Revolt RV-series motor review and comparisons

[larsb]

EG:s own summary on Revolt motor quality:

I did a bit more scrutinizing of the RV-100E and found a few things that are less than wonderful. Revolt is still failing on the details! GRRR! Horrific motors...no not really, but the entire product line does suffer from assembly and quality issues. The entire product line has things that are soooo easy to fix and aren't and then in other areas the attention to detail is really good. The stators are clearly dipped in lacquer after they are fully assembled, but the bearings are dipped too which means they get polluted with hardened lacquer. It's almost like machining and winding is done by one group and then assembly is done by uneducated dumb asses.

Just an opinion about Revolt...
I think they are cheap ass skin flints! They cut corners anywhere they can get away with it.
1. Sub-optimal hall placement.
2. Using pliers to smash a ferrol instead of a 6 point crimper.
3. Not using a crimp ferrule at all like in all the Regular motors.
4. Not using .3mm lams on the Regular motors.
5. Single short keys where dual longer ones need to be used.
6. Using unskilled and cheap labor for assembling motors.
7. Lack of attention to details.
8. Poor quality control.
9. Complete inability to fix or modify any existing design EVER to resolve known issues.
10. No included motor bases.

I pulled apart the phase ends on the RV-100E and all 3 of them were smashed together in a flattened ferrol and then soldered. Better than the butt soldering in the RV-120-regular, but still pretty pathetic. I took apart the 3 connections and redid all of them. From the factory, it will work, but Revolt could have put a bit more effort into this connection and used a crimper instead of pliers.

When I took apart one of them, I found that about 1/8" of wire ends were in the smashed ferrule. The smashed ferrol is 1/2" long and they had barely 1/8 of the winding end inside. Whoever wired up this motor really sucks!

I found this previously. Now I know why the phase end was so short. Pulling back those very short ends shown above, I was then at this bodge in the wind and was able to break it loose from the cured lacquer. This gave me the length needed to properly crimp and solder the winding end 1/4" deep inside the ferrol.

I used ferrules, crimped around the wires on 6 sides (not smashed flat) and then soldered to redo all the connections. They are each covered in 2 layers of heat shrink.

The halls and temp sensor come out in a 7 pin connector. I'm not overly fond of non-waterproof connectors so I replaced this with an IP68 connector.

The hall board is kind of cool how it mounts via 2 hidden screws that you access from the bottom of the motor.

I think I did this. The wires were super glued to the hall board and pulling out the board stripped off the wire insulation.

I pulled all 5 wires for halls, stripped new ends and soldered them back in place. While there, I repositioned the pull-up resistors and buried them in thermal glue. I can't imagine why Revolt thought leaving them unsecured was OK. The hall cable is silicon wires inside a silicon outer insulator...not bad!

 

[larsb]

Some other users threads on Revolt quality (i haven’t read a single user that was happy with the revolt motors performance):

Well, the threads will speak for themselves..

Revolt hall placement:
https://endless-sphere.com/forums/viewtopic.php?f=30&t=87305

Idiots guide to rewinding revolt:
https://endless-sphere.com/forums/viewtopic.php?f=30&t=85534

Johnnydrives thread:
https://endless-sphere.com/forums/viewtopic.php?f=28&t=86532#p1265709

Karolis unbalanced rotor:
https://endless-sphere.com/forums/viewtopic.php?f=28&t=80516&p=1251808#p1249066

Boisrondevens:
https://endless-sphere.com/forums/viewtopic.php?f=28&t=80516&start=75#p1237929

https://endless-sphere.com/forums/viewtopic.php?f=28&t=80516&start=75#p1238201

https://endless-sphere.com/forums/viewtopic.php?f=28&t=80516&start=75#p1240031

Wheazel:
https://endless-sphere.com/forums/viewtopic.php?f=28&t=80516&start=75#p1237956

Vasilisk:
https://endless-sphere.com/forums/viewtopic.php?f=30&t=75494#p1177299

Nextev review with excessive noload losses of rv160 motor:
https://youtu.be/6whYU1mLjlU
https://youtu.be/wRI_5cF3nUY
Read the comments on the second video..

This data was already on ES when you started your (quite positive) review. I hope none of the people reading here think that Revolt should be your next motor.. They need both new designs and new management to be a company that deserves business with their history.

 

[larsb]

I apologize to senior members who have technical disagreements with OP ElectricGod. The moderators of endless-sphere have a difficult and thankless job of parsing what to allow to be posted, and what to delete.

I do not agree with everything that EG posts, and yet...he posts pictures and experiments of motors that exist. Even if I disagree with any particular assessment he has made, the raw data of posting pictures of the internals of existing motors has a value that is difficult to assess. The value of these pictures is high and it is rare.

