RC motors.. an honest answer ?

OK, I have delved into these RC motor powered systems and built a test mule friction drive , but i am having doubts about the practicality of the RC drive system.
No doubt these systems "work" , but how consistent and reliable are they in practice.?
I have frequent "sync loss" screeching halts ! that no amount of ESC programming seem to resolve.
This is not a extreme set up, just a Turnigy 63 -54 250kv, with a Turnigy Kf100A ESC, running from 18.5 V lipo, ...so all simple stuff ?
I am used to a brush motor smooth power and control, not this unreliable, "all or nothing", frustrating RC system !
so, what am i doing different to anyone else with RC systems, ??
should i spring the $$$ for a CC ESC and hope things MAY improve ?
should i experiment with hall sensors in (or on, or near) the motor ??? ..and then look /make a suitable matched controller ??
do i need to forget these "model kit" systems and look for "real world" industrial / commercial equipment ?
Why do i suspect that any /all of these options are not going to result in the powerful, reliable, modern drive system that they promise in theory. ??

That esc is a pile of feces.

Castle does a much better job.

All rc escs struggle at very low rpms.

RC is not the easy way. Its the best way of you do it the best way. Its awful if you cut corners in the critical places.

Hmmm, I have well over 1000 miles on my current friction drive and it's still working great. Very smooth throttle too. I guess it kinda depends on your setup.

My gf's drive now has around 100 miles on it and it's still going strong with the cheap mystery controller. Go figure.

I personally think you need variable pressure and a small roller to really avoid problems.

EVTodd said:

I personally think you need variable pressure and a small roller to really avoid problems.

Click to expand...

Thanks EVTodd, i respect your experience.
however, i am sure the mechanical configuration ( or even the fact that it is planned for bike use) has no influence on the problem.
this motor will sometimes "loose sync" when simply free running, though normally it is when under load... and not even a low revs... mid speed, before there seems to be an acceleration "ramp" .
The acceleration curve is not linear, even in "soft start" mode, it seems ok'ish up to about 3k rpm , but then it seems to ramp up to max very fast.
LFP.. others seem to have used this ESC with apparently few issues, and i have read where others have had similar issues even when using the CC device !
But you both seem to be saying similar things,.. that RC systems are not good at lower rpm's ??
Thats fine but as i say mine seems to "fail at 2k-3k rpm... not exactly low speed.
Will a properly configured sensored motor & controller run better at low rpm ( <1krpm) ?..
...and what exactly is that motor & controller combo ? :lol:

If it's losing sync when there's no load then I agree, it's a controller (rather than mechanical) problem for sure. That cheap esc I keep talking about can run VERY slow with no load and never misses a beat. Heck, it can run really slow with a load and rarely loses sync.

What are you using for a throttle?

EVTodd said:

What are you using for a throttle?

Click to expand...

I have tested with the turnigy Servo tester and a modified Stester + throttle set up . Same result either way. :?

However ....
I recall this comment from Kepler on the throttle build thread ..

...I wonder if your setup has lost resolution and pulse width is increasing too quickly at the lower end of your throttle range now.

Click to expand...

I am beginning to think that this turnigy Servo tester may be part of the issue , since the throttle response is not very linear even when just using the servo tester ?
Is anyone else using this Turnigy servo tester ??

I've a fair bit of experience with motors similar to yours. In my view, RC ESCs are a very mixed bunch when it comes to being able drive motors smoothly and with good torque at low to moderate speeds. They aren't designed or optimised for this, as it's not something a model aircraft needs.

Certainly those running CC controllers don't seem to have too many problems, neither do some running some of the cheaper ESCs. The problem is that it's hard to predict which cheap ESC will work well with any particular motor - quite a few exhibit exactly the symptoms that you're experiencing.

I couldn't be bothered with this uncertainty over which ESC would work, didn't want to spend the money on a CC one that would almost certainly work well so I opted to try Hall sensors and conventional ebike controller. Having experienced the simply incredible start up and low speed torque when Halls are fitted there's no way I'd revert back to playing with RC ESCs again. Having said that, my milling machine is running a Turnigy Aerodrive 63-74 170Kv with a cheap 120A Swang Hobby ESC and seems fine - just luck I guess. I will fit it with Halls and switch it to an ebike controller some time though, as the lack of very low rpm torque is a very slight problem for some jobs.

