TIL. flying probe testers. cool. 
3 Likes
oh crap hey, speaking of tombstoning, my R22 on my current board was vertical. I assumed that was due to some issue on my part, but in retrospect it seems unlikely that that happened due to some issue I caused. I thought it had exploded and become vertical, but maybe it arrived that way?
1 Like
R22 is the pullup resistor for the fault signal from the DRV. Without it, fault detection from the DRV won’t work.
It seems that there’s a correlation between lot 88ASYCTG4 and faulty DRVs. @doomy, @Fungineers, @alwaysmohawk, @stratoglide, @Minimadness, and @Style all have had issues with DRVs from this lot number.
6 Likes
So I’ll be preparing a warning statement about the issue for this thread and the github repo. I’m also working on investigating this particular lot to see if it was flagged as a bad batch from the manufacturer and LCSC/JLCPCB got a hold of it somehow.
9 Likes
For those that were interested, I’ve got a pretty good chunk of the how-to guide up for the controller with LCD display and logger here:
I’ll be adding to it and hopefully “complete” it over the next few days. The plan is for all relevant information to be contained in the first 5 posts of the thread. So I’ll continue to update those posts as I get feedback.
4 Likes
how can I do more rigorous testing? do I need a metr pro? (why are they so expensive?)
It has hands down has the best customer support of any tech product known to man. @rpasichnyk Has done an amazing job, crushing bugs, and adding users feature requests.
Once you get yours you’ll understand it’s money well spent.
The thing is also built like a brick shithouse.
Everything else in your build might fail, but that thing just works.
*I am not sponsored or related to Metr any way, I just love that product.
More testing is always good, but don’t try to beat @shaman claimed specs:
Or else It will blow up! If you have attached a heatsink and also give the heatsink airflow then from my testing experience on 13s and around 2kw of power I could not get it over 45C. It ran so cool, which was my reasoning behind tring to push even more current, which ultimately lead to bunch of dead CFOC2s.
3 Likes
I know nothing of the metr pro, but you can certainly log data from the UART using an arduino or similar.
1 Like
You can use cheap nrf51822 modules from ebay and the vesc mobile app. When you want to start logging just put a checkmark at “Enable RT Data Logging” think of it as a start stop button. This will put a csv logfile in the VESCMonitor folder. If you want to analyze the logs and look at graphs, copy this file to you PC and open it in vesc tool.
6 Likes
I checked my three FOCers now. They are also with 88ASYCTG4
One of them has the beeping noise and can’t regulate 5V.
The other two seem ok
2 Likes
Just sent JLCPCB a stern email, will post reply’s here. I have ordered new DRV8301’s in the mean time.
1 Like
Thanks for reporting in! The audible ringing and failure to regulate seems to be the common symptom. Seems at least that particular lot number has a high failure rate. Still not 100% sure if that lot was bad from the factory or if JLCPCB mishandled that lot once they got it or something in between.
DRV8301 chips of lot number 88ASYCTG4 suspected of high failure rate
Several users and participants of the CFOC2 v0.9 have experienced failure from DRV8301 devices of lot number 88ASYCTG4 from JLCPCB’s SMT assembly service. This was originally thought to be inadequate soldering but has since proven to be something more complicated. This particular lot has shown to have a high failure rate though not necessarily always failing. Symptoms exhibited are audible ringing from the inductor close to the DRV (L1), deregulation of the 5V output from the DRV’s buck converter, and subsequent catastrophic failure of anything connected to the 5V rail (CAN transceiver, ESD diode, 3.3V LDO, ect)
@doomy has discovered a test to recognize bad units before powering on. On an unpowered CFOC2, perform a diode test with a multi-meter that has this feature from supply positive to supply ground and then reverse. Results should be close to the chart below
Good
NEG - POS + 2.615
NEG + POS - 0.555
BAD
NEG - POS + >3V
NEG + POS - 0.576
I am attempting to track down this particular lot number all the way to the manufacturer to see if this was a known bad lot that somehow ended up in JLCPCB’s stock. If you have been affected by this occurrence, I encourage you to contact JLCPCB’s customer support and seek a refund for the failed DRV units.
12 Likes
It is very interesting though that this issue manifests pretty much identically to that which is caused by an unconnected/badly connected ground pad (inability to regulate 5v/abnormal diode test measurement).
1 Like
And that’s why I though it was a soldering issue at first…
board is still running well, i finally charged my 14s lifepo4 battery to 3.65/cell (51.1V), and was getting some braking issues.
It was cutting out, so I would let off the brake, then back on, and that usually brought them back.
looks like I was throwing a bunch of over voltage, and a couple over abs current, codes:
The following faults were registered since start:
Fault : FAULT_CODE_OVER_VOLTAGE
Motor : 1
Current : -11.4
Current filtered : -9.6
Voltage : 57.14
Duty : 0.014
RPM : 2083.4
Tacho : 553775
Cycles running : 45172
TIM duty : 116
TIM val samp : 4200
TIM current samp : 4200
TIM top : 8400
Comm step : 0
Temperature : 36.29
Fault : FAULT_CODE_ABS_OVER_CURRENT
Motor : 1
Current : -10.6
Current filtered : 48.9
Voltage : 39.74
Duty : 0.512
RPM : 25431.2
Tacho : 1160140
Cycles running : 92970
TIM duty : 4299
TIM val samp : 4200
TIM current samp : 4200
TIM top : 8400
Comm step : 0
Temperature : 35.00
I guess it seems like the board was protecting itself, which is great, but I’m wondering why it was seeing such high voltages, and why it reported abs_over_curr when it was only 48.9A
motor: 60,-60,90
battery: 27,-7
3 Likes
I just opened a quality complaint through JLCPCB’s order page menu.
What are the chances we will receive the refunds?
I have just looked at by DRV batch, it is 88ASYCTG4. So on my side, 8 good (at least for now as I remaining components are still on the way), 2 are bad. However, I am just a bit worried that even though DRVs seems good now, are they capable of delivering intended performance? It would sad to throw 200€ out of the window.
Also, as I a not EE, I had discussions with some good ones and they also were interested in some PCB design decisions of FOCer 2, especially thermal reliefs as based on their opinion, these might lead to poor soldering during SMT process. Results also might highly depend on position of your PCBs on the conveyor (if you are lucky to be somewhere in the middle). If that is the case, there still might be some poorly soldered contacts that can deteriorate overtime. @shaman What is your experience regarding this? Do you think it is possible to have some hidden potential problems?
Thanks! and really appreciate the work you have done!
1 Like
They offered me to refund 25 DRV chips.
Just to clarify, when testing your DRVs with the diode tester method you need to use one that does 3V. Otherwise this test is useless.
Good
NEG - POS + 2.615
NEG + POS - 0.555
BAD
NEG - POS + >3V
NEG + POS - 0.576
Also the 0.55V measurements are just for reference, only a 2.6V measurement means a good DRV.
7 Likes