130v dc-dc pwm controller

[causalitist]

anyone know where I can buy a pwm motor speed controller for 0-130vdc that uses a 130vdc input?, or around there... i'd accept 90vdc

got a free treadmill motor 2hp, 17.5A, 130vdc. its already mounted ready to go on a 20 speed lightweight mountain bike. i have it geared for 10mph in 1st gear so peak amps wont be an issue. i have a 24v 1/2hp cim motor, it flys if i advance the timing and give it 36v, but its still too weak for me, and id rather not pull soo many amps.

i have lower voltage controllers, and dont see why i couldnt replace the mosfet(s) with a irfp4668pbf and use a voltage regulator between mosfet load voltage and the pwm part ... or seperate supply for the load/mosfet ...

i have a bunch of irf3710n i can experiment with and burn out before i use the good ones (200v, 9mOhm, 130A).

im only a 2nd year ee student, and of course they wouldnt teach anything that is applicable to an actual circuit.. its all double integrals in cylindrical coordinates inside a capacitor to determine the permitivity of the media... and that classes professor couldnt even tell me how to select a gate resistor for a mosfet!

for this question:
assume my basic 556 timer pwm controller i want to modify can supply a pwm signal with enough amps to run however many parallel mosfets i need
thanks!

 

[dknguyen]

Weird, why would a treadmill motor run off DC? But more importantly...where do you expect to get 120VDC from when you're on a mountain bike?

YOu can't simply replace the lower voltage MOSFETs with higher voltage ones. THere are other components that would have to be replaced. For example, the regulator that is powering all the low voltage electronics, as well as any large decoupling capacitors. The hardest thing, however, is the gate drive circuit that drive the gate of the MOSFET (by providing a current pulse to quickly charge up that gate capacitance so the MOSFET spends minimal time in the intermediate state between off and on where it dissipates the most heat unlike when it is fully on and conducting efficiently, or fully off so there is no power to dissipate). These have to be able to tolerate the same voltage levels as the power MOSFETs themselves and often consist the bulk of the power circuitry (the MOSFETs are just one major component, but usually much of the other power circuitry is dedicated to switching the MOSFETs).

And remember linear regulators work by burning off the extra voltage as heat. At 5V@0.1A output with even a 24V input, that works out to
P=(24V-5V)*0.1A = 1.9W being dissipated. Without a heatsink, most ICs heat up at around 40C/W. So that would be a 80C rise over ambient which is quite hot. You can imagine that at 100V, it's almost totally unworkable even with a heatsink. Wasteful too. YOu need a switching converter (a buck converter to be specific) which doesn't work by burning off heat.

You don't learn power electronics motor drives until 3rd and 4rth year (at least at my university). 2nd year is just an intro to the basic theory.

 

[Ubergeek63]

Weird, why would a treadmill motor run off DC? But more importantly...where do you expect to get 120VDC from when you're on a mountain bike?

They pretty much all do, it costs less. 90-130V motors are used for US and other 110VAC grids.

YOu can't simply replace the lower voltage MOSFETs with higher voltage ones. THere are other components that would have to be replaced. For example, the regulator that is powering all the low voltage electronics, as well as any large decoupling capacitors. The hardest thing, however, is the gate drive circuit that drive the gate of the MOSFET

Cascode

You don't learn power electronics motor drives until 3rd and 4rth year (at least at my university). 2nd year is just an intro to the basic theory.

I don't recall ever learning it formally.

Dan

 

[Ubergeek63]

anyone know where I can buy a pwm motor speed controller for 0-130vdc that uses a 130vdc input?, or around there... i'd accept 90vdc

Well Grainger has them, but they cost $350-$600 each.

We design and build them, but our customers pay $100 each for thousands a year. **broken link removed**

A 555 will not do it. A micro would do it by measuring the supply rail on a $10 capacitor and the current through the motor in real time (1000 times a second or more). The duty cycle needs to be

((Vrated ÷ RPMrated) × RPMdesired + Imotor × Rmotor) ÷ Vrail

Added to that is the PID loop...

Dan

 

[causalitist]

even if i used a 556 i will be using a push pull emiter follower to up the signal to whatever i want. i told u thats not the issue.

i will be using 35 lithium polymer batteries, 10 ah, 50A discharge.

i know i would have to replace the regulator. in fact i was asking if i can.
i wouldnt even bother though. im just going to have 2 batteries, one 12v one 130v.

that speed control is AC powered right? i have one of those.. worthless to me.. initially i thought it was just a dc speed control with a full wave bridge rectifier at the input... but apparently not.

 

[Ubergeek63]

even if i used a 556 i will be using a push pull emiter follower to up the signal to whatever i want. i told u thats not the issue.

i will be using 35 lithium polymer batteries, 10 ah, 50A discharge.

i know i would have to replace the regulator. in fact i was asking if i can.
i wouldnt even bother though. im just going to have 2 batteries, one 12v one 130v.

that speed control is AC powered right? i have one of those.. worthless to me.. initially i thought it was just a dc speed control with a full wave bridge rectifier at the input... but apparently not.

Most are, but they are more than just a speed control in the sense that you are thinking but your batteries will only last one to two hours on a charge. The motors take 5A just to run and are typically current limited at 20-35A.

I was trying to say before that the motor speed is proportional to armature voltage, which is the rail voltage times the duty cycle, minus the drop on the winding resistance. While you would still have mechanical and loop dynamics to contend with, you do not have to worry about line ripple running off of batteries.

Most treadmill motor controls us a AC transformer to supply 12V to the display and run the motor right off the rectified line and a 2000µF 200V cap. They generally have relays to drive the AC grade motor and PWM speed command. Our newer ones readily supply 1.5A for the display electronics from a switcher so you would not need the additional 12V battery.

Dan

 

[causalitist]

i just took apart another treadmill, and this one has what your talking about.. it takes ac runs it through a bridge rectifier and then to the boards inputs.
has two irf640(200v,18A,150mOhm) (ones burned out) , uses infinity sg3526bdw mosfet driver.

think i could replace these mosfets with better ones? the gate charge on the current ones is 55-72nC , it would work fine if i used something a little better, and with much lower Rds right? man would it be awesome if it could run two irfp4668pbf (Qg=151nC , 5mOhm, 200v, 90A, 520W) .. that mosfet driver can handle it right?

heck, anything 15mOhm and below will greatly increase current rating.

also...since the 130v comes in through a rectifier, it has a 1000uF capacitor to store the regenerative juice.. and a transistor that turns on when cap is full and dumps all regen into a few big resistors! nasty. the cap is right across the dc input to board.. so if i was just using batteries, i would have regen.


that other treadmill controller took ac input, had no transformer, no mosfets, only 2 lm324 and 2 SCR's and a few big diodes.. so i dont want to mess with scrs.. i want to figure mosfets out first.

what i dont get is that both of these seem to take 130vdc and put it right into all the mosfet driver, lm324 chips etc.. i cant tell very well, but neither board has a transformer.

oh, and pwm @ 20khz would go through a transformer right?
just popped in my head that i could just use a 48 volt battery, 48 volt controller, and a 3:1 transformer before the motor.. all those are cheap and buying 1/3 the batteries will save me $500 or so.... but the transformer would be huge right?