RC-car ESC as an option? need advice...

[spinningmagnets]

Just ordered an HXT 63-54-250 motor, making a friction drive similar to evtodd. I recently noticed these $100 ESCs for 1/8th scale cars, providing 150A continuous output. (2 different suppliers linked)

http://www.nitrorcx.com/1bresc1.html
http://www.hobbypartz.com/ezbrescfor18.html

There are several things that interest me. They have a built-in auto-start fan, there is a thermal sensor that ramps down current when it gets hot rather than a sudden cut-off, they have an adjustable "soft start" ramp-up of power...is there anyone here familiar with a similar ESC, who can warn me if I'm making a huge mistake to get this? I am an RC noob, so don't be shy to point out the obvious.

1.1 Output: Continuous current 150A, burst current 1080A.
1.2 Input: 6-18 cells NiMH/NiCd or 2-6 cells LiPo.
1.3 BEC Output: 5.75V/3A (Switch mode built-in BEC).
1.4 Resistance: 0.0002 Ohm.
1.5 Motor Supported: Sensorless and sensored brushless motors
1.6 Suitable Brushless Motor: 3.5T and > 3.5T
1.7 Suitable Car: 1/5, 1/8 on-road or off-road cars / trucks for competitive race.
1.8 Size: 68mm(L) * 55mm(W) * 45mm(H).
1.9 Weight: 150g(Without wires)


2 Features

1.1 Compatible with all sensorless brushless motors and most of sensored brushless motors such as Novak, LRP and Feigao, etc.
1.2 Excellent start-up, acceleration and linearity features.
1.3 3 running modes (Forward only with brake, Forward/Reverse with brake, Forward/Reverse immediately)
1.4 4 steps of maximum reverse force adjustment.
1.5 Proportional ABS brake function with 4 steps of maximum brake force adjustment, 8 steps of drag-brake force adjustment and 4 steps of initial brake force adjustment.
1.6 9 start modes (Also called “Punch”) from “very soft (Level 1)” to “very aggressive (Level 9)”.
1.7 Multiple protection features: Low voltage cut-off protection / Over-heat protection / Throttle signal loss protection / Motor blocked protection.
1.8 8 steps of timing adjustment by software.
1.9 Built-in switch mode BEC has a powerful output to supply all the electronic equipments.
1.10 Easily programmable with only one button, and also compatible with pocket-sized Program Card.
1.11 Firmware can be updated through an USB adapter (Optional equipment).
1.12 Splash proof and dustproof.

 

[Kepler]

If you can disable the reverse and brake function, I think this type of ESC would work really well. If can't then you are going to have problems. Car ESC use 2/3 of their range for normal foraward throttle and the bottom 1/3 for braking and reverse. Will be interested to if if can be done.

 

[Kepler]

Just thinking more about it. If you have a one way bearing on your friction drive, then brake wont be an issue.

 

[spinningmagnets]

Thanks, Kepler. I appreciate your input, especially with how busy you must be right now! I am pleased that I was right about the publics response to your friction-drive!

I didn't know that about how a car esc uses only 2/3rds of its forward capabilities, that is useful to have in the information base. I'm not sure what that means, especially as the motors they usually run sometimes spin over 100,000-RPM. I am very weak on electronics....

A recent response from Jeremy Harris has been helpful to clear the fog about the catalogue terms:

"...The main differences between the ESC types really seems to be how fast they will spin a motor (RC helicopters use very high rpm motors with gearboxes, cars, boats and fixed wing aircraft usually use slower motors) and whether or not the ESC has a built-in reverse (cars and boats often use a system where the mid-point of the control range is 'off' with either end being full speed forward or full speed reverse).

(edit: boat escs usually have a liquid-cooled heat-sink, and car escs often have an extra-large heat-sink with an integrated fan)

RC Radio-Controlled, Remote-Control

ESC Electronic Speed Controller

HV High Voltage, such as 8S (28V), 10S (36V), 12S (44V)....lower voltages, such as 5S (18V) and 6S (22V) are not considered HV in the RC world. However, they very common.