Jeremy

Jeremy,
how much influence on the motor "response" will the throttle + modded servo tester circuitry have ??
I know i am getting the full rev range .. 0-4700rpm ( 250kv and 18.5v) but for sure the response to the throttle seems more exponential than linear,
even in "Soft start" mode. !
I am curious as to if that is a characteristic of the ESC or the throttle interface (turnigy servo tester)
I am contemplating the Halls mod ( something new is always a challenge !) but then what will be my controller options ( model ESC's again , Yi yun, or Lyens etc?) for 5s (18.5v ) and 50-70 amps ?

The ESC should respond linearly to a change in pulse width on the input signal. The way these RC ESCs work is by looking for a pulse that is sent every 20mS (0.02 seconds) from the servo tester. If that pulse is 0.75mS wide, then the ESC interprets it as being zero throttle. If the pulse is 2.25mS wide then the ESC interprets that as being full throttle. The throttle setting for any intermediate pulse width between 0.75mS and 2.25mS should be linear, i.e. 1.5mS should be half speed.

It's most likely that the servo tester doesn't have a very linear response to the pot rotation. They don't need to be accurate for their designed purpose, as long as they can swing a servo from full travel one way to full travel the other way they will do the job they were meant to do. It may be possible to make one more linear, but to be honest I'm not sure it'd be any easier than just binning it and going for something like the thing I described over on the throttle controller thread. That has the added advantage of running from a Hall throttle directly.

Having said all this, I'm not sure that the throttle response has much to do with your motor losing sync - that's probably just a motor/ESC compatibility problem, as it seems to be common with cheaper ESCs.

Jeremy

so, is it "pot luck" putting these systems together ? or is there some kind of guide for guessing the best match of ESC for any particular motor ?
are low KV motors any more "sensitive" to the ESC ?
is going with something high $$, like a CC or Novak ESC, a sure solution , ? or just a higher stakes gamble ?
The Hall sensors seem like a safer bet, but then how do i know which controllers would be compatible ? :?

I can vouch for the Castle Controllers i cained the piss out of my HV160
280amps 10 000watt regularly never once has it lost sync (nor did my hv110 for that matter...not once)
and it didnt complain. Yes i blew it but that was user
error trying to make it run a motor that was shorted out DOH! :-(

I wouldn't bother even trying anything but a CC ESC if you stick wih rc escs...

Halls might be a better option for you ... still haven't seen anyone run a big Turnigy pushing big numbers on an Infineon successfully without smoking it though. Then i haven't seen a frock motor running big power that doesnt blow them either Doc Bass did 5 in a week on his X5.

KiM

AussieJester said:

I can vouch for the Castle Controllers i cained the piss out of my HV160
280amps 10 000watt regularly never once has it lost sync (nor did my hv110 for that matter...not once)
and it didnt complain. Yes i blew it but that was user
error trying to make it run a motor that was shorted out DOH! :-(

I wouldn't bother even trying anything but a CC ESC if you stick wih rc escs...

Halls might be a better option for you ... still haven't seen anyone run a big Turnigy pushing big numbers on an Infineon successfully without smoking it though. Then i haven't seen a frock motor running big power that doesnt blow them either Doc Bass did 5 in a week on his X5.

KiM

Click to expand...

Thats not true.. Kim I have run a big motor ( halls fitted ) with my nearly stock 12fet controller at 70A @50v and also with a baby highly modded 6fet without problems .

gwhy! said:

AussieJester said:

I can vouch for the Castle Controllers i cained the piss out of my HV160
280amps 10 000watt regularly never once has it lost sync (nor did my hv110 for that matter...not once)
and it didnt complain. Yes i blew it but that was user
error trying to make it run a motor that was shorted out DOH! :-(

I wouldn't bother even trying anything but a CC ESC if you stick wih rc escs...

Halls might be a better option for you ... still haven't seen anyone run a big Turnigy pushing big numbers on an Infineon successfully without smoking it though. Then i haven't seen a frock motor running big power that doesnt blow them either Doc Bass did 5 in a week on his X5.

KiM

Click to expand...

Thats not true.. Kim I have run a big motor ( halls fitted ) with my nearly stock 12fet controller at 70A @50v and also with a baby highly modded 6fet without problems .

Click to expand...

I said BIG power GWhy 50v 70amp is whimpy at best when a HV160 is 200amp rated at 50v and capable of near 300amp (actually Recumpence has recorded 400amp with his HV160 3220 setups) Stiull wainting for evidence of your 6fet controller running a Turniigy at 8 plus kw too GWhy...when i see that i shall change my opinion, seeing been waiting near ~3 months since you stated you would have results that weekend i have little faith its capable sorry buddy And as i have also said...sevral time, i would absolutely LOVE for you to prove me wrong.
I know Thuds tried beefed up 12 and 18fets and pops them with ease with his Turnigys how are you able to do it GWhy? Only thing i can tell is your running extremley EXTREEEEMELY low geared setups which will obviously be alot more forgiving on the controller, we need controllers capable of launchuing a bike hard and pulling hard to 70 clicks. Your 6fet do that GWhy?

