to the ELS to make threading to a shoulder easier?
Spend hours reconfiguing a new servo or $99.00 for the same one that fits. I'd do the same thing too, your time is money too. Now if it fails again then all bets are off as to if I'd take the time to reconfigure... I'd say you chose wisely. Oh, you forgot one tool in your list of tools used, like you my eyesight isn't what it used to be and I'd like to try the magnifier you used. Thanks for another great video James!
James. Thank you for this walk-through. Your explanations may be the only ones I can watch start to finish and truly understand the information. Have a great week sir.
Hallo The AMS11173.3 has Internally limited by 15V See datasheet, the Board support 24-36V DC input. This means that the voltage is regulated down beforehand
RIP STM32 ;-/ Perhaps guys in China are nice enough to provide you with new STM32 or firmware, but I highly doubt that. That's why I stick with 1980s and 1990s machinery, everything is in external EPROM ;-D
I'd suggest that running it 2V under its advertised absolute max supply voltage, 12V above nominal max, doesn't leave a lot of room for transients, e.g. from back-EMF. I'm going to run mine on 36V to start, and if that's not enough, start by turning down the voltage on a 48V supply to something less, likely 45V or so, to give more headroom. The datasheet says "24-36VDC recommended. Please leave reasonable reservation for voltage fluctuation and back-EMF during deceleration."
Look into the 42AIM30 or 57AIM30 digital servo. Seems to be better quality. Notably more stalled rotor torque, too. Needs the be reprogrammed before first use, factory is 32768 steps/rev
Please consider turning off the MSX feature on your flir camera if you don't want to take the time to adjust the parallax for every shot. It honestly makes it harder to understand the thermal content shown in the image. Yes, I know it looks really blurry with MSX off. But ironically, it's clearer that way in every sense that practically matters. I have the same FLIR camera as you, and I just leave the MSX off all the time. It's a harmful gimmick.
You just popped up in my UA-cam feed. Great work. Great subject mastery. Great video documentation. Great honest, human communication! (Thanks for the Dentistry Lenses tip for "old man glasses!) Masterful - novices or skilled pros can't help but learn from this video, sir! Great common problem of "buy replacement" or go through work of warranty, repair, research new source/parts - across all walks of life and lines of work. Of course your time and expertise are so obviously valuable, that the time saved in re-buying is wisest choice. Wow.! Thanks again!
I bought a 100$ pump, it failed after three years, i replaced it with the exact same unit, it failed after 1 year, i bought the EXACT same pump again, going on 2 years. It's competitor, the one "recommended", was $1300.00. Seams pretty easy to me. The time you saved with the copy, may be enough to rationalize the duplication. Imagine having to get the new programming cable, a new programming App, and then rebuilding a new mounting plate. Time, Money and sanity, check!
Maybe a more expensive motor in the first instance would have lasted longer - but as it was the u-controller that had given up, it probably wouldn't have made a difference. uC's are reliable if they are running in a benign environment. I find it really surprising that the microcontroller would fail unless something was pulling too much current on one of its legs. (they don't enjoy a leg-pull!) I would do a bit more research on the dead one (on the killed stepper killer killer) to see if something attached to an I/O pin on the STM32 is shorted or has too low a resistance. As a totally irrelevant aside, I find it quite surprising that such a trivial application uses an STM32. I guess they're quite cheap, but it seems like a sledge hammer to crack a nut. Could you buy a replacement controller board for it? Somebody suggested copying the code from the new one you have and flashing another processor, but getting the old one off the board and soldering a new one on, and getting it to work without damaging the solder-tracks on the board could be an adventure. I've never attempted it myself, but I've watched a lot of UA-cam vids on how it's done, and I'd put my chances of succeeding a job like that at perhaps 30%. I'm a retired electronics and computer engineer, but zillion-pin chips weren't a thing when I retired. Leastways, not on the boards I worked on.
And now that you have a working micro controller, why don't you "just"* dump the firmware and fix the other one too? *Extremely hand wavey over simplification 2000 O'matic - but should be doable
Construction and build quality looks decent. Shame it failed that badly. I mean, complete servo solution for $100. There's nothing wrong going for it; maybe it's just a one-of case, after all.
The maximum of the servo is 50V ... of course but when you stop the motor it will induce power to the power supply which itself might get damaged (recuperation). Don't run those motors with more than 36V PSUs. I think about decoupling mine and inserting an extra wire for the logic power supply.
This servo requires a damper. You can do a test - connect an oscilloscope to the power bus of the servo driver and watch the voltage spikes when changing direction or high accelerations. I connected a clone of this servo - ihsv57-30 through a 3:1 gearbox to the longitudinal feed of the lathe. I got a problem - the Fujitsu power supply went into error. The reason is the EMF that occurs when stopping at 3000 rpm is more than 10V. It was added to the supply voltage. As a result, either the servo drive went into error or the power supply. I put together a dumper circuit from the GeckoDrive website - it was a good solution.
I also had a problem with one of my Chinese stepper motors. For me it was a mechanical and I was able to fix it. Not before I purchased new one, but it's always good idea to have a spare if they are used in important place. It's funny how they appeared to be relatively good quality but there is still some details which fail. Mine stopped spinning but coils were powered and there was some sound when I drove it, though controller went immediately into error state. I suspected short circuit, but all coil resistances were OK. When I opened it up, I didn't find anything wrong but when I put it back together I realized what the problem was. Face and backplate were machined in a way they didn't lock into place. Alignment was maintained only by friction and it was big nema34 motor. It had been slipped and prevented motor from spinning. I redesigned my mount so the motor body is supported by rubber, and it has worked well since that.