"Optically coupled"
An opto ESC just means that it doesn't have a 5V power supply to run the servos or servo tester throttle (if used on an ebike). Many high voltage ESCs don't have a 5V supply, which means you need to find another way to power a servo tester to use as a throttle. (edit: from the catalogue: "The throttle signal is transfered through optically coupled system to avoid the electromagnetic interference")

BEC
A BEC is a Battery Eliminator Circuit, another name for the 5V supply that is needed to drive servos, an RC receiver or, in our case, a servo tester to use as a throttle. BEC is usually used to describe the built-in 5V supply inside an ESC. This "eliminates" having to use a second battery for the servos. A BEC could also be called by the generic term DC/DC-converter, as it converts the main batteries 18V/22V into 5V.

UBEC
It is a Universal Battery Eliminator Circuit and is a separate modulethat connects to your battery pack and provides the 5V needed to drive other parts of an RC system, like the servos (a servo is an electrically moved piston or turning-screw that might be used to move a planes rudder, for instance) or other accessories. The thing to watch when using them to drive a servo tester as a throttle on an ebike is the input, or battery, voltage rating. Not all of them will accept high voltages (over 28V) safely.

Rx
it is just an abbreviation for the RC receiver used in models.

If you buy a opto type ESC then you will also need a suitably rated UBEC to drive the servo tester throttle. It is actually possible to hack some opto ESCs to give 5V to drive a servo tester - I've done it to a couple - but it's not for the faint hearted, so the easiest solution is just get a UBEC as it will be virtually plug and play.

I'd advise getting an ESC that is as over-rated in terms of current rating as you can afford. Ebikes put a high load on controllers, particularly at partial throttle, low speed, and the cheapest low current ratings ones don't seem to last long. I've been running a cheap 100A 'Mystery' brand (it was around $35) ESC to drive my milling machine for the past year, which loads an ESC a bit like an ebike, and that is holding up well. As these very high current controllers are pretty cheap, they are worth trying, even if you only plan on drawing 30 amps or so..."

Note: for other newbs who stumble across this thread, if you add hall-sensors to an RC motor, you then have the option of using a conventional E-bike controller, such as the small and popular Infineon 6-fet
http://endless-sphere.com/forums/viewtopic.php?f=30&t=15686
http://endless-sphere.com/forums/viewtopic.php?f=31&t=18744

 

[spinningmagnets]

Other terms and descriptions in search of a definition, plus some answers and suggestions:

EVTodd: For the servo tester I just use this (as a throttle): http://www.esky-heli.com/esky-esky-servo-adjuster-tester-ek20907-000504-p-1532.html with an ev warrior throttle soldered in place of the original potentiometer. You can find the ev warrior throttle on ebay dirt cheap too. I see they have a new version of this servo tester with an led and an automatic function. Not sure if all versions are the same as mine. I power the servo tester with this: http://www.radioshack.com/product/index.jsp?productId=2062254 I could have just used a BEC but that battery holder works great. I just use rechargeable NiMH aaa batteries and they seem to last a long time between charges.

def215: "The only modification I did to the ESC was beef up the power (battery) wires to large 12 ga, from the small 16 ga that were on it"

EVTodd: "I took the sticker off the heat sink and added one big cap to the + and - wires (of the ESC) but that's it"

mwkeefer:
One of the reasons I suggested the 'cell log 8' was specifically so you can view all these (individual cell) voltages and even log them while using the WattMeter....
.....Maximum LVC (Low Voltage Cutoff) for 5S (a battery made of 5 cells in Series, =18.5V) is 15v - not 13.5V for the entire pack, but at 3v per cell. The trick is if you drag down cells that far, they will tend to get further and further out of balance. Going to 3v isn't smart for the pack and in some ways (with inaccurate meters and alarms), could be dangerous. A more appropriate LVC would be when the lowest cell broaches 3.5v. The absolute LVC is 3v, so when just one of the cells reaches 3v - you're 100% done....
.....These battery packs will last longer and fade less when you add more capacity, if indeed you are occasionally pulling 80A (you likely are, but only for a few seconds at most) that would be peak load of 16C - still well within your battery C-rating, but if you cut the load in half by adding an additional pack, then the individual battery packs current-draw only goes to 5C, and these packs are beautiful at that low of a discharge...(Mike recommends a battery rated for a current discharge capability of 25C or higher for E-bike use, as bikes draw more current than a relatively constant RPM model plane propellor)

Note: all ESCs perform a variety of functions to operate a motor, and they come from the factory adjusted with the most common function settings, however, a programming card (typically less than $10) can be bought that will allow you to input custom settings that may work better for an E-bike. Some ESCs have an interface that allows custom settings to be applied through a laptop computer, and some also have data-logging to record how the system has been operating.

