Motorboard project

[adrian_sm]

Nice work.

Just a question. Any reason you mounted the temp sensor on the end cap where you did?
I would have probably tried to mount it actually on the windings if possible.

- Adrian

[Stielz]

The only real reason for that is just that it was more convenient to mount it on the end cap. I figured that it should be the same temperature as the windings anyway except for perhaps any rapid changes in temperatures in the windings. I shouldn't get any too rapid changes when keeping it within the specified current rating.

Anyway, here are some pics of it all put together except for the chain. I have left that off for now because I will need to do more bench testing when I finish the motor controller PCB.

wheel assembly

CIMG0112.JPG (231.05 KiB) Viewed 1278 times

Bottom view

CIMG0117.JPG (134.35 KiB) Viewed 1278 times

Top view

CIMG0118.JPG (228.99 KiB) Viewed 1278 times

Over engineered?

Perhaps. We will soon see.

My only concern is that the clamping system used to attach the motor assembly to the axle may be a weakness. I didn't want to weld it on because the heat would deform the parts and make it very hard to get it all square (which is quite important because I don't have self aligning bearings). However that is plan B if the clamp does fail.

[fechter]

You could cross-drill the clamp and into the axle part way and place a pin or set screw in the hole to prevent rotation. A keyway would be stronger yet.

Is there a way to adjust the chain tension?

Have you figured out how to do the throttle? Wireless would be ideal. I'm thinking what happens when you need to bail at speed.

[Stielz]

I have already machined some parallel flat surfaces on the axle to prevent rotation so Hopefuly that will be enough.

It has sloted holes where the motor assembly bolts into the the clamp parts for tensioning the chain.

As for the trrottle, I have used a pistol type RC car transmitter. It isn't wireless because I have mounted a LCD screen into it. It has a plug on the bottom of the transmitter that will pull out if I have to bail. The motor controller will detect when this happens and hit the brakes so that the board doesn't run away. This is especially important with the cruise control function.

[fechter]

Excellent.

On the controller plug, it's a good idea if both plugs are mounted in-line with the wire so it will pull straight out during a bail. Just a short pigtail on the controller end is enough.

[Stielz]

I still haven't quite mastered the braking yet.. It seems to be either on or off with no levels in between.

I am using the technique where the braking is done by shorting all three phases together via the ground rail so that I can synchronously PWM all three lower FETs in an attempt to vary the brake force. However this doesnt seem to achieve any braking force until around 98% duty cycle at which point the motor suddenly becomes very hard to turn. I have tried it with a 21 kHz and a 7 kHz PWM frequency with the same result. Perhaps I need a much lower frequency? I know this technique is not ideal as it causes lots of heat dissipation in the motor windings and the FETs.

I have been reading up about the other methods of braking:

There is the reverse current method. From what I gather it basically just involves switching the gate driver output to the reverse sequence so that the force that the motor applies is in the opposite direction to the motion hence providing a braking force.I am a little hesitant to try throwing the motor into reverse while it is running forwards at high speed while under load. What I think will happen is that the back EMF which usually opposes the supply voltage, will stay the same polarity so when the supply voltage is reversed, there is effectively a short circuit between the BEMF and the supply for a short amount of time (until the magnetic field has reversed polarity) which will result in huge currents and probably burn out the FETs.

Alternatively, the retarded timing method. This just involves using a delay between when the motor would usually be commutated and when it is actually commutated. This method is a bit more complicated because the amount of time delay depends on the speed that the motor is running i.e to achieve a constant angle of timing retard, the time delay must be adjusted as the motor speed changes. This could be quite computationally expensive for the micro controller. It also involves two variables to control the brake force; how much the timing is retarded and also the PWM duty cycle.
I believe that the reverse current method is the same as this except with 180 degree retarded timing rather than variable retarded timing.

If anyone has advice or suggestions on this topic please let me know

[Stielz]

Obviously the best braking method is regenerative braking as it is the only method that doesnt just convert the kinetic energy into heat in the motor and the controller. However I need a second method of braking for when the battery is fully charged so that I dont have my $250 (NZD) battery pack go up in smoke when the lipo batteries get over charged.

So I think I will use regen brakes as the primary braking method and use the phase shorting method with a very low PWM frequency - around 100 Hz as the secondary method for when the batteries are fully charged. I have already built a voltage sensor into the circuit so having the microcontroller automatically decide what method to use is no problem.

