I’ve used your item 2133 – DRV8825 driver carrier board in two projects now and think that it’s fantastic. I was able to quickly and painlessly refine a microstep control strategy that would have taken much longer using conventional drivers.
However, I’d like to suggest that Pololu consider offering an enhanced version of the board. The enhancements I’d suggest would include, at a minimum, access to the current sense outputs (ISENA, ISENB) of the chip - perhaps with current sense resistors preinstalled or with pads for customer-selected resistors. It would also be useful to offer access to the index table related pins (nHOME, nRESET).
The current sense pins would be most useful in detecting pre-overload or pre-failure conditions. This would allow the device to react and compensate before a fault occurs. While having access to the nFAULT pin is nice, it is usually providing information to the uC too late to avoid a problem.
Access to the index table pins seems somewhat less important but would still be useful by allowing for fairly simple pseudo-closed loop control.
We are glad that you like our DVR8825 carrier, and we appreciate the suggestions. However, we do not think having access to those pins would be very useful. Could you be more specific about how you would use the ISEN pins to detect “pre-overload” conditions?
By the way, our carrier does bring out the nRESET line.
While my suggest differs from the OP, I think the most useful enhancement would be a way to bypass the potentiometer and to input your own voltage to the current limiting pin, rather, the part of the board that takes the potentiometer input.
I’m thinking the DAC on a teensy would be beautiful for that, or and other MCU DAC or digital pot.
It would probably not make much sense to redesign a whole different board form factor, but adding a pin somewhere along the pin-less edges, and traces that can cut the old pot out of the design, would work.
Thanks for your feedback. We discussed connecting a digital pot to the AVREF pin on the DRV8825 in another thread, and we have been keeping this kind of feature in mind. How many units would you use like that? Is removing the pot and soldering a wire to the pot pad too much hassle?
removing the pot is not a terrible hassle, but it is enough fiddly to discourage that as a means of approaching the situation (since it’s not just the work, but also the fact I can’t be certain a thin gauge wire wouldn’t be a weak point in the design.) It’s only doable for my at-home hobby work.
I’ve just considered it in the past but found other methods to deal with the situation. For example, where I work I’m occasionally doing one-off designs involving your boards so there are only a few units made yearly. That’s one of the reasons we use the breakout boards instead of rolling our own (and, your boards are great quality and hardly more expensive than rolling our own.) So in the end we only use 16-32 of your stepper driver boards in a year at this rate. Once you count all the other types of boards you offer it ends up being 25-50, I’m guessing. Not a big amount.
I tend to do something where I have a PCB socket for your parts, to simplify wiring and replication for the future units, if I’m not around next year and they want to make a few more units.
At the moment I just document things well, and have made some thorough “how to set the potentiometer” docs for assembly here. But a much nicer approach for me would be to just declare some trace that needs to be cut and plug the board into my master MCU pcb board, and control a digital potentiometer in code. This would be(much easier than desoldering anything, and much much easier than asking people to set the mechanical potentiometer which can be done poorly and can theoretically shift. So I’m not particularly inclined to think there’s a massive demand for another DRV8825 board with a digital potentiometer on board. Thanks for linking the other thread, I hadn’t noticed it. EDIT: Hehe, I had not noticed, but I responded to that thread a long time ago when it was first listed.
As I mentioned, I think it could potentially be done with just a clear trace to be cut ( representing 3.3V-mechanical pot leg high-side) to bypass the existing pot, and a through-hole along an unused edge, where that pin goes to where the output leg of the current mechanical pot goes. This way the board could 100% be used in the existing manner that people use them all the time, and also have the ability to be easily adapted for an external digital potentiometer. Or even just so that if people want to string multiple of the boards to similar stepper motors, they could cut the traces on them all, and set them with one external pot.
But I have no idea how much money it would take for such a design change, how much one extra hole would cost in production, and whether such (relatively limited) modification could fit on the existing form-factor.