12 pages of ESC selections from Hobby King:
http://www.hobbyking.com/hobbyking/...ucts.asp?idCategory=61&curPage=1&v=&sortlist=

 

[mwkeefer]

Hey bud,

Brake option should never be used for an eBike application or even a normal airplane model, traditionally that is designed for use with folding props to get them collapsed fast.

Another issue with RC Car ESCs... most of the RC car motors I've seen run the motors really fast but they are 2 pole motors as opposed to 8, 12 or 14 pole for common in/outrunners - this means to acheive 1000 RPM requires only 500 electrical commutation cycles or 500 electrical RPM on a 2 POLE motor - on an 8 pole (next best example) it would be upwards of 8000 electrical commutations and for a 14 pole motor - you would be looking at 14000 electrical RPM per 1000 mechanical RPM. In other words to reach the useable RPM levels between 3 and 7K RPM - your looking at the difference between say 14000 electrical RPM and somthing like 56,000 electrical RPM for an 8 pole (astro) or even up to 98,000 RPM for use with 14 pole motor to acheive 7K RPM.

I could be wrong on this - my experience is derived from the days before electronic speed control was a reality - rheostats with servos were my speed control (RC10, way back in the day) but the motor controllers I've looked at from the RC car world wouldn't serve us in eBike useage.

Now adding fans, active cooling and other bits to perhaps a Mystery 60 or 80A controller would be cool, but as a friction drive is what your going for... according to EVTodd and Kepler, they haven't blown one out yet - could be the lower voltage, that perhaps the lower voltage RC ESCs are a bit better at dealing or that they don't fire up from 0RPM in the case of Keplers design - even Todd say's he pedals first

Hope that answers your questions

-Mike

 

[Kepler]

Some good points ther Mike. Just in relation to the brake feature, if using a one way bearing on a friction drive, you only would be braking the mass of the roller which would probably be less then a prop. However, if the one way locks up and the brake function activates, its goodbye ESC. Having seen how often one way bearings fail and lock, having a any sort of brake function it not a good idea.

 

[spinningmagnets]

Found this about the RC-car ESC programming:

" 3 running modes (Forward only with brake, Forward/Reverse with brake, Forward/Reverse immediately)" so, it looks like the brake function 'can be' disabled by programming, but if there was a program failure, the ESC frying would negate any minor benefit to getting a car ESC rather than a plane-ESC...as far as the reverse capability, since it requires a specific signal to go into reverse, it shouldn't be too hard to figure out a way to only actuate a forward throttle signal?

The RC-Helicopter site listed their max RPM (I don't want to use the equipment anywhere near the "max" anything) for a 12-pole motor is 35,000 RPM. Also, the Heli-ESC was capable of 210,000 RPM for a 2-pole motor, and if you hadn't pointed that out, I wouldn't have noticed it (even with it right in front of me!)...thanks Mike!

Is there a formula as to what size and type of capacitor should be used? what is the major benefit of adding a capacitor to the ESC battery wires, is there a pic somewhere showing the proper attachment of the capacitor? EDIT: found the capacitor thread:

http://endless-sphere.com/forums/viewtopic.php?f=30&t=22194

 

[John in CR]

How about hacking into that temperature sensor for a cheap and easy means of controlling current? Isn't that the downfall of the RC controllers, no current limit?

 

[mwkeefer]

Temp sensor for shutdown? Why, they are going to go POP faster than you can react even automated without updating the firmware to accept I2C instead of PPM (google for more info, too much to explain). Temperature isn't so much the issue as the increased RdsOn as the temperature rises, that + the derating of the FETs.

For a quick example - check the specs on the 4110 FETs, good, solid pieces of silicone right? Except the huge ratings everyone assumes they have (some 120+ AMPS) is only at 25 degrees Celcius. Once you pass this thermal point, the derating scale begins - there is another factor on the specs sheets to help you calculate this, someone else is better to explain that part if needed.

