Just a generic Amazon one I got ages ago to check my lizard’s basking surface temperatures. It seems reasonably accurate when checked against a thermocouple.
Re:not having access to edit: what an incredibly stupid and strange restriction lmao.
2 Likes
Yeah the edit one is odd, in general the idea is that new users get limited abilities both so that spam bots can’t go wild and to not overwhelm new users. Moving from New to Basic (trust level 0 to 1) is quick and that gets most of the features. Basic to Member takes 2 weeks and Member to Regular is a lot of reading over a couple of months
1 Like
Probably calibrated for very high emissivity (flat black) surfaces so metals will read low. Black epoxy cases of components and solder mask on circuit boards will read well but bare copper/solder will read low.
Check the field-of-view of your IR gun. That will give you an idea of how close you need to get to fill that view with the component. Or just place the IR sensor opening up against the component or area you’re interested in. That will shrink as much as possible the area it’s looking at and get you the actual temp of just that component or area (if not bare metal).
For bare metals a thin layer of flat black paint (I use a little jar of hobby paint and small disposable brushes) is great or even electrical tape can work. For larger areas or entire circuit boards I use spray paint.
1 Like
Okay, so, after I finally got back to the poor FOCer, I discovered one of the big electrolytic caps was reversed, the left one. They seem to be supply filters. I put it back right way around, and I got continuity from positive and negative supply to both positive and negative on the caps, but not reversed, so that’s good.
However, when I give the board power, it now draws about 20ma. So I’m guessing this board is dungoofed. I’ll move the parts I didn’t order duplicates for(unfortunately including the electrolytic caps, but they both test within spec once I pulled them) over to my second board (parts were too expensive for me to get 5 populated boards). Here’s hoping I don’t blow anything up this time.
So I got everything soldered up on the new board except the MOSFETs. I wanted to see if I could get any appreciable current draw before installing them and I don’t think they’d keep the STM from booting. However, I still only get <20ma quiescent current and when I press the connect button in STM32CubeProgrammer nothing discernable happens. The programmer detects a target voltage apparently, it just doesn’t want to do anything about it. No components on the board are remotely warm to the touch, there’s just nothing happening there at all except the blue power LED. Not sure what to test next, other than I guess go ahead and install the MOSFETs. Don’t see how that would change anything.
The board does pest the diode check test mentioned in OP. I looked at the ICs under magnification and didn’t see any issues there, either.
I’ve seen that power draw on other VESCs at 30v with no firmware on the STM32. I think it’s pretty normal.
I’ve had better luck with their old software than the new cube programmer. Might want to try that without battery power and use the 3.3v from the stlink to power up the stm32.
2 Likes
Well how about that, it works! I’ll finish soldering up the MOSFETs and then I might even make a motor spin around or something crazy like that. Thanks for the help!
3 Likes
Just to follow up, thanks for all the help. The motor spins, and spins scarily fast, so that works, haha, so thanks again!
4 Likes
Hi everybody,
we are a group of volunteers and build a UAV/drone for search and rescue (SearchWing). Recently we got a nice test stand from Tyto Robotics and did some basic tests on motor efficiency. It seems that we can swap ESCs and get a efficiency gain of up to 15 percent. We are now of course curious what is possible with field oriented control and would love to get our hands on a FOC ESC. While searching the net we found the VESC project and after some digging this great DIY project.
Does it make sense to build a FOCer 2 for testing right now? Or should we wait for the next iteration?
Is there anybody in the community who could lend us a unit to do our tests? We work in Berlin and Augsburg, Germany. We have soldering skills and could finish a unfinished unit too.
Thanks for your help.
Philipp
3 Likes
The next iteration is is being beta tested with good results so far. There are several improvements with the CFOC3 so I would wait for that to hit.
4 Likes
Thanks for your reply. We will wait for the new version. If we can assist in beta testing somehow please let us know.
I’m trying to finish my spec of the Cheap FOCer 2 V1.0 and am not sure what the Schottky clamp voltage ought to be on D8 - D12. My guess is that it should be 10V but would like to confirm.
I’m a little confused about the FOCer 2 V1.0 build instructions. I’ve really been trying to manage this myself with searches but I’m at wits end and need some help.
I’ve ordered the boards for V1.0 but I am holding off on confirming the SMD build authorization until I understand what needs to be included. I just read that the Miller Clamp portion of the gate drive isn’t working quite right and that the components should be omitted. Fine, but now I’m not sure what BOM I should be using.
Looking at the rest of the circuit, there are a couple of discrepancies I am unclear about. First, Phase 3 doesn’t match Phase 1 & 2 (1nF caps vs 2.2nF caps). Next is the additional filtering on the hall sensors. What problem was that attempting to address? Finally, is the TVS diode required? I’m guessing there is a noise issue that was being addressed but I can’t find any mention of it.
Bottom line, I’m not going to hammer on these controllers. Is there anything in the changes for V1.0 that are basically edge cases for the performance envelope?
I’ve blown a few FETs on my CFOC2 (HW v1.0) lately. They are going in pairs (which I imagine is common), and I’m blowing the trace on the drain pin on the one on the capacitor side. Anyone have any tips to prevent this? Thanks!
I didn’t take any pics of my board. There’s no visible damage to the FETs. This is where the traces are blowing out…
That is a picture of a cheap focer 2 v0.9 right? In the v1.0 there seems to have been an attempt to reinforce this trace as it is the weakest point in the power current path. Only on Q3 though, not Q2 and Q1. Did you test the fets if they are blown out? Or is it just the traces that have burnt?
If it is just the traces you could run a wire to bypass the burned trace.
2 Likes
That is the v0.9 board in my pic. That’s what was still in the assembly documentation for the v1.0 at the time I did mine.
And yes, both times the pair of FETs on a given phase were blown. Based on the behavior, it seems very much like one failed open, I continued to ride it for a few days, and then the other failed short.
When replacing the FETs, I did add a small piece of copper to replace the trace.
1 Like
Agreed. But I’m guessing I’m having bigger problems if I’m blowing FETs and traces. I have a theory, but maybe not a good one. At one point I had the “DRV8301 OC Mode” set to “Current limit”, and my board struggled mightily to go uphill. Disabling that has turned it back into a mountain goat. Perhaps I should set it back to “current limit” and increase the value?