When dealing with electronically populated PCBs by using a pencil around the circuit (CCt) you can place carbon particles on the PCB which can cause short CCts in the most extreme case
That seems like the typical failure mode of an STM32 MCU when subject to transients 5V, indicating a probable weakness in the circuit design. It would be interesting to get a trace of each of the pins (on a working servo) during normal operation to find the culprit.
There is a rising "Right To Repair" movement. You just ended up making a lot of waste because you couldn't replace a (my guess) less than $10 micro-controller. There has to be a better way.
9:56 - Isn't a buck converter meant to step UP voltage? A step down converter would use PCM and filter caps to lower voltage using a variable duty cycle (and current sense to maintain voltage). The INDUCTOR looks to be a buck converter, but must step up the voltage to a higher level (as when you switch an inductor, as the magnetic field collapses you get a spike of higher voltage, and then you use smoothing caps to give a consistent voltage level).
James not wishing to question your choice of motor ... but other ELS builders have quoted using the 3Nm version of the Nema23. Have you ever experienced issues with insufficient power when turning/thread cuttting tougher materials?
100 bucks to save a day and make 300 is what you call 1 step back to gain 2 steps forward. The opposite, would be investing 3 hours to save 1 in the future, or what you call Government Bureaucracy.
There is a way to pull firmware from those little ROM chips. Get an RT809F chip programmer for about 60-80 bucks and it comes with software that allows you to read and write chip software.
I am using this motor for my lathe that is a 12 x 24 with a 3mm pitch metric leadscrew i am thinking of using a 90 deg. angle gear reducer with a 5:1 or a 10:1 reduction. There will be backlash in the reducer but I think it is not a problem since exact positioning is not required because when threading the backlash is taken up before the cut begins. Has anyone done this or fore see a problem.
I think at that price as long as the failure rate of the Chinese Servo is less than 3 times as high as the more expensive option you are a making a smart choice. Especially when using multiples of the same one. Ie. if you have 4 machines all using the same servo even if the cost of replacing the servos every few years comes out to exactly the same price as the life of a more expensive option, it would make sense to keep a couple spares on hand which would save you money on downtime while waiting for shipping on a replacement. In the 4 machine example: If an expensive option lasts 6 years and costs $300 You would need to invest 4x300 + 300 to have a spare ready. Or you can buy 12x$100 + 100 and have a spare ready. Giving you the same redundancy at lower overhead or even better redundancy if you want to buy multiple spares. Since you don't need to buy all replacement sets at once the cost advantage becomes even greater 4x100+100 and 400 every 2 years for maintenance. This gives you a capital advantage, assuming even 5% return on your money, you would save $120 in 4 years + $10 every year on the backup motor. Although I think this was a freak accident because it seems to use quality parts and STM is a reputable company which is unlikely to have the same failure happen that often.
Servo to the power supply " Hay you know all that power your are sending me, don't need it any more got to send it back" Power Supply to the servo: "Sorry all sales are final, I have no facilities for returns, you're just going to have to let out some blue smoke, sorry" Like a couple of others my bet is on back emf from deceleration, raised the rail voltage above what the primary switching regulator could manage. When those simple buck switching regulators fail with the pass transistor on you get full input on the poor 3.3V linear. What you need is the equivalent of the braking resistor used on a VFD. When main DC bus rises above the nominal value they turn on a FET that connects the main power bus to an external power resistor to dissipate the energy. If the motor is running at fairly high power, there are considerable magnetic energy stored in the fields of the windings. If you suddenly mechanically stop the motor that energy needs to be dissipated someplace. There is likely some open source servo code you could port to a new STM32. and get it working again. Years back the was the HUH servo dirver board on CNC zone.
I'm guessing that just buying a new circuit board was not an option? If you ever do find one you'll have a back up motor should you ever need it. P.S... your explanations are so good I almost understood them. This is from a total ignoramus about electronics, so: "Good Job".
Welp, buying the same motor is a reasonable move. It lived long enough to warrant its price. But... I'm wondering... Does STM32 contain any anti-read measures? Because if not, then how about dumping the firmware from the new motor (if it is the same board, same revision), and then buying a new microcontroller, flashing it and replacing the burnt one? Just so next time You land in such a spot, You can just replace the motor with a "refurbished spare", and then have all the time in the world to repair the broken one, while still being able to use the lathe in that time.
Would be cool, if it was worth it, to send the servo to Northridge FIX to check out what really caused the problem. Besides that, I think going the cheaper route is the smarter way of handling things. The metric shows that you get a lifespan of two years from that servo, let's see what happens in two years with the new one.
Can you pull the program from the new chip, then program a new one? Or is it proprietary/encrypted programming? If you could replace the chip. Then shelf the new one. (To use if it breaks again.)
