to clarify, You’re saying that on VESC6 with the latest firmware, you are still getting this crazy delta of current draw on unloaded 63xx motors of different lengths. But you are not getting that on the stormcore.
4 Likes
@longhairedboy exactly that, one thing that had to be tested but I don’t have the means is if that these errors are proportional to current or just some small value that is always there. My guess is the first one, just on small currents, by small I mean even the currents that are seen when just cruising
My entire graduation thesis is a bit flukes because of that lol, glad I finished and already sent the final versions before finding this or I would have to do all again
@rich assuming that the value is right, it’s a constant. For example, a motor that need 4 A to spin at 5 k rpm, will always need that, be it running 4 and just using power to spin the motor or 80 A, and just 74 A would be used to produce actual torque, and on top of that you have to add the resistive losses due to the total current
So I’m your example, the motor that need more current just to spin will have less torque available to actually move you, but you have to consider that usually the motor that need more power to spin, is bigger and can sustain higher load for longer periods of time
As someone said in a similar topic this days, the most important factor in any board efficiency is the rider, and this is why data is needed, and good data, not like me that did 2 years of work on bad data lol
To be honest is not that bad that the results are completely wrong, since even with higher losses, the power used to move is a way bigger fraction compared to the power lost in friction, resistance and all that
6 Likes
if my ubox remote is to be trusted, at 170 pounds on a 4.4:1 ratio on 8" tires with my 6374s I’m usually drawing anywhere between 10 and 16 amps per motor maintaining speed. The numbers jump around too much when i jam the throttle so who knows.
I really need to get my robogotchi in there.
1 Like
do you have a 6389 sealed motor or a torqueboards 6355? If you just spun it at top speed with no load at any voltage we can see if the current reading is good
Currently the only motors i have that are not in the “motor parts” drawer are sealed 6374 maytechs, specifically my Totally Bitchin’ 190KVS which I’ve been using for like 3 years straight on every build that’s left my shop.
with the exception of the occasional TB DD build which i won’t do anymore.
2 Likes
Unfortunately that won’t work unless you do the whole test of putting a current meter on the battery wire since I have no idea what the currents should be for them
Sooner or later I will try to get my hand on a recent vesc 6
1 Like
I’m seriously curious about the results of this. I wish i was in a better position to help you figure this out.
3 Likes
Crowdsourced, we might be able to figure it out. I will respond the next time a 6374+ is on the bench, which should be soon. Right now I’m working on some 6355s.
2 Likes
I did spin it by hand. Both are chain drive setups.
The 56114 continued to spin around 1/3 longer than the 6374.
I did not measure any currents.
1 Like
I would be careful to compare values from a phase shunt design and a battery side shunt design in a simple test bed.
No load current depends on many factors, for example magnet strength and air gap, magnetic pole pairs vs coils, stator lamination and material etc. A motor with lower no-load does not necessarily need to be more efficient under a certain load scenario.
There are also different kinds of inefficiencies. Copper losses, Core losses, mechanical losses etc.
Depending on the RPM and load scenario the efficiency will differ a lot. Some torque rich motors have less copper losses but higher core losses at high RPM. Others have higher copper losses and less core losses at speed. It is all about the right balance for the application, average RPM, max torque output etc. There is no simple test bed without knowing all parameters. This could be average speed, gearing, ambient temperature, torque demand, desired form factor etc.
We just re-designed our motors to optimise everything as much as possible. Last summer when it was boiling hot we used the climate to do systematic tests under hot conditions and logged everything to evaluate the data. We then used the data to get all details right.
3 Likes
So . . . Current measurement errors on vesc 6 hardware is to be expected?
It is no error, you just measure either on the motor (phase) side or the battery side, depending on the design. VESC 6 is a phase shunt design, Stormcore or VESC HD are battery side shunt designs.
2 Likes
But phase side or battery side shunt doesn’t really explain why the VESC 6 reported battery current is so far from the multimeter measured current.
I’m no motor expert and I’m asking questions cause I’m curious to understand, but it seems to me that the reported battery current should match the actual current being used from the battery right?
4 Likes
The “problem” is that there is no dedicated shunt that measures all the overall current, I think I get what Frank means but I’m no expert so won’t make comments on it, on VESC 6 the shunts are after the FETs, on StormCore and HD they are before, but each phase still has its own shunt
The battery current given by the firmware is just an estimate based on the phase currents and duty cycle, so if the first is wrong, it will also be wrong
To be fair in a complete system the job of measuring the total current should be from the BMS, but most setups don’t do that yet
3 Likes
My 6374’s take around 3.5-4A to spin at full duty, but they are also quite unbalanced.
I want to say that motor can inertia also plays a role in real world efficiency. Switching from 6355’s (~550g?) To flipshit 6374 (960g) the acceleration and braking is overall a bit less snappy, even if the power is about 1.5x, and the efficiency took overall a 2-3wh hit, but that could also be because the additional available power to spend
1 Like
If you want to see no load currents, I usually look at the battery current at 95% duty. For precise experiments it is probably best to look at a dedicated shunt for batt current as you suggested. We can do that with the 18650 /21700 packs for example, since those have a dedicated shunt and are CAN linked to the system. I will try to figure out the differences in between what the VESC motor controller calculates and what the VESC BMS measures. A small diference will not really matter in real life applications though. Other factors will play a much bigger role in accurate real time data.
2 Likes
Cool, looking forward to that test
I don’t have the tool to compare as I said at the moment, but from all the data I gathered, I suspect my VESC 6 underreports energy usage by quite a margin, and that explain why I’m one of the few who get such a good mileage on 8" tires and why I never got close to what my battery should have in energy capacity, but won’t conclude anything until I can test that
The newer ones are constructed a bit different, yes.
1 Like
Did you ever do something with that football sized motor you showed last year (or 2019?)
1 Like
dont forget that to respect the safe limit of a cell its usable capacity is only about 80% of its total capacity and this will be even less under high load conditions.
1 Like