Hello all.

Im new to the forum and have only had time to look through it briefly, however, Im glad to have stumbled upon it. I am also very new to the electronics world. Though I have dabled in electronics my whole life, I have only recently begun to really study the subject.

Basically I have recently begun to build myself and E-Bike, a mountain bike with a 600 watt dc brushed motor built into the drive chain.

One of the critical components is a pwm speed controller, so that i can finely adjust my speed rather than just having an "on" mode that runs balls to the wall. That way may be fun for a little while but I value my life. lol

So.....Here is what I have done / researched so far:

PWM - I bought a basic stamp microcontroller a while back that came with a breadboard built into it. This is the sort of controller that they sell to schools to learn microcontrollers and all of their functions. I have had a blast tinkering with it, and recently realized that it had a pwm function on it. so I set up a circuit using pwm that references a capacitor timer circuit controlled with a pot. So as I turn the pot, I increase or decrease the pwm. I threw a LED into the mix to see the results, and it works like a charm. Very simple, but to a newb like me, I was quite pleased with myself. When I finish my project I will buy a more permanent microcontroller that doesn’t have the bread board, but its a good starting point to see what I can do.

So now I have a signal that is adjustable through a pot, now what I need is an amplifier to increase this signal to a useful output. This is where it all gets a little hazy for me and my ignorance really shows through. I posted on another forum more specialized to E-biking and got a good response to some of my questions, however they raised more. I understand that the next stage after the pwm mcu, is a driver" circuit (an IC) called FET drivers, or gate drivers, WHAT IS THIS? lol, more aptly where can I read up on this subject.

I also understand that many speed controllers use an H bridge for forwards, backwards, regen braking forwards, and regen braking backwards. This is getting way ahead of me and not really worth my investigation. I really only need forwards. Can I use just one driver, and one large power switching mosfet connected to my mcu?

My motor runs around 90 amps max, so Im thinking a good safety margin would be a 120+ amp mosfet. Hard to find.... can I use several in parallel?

I also ran across this schematic for a half bridge class D amplifier

This is of course for audio applications and so employs two mosfets, can I employ this same setup and simply use one driver and mosfet?

Please let me know if I can provide any more information.

Thanks in advance for all your help. I know its annoying to explain things simply, but please remember I am no electrical engineer (one day maybe

 

The output filter (L & C ) in that class D amplifier, low pass filter the output which, is good for Audio because one doesnt want the high frequency switching noise to get to the speaker. Notice also that the supplies to that Half bridge circuit are + and -. You likely wont have that because you have just a single battery right?

Is your motor DC?

Yes, you can use several power mosfets in parallel to get to your 120A rating..

The gate drivers refer simply to the circuits that drive the gates of the power fets. In your case, you wont be able to drive it directly from your microcontroller. You will need something with a stronger output than that. You can find pre-packages gate driver IC's to do the job if you dont want to design the drivers with discrete parts.

A word of caution so you dont blow all your hardware up. You need to be able to calculate your power losses and protect your FETS from all the nasties that come with switching 90A as fast as you can. Also fo rthe batteries sake, you want to minimize these losses so that you can run longer - they can be significant if you are cavalier about the whole thing.

Work out a schematic for your new output stage and post it up here with details on the motor you're driving. People on this board are very friendly and willing to offer suggestions and help you get it working.

 

I was also thinking......I remember back to the days of physics class, and when a field colapses in an electromagnet, it generates a field of equal intensity and oposite polarity for a short duration. Should I be protecting against this on the outputs of the controller? I searched a little online and came across a short blurb about freewheel diodes. Is this necessary, or negligable

 

In this forum recently somebody said the collapsing field can create a voltage spike of 4x the input voltage... in the case of a model railway I work with which uses 48vdc telco relays, this is 200v and really makes you jump! I rue the day the folks at the club left the snubber diodes off every time I have my arm on the track as a relay shuts off!

That said I believe some MOSFETs are protected internally...

 

It is dependent on dI/dt.. 20x, 50x is not uncommon.

 

You need to protect against this "kickback" voltage. Diodes are the usual method.. they need to be fast an not necessarily powerful (they do not burn power continuously, only briefly) other faster devices like MOV & TVS diodes also come to mind.

 

Any particular reason you're using both N-channels on the high and low side?