In terms of failure of the STM microcontroller and V-reg... the root cause might be poor or broken ceramic capacitors on the V-reg, or a V-reg failure in general. My expectation would be that the voltage to the microcontroller exceeded the parts maximum, and the microcontroller shorted out internally. A microcontroller running for 2 years usually doesn't just die, but capacitors might. Voltage regulators require a minimum capacitance for stable operation. The most common place to save money on cheap electronics is to remove capacitance or put bad quality capacitance in there. In terms of preventative actions, you could look at the V-reg circuit and add a few microfarads of quality capacitance to the input and output of the voltage regulator. This could be achieved by stacking SMT ceramic capacitors, or soldering on a radial 10uF electrolytic cap across the output capacitors, being aware of polarity and voltage rating of the caps. Other areas of failure could be ESD static discharge onto unprotected microprocessor pins... this is harder to fix, but during installation and programming you should make sure you are grounded, just like when working inside a PC case. Once everything is connected it should be relatively hard to zap things. Unless the failure corresponded with unplugging cables, I'd suspect bad capacitors over ESD.
Ok, I have seen both now. The coil windings quality is different. The electronics are different. The bearings are way different, and ClearPath comes with 3-year warranty, the price is only two to three times based on your configuration. If you are looking for a reliable uninterrupted business processes for at least till the warranty expires? Go for local companies with good manufacturer warranty. Not to mention that these knock offs look exactly like the ones from other companies in the west. ;-)
so who was the killer- killer- killer ??? Until the jury isn't out anymore you're a genius idiot, me thinks? :D I would definitely do the same and buy a new cheap thing that works. It kinda shoulda oughta work since it actually did until it didn't. And what's the alternative? Get an expensive time consuming thing that also may or may not work some day. Or cut the costs a little and buy two. Yeah, you did the right thing!
Buying the same servo DOES save time and money for the "home hobbyist". In industry, they would have re-engineered the whole thing, taking lots of time and really increasing the cost. I would try and find what might have caused the failure, but with the chinese stuff... that could be anything.
If the new servo has the same PCB inside you might be able to exploit the ST-Link firmware-dump bug to extract the firmware from the new servo and flash it to a fresh microcontroller for the old servo. That's assuming the microcontroller's read protection is actually enabled; if not it's even easier to dump the firmware.
Had the same idea... I definitely would try it.
I was curious if something like this was even possible considering he had a "copy" in the new version.
Thanks
How can one typically extract firmware from an existing STM32? What kind of interface or dongle is needed to connect the the chip's pins to the computer?
@@pitot1988 I haven't specifically looked into this board but one thing I'd suggest researching is search for UART. It seems to be a serial connection and programming standard interface. The other you could try was searching for JTAG.
@@pitot1988 you could likely do it right through the programming header on the board, if there's no protection enabled, dumping the ROM is pretty straightforward, I'm sure someone has a video about the process here somewhere.
Do you have a lot of electronics background? As an EE of many years, I think you did a great job of explaining that whole process and identifying the operation of the various circuitry. Awesome job!
And I think you made the right move buying the same motor again, at that price point it would be silly to re-engineer it.
I agree with Stu completely and in both respects.
...and another point is spare part if somethin else fails in it...
unless its the same chip ofc.
"buy cheap, buy twice" ... in this case let's wait for part three ...
Jokes beside, nobody can tell if the expensive branded option would also have failed. Because of the circumstances that lead to this failure ... like over-voltage, shocks or the usual "produced on a Monday-product" ( hehe ). But all this is speculation and only one kind of person can exactly know what is the right thing to do: UA-cam commentators ... :P
But what you CAN say is that the brand-manufacturer provides guarantees. F.E. exchange in-between 2 or 5 years, or lifetime; Customer support and possibly help with the design of your contraption if some hurdles occur. You just have to RTFM!:) I don't see any reason to hold court about Clough42's decision to go for the cheaper exchange part. One way may be as good as the other. But one thing he said in the video made me think: that it is practical now, that he can just exchange the motor and has nothing new to adjust. Branded and certified products (this is why we in the industry use them ... and also to not land behind bars ... **g** ) guarantee a time-span when they are produced and can be supplied[1]. This is also guaranteed for spare parts and software for such systems, which are often modular and extensible. My opinion to the question if that was a good choice: It depends ... and you do you!:)
[1] Edit and Example: For example and out of the field, the not so cheap motor and motor driver parts including all connection systems and cabinet parts from Siemens come with extensive personal advice and guarantees on availability and service. Some parts will be in stock for 20 years, some are even guaranteed to be in stock and with maintenance support for 50 years. You will certainly look in vain for this at the "Alibaba" dealer.
Mtbf is 2 years. I can tell you from experience that the clear path would have lasted significantly longer as long as it was sized and installed correctly. Just my opinion, but decisions like that add up, and you can find yourself “going broke saving money” quickly. I watch people do it every stinking day.
@@VoteForBukele normally i would agree...but its obvious he does know a bit in electronics and is likely to solve many possible cenarios... guys like this are very valuble for any company...
I think you did the right thing, hopefully this servo will work just fine. I'm saying it because I recently finished building your ELS for my lathe and I used the same servo. Hope you are reading this comment, I cannot thank you enough for your brilliant design and fantastic videos that guided me through the entire build process.
Hi James - I’m very interested in knowing more about the dental glasses. Source? Cost? Would you recommend them? Pros? Cons? Optical clarity?
FWIW, please do a video on the glasses. There are a bunch of us old guys out here, and not much info in YT machinist land for vision aids. OptoVisors work only up to a point.
I'm not old, i'm only 40, but i need all the vision help i can get as well!