I am getting stuck into the PCB design now so hopefully not too long now before I can get it up and running

[adrian_sm]

However I need a second method of braking for when the battery is fully charged so that I dont have my $250 (NZD) battery pack go up in smoke when the lipo batteries get over charged.

Is it really needed? The only scenario I can think of that it would be needed is if you fully charge your batteries, then immediately head down hill for a decent way with regen braking. All other scenarios shouldn't overcharge your batteries.

[SamTexas]

What happened to the pictures? Most of them no longer exist.

The selected attachment does not exist anymore.
The file ./../files/20354_703f6dc0752d408b09db09e4562e1f82 does not exist.

[Stielz]

I think you are probably right there Adrian, however I would still like to have the second option there just in case that scenario does occur. I also think that the braking force will depend on he state of charge of the batteries with regen braking which could mean that I dont get very strong braking when the battery is near full charge. I wont know until I am able to test it so my first priority is to get it up and running, then I will see if a second braking option is necessary.

As for the photos, Im not sure what happened to them. I will try uploading them again

[Stielz]

I finally have my PCB design finished after about 15 hours of working on it. I am still just a beginner at PCB design so there will probably be some flaws in it. I should have it printed by the end of the week

PCB layout

[fechter]

So I think I will use regen brakes as the primary braking method and use the phase shorting method with a very low PWM frequency - around 100 Hz as the secondary method for when the batteries are fully charged. I have already built a voltage sensor into the circuit so having the microcontroller automatically decide what method to use is no problem.

If you short the phase wires at 100Hz, the energy will still circulate to the batteries through the body diodes in the FETs.
You really need a big load resistor to suck up the excess energy in the event the batteries reach max voltage during regen.
If you had such a resistor, it could just go across the pack anytime the voltage goes too high.

Some packs have an active cutoff for both charge and discharge. If you hit the charge cutoff (a cell went too high) and the pack connection goes open, the controller would most likely self-destruct as there would be no sink for the current.

[AAAkings]

Just wondering how the project is turning out? Thinking of doing something similar thanks to this thread.

[Stielz]

It is going reasonably well, I am just having problems with my home made motor controller when I run it on the 10S lipo pack as it keeps blowing transistors. I have just had my fist go on it today just running it on 5S and it goes well, not overly fast but it has plenty of torque and acceleration.

One thing I can tell you is that if you have a live axle ie. two wheels locked together then turning while on tarmac is near impossible. It takes awhile for it to start turning and awhile to stop turning. Its not too bad when going at a decent speed but at low speed it just doesn't work, the tyres are too grippy. I will take one of the chains off tomorrow and try it out in one wheel drive. It turns fine with a live axle on grass or gravel.

I would suggest that if you want yours to be good for on road and off road use, to use two motors and ESCs or even four if you go for 4 wheel drive.

You should definitely have a go at making one, it is extremely fun to ride. It will be good to see another persons take on it

[Stielz]

If you short the phase wires at 100Hz, the energy will still circulate to the batteries through the body diodes in the FETs.
You really need a big load resistor to suck up the excess energy in the event the batteries reach max voltage during regen.
If you had such a resistor, it could just go across the pack anytime the voltage goes too high.

Thanks for the advice fechter, you are quite right. I have tested this by unplugging the power battery and pushing the board along while holding the brakes on. when I do this I can watch the voltage on the capacitors increase and if I do this long enough it generates and stores enough energy to turn the motor a few ticks

[Stielz]

Progress update:
I am still in the process of developing my motor controller program. I am having some issues with it blowing transistors when I run it on 25+ volts.

It does seem to run fine on any voltage less than that. I have had a ride on it running on 5S lipo battery and it handles the current ok. It can produce some pretty decent torque, enough to throw me off the back if im not prepared for it.

Also, the brakes are very effective. I think I will have to limit it to about a 20% duty ratio when braking otherwise it just locks up the wheels and throws me off the front.

Anyway, here are a few more photos

Assembled circuit board

CIMG0123.JPG (230.5 KiB) Viewed 712 times

Enclosure/heat sink

CIMG0130.JPG (216.72 KiB) Viewed 712 times

Motor controller

CIMG0135.JPG (231.45 KiB) Viewed 712 times

[amberwolf]

Best guess is inductive kickback (flyback) spikes at the FETs during switching off, overloading the body diodes.

Stick little caps right at the FETs themselves, to help absorb those.