The only thing that is more valuable than tear-down pics of available motors, is a posting of the technical data of motors/controllers/batteries under load.

If you disagree with EG's assessments, the best course of action is to purchase the same motor, and start your own thread. Then post the data from your experimental variations in order to prove your point.

After you have stated your case, trust the reader to assess who is correct.

I purchased the same motor a long time ago. Twice actually since the motors didn’t work. Read threads in posts above, make your own judgement. It’s pretty clear what kind of company Revolt is.

 

[ElectricGod]

dead thread

 

[John in CR]

ElectricTrump strikes again. He obviously didn't read Larsb's new posts and immediately started his tirade.

The bottom line with the Revolt motors is that they're probably only useful at half the rpm and power claimed, and even then you're likely to need to put some significant work into one to make it reliable. Thank goodness I bought only one, and I'll replace a gas weedwacker motor with it and run sensorless with a 37V pack.

 

[ElectricGod]

dead thread

 

[ElectricGod]

dead thread

 

[ElectricGod]

dead thread

 

[major]

I have the RV-160-SH running on a 24 fet Nucular controller now.
After a bit of set up, the motor is running great.
At 82v I am seeing 6 battery amps and 10 phase amps under no load.
It's not the best for efficiency as that's 820 watts just to run the motor.
I'm sure with 3X the original copper cross section it will run a lot better than it did from the factory.

Hi,

Battery power is 82V * 6A, or 492 watts. Your figure of 820W is in error.

Regards,

major

 

[John in CR]

The Nuc seems to have it running more efficiently in terms of iron losses, and your extra copper will definitely help current handling while under load by reducing copper losses. What's the Nuc display saying is the new Kv with the rewind, so we know the rpm to go with the 6A no load?

Were you worried about copper strands singing? That's what the thick black epoxy is for on the long stretches of end copper and terminations in the 6 phase motors. I don't understand why so thick with the red stuff on all the copper on yours though, as it reduces surface area for heat dissipation on the part of the motor that generates the most heat. There's a high temp spray coating for motors that's easy to find, pretty much the same color too. I use it on the stator steel to prevent corrosion, but typically don't put anything on the magnet wire, since any additional coating is heat insulation, so I've never agreed with the guys who spray a thick coating on motor copper. The varnish on motor wire is pretty tough stuff, and you had it tied down really well with the string. I guess better safe than sorry, and no copper is moving now, but it will hurt cooling somewhat.

 

[ElectricGod]

dead thread

 

[amberwolf]

82v x 10 amps = 820 watts (phase amps)
82v x 6 amps = 492 watts (battery amps)

i'm often bad with math, but i'm pretty sure you can't get more watts out than you put into a circuit. (can't have more watts in a part of the circuit downstream from a section that has less watts)

you can't calculate motor watts using a simple "phase amps x battery voltage", because those phase amps are not *at* a continous battery voltage, they are at a pwm'd voltage that averages less than that voltage (usually a lot less).

i don't know what the actual formula is to calculate the motor watts based on voltage and phase amps (or if there is any simple one)...but battery voltage x phase amps can't be it.

I have the RV-160-SH running on a 24 fet Nucular controller now.
After a bit of set up, the motor is running great.
At 82v I am seeing 6 battery amps and 10 phase amps under no load.
It's not the best for efficiency as that's 820 watts just to run the motor.
I'm sure with 3X the original copper cross section it will run a lot better than it did from the factory.

Hi,

Battery power is 82V * 6A, or 492 watts. Your figure of 820W is in error.

Regards,

major

82v x 10 amps = 820 watts (phase amps)
82v x 6 amps = 492 watts (battery amps)

I listed both in case people cared.
It's not an error, it's being thorough.
<snip>

 

[ElectricGod]

dead thread

 

[amberwolf]

PWM doesn't adjust voltage.

it does--it adjusts *average voltage*. see below the next quote for how.

Each pulse is always at battery voltage. What changes is the duration of each pulse vs time when there is no pulse. In phase current, the PWM pulses (yes I know that's pulses twice) make more or less current flow in the phases depending on the pulse widths. The wave form of the current in the 3 class D amplifiers that make up the motor controller is a sinusoid. It seems logical that phase amps X battery volts ought to be correct. I could be completely wrong, but that's my thought process.

while pwm pulses are individually at or near the source voltage level, on average they are not.

if you had a single current path, like in a brushed motor, then the average voltage (that you could use to calculate watts) is the duty cycle of the pwm times the source voltage. for a pwm ("speed") based throttle control method, at half throttle, that would be about 50% duty cycle, and if that was 82v source voltage at that moment, no voltage sag, then it would be about 42v average voltage. you would then use 42v times the current thru the motor ("phase current") to get motor watts.

