Many vendors are now making ESC’s and also making claims about the power output of these devices. At the moment were kind of trusting that vendors specified ratings are accurate however it seems there is not a standard test method for direct comparison between vendors.
Can we trust these ratings?
In a thread for a new ESC i noticed this coment by @Trampa
Trampa claims “We recently tested a 200A continuous rated clone device and it could only measure 120A reliably.”
This may or may-not be true. As it stands this is an anecdotal and unsubstantiated statement about a competitors performance. The statement is from a trustworthy source and @Trampa obviously know what they are taking about when it comes to high current VESC design. However they could also be seen as a biased source with a conflict of interest.
The solution . . .
The reason i started this thread is that as a community i believe we should be able to independent verify the data we are provided by vendors. Between us there must be people with the expertise or equipment necessary to test ESC’s and create some sort of standard to evaluate their performance.
It would be good it we could figure out a way to generate a pier reviewed method for rating ESC’s and making comparable datasheets based on real world testing, kind of like Mooch has done for 18650 cells.
I also believe that vendors should be held accountable for claims make about performance.
@Trampa
Do you have any data to substantiate your claims about “clone device” performance? Can you share this data with the community?
If your not able to back up your claims with data then are you willing to share your test methods or detail the equipment your using to perform these tests? This would mean that other community members would be able to replicate and verify your results.
Other vendors . . .
Are any of you willing to share information about how you test your devices and verify their ratings?
Just some telemetry and metr. reports ride after ride, itch set at the max recommended.
Right now I only keep V4.12, ESCape and VESC 6.
I would recommend ESCape anytime, this v6 like is like a v6 yeah, with fair price, and the dual ESCape case is a plus (imo) for a cool looking custom build.
Unfortunately, we all know that is no longer made.
Somebody have to contact Stewii and take his project back !
If you will run bldc V4.12 can do the job if you well refresh them.
Remember what Duffman pulls out of some 4.7 or 4.10 on his mad 4x4 monster (something like +125Amps, even more, when Jenso test it if I’m right) … yes, he make some specific copper heatsink but wow. I’m definitely going with 4.12 when I will try 4x4 for off road use.
The VESCs 6 are from the 1st Beta tester bach(summer 2017 if I remember well), still working perfectly never get real trouble, except 2 times. Once the slave lost is parameters, and another time it was both (it was at an event, what a frustration, it was almost the end but … ^^)
Can report the same on the ESCape I haven’t ride them a lot over this 2 yrs. So they sound better in a value for money to me.
VESC4.12 clones throttle hard when driven at 29A (battery)
FOCBOX 1.6 doesn’t throttle at all at 29A with TB6384 motor
FOCBOX 1.7 throttles some at 29A with Turnigy sk8 “6374” motor
@Kug3lis Now this is how you cool down an ESC with plastic body MOSFETs, maybe you can start turning your aluminum cases into something useful given the pretty price.
@Riako do you know if the guy sells them or the files are available eventually for custom CNC?
So I’ve been thinking about this further. if i was buying an ESC i would want to know . . .
What the maximum burst current rating is.
What the maximum short term current is (when the heatsync is cold)
What the maximum sustained current is once the heatsync reaches the thermal throttle
Maybe if we measured temperature at different current loading then we could generate graphs of load vs temperature allowing direct comparison between esc’s.
May I prompt for a foc_openloop test as a kind of lab test.
We basically only have set as a community on some testing variables like which voltage to test with a high amp constant voltage supply, how high the ambient temperature may be and it’s diviation (should still be measured tho) for example.
Other things we have to set in stone are if we want to measure time till thermal throttle, max amp till throttle, or just try to validate vendor rates …
Like: “50A continuous for Xx VESC based ESC xX seams reasonable it hold up without thermal throttling in a 10 min load test, reached 78°C at the end of the test”
Sure it will only be a “lab” thing, no real world but easily to do since foc_openloop is VESC Tool build in (you should mention foftware/firmware version while testing btw) and quite easy to be validated by other people.
My thoughts would be to call something worthy to be called “reasonable rated” by vendor spec validation in a 5 min continuous load test (2 confirmations not more than 5 min cooldown time in between) .
Other parameters:
voltage: 36V
motor: 6374 (some common motor like a SK3 submerged under water, we may have to swap to better connectors than 4mm bullets tho since they may melt on high current. Maybe go right away with 5.5 mm ones)
ambient temperature: 20+/-3 °C (still measured by an external thermocouple tho)
temperature reading taken by VESC Tool MOSFET temp + external case thermocouple on the heatsink if applicable.
For those that don’t know about VESCs foc_openloop
you may jump to 7:45 min
I did something like that on the MakerX Mini Foc (one of the first versions). Sorry my english is embarrassingly bad, I really should had done some preparation like scripting and a teleprompter maybe.
Not a very scientific data but maybe you can narrow some inputs from Metr recording based on how much amps you can pull before you meet thermal throttle?
I recall vidéo tests back in the day showing the differences in various conditions.
What I got more or less is : if you wanna hammer your VESC continuously keep it under 60 degrees Celsius. Now about the max absolute amps, I’m interested for sure in this topic.
Yeah this is a good thought the metr data could be a good starting point only things that make it hard to normalize that data is the atmospheric conditions (outside temp primarily) that will throw all the numbers off.
Also not sure how accurate the altitude data is in there… Guess if just comparing amps and heat it’s sort of irrelevant.
this is very interesting. i have never head about the “foc_openloop” command before but this seems like a very easy way to find out how much current an esc can deliver for sustained periods of time.