Thanks for letting us know more about your application; we will keep your feedback in mind. We would eventually like to make larger versions of our stepper motor driver carriers with more of the pins brought out, though that is not a high priority at the moment. If anyone else is interested in this kind of feature, we would love to hear from you as well.
If having better access to the voltage reference node would make things easier for you, you might take a look at our DRV8834 carriers. They have a lower input voltage range, but the chip’s VREF pin can be connected to a header pin along the outside of the board by bridging a surface mount jumper. If you plan to control the current limit with a microcontroller, you would still probably want to remove the potentiometer though.
You know, I just found [looked into] the “surestepper 8825” copy/inspired derivative of your 8825 driver board. I realize it came out a long time ago (a least a year ago) but I noticed it does have a better vref layout that allows for a jumper for external control of vref and easier potentiometer adjustment. It also has a slightly better thermal layout, it seems.
I’m still recommending Pololu for work, because you guys are 10x more reliable than any of the surestepper vendors I’ve found, and your price is still more affordable and reliable than us running our own boards, but it is surprising that there’s an alternative that has good benefits.
I’m not entirely on board with your talking about making newer bigger boards with more pins, and I understand why that’s low priority (it’s not a huge demand.) But i was not talking about a different form factor, and I think it might be possible within the existing form factor. And it turns out the “surestepper” manages to do at least some of that.
But, of course, this is a major product line and I’m sure it’s still a lot of work to make changes to it in the existing size.
The DRV8834 boards are sweet too, and good for some projects, but the low voltage limits their ability to drive most motors at a fast speed. Well, it’s great, but just imo a very different board for a different sort of application.
Thanks for pointing out that example. Why are you against bigger boards?
Bigger boards are O-K, especially when they have mounting holes. I’m not too averse. I’m sure you get people comparing boards to sparkfun boards with mounting holes and they ask why your boards can’t have mounting holes [where many sparkfun boards tend to be oversized and have mounting holes more often.]
But one of the appeals of Pololu boards, besides quality and often price, is that they are so compact. They’re well designed, and I couldn’t easily do that myself. Even if I could, I could hardly get the <0805 parts down onto the tight spacing by hand :p. Morover, because they are small, i can still use them in socket-designs where I make up a PCB that the Pololu board plugs into, and since I’m paying per Sq.inch usually, I don’t want a big board. Sometimes, a big board is necessary. But when it’s not, mehhhh. Plus, a small board is also better for reusable breadboarding, though I don’t use removable breadboards usually (perf board and female header, instead.)
I was just using the micro-motor encoder board and that thing is tiny. Pretty cool. A little hard to reliably wire to, though. but that seems to be the only way to have made the board so flush.
But in this specific case (DRV8825 board), I thought you [/Pololu] would be averse to a bigger board, because I know the DRV8825 was originally thought up as a replacement for the widely used reprap a4988 board (that pololu also made first, I think.) So you needed to keep the socket-design as a form of legacy. There are two detractors to this now
(1)A number of reprap control boards no longer use the socket designs, though a lot of people still use the older boards that do.
(2) Maybe reprap/3d printing is a small component of your sales (though for this board I feel it must not be that insignificant, even if it’s not a majority.)
Anyway, just two cents :). Thanks for asking!
While setting a DRV8825 board today, it occurred to me that there is still another perceivable flaw to the board.
At the moment the current setting goes 3.3V --> Pot–> RefInput, if I remember right. This varies the voltage from 0-3.3v. But the maximum possible useful voltage you’d want to apply would be 1.25V, which even then is hardly realistic for the board. Is this intrinsically that bad of a problem? no, but I find the potiometer to be extremely touchy how it’s scaled. I set to .6v, and the slightest nudge makes it .85 volts, and practically speaking an amp limit of 1.4 and 1.7 is a big difference. Often the limit is set very carefully for a given application. There are a bunch of solutions (different pot, nonlinear voltage divider, etc) but I’m not sure if any is viable. I just wanted to remark of it as a minor but meaningful nuisance when using the board.
Thanks for all the input; please keep the feedback coming.