EEVblog 1560 - TIP: Use Your Arb Gen as a High Resolution DC Voltage Source!
Вставка
- Опубліковано 16 сер 2023
- Tech Tip: How to use you Arbitrary Waveform Generator as a High Resolution DC Voltage Source!
Siglent teardown: • EEVBlog #805 - Siglent...
Forum: www.eevblog.com/forum/blog/ee...
If you find my videos useful you may consider supporting the EEVblog on Patreon: / eevblog
Or with crypto:
BTC: 33BsprBQNBtHuVzVwDmqWkpDjYnCouwASM
ETH: 0x68114e40ff4dcdd384750500501e20acf3875f8c
BCH: 35n9KBPw9T7M3NGzpS3t4nUYEf9HbRmkm4
USDC: 0x68114e40ff4dcdd384750500501e20acf3875f8c
LTC: MJfK57ujxy55su4XicVGQc9wcEJf6mAoXF
Web Site: www.eevblog.com
Other channels:
EEVblog2: / eevblog2
EEVdiscover: / eevdiscover
T-Shirts: teespring.com/stores/eevblog
#ElectronicsCreators #TechTip - Наука та технологія
This is such a great example how engineering works: pretty much exactly 2 minutes of WHAT it is, HOW it works and WHY it is great... And then 6 additional minutes about what you have to know to not run into problems 😂
I will adopt this 25%/75% ratio into my rhetoric when I have to explain outsiders, why complicated things are complicated
Try reading the manual for Keithey KEI-260. It is just that same way, too.
This! Now if only CAD tutorial (looking at you FreeCAD community!) were like that. Most of them are 10 minutes of some dude mumbling without giving the big picture.
So, are you saying engineering is working the way it should when the final product misses the initially promised target by 300% ? 🤔
(That would explain a lot) 😅
@@RottnRobbie oh, not at all. Thr given value of 2 minutes is either project management or marketing and has little to do with engineering
We literally did this in school when we were all remote because they couldn't afford to send us actual bench supplies
Couple years ago I discovered this trick when I was tracking down an issue with a particularly weird brownout detect implementation. There were multiple thresholds, holdoffs, etc. I was able to create an arb profile in excel that matched exactly what i wanted, load it on my flash drive, drop it on my scope and test EXACTLY the profile I wanted. Worked great for a little low power micro.
Doing it often would probably be a good idea to build up a power stage, using one of the power capable DC stable opamps, and simply put in a gain of 2, and use a 50R input termination, so the opamp gives current gain and some isolation, and the ARB provides the precise input. You could even use a regular power supply, and feed the ARB output into the control section, giving you a power supply that you can program, using the ARB waveform generation to give power up and down ramps as well. Not going to be 4 quadrant all singing and dancing, but even a single quadrant one, that can emulate say a battery discharging, can be useful.
I love these shorter tip videos, please keep making them!
If you want a good affordable precision voltage source people should look at Ian Johnston’s PDVS2mini adjustable reference.
Yes, I have one, should have shown that.
Good to see the 121GW being used in a video again :) This was a good video! Would enjoy more tips about how to use equipment for things outside the primary purpose like this.
"Tech Tip".. I see what you did there! Great vid, Dave!
2 minutes video evolving into 8m rant, I love you Dave for that, the normal way, not that QWERTY rubbish way
Keep the tech tips coming Uncle Dave cheers!
I enjoy your presentation's, They are always educational. I have been in electronics, sense the early 60's, first as a T.V. tech, then in later years, as a land mobile tech, for Motorola, and at 80 Yrs. i still find that I don't know every think, so thank you for the knowledge you share.
Two minute tech tip video? I see some pretty advanced time dilation here, haha!
Looking at it this way, an AWG could be used for setting the bias point for vacuum tubes and transistors, testing the output waveform for optimal distortion.
Many thousand thanks for the tip on using your Arb Generator as a DC. mv source ! Greetings from Stockholm ( Sweden ) and Jan Eklöf
Man, I quite literally figured this out last weekend
Thank for tip on the Sig's output impedance, I had no idea I wasn't getting hiZ output.
Tech tips without a segue to a sponsor. Approved.
I love these short tech tips
A small tip for using signal generator as a dc source: some source will automaticly change range when you adjust the dc offset(My Keysight 33622A), it will quickly jump from 0 to offset volt, almost unnoticeable . for some devices has hysteresis input voltage, be careful, because it may performance no hysteresis at all.
wow thanks Dave never thought of that
Damn good idea! I've got a couple of Agilent generators, too! Thanks.
Great tip. I've been using the two channels of my Analog Discovery 2 like this for a while. Now following up on your video I hooked it up to my 6 1/2 digit multimeter and found accuracy to be quite decent. It seemed like it has less jitter and voltage error in the negative region. The step width is roughly 1 mV
Sweet test equipment. The stuff is getting so cool, even if you just want to bench it and look at it. 😊
Thanks for the video. I was just discussing using my waveform gen to do some testing as a sub for my now blown up DC power supply (lol).