I guess we would just have to standardise some values for thermal cutt-of to make comparison fair between different ESC’s. do vesc 4 and vesc 6 ship with the same default thermal throttle settings?
You have to keep in mind that there are many canceled variables in the test it’s not 100% applicable to the real world, however it should be valid to test the max Power the heatsink (well how much the setup, but as long we only change the ESC it makes them comparable) is able to dissipate. We have to make sure however that
there is limited air flow next to nothing is best since most people enclose them anyway
the ambient temperature should be as steady as possible
the power source needs to be stable, in theory a capable battery is fine but PSU prefered obviously
we should also be fair enough to agree that rounded amparage by 5A increments are probably the max accuracy we can deliver (meaning that if a ESC fails the test we may try again ,cooled down ofc reducing the current by 5A)
during my testing I figured that attaching a thermalsensor to the outside of the heatsink may be, even though being useful, kinda subjective without a themalcam to determine the hottest spot. So has to be taken with a grain of salt.
Overall all this type of testing has to be taken with a grain of salt without various people more or less agreeing on similar results (as always), also as this isn’t any official stuff it’s fair for any vendor to call this BS there is no official guideline what “continous” or “burst” nor is for testing just a kind of thing to compare nothing set in stone.
yes i guess this test shows how well the factory heat sink is able to dissipate power and gives an indication of the maximum sustained current a user is likely to see if they enclose the ESC without any additional cooling.
this is very useful as 2 ESC’s with the same current rating may perform very different in real work conditions if one has a large factory heat sync and the other does not. i feel like this may be a good metric for how well an esc performs under stress.
Its also easy for a user to replicate at home, and could be used for testing the effectiveness of an additional heat-sink.
If we were to use the watt value rather than the Amp value then would voltage fluctuations be less important whilst maintaining repeatable results?
Also if foc_openloop is set to something ridiculous like 500A are we likely to damage the ESC or will the vesc stop its self from operating outside safe limits?
I think the max Amp limits (as long you don’t flash unlimit firmware) should still be present.
BTW I said high stable amparage PSU, but actually it’s reasonable, you use and blow the Power the VESC dissipates + a bit headroom so a 36V 600W PSU (of decent quality) for example should be fine for +90% of the ESCs out there. (but the cheap usual 5A lab PSUs will be to weak)
I wouldn’t change the 80°C MOSFET temp limit for example, you shouldn’t use it regularly above that anyway and it’s after all the stock value.
But sure if the temp sensor is broken, the current measurement bad it might be killing the VESC but if you tend to push the boundaries of this kinds of ESCs they soon fail anyway.
PS: By any means I’m no expert for any electrical stuff (just a usual DIYer) and anyone with more knowledge should correct me.
I’m just someone who deals with testing stuff for work (mostly standardized tests, but as not everything even in Germany has a official guideline, DIN or agreement we sometimes have to come up with our own at work)
I will not tell you which model it was since that would be not fair. It’s commercially available.
We will not test third party ESCs and publish results. In this case a customer had problems and finally switched over to the VESC 75/300. We found out that the 200A ESC was actually less capable than the regular VESC 6 and that was simply not enough for the customers project. So we found out by accident…
If you have a look at the PCB layout and shunts used, you can pretty much tell what to expect max current wise… It was no big surprise in this case.
Testing ESCs is not easy. Driving nasty motors helps to get an idea of the capabilities. Benjamin Vedder has quite a selection of motor to hand, so he can strap a new ESC to all sort of motors. For FOC-Openloop testing it is good to have a 15KG motor that does not heat up fast and can swallow 300 to 500A. Or take a smaller motor and run it in a water tank.
It is also good to have some motors that are not easy to detect and run. Motors with high pole count and Inrunners with low pole count, motors that need high switching frequencies, low inductance motors, and finally you need a load that you can strap to the motors.
It’s also interesting to push ESCs to the limit at very high ERPMs.
It’s interesting to changes loads quickly (e.g. full throttle forward to full throttle reverse)
Running a ESC as a generator for a while is fun.
Thermal cameras help to see hotspots and judge temperatures of individual components.
EMC test chamber helps to check for EMC problems.
Finding the worst case:
For the VESC 75/300 it was not super easy to find the worst case to test with. Finally we got hold of a turbine with super low inductance motor and preferably 70Khz switching in FOC. These tests made everyone want to keep some distance to the equipment. Really scary!
However, bold claims of fantastic BURST ratings, far beyond the continuous Amp ratings should make you sceptical. It simply makes no sense. This is typical RC-Equipment kind of advertising. Divided by two is about right.
Thanks for the input here, I think everyone can agree on your point about third party ESCs.
foc openloop is just the easy thing that came up to my mind, something everyone could do, so easy to verify if they wish. @ducktaperules point about some kind of community “validation” about the ratings of all the upcoming ESCs is absolutely reasonable though.
I think it’s fine to discuss some kind of guideline beforehand, than let it be tested by a few members and than make a list about the result and if they hold up (at least vaguely) to spec.
You are right that there are far more parameters than just heat built up for performance of a ESC (at least for me). But testing isn’t easy, also the guidelines unset so something to start with.
Sure a close up look on the product by a somewhat experienced DIYer and he should be able to determine if claims are able to hold up by some degree, ridiculous claims at even more ridiculous prices are always suspicious. At the other hand people will always try with the least expensive option, probably at least once or twice.
BTW
Burst ratings overall are bullshit just as they are for batteries.
yeah, with fair price, and the dual ESCape case is a plus (imo) for a cool looking custom build.
never get real trouble, except 2 times. Once the slave lost is parameters, and another time it was both (it was at an event, what a frustration, it was almost the end but … ^^)