Probably even better advice from LiveForPhysics, Bigmoose, or anyone that's built a motor controller--I'm only going by what I ahve already seen fix certain things for other poeple's threads/designs/problems. (I never got this far on my own controllers )

[Stielz]

Thanks for the advice, that idea could be worth a try.

I am using a more advanced control technique in an attempt to overcome this inductive kickback problem. When the PWM on one of the lower FETs switches off, the upper FETs are inverted - so the one that is usually held on the whole time is switched off, and the other two are switched on to circulate the freewheel current and prevent voltage spikes. however there is some dead time between when the lower PWM switches off and when the upper FETs are inverted - around about 1us. I am wondering if this is too much of a delay to prevent this method from working..

[AAAkings]

Thanks for the reply and tips. Much appreciated.

Now one final question. Would using an motor like yours be recomended (RC engine) or should I look for motors meant to drive a persons weight like a scooter motor? Mainly hoping that the engine is not overworked due to the weight of the rider.

[addicted2climbing]

[quote="One thing I can tell you is that if you have a live axle ie. two wheels locked together then turning while on tarmac is near impossible. It takes awhile for it to start turning and awhile to stop turning. Its not too bad when going at a decent speed but at low speed it just doesn't work, the tyres are too grippy. I will take one of the chains off tomorrow and try it out in one wheel drive. It turns fine with a live axle on grass or gravel.

I would suggest that if you want yours to be good for on road and off road use, to use two motors and ESCs or even four if you go for 4 wheel drive.

You should definitely have a go at making one, it is extremely fun to ride. It will be good to see another persons take on it[/quote]

Hello Stielz,

I am also working on an electric mountainboard (flexboard) as well but I went two motors to avoid the live axle issue. I am unsure if this would work, but any chance you could swap out your drive sprockets with the type with the one way clutch bearings? If it worked, you could just coast through your turns and maybe the wheels would be allowed to rotate at different speeds...

I am still in the design phase, but have all my parts but my batteries. Tonight im working on my battery box...

Its great to see many different ideas to electrify these boards. I wish I had your electronic skills as I am relegated to use a HK ESc's on this project.

Take care,

Marc

[Stielz]

Thanks Marc.

I have never seen a flexboard before, your design looks good though. I will be interested to see how your project turns out.

As for your suggestion about the one way clutch bearings, I have considered having something like that. the problem there is that I wont be able to do the braking through the motor (or go in reverse) so I decided to just leave it as it is.

Cheers

[Stielz]

Thanks for the reply and tips. Much appreciated.

Now one final question. Would using an motor like yours be recomended (RC engine) or should I look for motors meant to drive a persons weight like a scooter motor? Mainly hoping that the engine is not overworked due to the weight of the rider.

I would definitely recommend using an RC type motor over a scooter or other motor. These big RC motors are easily capable of driving a persons weight. My motor barely gets warm at all when running on 18.5V and I don't think it will get too hot when running on 37V either unless maybe if I was riding up a long hill. You would only have problems with overheating if you are running the motor at max power for a long period of time.

Also I found that RC motors are more powerful, smaller, lighter, and less expensive than specially built EV motors (unless you can get second hand ones).

You will want one with a low kv value which is why most people use outrunner motors for EV projects. A low kv value means you can do all the gearing in a single chain or belt drive.

[Stielz]

I have just come across this ESC on hobbyking http://www.hobbyking.com/hobbyking/stor ... OPTO_.html

looks very impressive, however it says it is not suitable for mountain boards or other similar EVs. I was wondering if anyone has tried using this ESC with a electric bike or skateboard or other EV?

[Lebowski]

I'm wondering whether this is because it's a sensorless controller. There's a good change the forced startup commutation
(before it switches to true sensorless) cannot be performed correctly when a vehicle is attached to the motor. A prop can
easily be spun on startup.... Maybe this type of controller would be OK if your motor had a clutch somewhere.

What IC where you using in your home-made controller ?

[Stielz]

I think you are right there. Im wondering how fast the motor would have to be going for it to switch to sensorless operation.. like if someone were to give a little push off before giving it some throttle would that be enough to get it straight into sensorless commutation?

Im just wondering because this would be ideal for high power electric skateboards if only it was a sensored ESC.

The IC on the board is a 3 phase gate driver IC (HIP4086), it takes care of stepping up the voltage from the micro controller for the gates of the FETs and also boosting the voltage to above the control battery voltage(12V) for driving the high side N-MOSFETs