PWM control problems

[cincboss]

Hello all,
new on here so please be gentle .

Having problems with bought in pwm controller boards for a new product which MUST have reliable motor speed control.

Motor: 24v dc 3.3a 60w pm.(Parvalux)
Circuit is 12+12v (24v)toroid transformer,rectifier, smoothing cap,pwm board, speed pot in hand held unit.

First board tried was Velleman K8004 which worked fine for a while then would only give full speed(I assume one of transistors failed)at 6.5a max,8-35v it was probably running close to it's limits.

Second board was US made 8-36v, 24a cont. inductive load which appears to be very well made (currently on ebay) and also worked ok for a while (tested by varying speed up/down switching on/off numerous times over a few days to simulate customer use).This also failed after a while but motor was not getting any current although meter showed 38v!!!! in circuit before pwm, with pwm disconnected meter gave 25v.

There is a diode on both boards near the output which I assume is to prevent damage to the transistors, would a second diode close to the motor brush leads be more effective or is there something blindingly obvious (once you know) that I have overlooked?

Does the smoothing cap (10000uf) need a bleed resistor, if so what value?

Would a simple rheostat generate a lot of heat, I could live with the loss of efficiency?

Would a board with a higher voltage rating solve the problem?

If anyone can shed any light on this I would greatly appreciate it as this is getting expensive.

Many thanks.

 

[shortbus=]

Hi, you might want to check out; CNCzone.com-The Largest Machinist Community on the net! - Powered by vBulletin

That forum is dedicated to CNC. If they can't help you, no one can!

Cary

 

[tblo163]

PWM Motor speed control

I suggest you build your own PWM,then you can rate it to your requirements.
Use LM324N for the driver and 2 high powered mosfets,allways incorporate a large heat sink & fan.and mount the LM234N on a 14 pin socket,to avoid over heating when soldering.If you need any data sheets,let me know,here's one for starters;-

 

[cincboss]

Thanks for replies, but what I am trying to ascertain is why a pwm controller which is more than adequately rated (the US board) according to the specs should fail after a day or two working ok. (Parvalux and others do offer controllers but the prices make a nonsense of the economics).

 

[tblo163]

As I suggested,build your own PWM,I've been running mine at a constant 30amps for about 2 years,and it's never missed a beat.The most common reason for failure is inadequate cooling of the mosfets,I make my own heat-sinks,but a very usefull one is a meaty CPU heatsink and fan from an old mother board.you can mount the mosfets directly on to it.With regards to cost:-the LM324N & socket is less than £1, resistors,caps and diodes,are also very cheap,if you "scavenge"any disused circuit board,you can accumalate lots of very usable components.I use strip board or verry board to populate the components on,the tracks are ideal for the layout,with the use of added jumpers,and you can beef up the source & load tracks with solder.Don't buy chineese PWMs from ebay,they are usless!I would be happy to send you a schematic for the lay out.With regards to rheostats,they will burn out very quickly,and get very HOT!

 

[cincboss]

Thanks for that, I was considering the chinese products but having heard stories of components mounted wrong way round, etc I think they are best avoided as you say.Will look into making my own but just wish I could pin point what is causing the problem, possibly inductance in the motor, heat or back emf? The heatsink seems to stay cool but the cpu & fan sounds a good idea.
A circuit board schematic would very useful, thanks.

 

[tblo163]

PWM.Diagrams

I attatched a PDF for you to look at. Pages 1 to 3,show wiring diagrams from different manufactorers, Page 4,shows my design,with the following pages,with info,components list and assembly.As you will see,pins 2,8,9&12 are connected together,the 14 pin socket gives space to fit these jumpers while it's being installed.Pin 4 is connected to VCC + & Pin 11 is Grnd;If the socket is mounted astride the break in the strip board then the rest is easy.
Any more info? just ask.Terry.

 

[cincboss]

Hi Terry,
many thanks for that, your circuit looks as though it should solve the problem.
I note you have two caps and a diode across the motor terminals, I wonder if the lack of them was part of the problem with the original boards?
Thanks again
Dave