KiM

Can anyone imagine a motorized bottom bracket with a high powered ASTRO motor?

Me thinks that approaches the holy grail of ebikes.............

Tom Tom said:

Can anyone imagine a motorized bottom bracket with a high powered ASTRO motor?

Click to expand...

Do you mean powering the bikes original crankset and thus using the bikes original chain to
drive the bike... that would be clever wouldnt it i wonder why nobody has done that with an Astro motor?
you would need some sort of freewheeling crank mechanism so the pedals didnt spin around fast when the motor was at full speed though
something like a freewheel whithout teeth on it that would someohoe 'screw' onto the crank itself to which you could
attach two sprockets, one sprocket could have the astro motor hooked via chain and the other the standard bike
chain driving the rear wheels!! Wait WOW that would even allow you to have MULTI RATIOS thanks to the bikes gearing!!!!
Mr Recumpence perhaps thats something you should look into doing? Miles
perhaps you could do us a CAD drawing of this concept! Ill get onto White Industries and see if they can develop
some form of freewheel that we can use that will stand upto this power :lol: Just playin Tom Tom
i gather you mean something like this right



http://dvice.com/archives/2009/05/gruber-assist-e.php The guys over at Motoredbikes.com
still think the Guy was using it in the tour de france race to the poor dumb bastards everyone
knows Tor riders don't use mechanical aids, they don't need it with all the roids pumping through them

KiM

AussieJester said:

I said BIG power GWhy 50v 70amp is whimpy at best when a HV160 is 200amp rated at 50v and capable of near 300amp (actually Recumpence has recorded 400amp with his HV160 3220 setups) Stiull wainting for evidence of your 6fet controller running a Turniigy at 8 plus kw too GWhy...when i see that i shall change my opinion, seeing been waiting near ~3 months since you stated you would have results that weekend i have little faith its capable sorry buddy And as i have also said...sevral time, i would absolutely LOVE for you to prove me wrong.
I know Thuds tried beefed up 12 and 18fets and pops them with ease with his Turnigys how are you able to do it GWhy? Only thing i can tell is your running extremley EXTREEEEMELY low geared setups which will obviously be alot more forgiving on the controller, we need controllers capable of launchuing a bike hard and pulling hard to 70 clicks. Your 6fet do that GWhy?

KiM

Click to expand...

Your not going to let this rest :lol: the 6fet will only be reialble upto 4200w or it can be pushed upto 6-8kw but could pop ( this is with the standard fet package installed ). A 12fet will be reliable upto about 7200w but could be pushed upto just under 11000w alas the 18fet can only be the same as the 12fet but will remain slightly cooler, Now the 24fet should be relaible upto around 12000w and could be pushed to around 20000w but the mods on the 24fet would need to be very major to handle these sort of currents but in general these are what the absolute max that can be acheived with these controllers. I agree the gearing on my bike do play a major role as to why I have no reliabilty problems with any of my controllers, Thats why Im making a 2 speed gear box so I can make my cake and eat it 0-25mph in around 2sec then another 4 secs to get upto around 60mph oh yes happy days all on a 3250w motor why would I want more than a 5kw controller 8) . But I have bought myself another big motor to play with and test whats going on with the controller with these big'uns which I have already said a while back and it just getting the time to do everything I want to do.

Gwhy,
How about giving us a little math to support your statments?.....It looks like the limiting factor on the TO-220 FET's are the current carrying capacity on the legs themselves.....after a few seconds amp draws drop from the whopper peaks settle into the managable realm, But for a "controller smoking" run of 30 minutes we prolly need some active cooling.....I am liking the submerged in mineral oil & flowing coolent senario the more I think it over.

So far with my sensored controllers, I am not impressed with the heat I am generating in 10-12 minute runs, On motors geared soft enough to let the controllers survive for 20 minutes. Am I a complete idiot with speed controllers? Or I am riding far more agresivly than any of the sensored controller guru's offering free advice are comprehending.