So... now we have some SMT FETs clustered together and nothing but a wedge of aluminum epoxied down or screwed down onto it, add to that our type of loads (200-300 lbs seems a good estimate for bike + rider + etc) and well... how long do you think at NO PWM or limiting that will survive before the FETs heat up and blow ? What is the MAXIMUM peak current of the controller and the durration of the peak. Also take into account mounting location - car controllers are sheilded from outside air hence the need for fans... should be atleast a hint to us all.

Ok, now add partial throttle which causes PWM or Pulse Width Modulation in order to limit the current being passed from battery into motor phases each time, as these FETs open for shorter and shorter durration, the amount of time they take to totally turn on or off, however long it is also needs factored in - especially since it's durring the Partial ON or Partial OFF (look at it either way you want) when the FETs will shed or attempt to shed the most heat.

Now pull the HEATSINK on the FETs (Mystery Controller or Car Controller) - what FETs are used, get the data sheet...

What is maximum continuous current and how many are in parallel - also what is each FET rated for (given the SMT and cooling provided) in terms of heat dissipation? I want to say the 4110s are rated at 325w, so each FET could in effect handle dissipating 325w of HEAT or waste... the rating on the SMT fets is going to be lower and without FAN or active cooling of some sort - you need to look at worst possible performance ratings - not the 50C rate.

Ok - so brake is bad, we've established that
RC controllers want either FANS or Airstream cooling them and drawing away heat

Honestly the short wiring and lack of drop likely don't hurt too much, coupled with a "properly" geared or mechanically reduced (ie: friction drive) they should be up to the task - heck the HV series have been working for recumpence forever now.

With regard to Caps - It would appear that the factory equips most RC based controllers with 100uf per 10A of power passed so a 60A controller I have has 660uf (2x 330uf). Far more important that size I believe, is the LOW ESR - the whole point of Capacitors is to handle the ripple voltage created by CHOPPING the DC into 3 PHASE AC to power your BLDC motor - many have said this on forums here and elsewhere. Most of the time with HIGH END controllers (read: CC HV or better) the power and phase wires are quite stout and short - they are intended to stay short (<12" is most recommended length) when this isn't possible - instead of 1 set of 12AWG, run 2 sets for power which should work out roughly to a 6G power cable. I believe this is how recumpence handled some HV110 issues on his power drifting yellow trike pulling 5kw per controller (I think).

So what size caps should you use... I would personally use as many and as much as I could fit, they wire up so the (-) side of the cap is connectted to the ground net or BLACK - the other side of the cap is connected to the positive net or RED - if it were wire, you would have KFF right now

One thing I've learned (gleaned) - it's better to use multiple smaller caps if possible (smaller UF) even if increasing the stock provided capacity by 2X in parallel to acheive total uf desired. Provided sufficient capacitors of fast enough type, they will smooth the RIPPLES caused by the aforementioned chopping... this is what kills controllers which would otherwise have lived a good healthy life.

My new 100A kForce ESC is already modified with 4 x 470 uf Caps attached and I doubled up the power wires with 12 AWG at 8" (just long enough to reach intended location) so really it's an 8" 6 AWG (if I ever use more than one set of connectors) - in addition I found an older 486 CPU fan and modified it to run from 5v (the BEC onboard)... Eventually I will replace with a cooling system that runs on 3-4.2v and works very similiar to backup AC inverter - this will enable the FAN to run after power off and cool the unit.

I have not yet tested this ESC - my son is with my this weekend so my time for R&D is limited but I will get to it shortly.

The caps I used were some 100v nicholson low ESR I got a long while back.

-Mike

 

[spinningmagnets]

Thanks everyone for the great responses. I have read each one several times. I'm sure a year from now I will have a different opinion, but I want to do something now, just to get rolling. I'm ordering the $37/85A ESC listed below.

Lyen makes a $90 sensorless Infineon 6-FET using 4110's, so it "should be" able take a lot more amp-abuse than the tiny ESCs that most here are experimanting with. I have no experience with them, and you must specify the voltage you want to run at before ordering. (they also might be limited to a lower RPM than you need)
http://endless-sphere.com/forums/viewtopic.php?f=31&t=19587

My interest in the $100 RC-car ESC is that it is much smaller than Lyens, and other than the whole "blowing up from amp-spikes" thing, they provide a tremendous value of features for a very reasonable price.