+1. Looked for this comment to see if there was anyone else who is also old and blind. I tried one cheap rando dental loupe from Amazon and it was useless rubbish. I know some must be better than others, but I aborted my search.
I'd like to see a link for the dental loop glasses too.
I’m also interested in these. Can you provide a link for them?
+1. I d like to know about them , too. Real dental magnifiers are screamingly expesive though. I m hoping this has a more reasonable price tag.
James I’d say it was a good idea to go with an even exchange. Now with that said, if this one craps out in a year or two, it might be worth stepping up (pun intended! 😂) to a more reliable long term solution. Thanks for the video. They are always a pleasure to watch.
These servos are typically great but i would’ve thought a 300w model would’ve been more appropriate for his application.
There are usually some internal regulators inside the microcontroller, outputting internal core voltages onto lowish value ceramic bypass caps near the micro. Those multilayer ceramics are some of the most common things to fail - they can crack when the board flexes or vibrates. If one of those caps is shorted, I'd expect the regulator in the micro to get hot. You might be able to identify the culprit with a multimeter, or by just pulling the bypass caps off one by one.
That sounds like a reasonable possibility. A servo is certainly something that can have some vibration, especially mounted on a large vibrating machine. I don't know how much the boards would flex, they looked reasonably well mounted. But it is certainly a possibility. The interesting question is whether the microcontroller would have survived after the internal heating.
An ohmmeter with Kelvin connections might get you into the ballpark.
The hard part with those ceramics is that if they're broken, it's basically impossible to know the value without the schematic.
@@Teklectic Could remove two, the bad one and a good one, then measure the good one with a capacitance meter.
Thanks for taking the time to share this. Are you considering a Clough42 redesign of the board that failed?
Getting the same servo isn't lazy... it's strategy! :D
I decided on mine to go with a 36 volt power supply because the technical book that came with the servo says to leave room for back feed EMF during deceleration. I wonder if your pushing it past what its little buck convertor can handle over time. Just thinking out loud.
You might be onto something...wonder what voltage the buck converter is supplying to VCC at the micro when you add reverse EMF to the 42 volts on his input? There is a 3 terminal regulator before the micro, but the manual does specifically mention leaving headroom for back EMF
I was thinking that! I didn't see a whole lot of input protection on the buck converter, they step down the voltage but they don't isolate anything
I had that same thought. The motor may be rated for 48V but that drive circuitry is rated for much less. It looks like they depended on the body diode on the FETs to prevent back EMF instead of a robust design that uses an external diode.
@@PNWPrototyping Buck converter (LM5007MM) is rated up to 75v. Back-efm from the deceleration will simply charge the cap on the 42V rail to higher voltage, no matter what diode will be used. Body diode is enough in bldc motor drivers, you can reuse mosfets shown on the pcb as a ideal diodes.
@@kermitfrog1897 Voltage after buck converter don't change. Should be approx 8v
Put a 3.6 zener on 3.3v rail and it might save this servo. Often over voltage on VCC kills processor. For example turning motor by hand making generator might do it.
i think it is designed for that purpose ... ergo it will have its own protection
Hi, would please list the 'old-man-glasses' source? They sound like what I need! Thanks.
I watched this with keen interest (and was "rooting for you" through the diagnosis, hoping you'd find a simple fix) because I have the exact same servo driving your ELS on my MX-210V. Mine has been working perfectly well for about a year now; hoping for many more. Still, for $99, it's a terrific bargain -- unless these failures become a pattern. Thanks for the excellent walk through of the circuitry; very informative and entertaining.
A wise man once told me that Chinese tools and equipment are not purchased,
They are rented.....
I built the ELS about 1 year ago and love it , no more gear changing , I was wondering if you had any plans to add any features to the firmware, such as automatic stop and start and return to the tap cycle Just a thought Thanks for the great projects looking forward to the future
Leadshine was the company that ripped off the stepper drive design from Gecko years ago. They sold them so cheap that Gecko developed an improved design that used programmable chips to slow down the Pirates.
I wouldn't be surprised if your servo drive was also stolen. That's the world we live in and the Chinese have different ethics than we're used to.
You may try and contact them to see if you can buy the replacement board without the motor to save a few pennies on a backup motor.
Dead on with the Chinese ethic on IP. Looks like a logical reason to stay with a replaceable controller but at $100 a copy it may not be worth it.
A gremlin let the magic smoke out of that servo. I can honestly say that I have done the same, bought the same thing that betrayed me and sometimes it betrayed me again. Thanks for the video keep on keeping on.
If it's not broken, don't fix it. If this one drops you can upgrade. With the way tech goes it'll probably be smaller and have more power by the time that one farts and let's the smoke out.
100$ for 24 Month of work. Comes down to about 0.14$ per day. I could certainly live with this, too. And maybe this STM32 will last a little longer. We'll know in 2 years time.
Stops video at 6:57 and goes to Amazon to find dental magnifiers.
I vote for genius :) But where is link to your dentist visor??? :)
HAHA...又是萬惡的"雷賽智能"
Bro You can try Odrive to replace that Shenzhen company's PCB
Have you got a link for those dental glasses if they work well?
Remember, you now have a bunch of spare parts for the new stepper. Work that into your cost/benefit ratio.
It's rated voltage is 36 V. You can see at 2:36 that you're operating the device outside of it's rated operating range. This is most probably why the device failed. An avalanche breakdown of the buck regulator would lead to avalanche breakdown on the linear regulator which would expose the STM32 to the full supply voltage, causing internal breakdown and latchup. This would explain why the microcontroller is drawing high currents at roughly two diode drops.