I don't care what a Cycle anylist says my peaks are....I am searching for a hard 30 minutes at or near peak levels...the TO-220's won't perform in that regard by my math. (without some active cooling on the whole board)

On topic: Frankly, for a general ride around Bike Rc is not the way to go. They are for wide open riding & require another level of comitment to fit & finish. For a Successful rc build you need 3 parameters to perform in harmony.
#1 realistic expectations (regarding your budget for batteries, controllers, motors & performance desired)#2 A full understanding of gearing your chosen components to suit their operating range. #3 The comitment to only operate the set up inside the parameters you designed.(not as easy as it sounds)

Don't bother with any thing other than Castle controllers on an rc build (unless your doing a Keplers style drive...the only exception I will allow ) The inrunner Astros are far more controller friendly regarding durability of rc set ups. (has any one burned up an astro yet? or Blown the crap out of a CC controller with an astro, without going crazy with gearing?) Like any thing, you pays your money & take your chances.

I burned an Astro up. But I was trying to do something outside of what it really should of been. Even though I discussed it with Astro-Bob first. I tried to run it on 24 volts at power levels that only an Etek would be happy with. But again that is running it outside of where it should be. On my Mountain bike it is very happy. I only have maybe 100 miles since switching to the Castle Ice160 over the Turnigy's, but so far it is superior in all aspects. I still try my best to run it with design specs...and usually do...so far no issues.

Thud said:

.....It looks like the limiting factor on the TO-220 FET's are the current carrying capacity on the legs themselves.....

Click to expand...

Nope. The limiting factor is the ability of the FET junction to conduct heat away into the heatsink fast enough to keep the junction temperature within safe limits.

This is way lower than either the lead limit or the semiconductor limit and is the real killer for all controllers that use the puny TO220 package.

For example, a good TO220 FET *may* be able to lose around 0.5 deg C per watt from the junction to the heatsink. A really good heatsink *may* be capable of dissipating around 0.5 deg C per watt. This means that the junction will, at best, reach it's limiting temperature at somewhere less than 120 watts dissipation, most probably a lot less than this for a typical heatsink.

I've yet to find a FET that will truly deliver it's spec sheet current rating in reality - even the mega packages still end up being thermally limited more often than not.

Jeremy

That is good to know Jeremy,
Thank you for the correction.
Care to opine on a flooded system? is it fools play?
I can fab polycarbonat enclosures & fluid tight wire junctions....& have a poor-mans understanding of water cooling from my racing days on the motocrossers.
I guess I am just looking for someone to double dog dare me to attempt it. :lol:
How far off track can we lead this thread?
sorry for the detour HH.

An honest answer

Well firstly I think coupling an RC motor to any bike has issues from what I have seen and experienced first hand for myself, I am beginning to think that I missed a meeting or is it a case of the emperors new clothes but how on earth have we all gotten back to friction drives? :? :? there is only one place for them IMHO and thats in an e-bike museum? I can understand folks wanting to fit them to bikes where its impossible to fit any other kind of assist but why on earth fit them to anything where any other kind of system can be fitted is beyond me.

A few of the top guys on here have managed to eek out some decent performance from RC set-ups however the controllers do seem to be the weakest link here, I see far more pictures of burnt controllers and other components than I do of happy ride reports, the Jury is also out with me with regard to the overall efficiency of an RC system, when you take in to account all the issues of the high rpms and reductions needed you are wasting a whole lot of energy before you even put it to the back wheel, I havent seen anything to convince me that RC systems are reliable as bike drives or efficient even when compared to direct drive hub motors :?

I think folks are expecting a lot from these little motors and controllers and I don't think that folks can be too hard on this kit as it was never designed to do what it is being asked to do, I think if the manufacturers knew how the kit was being used they would not likely warranty its use, these tiny little motors may be able to run whilst bolted to the front of a plane with 100mph cooling winds but sit them at 40 deg under the seat of an ebike and then try and haul a 300lb load you cant really expect anything else.

I admire the folks like Matt and Thud for turning out some amazing kit esp Thud with his hand tools 8) and folks like AJ are truly pushing the limits on this stuff however the jury is still out with me until I see more, I think the sensored route using bike controllers is the way to go if you want to go this way, I still think if you really hate the idea of a hub motor ( I cant think why mine are fast and very reliable) then there are other options that are much better, Miles approach I think is the best option period, suitably sized motor direct driving a 3 speed rear hub gear, I would like to see more of these type of drives.

But please can someone help me with the friction drive thing??? noisy, not efficient, slips when wet, cant be ridden off road and if you couple in the RC motor controller reliability factor then you have IMHO a recipe for putting people off this hobby pretty quickly, there is a good reason why these systems disappeared in the first place :?

Asked for an honest answer and so there you are thats what I think

Even if only reliable at half the rated power, (and I'm not saying its the case) even then you gotta admit the RC stuff is pretty amazing for the size. Half of one of the 7500 watters is still 5 horse power. And the sensor issue is getting ironed out thanks to a few awesome ESer dudes. Its going to be a non-issue.