I am hoping that by using 6S instead of 5S, I will need fewer amps to accelerate past the low-RPM danger zone without bogging down as much. That is also why I chose the 63mm motor rather than the 50mm, as I feel that the larger diameter will give it more leverage to pull past the danger zone in a shorter time span (like the 9-continents, which has lots of torque per Watt due to larger diameter?).

The 250-Kv will give me a slower top speed, but I also chose that as a reasonable compromise to help with low-RPM ESC-synching, and low-RPM heat . I am hopeful that the small one-inch roller will also help with the low-RPM/partial-throttle amp-spikes that seem to be the ESC-killer. (there is room in the current drive for a 1-1/2" roller, but thats a future experiment)

http://www.hobbyking.com/hobbyking/...ct_Name=Turnigy__Brushless_ESC_85A_w/_5A_SBEC

 

[spinningmagnets]

From another thread: http://endless-sphere.com/forums/viewtopic.php?f=3&t=21173&start=30#p309535

"...I was doing just fine on gentle hills and mostly flat roads, then I just HAD to try it on our local steep hill (5-7%) and since I went outside of "Lines" defined by Kepler, and used a much more powerful motor and battery, I believe I have let out the magic smoke ....

....Oh, I checked my Turnigy Power and Watt meter, and it showed peak amps at 125!, Also peak watts was 3432!!

I think however that the Castle Creations info from my ESC (data logging) is more accurate, it shows peak amps at about 90 and the peak watts at ONLY 2600 .... "

This quote is from a Kepler-drive user (the 2" motor shell is the roller) using 8S (28V-30V?). so, I feel pretty comfortable with the 85A-continuous ESC I just ordered. I don't know where the line is, where if you cross it, your ESC will fry (so many variables!) but If it does fry, I will likely buy the RC-car ESC in the first post as a spare ESC, as the data-sheet claims its rating is 150A continuous, and 1,000A for 10 seconds.

The unusually high peak-Amp rating wouldn't help on a long hill (1000A only available for 10 seconds), but it may be enough to allow customers to accelerate from a stop without having to pedal first. Just a thought for the info files...

 

[mwkeefer]

It shouldn't need to help on a long hill - if your not at partial throttle and you are geared right.

Using info from the other post (LI), it's a 26" bike with a 1.978" diameter + grip motor, call it 2" so total physical reduction is 13:1

For comparison 20mph on 26" would require wheel RPM of 258.5

Running on 8S (as he was) hot would be roughly (assuming 4.1v per cell initially on first load) 32.8/33v, for this argument let's say it's only 32v.

The motor in question was an eFlite Power 110 - rated for 16 lb models maximum and up to 2000w with a kV of 295, Max Burst Current of 65A for 15 seconds and continuous max of 55A.

First how fast was the motor spinning?

According to poster, I think he reached 58 mph no load speed driving the wheel off ground... some quick math:

749.53846153846153846153846153846 RPM / 336 * 13 = 9744

No efficiency information about that motor but I think we can assume 85% without being too far off mark.

Now, let's do the math the proper way...

295 * 32 = 9440 RPM

Wow... that's not bad, very close to on mark: 9440 by the math and 9744 with the wheel up and likely a few amp load so no voltage sag of the 8S pack from full charged.

There in lies the problem.

Assuming a top speed of 49.3 MPH (85% of 58mph) means to acheive anything but about 48mph at nominal voltage would require partial throttle. This is one reason I'm so big on gearing motor output through multiple ratios.

A better fit for that given setup would have been 35mph @ nominal voltage which we would figure out by:

35 MPH * 336 = 11760 / 26" = 452.31 RPM
295 kv * .85 (factor efficiency under load) = 250.75 loaded kV
452.31 * 13 (multiply target RPM by reduction ratio) = 5880.03 RPM Required @ motor (load accounted for)
5880.03 / 250.75 = 23.45v nominal is perfect
23.45 / 3.7 = 6.3 S Lipo

So for this setup 6S would have been perfect and would give a nominal speed of approx 35mph.