Good tip!
Excellent idea! 👍 Never thought about that..
And first thing what I wanna do now is to check out my signal generator options.. 😂
Thanks for sharing
You don't even need a fancy Arb Gen for most stuff like this if it's just DC, I regularly do this with a standard lab PSU and a voltage divider, requires a bit more planning to make sure your DUT doesn't affect the reading through it's internal resistance but you can often get high resolution (smaller steps), lower offsets and lower noise with this method and for much less money in almost any lab.
I often need tiny currents when characterizing charge sensitive amplifiers (1 - 100s of nanoamps), I could buy a $30,000 SMU but a lab PSU and a handful of large value series resistors gets the job done. And in cases where I need precision, I find I don't actually need exactly 1.000 nA, almost always I just need several values between 1 and 100 to form a trendline to get the value at 1.000 nA so I can just put in a random big resistor and measure the actual current (I do use a reasonably expensive DMM6500 for that part). I also often find coupled leakage currents become an issue for measurements like this (Marco Reps has a great video on that topic), instead of again buying a fancy SMU I just switch the PSU out for a 9V battery with none of those problems.
I was just going to mention this myself! As long as you're ok with the output impedance the voltage divider gets you, you can get really precise with your voltage. If you don't have a multi-decade divider, then you can use one resistor for one side and a resistor decade box for the other. Voltage dividers are included in some hardware, my Advantest 6161 has a voltage divider output option to get into the nV.
When you are chasing pA and fA, even a battery source can have leakage current over your mat and you have to get creative with placement. It's really fun to do those measurements.
I just figured this out when repairing an Owon Arb Sig Gen. Instead of good products, they don't share the procedure of how to calibrate the instrument. Any small changes in the op amps, these offsets can be just too high and the correction is done by software. Even not calibrated, it's still a good source of dc and ac signal, once you check it first.
Ah yes. We do tend to forget that DC is a waveform too :-)
Great for testing voltage regulators and especially those used in a negative DC rail.
i would really like to see you do a NEW shootout between the three eevblog meters. i get asked by newer users all the time which one or whats the difference etc?? can you do a new video shooting these three out and and showing the differences between them then i can direct them straight to the video. even myself who has years exp may learn something new about them never know lol. most new users will not spend fluke money and even people like myself i dont take my flukes out in the field they either go missing "somehow" or they drop, fall, or who knows what. i really think that would be a great video. maybe you have one and i just couldnt find it but if i searched and couldnt find it then others cant either. and yes these videos are great especially for the newer guys who dont have the patience to sit and watch an hour long video they want a quick 5 min one so they can learn something and be done. appreciate all your work.
Awesome, thanks for this. Amateurs like myself definitely appreciate these tech tips.
Side note and hopefully a segway, but the kids and I are starting to mess with small DC motors, up to 540 size and would love to see how to "not mess up your PSU" lol.
Maybe a vid on how to safely connect a PSU to different size DC motors so when they spin down, they don't backfeed the PSU and damage it.
Also measuring how much current they can spike when backfeeding.
But either way, love the content!
I have seen a few vids on this, but not many with what diodes if any are needed and the best way to connect them.
Most decent quality PSUs should be safe from smaller motors, as the role of a benchtop PSU is to power anything and everything within its range. And aside from old school linear PSUs, switching supplies are undoubtedly using MOSFETs which have a reverse diode built in.
But to be careful just solder a flyback diode across the power leads on your motors. Will have 0 adverse effect on them during normal use and will only protect the motor and the PSU itself. Unless you are trying to switch them on and off at a high repetition, which will cause a delay in their turn off time (you'd have to be switching them in the 100s of Hz for this to be an issue, such as in PWM). Any 1N400x diode will do the trick.
smaller and floating, you could use a loop calibrator/process meter - see e.g. the RS Pro RS-135
For those that don't know, I found out on my siglent if you hold either of the "OUTPUT" buttons for a couple seconds it will switch from hi-z to 50 ohm or vice versa
Good tip Dave.
I checked with my UNI-T UTG932 generator does 1mV offsets but you can't have zero amplitude waveforms. However, you can program a square wave with a very long period (1 uHz should be long enough) or upload a zero arbitrary waveform. Only 14 bit and 200Ms but much cheaper. I paid £80 two years ago, like most China test gear, that price has doubled since then but still a good deal, especially as the Siglent is approaching £1k.
You forgot about the best part - if you have a half-decent linear lab power supply then you can combine it with the sig gen and get the best of both worlds! Just remove the shorts from the external voltage control lines on the back of your PSU and hook up the output of your generator to them! This way you will lose a bit more precision due to PSU control input internal nonlinearities but the nice resolution of your generator will remain and now you can forget about all the 50Ohm output impedance nonsence and power output limitations!
Even ancient HP Harrison linear PSUs or proper soviet made ones from the 70-s had the option! And most of the ones I have encountered even conveniently had 0..10V control voltage range corresponding to 0..FullScale on the outputs. And the more fancy ones even took -10..10V control so they could freely swing the output from positive to negative and some had a decent frequency response going up to a couple of kHz!