You should get a different power supply if you want to avoid this $99 mistake from repeating.
HolUP HOLUP the thermal pad is connected to the insutator card and connected to the back cover with thermal past and the back cover looks to be made of PLASTIC?
I think that maybe the problem. cut out a 40/40mm hole in the cover and add a ALLOY sink to it directly to the thermal pad and it maybe a 40mm fan as well.
Those mosfets must be cooking ?
Purely anecdotal at this point but I swear there's a design flaw in those STM chips. I repair arcade machines, and one manufacturer *loves* those things, there's one on practically every single board... and every time a board fails in any of the games from them, it's *that.* EVERY time.
Hi James, I thought it was a good idea to replace the servo with the same type as you will have an opportunity to extract the firmware from STM32 on the new servo. Could be the making of another great video for the future. Never a bad idea to replace a part with a known good one. I know lots of mechanics that do that very thing each and every day. Thanks again for all your efforts in producing great videos.
They recommend 35 but I turned it down to 45...wonder why it's dead?? Is that what he just said?? Lol oook 👍
Evil genius: showed us installing this cheap chinese crap and then replaced it with Siemens 999$ servo off-camera.
All my years of buying inexpensive chinese whatever that fails all too often, it's cheaper in the long run to buy the more expensive option. You don't know the quality of the parts going in to those boards , as they might be factory seconds or thirds which will fail over time. Buying the same part again would be the faster - easier option but if the servo fails again then not the cheaper option over time. The servo did have voltage limits and maybe for a reason.
How about asking leadshine for the price of a replacement circuitry? Some companys are very helpful and offer cheap replacement parts if asked.
Hi, great vid. Would you please be so kind as to publish the urll for those magnifying specs. Thanks
Link to the dental magnifiers please.
If the new servo's microcontroller doesn't have an internal security fuse (or if it does and the fuse hasn't been burned) you could extract the firmware and possibly repair your dud. You just have to decide whether it's worth going down that rabbit hole for a $100 item. At those prices a one for one swap seems like a wise move. If you find yourself faced with having decide whether to buy a third one at some point, then go for a higher quality one. My $0.02.
*You're* the stepper-killer-killer-killer?
I have 3 of "those" servos but made by jmc (IHSV57) The PCb looks nearly identical. only thing is that the LDO is on the bottom and the Buck is on Top. The Failure stayed the same. There seems to be some sort of problem with this kind of Servos.
Can you elaborate more on what you know about these failures? Sounds like you've had the same experience
Time is money bro.
Also no need to use other motor if this one works good. Accidents can happen with any motor. You can shift to other motor if this happens more frequently.
You could read the flash to extract the binary file from the new motor's STM32and then replace the uC on the old motor, to then program the new uC.Will cost you a second servo anyway, but you'll have a replacement, at least.
Yes, I was looking at that debug connector on the main board.
Never hurts to have a backup.
Possibly... most likely scenario is that the flash read back lock bit is set. There are ways around this but takes time and energy.
Highly unlikely that the micro is unlocked for reads, but worth a try as the STlink is such an inexpensive programmer/debugger.
STM32C031C6T6 actually possible to buy one of these, ST Micro's parts have been out of stock for ages due to supply chain and stupid management issues.
So how do you find those glasses, I have been looking at the same thing that is popping up on my FB all the time. Im in New Zealand so by the time I get it here its not as cheap as I suspect it would be in the US so Im a little hesitant having been bitten by some previous adjustable glass's I purchased that were useless. I do like the look of these and I assume I can fit them onto my prescription glass's and they dont fall off etc. Any suggestions/recomendations
Since when do ST's go bad? I'd say it was a safe bet to buy the same. I'm glad you took out the thermal camera, you watch Northridge Fix too?
Get what you pay for I guess, but if you can replace it 4 times for the price of a "premium" device... I can't judge. If it occurs again - I'd be inclined to lower that supply voltage some more, to under 40V. You're current 20% over the rated voltage, so there might be some undue stress there, who knows - may even be a firmware bug that allowed some I/Os to trigger and cause this failure that you're not going to see with the new hardware.
One failure does not establish a pattern. If the second suffers a similar failure, it will call for an upgrade.
BTW, having lived with your solid tool post base for a while now, do you like it? Is it worth the expense, effort and loss of versatility for a general purpose hobby lathe? I know Stefan and Robin like theirs, but they kinda work on a different level from most of us. Me, at least.
WHY WHY WHY Didn't you disassemble the new one and thermal image the STM32? simple heatsink safety?
Ain't broke. Yet.
FYI, some Leadshine resellers or volume users can obtain new spare parts for the motors including the PCBs. I have replaced some of their encoder/hall sensors before now on their regular servos.
This was the comment I was looking for. Contact Leadshine and send a link to your video. Tell them you will do a video of the repair when they sell you the spare PCB or Microcontroller.
Just checking, I recall learning that pencils contain graphite and clay. A pencil can cause traces of its own, can't it?
Read the code off the new motor. and resurrect the old one.. You could even post the .bin file so others can fix them..
These STM32 are extremely sensitive to EOS and EDD stress. If for the smallest reason your voltage supply went too high, it could burn up the MCU. Typically I give STM32 good isolation from peripherals and power supply for this reason.