It's important to note, were talking about a delta wound motor and a difference of RPM at the shaft of roughly 3600 RPM or 21.42 MPH difference between the failure scenario and this calculated scenario.

Also... 2000w is the rated power so at 22.2v we could provide 90A maximum with the 6S, on a full charge that would drop to 80A (assuming 24.96 for 6S fully charged).

And I am not sure if the CC ICE or the other logger are more accurate - honestly for peak capture, the sample rate needs to be pretty darn high - spikes like that (factor of difference, not really that level) I only see with my EagleTree logger setup to the Highest Sample rate which gives me about 8 - 10 minutes maximum of logging (as opposed to 20 hours at 2 samples per second)... I would tend to believe the higher number of Amps.

In either case just 100A surge at full charge 8S would be > 3kw, 33% more than the motor is rated for at it's maximum

Add to all the above - the attempted partial throttle uphill, overgeared by 20mph and

BTW: If that motor can handle a 16lb model, then with a 13:1 reduction the max load (in theory) should be 208 lbs - LI is about 165 + bike + bits so probably right in that range.

Anyone as always, feel free to correct, clarify or otherwise make sense of my ramblings

-Mike

 

[Hillhater]

Mike,
sorry to trip you up after all that maths but, ....
Li ghtcycles motor (eFlite Power 110) was a 63mm dia ( 2.5") not the 2.0" you used for your calc's :wink:
Which means that his gearing ratio was effectively 10.4 :1 isn stead of the 13:1 you used.

.. so at 58mph with the wheel up, the motor rpm was only 7800 rpm ... some way off its theoretical no load speed of 9440.
He would expect to see nearer 70mph with the wheel up, to be close to the motors designed no load speed !!
and at 20 mph , the motor would have been only turning 2700 approx rpm ....not a good speed for max load.

 

[spinningmagnets]

A properly set-up Kepler-drive should work well (especially with his pedal-first configuration reducing amp-load), but one of the things that drew me to EVTodds set-up is all the variables that are allowed.

For instance, there is a much wider range of roller diameters that I can use, and some sizes (Like my current one-inch roller) are not even possible with a Kepler. A Todd-drive can also start out by using a larger in-runner motor (such as a common Kollmorgan), and you can upgrade to a smaller and lighter motor as your funding allows.

Once I get rolling, I can document the real world performance, and evaluate the unexpected results (if any). Then I have many options available to modify the drive. A voltage change, a differently-wound motor (higher/lower Kv) and a different diameter of roller (perhaps others?).

I'm a caveman when it comes to electronics, so this has been a difficult learning curve for me. Any advice and help is greatly appreciated!

 

[Hillhater]

Todd's drive certainly has some advantages. in that you have a wider range of motors and battery voltages to chose from, but keplers has the benefit of simplicity and fewer working mechanical components, ie just the motor and a pivot bearing, no clutch or slide mount to maintain.

 

[bandaro]

using a car asc would make the car/plane esc difference (if there is one) less of a problem.

i say this because when you compare the amps drawn on a graph between a car, bike and plane, the car will match the bike far more than the plane. the bike and car have frequent peaks in amp draw when starting, while a plane only has a single peak. this peak is what will kill the esc more than anything, a plane one simply wont hold up to that abuse nearly as well, as its designed to sit at a level of amps, not bounce around a lot like the cars.

the main problem with a car esc in a bike is the voltage, cars simply dont use the high volts that we like to run in our bikes for efficiency and power reasons, so finding a car over 6s and a bike under 7s is going to be tricky. that being said, keplers drive uses low voltage, as do similar builds, so if the lower voltage suits you, then i see no reason why you shouldnt use a car esc. they are generally better suited for bikes anyway, cooling, bumps, load, etc.

 

[EVTodd]

Todd's drive certainly has some advantages. in that you have a wider range of motors and battery voltages to chose from, but keplers has the benefit of simplicity and fewer working mechanical components, ie just the motor and a pivot bearing, no clutch or slide mount to maintain.

Thanks for the kind words.

Keplers drive is certainly direct in it's approach with the motor actually being the roller but keep in mind that my drive now only has 4 basic machined parts, 14 or so if you include bolts and washers (I think Kepler's system has around 90, but that's including hardware and a way to mount it to your bike). I do understand that you are talking about just the working components but I think I've gotten it down to a pretty bare minimum.