The most important lesson here is - read the manuals and learn how to use your equipment! Cheap PSUs may be more or less the same but good ones are not all created equally and for a specific slighly more complex than the ordinary task one good PSU can be much better than another.
I can only find the regular thumbs up button. Where is the BIG thumbs up button? I don't want to let Dave down this time...
This is something I tried a long time ago on my Rigol DG1022, but __annoyingly__ I could never get rid of the AC component completely (I think the min value was 1mVpp or something). I suppose I could have set it to like 10MHz at 1mVpp or something and put a LPF on the output but it was just too much messing about (I ended up using another PSU via SCPI). I did want to see if loading a DC value into the arb gen would do it, but that too was a lot of messing about to generate DC.
My cheap Hantek DSO2D10 let's me adjust the generator offset in .1uV steps.
I dunno if the ADC in there even has that resolution (thinking maybe the software let's monkey me play with the knobs and decides what the closest voltage is it can output)
But now i am intrigued.
Edit:
Turns out my Monkey theory was correct and also there is a 4mV offset to 0V and a 2mV non linearity in the range of 0V to 500mV.
To be fair, i did not let scope or meter properly warm up. It was just a quick "Oh, gotta try this"
I've already used that "trick" more than once, i had my power supplies in use and i started to think were i could get just a signal voltage, and that's when i remember i had an arb gen with DC option. I have one of those cheapies ones (feeltech FY6900) and unfortunately it has only 14bit resolution, the adjustment is made in 1mv steps but the output only changes in 2 or 3 mv steps, but nonetheless it worked for what i needed to do. The only plus side is that is dual channel so i have two different voltages from the same unit, however ground is the same for both
What's the use for the nano volt source? I'm thinking of some biology lab or counting electrons.
Two AUSSIE minutes. Very similar concept to Windows minutes.
Yes please
Interesting, they always told me that you shouldn't really use signal generators like that if you can help it. Since as soon as you start drawing modest amount of current the signal will get heavily distorted. But I've never actually tried doing that.
You have to buffer the output of the arbgen with a power opamp, function generators are not design to provide any significant power
Got a vid Idea. I dont know if you did one like that before but Id be really interrested in a video explaining Transistor types, brief usescases and explaining the specifications/datasheets and correlations between the specifications. If I need to replace a transistor, mostly lower speed stuff I work on, I just compare the specs and use one that is close enough to the one I wanna replace. It usually works out but that cant be the best way of going about it. And if Im using a transistor for something I design Im just going by gut feeling after looking at the Datasheet for an unhealthy amount of time.
Now I have to design a power output stage which can be controlled by my cheap arbgen to add some current capability to my newly discovered "high resolution DC voltage source". 🙂
One power opamp with unity gain and done.
If you don't have a DC offset, how about a square wave with 100% duty cycle?
2³ minute tech tips
Are voltages below 1mV even real?
Yes, but do you repair cockrings?
Can't we do much better by actually generating a signal and lowpass filter it?
A bit like what you do with a pwm output of an MCU but with much more flexibility and granularity.
If you can set the 16bits DAC output for 2^10=1024 samples, you can (in theory) have an output resolution of 16+10=26 bits.
Of course in the real world it will not be that good but it would probably be much better that 0.1mV...
You've just kind-of rediscovered the Delta-Sigma DAC...
At that point, just get a decent voltage reference IC, a 24 bit DAC, an op-amp and build the precision DC source yourself.
"Just a quick 2 minute tech tip video"? Out of Dave jones? BWHAHAHAHA! There is no way this is a 2-minute video!
I just read the manual for the Keithey KEI-260. I had an orgasm! They certainly don't write manuals like they used to!!!
The sourcemeter for the poor. Pretty clever.
The Sigilent Arb is totally ripping off Agilent 33520 series' UI
Yep, Siglent have ripped off several Keysight UI's almost 100%.
Better a good copy than a terrible design.
Imitation is the highest form of flattery 😉
Wait until you see products from Hantek, they're arb gen is certainly heavily 'inspired' by Keysight.
If I need something at 0.1mV increments, it'd be calibrating my power supply meter or my THS handheld oscilloscope, which a crappy made-in-china ARB will certainly provide a terrible reference even if it had the resolution. Well, I'm sitting on a pile of Data Precision 8200 so my view could be biased. Lol
i like EEVblog techtips, but I dont like the aspect of 2 minute videos, they come acroos as speculative & that tic tac minute vibe to them and for kids with short attention spans... needs to be way longer, so put numeorus techtips into a longer video..
The Riden with a P has 1mV and 1mA resolution.
For example RD6012P.
It's a programmable power supply, so you can do sweeps and whatever else you want.
The problem with the Arbitrary Function Generator is that it hasn't any power behind it.
It's not like you could draw 20-100W from your Arbitrary Function Generator.