^ This right here. I've also had designs where the input protections were marginal, and had the uC fail in the same manner as what happened here in the video. Hot chip, low voltage on the VCC line, presumably due to an internal short. They seem to be a little more sensitive than some other manufacturer's chips I've worked with.
You were correct to replace with the same, stm32 blowing seems a random failure, unless there is back emf appearing, very unlikely with what looks like optical isolators before the power electronics… would have been interesting just to break the power supply to the microcontroller and see if the psu and regulator recovered to 3.3v more out of interest then anything else.
The opto is between the STM32 and the outside world. But the stm32 is not isolated from the drive circuitry or the motor.
man i hope you have kids cause you would be a dream father....
Lazy????IMO…simple/smart!!! Enjoyed discussion/demonstration
The only issue I see here is you didn't say you bought 2. Put one in and keep one as a backup. For low cost mission critical parts like this I always try to keep at least 1 extra on hand so I can get back up and running ASAP then order the next spare with a little bit less pressure to have it right now.
Two is one and one is none.
my question is what do you do with the old one now it don't work ? have you a answer Or a work around ? !.
Seems legit. 2 years and a failure that probably could never have been found in QA, I would by the same again in a second.
Genius! Thank you for posting this video. I’m configuring these servos with a masso. Got the db9-usb cable from A..zon and finally have a connection. I’ll try 500 steps too as a starting point. I greatly appreciate your channel and hopefully your getting rewarded for your hard work!
Hi James, useful walk through to show how to diagnose the problem. Thank you for spending your hard earned money to let us know if it is a generic fault! I've still got your wonderful ELS on a breadboard ...... one day, one day.
If you go to the settings in the FLIR software, you can adjust and remove the paralax issue.
20:05 ask me.I married my wife twice after a break of 10 years......less adjustment and and only known issues.
I like that metaphor, and I hope you're doing well.
Is it possible to cut rifling using the electronic lead screw? I collect and shoot antique firearms. Every now and then i have to put a barrel liner in a gun. Colts originally had Progressive rifling. so it would start off straight and then progressively add more and more twist. I would love to be able to replicate this rifling. How difficult do you think this would be using electronic lead screw?
Hi James, i would have done the same as you and bought the cheaper servo. You were just unlucky as STM micro processors are used in cheap as well as expensive equipment and could have failed in either. The other point i have is the cost versus benefit of only $1.00 per week based on $100 for 2 years use, i think that is a win in my book. Cheers
I mean there are definitely poor design choices that can be made to increase the likelihood of killing a microcontroller, but otherwise, I agree that as long as it isn't a recurring fault it's worth getting the same motor.
You should put a heatsink on the new stepper controller.
I too have this same servo and all the components to build an ELS, I simply have not had the time to put it all together yet. I hope this new motor lasts longer than a couple years as well. Have you considered adding stops to the ELS to make threading to a shoulder easier?
Spend hours reconfiguing a new servo or $99.00 for the same one that fits. I'd do the same thing too, your time is money too. Now if it fails again then all bets are off as to if I'd take the time to reconfigure... I'd say you chose wisely. Oh, you forgot one tool in your list of tools used, like you my eyesight isn't what it used to be and I'd like to try the magnifier you used. Thanks for another great video James!
The channel gets a new nick name... The Lazy Genius
I am more interested in seeing if this one fails in the same way.
James. Thank you for this walk-through. Your explanations may be the only ones I can watch start to finish and truly understand the information. Have a great week sir.
Hallo The AMS11173.3 has Internally limited by 15V See datasheet, the Board support 24-36V DC input. This means that the voltage is regulated down beforehand
RIP STM32 ;-/
Perhaps guys in China are nice enough to provide you with new STM32 or firmware, but I highly doubt that.
That's why I stick with 1980s and 1990s machinery, everything is in external EPROM ;-D
I'd suggest that running it 2V under its advertised absolute max supply voltage, 12V above nominal max, doesn't leave a lot of room for transients, e.g. from back-EMF.
I'm going to run mine on 36V to start, and if that's not enough, start by turning down the voltage on a 48V supply to something less, likely 45V or so, to give more headroom.
The datasheet says "24-36VDC recommended. Please leave reasonable reservation for voltage fluctuation and back-EMF during deceleration."
Look into the 42AIM30 or 57AIM30 digital servo. Seems to be better quality. Notably more stalled rotor torque, too. Needs the be reprogrammed before first use, factory is 32768 steps/rev
Please consider turning off the MSX feature on your flir camera if you don't want to take the time to adjust the parallax for every shot. It honestly makes it harder to understand the thermal content shown in the image. Yes, I know it looks really blurry with MSX off. But ironically, it's clearer that way in every sense that practically matters. I have the same FLIR camera as you, and I just leave the MSX off all the time. It's a harmful gimmick.
I am interested in how you get on with the loupes. I need some ocular assistance too these days. 👓
They work great, but the experience is a little odd. If you get motion sick, they're likely going to be a problem.
You just popped up in my UA-cam feed.
Great work.
Great subject mastery.
Great video documentation.
Great honest, human communication! (Thanks for the Dentistry Lenses tip for "old man glasses!) Masterful - novices or skilled pros can't help but learn from this video, sir!
Great common problem of "buy replacement" or go through work of warranty, repair, research new source/parts - across all walks of life and lines of work. Of course your time and expertise are so obviously valuable, that the time saved in re-buying is wisest choice. Wow.!