__________________
"Everything that is done in the world is done by hope." -Martin Luther
"There are two ways to live your life. One is as though nothing is a miracle. The other is as though everything is a miracle."-Albert Einstein

 

This schematic is not mine, I "stole" it off of a site explaining class D amplifiers for audio applications. I dont know why they chose to use that particular configuration, as I am not well versed in this subject, hence all the questions.

I am finishing up a diagram of my circuit to post up, but here are a few more specs. I am using a 12 volt "winch" motor off of a deep draw lead acid battery. According to the specs, the brushed motor will run a max of 90 amps. I would like to run more voltage through it and less amprage, increase speed and decrease torque, as I understand it a batteries Ah rating falls more quickly when major amprage is being pulled from it, i.e. it is not a strict linear relationship. According to my calculations and research, voltage will go farther than amprage. Again, this could be mistaken, but well see.

 

Okay, so I have been reading up alot on transistors and had a couple more questions before I come up with a schematic.

1) can I use a darlington pair of mosfets to replace a driver circuit for the one mosfet? I hear that there is a loss of speed, will this be too detrimental to a pwm signal?

2) Since the pwm signal is not a sinusoidal wave that goes negative and positive like an ac wave, I would assume i am not going to bias the mosfet at a point Q right at the center of the base vs current curve. Where should I bias to?

 

oh, also, I understand that class A amplifiers are very inefficient, due to the fact that at v=0 on the signal, they are using 50% of the max useful base current, i.e. they are biased midway on the range of useful currents. BUT...could I bias this lower on the range due to the fact that I am not worrying about an AC wave (or more simply the negative on an ac wave), thus making a class A amp not so inefficient?

 

A class-A amplifier is ALWAYS less than 50% efficient, with 50% being the theoretical maximum effiency (but obviously never acheivable).

But why on earth would you want to use class-A for such a device?, it's probably the worst possible solution, with NO advantages.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

well ill explain my reasoning. Its theoretical 50 percent efficiency comes from the fact that the transistor is biased at point Q (the center of the load line) to allow for rises that fall above and below 0v, which is good for a sine wave such as used in audio applications. This way you are taking advantage of the full range of the wave without clipping (where the wave falls above the saturation point wasting energy, or below the cutoff and not using energy). However if I am using it as a on off switch such as with PWM, I believe I can bias way way below this point (Q), slightly below the cuttoff point and have the high portion of the pwm wave be at saturation. This way it is only expending energy when the transistor is "on". This would never work for a sine wave because it would clip the bottom half of the wave completely off, sort of like a half wave rectifier would do, but since i am only dealing with half of a "wave" I dont really need to worry about it dipping into the negative. So class a is very simple, and efficient if used as an on off switch.

I could be completely wrong, and thats why I asked what I did. Please correct me in my logic if this is the case.

 

???? - you've just described (I think?) class-C, it's certainly NOT class-A or anything like it.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Hello
Reading you reminded me of my recent design where I made my mum's sewing-machine a nice PWM controller. It used tacho for feedback which kept the rpm close to 4-5% error from the commanded. I used TL494 chip which is well known and a IRF540 mosfet which drives at 24V 5-6Amps even without a heatsink mounted on the device. I put MBR20100CT diode at motor terminals for the kickback. That's all, circuit runs around 16Khz.

You may need to use different output device as your needs are much higher. Parallel mosfets is needed. Some kind of PID regulator is also necessary to keep the motor smooth running. I did this with the 2 OP-AMPs that are in the input of the 494.

There is a lot of schematics on the web about DC control.

Here is a 40A RC motor controller I found recently
http://norbique.rchomepage.com/esc/index.htm

 

Thanks for all the replies, I have done alot of reading and research, (learned almost everything from scratch) and feel pretty confidant I can give this project a go now. One question still remains. Many of the mosfet drivers out there seem to be of a five pin design, and seem to have the same outputs as a op amp, so I am assuming that they are just op amps tuned to better suit driving a mosfet.

Well I want to run this whole thing off of a large lead acid battery....just one battery, but most op amps require a dual power supply with a ground somewhere in the middle... you see my problem. how do I get a dual powersupply effect out of one battery?

My set up is as follows, 12 volt 60ah lead acid battery voltage devider to a BS2 microcontroller, pwm signal goes to a driver to a couple mosfets in parallel, which connect to the battery. There are a couple other stages, such as automatic temp shutoff and am working on a current sensing circuit to protect in the case of a motor stall, but thats the basic set up.

If anybody has experiance with this little problem of mine I would greatly appreciate the help. thanks all.