a 3phase motor is more complex, because you not only have the pwm, you also have commutation that turns the pwm itself on and off (and also reverses the voltage and thus current direction) to cause the rotation of the field that pulls/pushes the magnets in the rotor to spin it. i'm sure there is a formula to give you the average voltage for this, but i don't know waht it is. most likely the result still comes close to the simple brushed motor math, where about half throttle gives about half battery voltage as the average phase voltage, which would then be multiplied with the phase current (whcih itself is also an average, because it's also pulses), to give an average motor wattage. voltage is also out of phase with current, if you look at the waveforms in an oscilloscope, which also figures into this, though again i don't know the formulas for that.

Phase amps vs battery amps:
These 2 values can be wildly different. You can have 100 phase amps and 9 battery amps.

yes, you can. but the voltage changes to compensate for this, because power out can never be more than power in, for a total system (battery to controller to motor).

something you can pretty well bank on is that if you ever do math that shows you have more watts at the end of a system than you put into the system, there's an error in the math somewhere.

 

[madin88]

82v x 10 amps = 820 watts (phase amps)
82v x 6 amps = 492 watts (battery amps)

I listed both in case people cared.
It's not an error, it's being thorough.

If you want to convert 82V DC into 3-phase AC which is what the controller is doing, it cannot be 82V RMS on the output even at 100% PWM or WOT
It would be 82V peak to peak, but RMS voltage, which is what counts, should follow the factor 1.414 (square root of two) i think, at least thats how it happens the other way round if you convert AC into DC.

Additional to that on motor side there will be some voltage and current ripple which could lead to higher readings.

 

[sn0wchyld]

I have the RV-160-SH running on a 24 fet Nucular controller now.
After a bit of set up, the motor is running great.
At 82v I am seeing 6 battery amps and 10 phase amps under no load.
It's not the best for efficiency as that's 820 watts just to run the motor.
I'm sure with 3X the original copper cross section it will run a lot better than it did from the factory.

Hi,

Battery power is 82V * 6A, or 492 watts. Your figure of 820W is in error.

Regards,

major

82v x 10 amps = 820 watts (phase amps)
82v x 6 amps = 492 watts (battery amps)

I listed both in case people cared.
It's not an error, it's being thorough.

The many 28 awg strands need securing so they don't move about. I used lacquer first and then coated everything in several layers of electrical paint. The paint will get VERY hard when it cures from heating. This motor will be exposed to the weather and dirt and mud. The windings need protection from damage and yet still allow air flow. It's not perfect, but I'm going for protection and less cooling. This is still far better than how the motor came to me. It had covers under both bell ends so that zero air flow happened. I posted a picture a while back in this thread of the plastic insulators that used to be on the stator teeth. They are all melted. I doubt that will happen now.

In my test runs on the bench, at any RPM, the motor does not warm up at all no matter how long I run it. Checking the temp sensor, it just stays at room temperature. Obviously under load that will change. I'm pretty pleased with the improvements in motor performance I am seeing so far.

I'll post a video later...

The calculation of x*y=xy may be right, but the interpretation of this values 'value' is definitely in error...

specifically you're multiplying the output current of a buck/boost converter with the input voltage... In terms of power calculation they bare no relation to each other. There's <500W going into this motor, so there's no possible way short of re-writing our fundamental understanding of thermodynamics that the 'phase power' or 'copper loss' or whatever your trying to calculate can be >500W. If you still doubt it then measure and integrate the phase voltage - ill bet its not 82V. Put simply its not physically possible to have higher output power than input power. Pointing out the 10A phase amps is useful info - multiplying it with the battery voltage yields nothing of value. Its like multiplying the flow rate out of a water pump with the input pressure to that pump... it doesn't make sense, it has no use, it reveals nothing about whats going on in the system - despite each parameter (ouput flowrate/current, and input pressure/voltage) being potentially useful numbers in of themselves.

The degree of improvement in this motor are pretty surprising/interesting - have you tried measuring the no load with the same controller as last time? it'd be interesting to see how much of it is the rebuild and how much of it is the NUK. May well be that the re-build will only yield significant improvements in the loaded efficiency (though the NUK will probably help there too)... unless you also changed the bearings/#turns (i may have missed that bit)?