Thanks again!
I bought a 100$ pump, it failed after three years, i replaced it with the exact same unit, it failed after 1 year, i bought the EXACT same pump again, going on 2 years. It's competitor, the one "recommended", was $1300.00. Seams pretty easy to me. The time you saved with the copy, may be enough to rationalize the duplication. Imagine having to get the new programming cable, a new programming App, and then rebuilding a new mounting plate. Time, Money and sanity, check!
Maybe a more expensive motor in the first instance would have lasted longer - but as it was the u-controller that had given up, it probably wouldn't have made a difference. uC's are reliable if they are running in a benign environment. I find it really surprising that the microcontroller would fail unless something was pulling too much current on one of its legs. (they don't enjoy a leg-pull!)
I would do a bit more research on the dead one (on the killed stepper killer killer) to see if something attached to an I/O pin on the STM32 is shorted or has too low a resistance.
As a totally irrelevant aside, I find it quite surprising that such a trivial application uses an STM32. I guess they're quite cheap, but it seems like a sledge hammer to crack a nut.
Could you buy a replacement controller board for it? Somebody suggested copying the code from the new one you have and flashing another processor, but getting the old one off the board and soldering a new one on, and getting it to work without damaging the solder-tracks on the board could be an adventure. I've never attempted it myself, but I've watched a lot of UA-cam vids on how it's done, and I'd put my chances of succeeding a job like that at perhaps 30%. I'm a retired electronics and computer engineer, but zillion-pin chips weren't a thing when I retired. Leastways, not on the boards I worked on.
And now that you have a working micro controller, why don't you "just"* dump the firmware and fix the other one too?
*Extremely hand wavey over simplification 2000 O'matic - but should be doable
Construction and build quality looks decent. Shame it failed that badly. I mean, complete servo solution for $100. There's nothing wrong going for it; maybe it's just a one-of case, after all.
The maximum of the servo is 50V ... of course but when you stop the motor it will induce power to the power supply which itself might get damaged (recuperation). Don't run those motors with more than 36V PSUs. I think about decoupling mine and inserting an extra wire for the logic power supply.
This servo requires a damper. You can do a test - connect an oscilloscope to the power bus of the servo driver and watch the voltage spikes when changing direction or high accelerations.
I connected a clone of this servo - ihsv57-30 through a 3:1 gearbox to the longitudinal feed of the lathe. I got a problem - the Fujitsu power supply went into error. The reason is the EMF that occurs when stopping at 3000 rpm is more than 10V. It was added to the supply voltage. As a result, either the servo drive went into error or the power supply. I put together a dumper circuit from the GeckoDrive website - it was a good solution.
I also had a problem with one of my Chinese stepper motors. For me it was a mechanical and I was able to fix it. Not before I purchased new one, but it's always good idea to have a spare if they are used in important place.
It's funny how they appeared to be relatively good quality but there is still some details which fail.
Mine stopped spinning but coils were powered and there was some sound when I drove it, though controller went immediately into error state. I suspected short circuit, but all coil resistances were OK. When I opened it up, I didn't find anything wrong but when I put it back together I realized what the problem was. Face and backplate were machined in a way they didn't lock into place. Alignment was maintained only by friction and it was big nema34 motor. It had been slipped and prevented motor from spinning.
I redesigned my mount so the motor body is supported by rubber, and it has worked well since that.
When dealing with electronically populated PCBs by using a pencil around the circuit (CCt) you can place carbon particles on the PCB which can cause short CCts in the most extreme case
That seems like the typical failure mode of an STM32 MCU when subject to transients 5V, indicating a probable weakness in the circuit design. It would be interesting to get a trace of each of the pins (on a working servo) during normal operation to find the culprit.
Have you tried dumping the firmware from the new µC? You never know, you might get lucky.
Or contacted the original manufacturer?
There is a rising "Right To Repair" movement. You just ended up making a lot of waste because you couldn't replace a (my guess) less than $10 micro-controller.
There has to be a better way.
9:56 - Isn't a buck converter meant to step UP voltage?
A step down converter would use PCM and filter caps to lower voltage using a variable duty cycle (and current sense to maintain voltage).
The INDUCTOR looks to be a buck converter, but must step up the voltage to a higher level (as when you switch an inductor, as the magnetic field collapses you get a spike of higher voltage, and then you use smoothing caps to give a consistent voltage level).
James not wishing to question your choice of motor ... but other ELS builders have quoted using the 3Nm version of the Nema23. Have you ever experienced issues with insufficient power when turning/thread cuttting tougher materials?
100 bucks to save a day and make 300 is what you call 1 step back to gain 2 steps forward. The opposite, would be investing 3 hours to save 1 in the future, or what you call Government Bureaucracy.
There is a way to pull firmware from those little ROM chips. Get an RT809F chip programmer for about 60-80 bucks and it comes with software that allows you to read and write chip software.
I am using this motor for my lathe that is a 12 x 24 with a 3mm pitch metric leadscrew i am thinking of using a 90 deg. angle gear reducer with a 5:1 or a 10:1 reduction. There will be backlash in the reducer but I think it is not a problem since exact positioning is not required because when threading the backlash is taken up before the cut begins. Has anyone done this or fore see a problem.