I agree on the friction drive, looks like it works well as an assist on road only bikes. Not really the thing for off-road or powerful bikes but has its place imo.

Thud said:

I guess I am just looking for someone to double dog dare me to attempt it.

Click to expand...

Thud, I DOUBLE DOG DARE YOU to build a water cooled controller.

I've no experience on a totally flooded cooling system, but doubt that it'd make a big difference to a TO220 packaged FET's ability to get heat out of the junction, as that's dominated by the die to base plate thermal conductivity.

The best solution is really to go for a FET package that has the lowest possible junction to case thermal conductivity figure. For example, a good TO220 FET, like an IRFB4110, has a junction to case figure of 0.402 deg C per watt, a case to heatsink figure of 0.5 deg C per watt and you may possibly be able to fit it to a heatsink with a thermal conductivity of 0.5 deg C per watt, so your total junction to ambient thermal conductivity is around 1.4 deg C per watt. The max junction temperature is 175 deg C, so the real maximum power rating isn't the 300 watts on the data sheet, but around 100 watts at about 20 deg C ambient.

A good FET package, like the Hiperfets that LFP, Bigmoose and myself are playing with has a junction to case thermal resistance of 0.08 deg C per watt, a case to heatsink thermal resistance of 0.12 deg C per watt and again a heatsink to ambient thermal resistance of maybe 0.5 deg C per watt. These can dissipate around double the power in the junction before overheating, plus they have less than half the Rdson of an IRFB4110, so in reality they will take at least four times the continuous current. Add in the fact that they have a nice high thermal inertia, meaning that they can absorb high current spikes more readily, plus the very much reduced resistive heating from the leads and interconnections and you have a device that will comfortably outperform even a bank of half a dozen paralleled IRFB4110's. in effect, three of these would equate to maybe a 36 FET controller using IRFB4110's, but without the risk of uneven current sharing and with an easier gate drive requirement.

Jeremy

Thud said:

Gwhy,
How about giving us a little math to support your statments?.....

Click to expand...

Theres not much math to go with this, with the package used the fet leads are one factor ( I have always looked at these as carrying upto 70A per leg they will take more but at a cost of long term reliability ) ,

Jeremy Harris said:

The limiting factor is the ability of the FET junction to conduct heat away into the heatsink fast enough to keep the junction temperature within safe limits.

Click to expand...

This is a major factor and what some people seem to over look.

Also stacking fets is not always the answer as Luke have said in another post somewhere , All the fets would need to be identical in spec and identical in temp in operating conditions which with the component layout of the controllers is near impossible so you need to de-rate even more

When I was designing circuits many years ago ( which I done for around 10years this also included fault finding on other peoples circuits) I was told to ALWAYS Knock at least 10% (20% for critical circuits) off of the absolute max ratings when selecting components and this practice have always stuck and served me well. But I have been out of the industry for what seems like a life time and things have really changed a lot on the electronics's front.
also think about the ripple voltage that the fets see from these controllers even the CC needs some help at these high currents. The values I gave about controllers are just from the spec sheet of the fets I am using with the 10-20% knocked off. Main controller killers are the phase currents, temps and over voltage ( due to ripple ) .
The phase currents can be brought down with e.g shorter gearing or slower throttle input but there are many factors that can effect phase currents the temps can be lowered by getting more heat conducting mass as close to the fets as poss beefing up the current carring tracks on the pcb will also help keep the fet legs cooler. More caps to reduce ripple.

I still havent popped a controller yet but something tells me I will when I start playing around with one of the bigger motors .

Thanks for the feedback everyone ( even the off topic is interesting), but lets not get into the Friction drive debate, just accept that so some of us it is an effective solution to a specific situation ( road only, lightweight, minimal parts, quickly removable, etc), and that for a "kepler" type drive an RC outrunner is key to it design.
My problem is figuring out why the RC "electrical" systems seem to be unpredictable in operation. some people have good success, whilst others (myself included) have a hard time even making them operate effectively (ignoring the potential for terminal failure), when using similar components, .
I would happily accept the responsibility for equipment failure due to miss use or abuse , overloading , etc...but for basic components not to function as designed ..is very frustrating.
Hall sensored motors would appear to be more appropriate but i dont think they are readily available in the sizes required ( 50 -70 mm dia outrunner) ..or am i not looking in the right places. .. Is there a "commercial" package of sensored RC outrunner and matched controller that would suit a "Kepler" friction drive?
Why do the (brushless) hub motors & controllers appear to be more consistent with regard to motor/ controller functionality ?, is there a fundamental difference (sensors ?) in their electrical designs compared to RC systems ?