 

[major]

3-phase power.
Here's a short study I did a while back, for a different reason, but it shows some basic formulas.

https://www.diyelectriccar.com/forums/showpost.php?p=376274&postcount=9

Regards,

major

 

[John in CR]

To calculate power in using phase amps I think you'd have to multiply it by the "apparent voltage" that the motor "sees", something I have no idea (or need) how to measure. It's simple... battery Voltage while running times battery Current. Since no load current is so low the copper losses are minuscule and can be ignored, so the no load power (as long as you don't have gearing or a tire to turn, so no windage losses) represents the iron core losses (also ignoring friction losses of the bearings, which should also be minimal) at that rpm. That amount of heat is incurred whether or not the motor is under power. The number of turns of copper on the stator or copper fill % have nothing to do with this. These losses increase in a fairly linear manner with rpm after the initial plateau relating to (I believe) cogging torque is overcome. While this is an over-simplification of the hysteresis and eddy current losses in the iron and magnets it gives a basic idea of why no-load losses are important to us as they give us an idea of the maximum rpm we'd want to run the motor and peak efficiency.

Peak efficiency is the point where while under load the copper losses and iron losses are equal.

If the no load current has improved, then other than less friction losses in the bearings if they where changed, the improvement is coming from the controller...probably mostly due to the Nuc controller's hall placement correction, since there's been no change in the magnetic circuit (steel and magnets) of the motor. The big operating benefit of the motor rework is in all the additional copper fit on the stator teeth, because that reduces copper losses, the primary source of loss in our motors while in use. That means it runs cooler at any given power level (assuming heat dissipation is unchanged), so it's capable of handling more current and therefore make more power at a given rpm (as long as it remains below the current saturation point of the stator, which is unchanged from before).

 

[major]

...
Peak efficiency is the point where while under load the copper losses and iron losses are equal.
...

Hi John in CR,

I've heard this statement a number of times recently. I don't recall having ever encountered it last century when I was learning motor theory. Do you have any idea of the source or where I can find proof? Sorry to stray off topic.

Regards,

major

 

[sn0wchyld]

To calculate power in using phase amps I think you'd have to multiply it by the "apparent voltage" that the motor "sees", something I have no idea (or need) how to measure. It's simple... battery Voltage while running times battery Current. Since no load current is so low the copper losses are minuscule and can be ignored, so the no load power (as long as you don't have gearing or a tire to turn, so no windage losses) represents the iron core losses (also ignoring friction losses of the bearings, which should also be minimal) at that rpm. That amount of heat is incurred whether or not the motor is under power. The number of turns of copper on the stator or copper fill % have nothing to do with this. These losses increase in a fairly linear manner with rpm after the initial plateau relating to (I believe) cogging torque is overcome. While this is an over-simplification of the hysteresis and eddy current losses in the iron and magnets it gives a basic idea of why no-load losses are important to us as they give us an idea of the maximum rpm we'd want to run the motor and peak efficiency.

Peak efficiency is the point where while under load the copper losses and iron losses are equal.

If the no load current has improved, then other than less friction losses in the bearings if they where changed, the improvement is coming from the controller...probably mostly due to the Nuc controller's hall placement correction, since there's been no change in the magnetic circuit (steel and magnets) of the motor. The big operating benefit of the motor rework is in all the additional copper fit on the stator teeth, because that reduces copper losses, the primary source of loss in our motors while in use. That means it runs cooler at any given power level (assuming heat dissipation is unchanged), so it's capable of handling more current and therefore make more power at a given rpm (as long as it remains below the current saturation point of the stator, which is unchanged from before).

main thought re rewind itself helping is a higher turn count will lower the rpm (presuming testing done at same voltage) which will also lower no load loss, if only because of reduced rpm. dont want to go saying the nuk has made x improvement if the turncount is doubled form the last 'stock' test

 

[ElectricGod]

dead thread

 

[ElectricGod]

dead thread

 

[ElectricGod]

dead thread

 

[ElectricGod]

dead thread

 

[John in CR]

...
Peak efficiency is the point where while under load the copper losses and iron losses are equal.
...

Hi John in CR,

I've heard this statement a number of times recently. I don't recall having ever encountered it last century when I was learning motor theory. Do you have any idea of the source or where I can find proof? Sorry to stray off topic.

Regards,

major

I should have included "for a given voltage" to that statement. I'm not sure where I learned it here on the forum, but it's also true for transformers, which comes to the top in google searches. It can be shown playing around with the numbers on Miles' motor spreadsheet. I'm unsure how to prove it. For me it's one of those things I thought about when stated by guys who know tons more than I about this stuff and burned into my head, just like the fact that as long as current is limited by the BEMF of the motor (not controller current limiting) peak power occurs at 1/2 of no-load rpm.

Regarding peak efficiency, it make sense because it's 0% efficient at no-load. As you load it down torque and power go up as does current, but copper losses go up by the square of current. When the copper loss graph crosses the iron loss graph is peak efficiency, because copper losses go up so much faster while power goes up with torque and current.