I think at that price as long as the failure rate of the Chinese Servo is less than 3 times as high as the more expensive option you are a making a smart choice. Especially when using multiples of the same one. Ie. if you have 4 machines all using the same servo even if the cost of replacing the servos every few years comes out to exactly the same price as the life of a more expensive option, it would make sense to keep a couple spares on hand which would save you money on downtime while waiting for shipping on a replacement.
In the 4 machine example: If an expensive option lasts 6 years and costs $300 You would need to invest 4x300 + 300 to have a spare ready. Or you can buy 12x$100 + 100 and have a spare ready. Giving you the same redundancy at lower overhead or even better redundancy if you want to buy multiple spares. Since you don't need to buy all replacement sets at once the cost advantage becomes even greater 4x100+100 and 400 every 2 years for maintenance. This gives you a capital advantage, assuming even 5% return on your money, you would save $120 in 4 years + $10 every year on the backup motor.
Although I think this was a freak accident because it seems to use quality parts and STM is a reputable company which is unlikely to have the same failure happen that often.
So, what killed the servo motor? Looks like the micro controller was zapped, probably by not having enough protection against back-emf.
Servo to the power supply " Hay you know all that power your are sending me, don't need it any more got to send it back" Power Supply to the servo: "Sorry all sales are final, I have no facilities for returns, you're just going to have to let out some blue smoke, sorry"
Like a couple of others my bet is on back emf from deceleration, raised the rail voltage above what the primary switching regulator could manage. When those simple buck switching regulators fail with the pass transistor on you get full input on the poor 3.3V linear. What you need is the equivalent of the braking resistor used on a VFD. When main DC bus rises above the nominal value they turn on a FET that connects the main power bus to an external power resistor to dissipate the energy. If the motor is running at fairly high power, there are considerable magnetic energy stored in the fields of the windings. If you suddenly mechanically stop the motor that energy needs to be dissipated someplace.
There is likely some open source servo code you could port to a new STM32. and get it working again. Years back the was the HUH servo dirver board on CNC zone.
I'm guessing that just buying a new circuit board was not an option? If you ever do find one you'll have a back up motor should you ever need it. P.S... your explanations are so good I almost understood them. This is from a total ignoramus about electronics, so: "Good Job".
Welp, buying the same motor is a reasonable move. It lived long enough to warrant its price. But... I'm wondering... Does STM32 contain any anti-read measures? Because if not, then how about dumping the firmware from the new motor (if it is the same board, same revision), and then buying a new microcontroller, flashing it and replacing the burnt one? Just so next time You land in such a spot, You can just replace the motor with a "refurbished spare", and then have all the time in the world to repair the broken one, while still being able to use the lathe in that time.
Would be cool, if it was worth it, to send the servo to Northridge FIX to check out what really caused the problem. Besides that, I think going the cheaper route is the smarter way of handling things. The metric shows that you get a lifespan of two years from that servo, let's see what happens in two years with the new one.
He will probably get more than two years if he mounts it on a proper metal bracket lol. 😁
Although Alex has the skills to fix it, I doubt he would take it on (not worth the repair cost for something that is so cheaply replaced).
You led pencil is a conductive pointer. Leaving conductive duzt??? Probably.
With a little legwork you could probably source a replacement board; giving you a spare servo for the shelf.
Can you pull the program from the new chip, then program a new one? Or is it proprietary/encrypted programming? If you could replace the chip. Then shelf the new one. (To use if it breaks again.)
In terms of failure of the STM microcontroller and V-reg... the root cause might be poor or broken ceramic capacitors on the V-reg, or a V-reg failure in general. My expectation would be that the voltage to the microcontroller exceeded the parts maximum, and the microcontroller shorted out internally. A microcontroller running for 2 years usually doesn't just die, but capacitors might.
Voltage regulators require a minimum capacitance for stable operation. The most common place to save money on cheap electronics is to remove capacitance or put bad quality capacitance in there.
In terms of preventative actions, you could look at the V-reg circuit and add a few microfarads of quality capacitance to the input and output of the voltage regulator. This could be achieved by stacking SMT ceramic capacitors, or soldering on a radial 10uF electrolytic cap across the output capacitors, being aware of polarity and voltage rating of the caps.
Other areas of failure could be ESD static discharge onto unprotected microprocessor pins... this is harder to fix, but during installation and programming you should make sure you are grounded, just like when working inside a PC case. Once everything is connected it should be relatively hard to zap things. Unless the failure corresponded with unplugging cables, I'd suspect bad capacitors over ESD.
Ok, I have seen both now. The coil windings quality is different. The electronics are different. The bearings are way different, and ClearPath comes with 3-year warranty, the price is only two to three times based on your configuration. If you are looking for a reliable uninterrupted business processes for at least till the warranty expires? Go for local companies with good manufacturer warranty. Not to mention that these knock offs look exactly like the ones from other companies in the west. ;-)
so who was the killer- killer- killer ???
Until the jury isn't out anymore you're a genius idiot, me thinks? :D
I would definitely do the same and buy a new cheap thing that works. It kinda shoulda oughta work since it actually did until it didn't. And what's the alternative? Get an expensive time consuming thing that also may or may not work some day. Or cut the costs a little and buy two. Yeah, you did the right thing!
Buying the same servo DOES save time and money for the "home hobbyist". In industry, they would have re-engineered the whole thing, taking lots of time and really increasing the cost.
I would try and find what might have caused the failure, but with the chinese stuff... that could be anything.