I do recommend a little oil on the slide from time to time but that's about it. And the one-way bearing (clutch bearing) is actually a very simple device. It's just a bearing that locks out in one direction. Really nothing to maintain (maybe a little light oil on the shaft from time to time). The roller is made up of a hardened shaft, the clutch bearing, two spacers, and the roller itself. I'm working on a version of a roller with fewer parts too. I think it was you that suggested having the roller completely off the tire so you don't need a freewheeling roller. I may still mess around with that a little too. It would be nice to have even fewer parts if possible.

I've said it before, I absolutely love what Kepler has done with his drive. I personally still like the idea of smaller rollers though. Using the can as a roller you have no choice but to have a bigger and bigger roller as you move up in power. I'm currently also messing around with just how small I can get with a roller and still have enough traction. I need as much hill climbing ability as I can get around this area.

Anyway, I don't want to sound like Kepler and I are in competition because I don't think we are at all. My plan is to sell basic components of my drive to people that have the ability to build one themselves to their own specs. Kepler seems to be going more for the complete package.

 

[EVTodd]

Some good points ther Mike. Just in relation to the brake feature, if using a one way bearing on a friction drive, you only would be braking the mass of the roller which would probably be less then a prop. However, if the one way locks up and the brake function activates, its goodbye ESC. Having seen how often one way bearings fail and lock, having a any sort of brake function it not a good idea.

That's a good point, I was going to mention that on my bike where I've got the roller mounted a little too high on the tire I sometimes notice the motor spinning when I'm not using the throttle. My gf's motor doesn't do this since it sits further forward on the tire. I should move mine a bit but I used existing holes in the rack.

Anyway... My point is the same as Kepler's. I wouldn't totally rely on the clutch bearing to spin freely 100% of the time. I would never have the brake enabled. Good thing the mystery controllers don't default to that since I have no way to program them.

Speaking of this issue. I recall Matt saying any regen going back into an rc esc can fry it. My drive has probably been doing this on and off for around 1000 miles now and it hasn't fried yet. Should I be concerned that it will?

 

[Kepler]

I've said it before, I absolutely love what Kepler has done with his drive. I personally still like the idea of smaller rollers though. Using the can as a roller you have no choice but to have a bigger and bigger roller as you move up in power. I'm currently also messing around with just how small I can get with a roller and still have enough traction. I need as much hill climbing ability as I can get around this area.

Anyway, I don't want to sound like Kepler and I are in competition because I don't think we are at all. My plan is to sell basic components of my drive to people that have the ability to build one themselves to their own specs. Kepler seems to be going more for the complete package.

No problem here Its your fault I built a friction drive in the first place. Thanks mate

I am actually protyping a roller arrangment at the moment but still using my basic setup. There are definite adavantges to be had with a smaller diameter roller and I am keen to see if the positives outweigh the negatives.

 

[Hillhater]

Guys, resurrecting this topic to sound out opinion on this car ESC being able to operate with a hall sensored Turnigy 63-54, 250kv motor.
This particular ESC claims to function with sensored motors ( but i dont see the required connector point ?) and is also one of the few car units that will operate on 5s or6s lipo.
Could this really be an option for our Ebike drives using sensored motors ??

 

[spinningmagnets]

I started this car-ESC thread after reading some of Jeremy Harris' and MwKeefer's posts about RC models (boats/planes/cars/heli's) and they had years of experience with many RC issues in their original design applications.

I liked the car-ESCs right away, and would have been happy with the much cheaper $50 80A unit, especially with its high 10-second 270A peak-amp rating. Both the controllers and motors have factory aluminum fins and integrated auto-start fans (which I assume help the peak-amp ratings). I didn't buy the 80A because it would only run up to 5S (18V) and I was looking for a 6S (22V) ESC solution as the lowest voltage I would want to use for my particular goals.

For any readers who are new to electrical stuff (as I am) the more volts you start with, the fewer heat-producing amps you have to use to achieve your power goals. The $100 150A car-ESC is still my second choice if the $40 85A unit I ordered fries when I test it on my particular loads.

