I'm setting up a quad with a Holybro Kakute H7 v2, and a Holybro Tekko32 65 metal 4in1, 4.4.1, and can't for the life of me get the current sensing to work properly. I did a test flight and throttle pumps with biblades got pegged at 182A (46 per motor). Put tri blades on and got the same result. This is on a setup that should draw like 60A (avenger 2808-1500, 6s, hq 7040 triblade). This is using the FC adc, with the amp scaling set to 180 per holybro website for my esc that is marked "v1.7":

https://docs.holybro.com/esc/current-sensor-scale

I tried as well using esc telemetry, but that doesn't seem to work at all (zero current). I think there's a new bug with esc telemetry in betaflight due to tx/rx switching. However I also haven't read across anyone ever getting the 65 metal esc current telemetry working anywhere.

I figure I may be doing something wrong... but the math seem to line up with 3.3V limit on the adc (18mV/A at 182A). If it scaled to 5.0V then yeah, 65A would be in range. Can someone please shed some light and help straighten my brain out on this one? Anyone else deal with this problem on their setup? Anyone have an h7+metal stack with current functioning properly? Anyone have esc telemetry working on their Kakute h7 with 4.4.1?

Thanks in advance,
Kev

You may just need to calibrate ESC current.If you are getting a current reading it could be wrong and just need calibration.
  
 I have found that the current calibration may need to be adjusted after a betaflight update.
 


  https://oscarliang.com/current-sensor-calibration/

I just verified calibration with my BK Precision Autobench CP-3 current clamp, and current seems to be showing correctly using 180 scaling. I only checked it up to 20A to avoid injury at my bench lol. Maybe if I had an assistant to work throttle so I could hold the quad with both hands, then I'd feel less sketch about measuring WOT. It tracked 1:1 with my clamp meter from 0-20A though.

I have a background in circuit design... I'm beginning to wonder if this is a hardware design flaw. Maybe the instrument amp is misconfigured with too high a gain to stay under 3v3 for the required range of 65A. Doing the maths, like I said above 180 would be perfect if the bridge scaled to 5V output, but that doesn't appear to be the case. Maybe the h7 adc vref somehow can be changed to 5V in betaflight? IDK... I'm only familiar with fixed input references on older uc's.

I've seen others write about the same esc actually showing ~60A draw per motor in logs, though not sure which version they had. It seems totally unreasonable for a 65A rated esc to top out measurements at 46A. I could see this for early prototypes of early >45A esc's, but not the latest revision of a top shelf esc.

[edit: Speaking of Oscar, he did a recent review of this exact stack; H7+metal. I may have to ask him how the amps were working on his review unit.

...NM, just found it and it's really old, like Jan 2022. He probably was reviewing an older version. HBro made some changes to the current sense for 1.7, as indicated by the changed scaling factor.]

I'm giving up on this for now and just tuning/enjoying my quad... I sent Holybro tech support an email explaining my findings. We'll see what they come up with, but all the signs are pointing to a simple design flaw. I just scanned the h743 datasheet, and yeah everything is supposed to be <3.6V max. Sigh... all I know is I hope the expensive H7 doesn't get cooked by overvolting the adc pin! I feel like I'm in a 3-stooges episode lol.

I still haven't heard back from Holybro, and I have a feeling I know what the response may be anyways... "It is working as expected. Most people don't need more than 45A." I would be totally surprised if they said, "That is not OK, we will make another revision and send it out to you ASAP." Since that's not likely, and I really wanted the metal fets on my beast of a quad, I decided to take matters into my own hands and make a diy solution. I'll share it here in case anyone else is interested.

The 4in1 comes with 2x wiring harnesses for fc/esc, one long and one short. I cut the `cur` and `gnd` pins on the long harness and added a voltage divider with some spare 0.1% 1/4W resistors I had laying around. A 2210ohm on top, and a 3390ohm on bottom, forming an 0.605:1 divider for the current signal. I adjusted current scaling in betaflight to 109 to match the reduced signal. Now it reads accurately, and maxes out ~300A instead of 182A. Resolution is reduced as a result, but that's OK with how I fly it, lol. I heatshrinked the resistors into a ~1/4" cube and used some 3m VHT tape to stick it to the frame so it won't add vibrations. The coolest part is I still have the short unmodified cable for higher resolution if I ever use the 4in1 on a smaller quad that draws <182A.