I will now be ordering the 63-64/280-kV motor, which also has the same rare skirt bearing as the 63-54/250-kV I have. 8S (28V), 10S (36V), and 12S (44V) remain an option, but the controllers and chargers are MUCH more expensive, and I'm waiting to see how 6S turns out after all the parts arrive.

 

[spinningmagnets]

For those still interested in a 150A car-ESC for an RC E-bike, Hillhater has purchased a similar unit to this one by a company called Aeolian, and so far he is reporting good results. The list price is $88 (plus shipping and tax). He is using 5S (18V-20V) and the caps are rated for 25V so it should be fine at 6S (22V-24V). If a cap does blow at 6S, they should be easy and cheap to upgrade (IF,...the bad cap is the only damage...)

http://www.leaderhobby.com/product.asp?ID=9394001220262
http://endless-sphere.com/forums/viewtopic.php?f=28&t=21771&p=324560#p324312

 

[spinningmagnets]

I have been asking several builders with experience in electronics about what catalogue specs are important for selecting a capacitor to add to the ESC to improve its performance, since we are using ESCs in a way that they were never intended. You can remove the old Caps (if your ESC already has some) or you can leave them on and just add another cap, (or as many as you can fit).

http://endless-sphere.com/forums/viewtopic.php?f=30&t=22194&start=0

There seems to be two problems, the main one is they need to have a low ESR (Equivalent Series Resistance), the lower the number the better. Bigmoose has recommended 10 milli-Ohms as the best performance, but those are much bigger and more expensive. If the other options didn't work, I would use one of those, but I am reading reports of success with much smaller caps that only have a slightly higher ESR. Ripple happens quite rapidly and average-ESR caps cannot absorb the ripple well enough to do the job.

The other problem is that the caps occasionally pass current (from clipping voltage spikes) during high throttle at low speed, which results from the ESC using Pulse Width Modulation (PWM). Its like flipping a light switch on and off rapidly so the room will only seem "half-lit". If you are presented with two cap choices with very similar specs, the one with a higher ripple-amp rating will be better, and adding a second cap (or even a third) in parallel will increase the cap-sets ability to pass even more current without overheating, as overheating can cause it to literally burst.

Once you have spent a few seconds in heavy PWM mode, and you stabilise near your top-speed, the caps will have some time to cool off (until your next hard acceleration, or steep uphill with heavy cargo climb). Two long and thin caps will cool off faster than a single large cap that is short and fat, even though their performance specs are identical.

The NIC web-catalogue has some caps with slightly better specs than the Digikey catalogue, (14 mOhm vs 16 mOhm for the 50V). However I am having difficulty ordering from NIC, and Digikey is easy and fast. Thanks to Mwkeefer, Bigmoose, LFP, Fechter, etc I now know what I'm looking for (though I must confess, I still do not understand electronics)

The volts that a cap is rated at (for our uses) should be higher than your battery packs highest voltage, usually when the battery has just come off the charger. It does not hurt for the caps voltage rating to be higher, sometimes MUCH higher than the system voltage (100V cap on a 44V system).

However, when I was comparing many caps specs, I noticed a trend. It appears the closer the cap voltage is to your system voltage, the lower the ESR and the higher the ripple amps were (although, sometimes there was no benefit, so in that situation, I would defer to the higher voltage cap with the same performance spec).

The actual capacitance (uF) of all of these choices far exceeded our needs, so we are looking for low ESR (impedance measured in milli-Ohms), and high ripple amps. Plus if possible, a long/thin form will shed heat better to cool off faster. If the loads you apply still overheat and burst your cap, add more in parallel (+ to +, and - to -) until there is sufficient mass to absorb and shed your typical heat loads.

All choices listed here are under $2 per unit. All battery series designations and voltages are for LiPo only. All corrections appreciated (I am still learning), and any links to other capacitor options with better specs that are reasonably sized and less than $10-each will interest me.

5S (18V-20V), 6S (22V-24V), and 8S (28V-30V)
35V Nichicon, 12 mOhm, 4.28A, 8 X 40mm
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=493-1602-ND

10S (36V-38V)
50V Panasonic-ECG, 16 mOhm, 3.32A, 16 X 25mm
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=P12393-ND

12S (44V-48V)
63V United Chemi-Con, 17 mOhm, 3.50A, 8 X 40mm
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=565-1731-ND