PDA

View Full Version : 90VDC Motor driver?



Martin Hebel
02-18-2007, 01:01 AM
In lab we have a 90VDC motor that pumps oil around a loop with a flow sensor.· It had an industrial speed controller we use to interface a BASIC Stamp to for closed loop control.· Sadly, a ground loop issue cause the speed controller to go poof.· The instructor (ummmm) got in a hurry and connected the 2 with the opto-isolators by-passed (ooops!).

What I'd like to do is design a new controller for it using the Propeller for some PID control. What I am looking for is a relatively inexpensive drive that takes 120VAC, and a digital, analog or PWM input to control the 90VDC, 3A motor.

It would make a cool inerface with the propeller, monitor and keyboard to adjust the drive parameters and monitor flow.

Thanks for any recommendations as to an appropriate drive to use,
Martin


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
StampPlot - GUI and Plotting, and XBee Wireless Adapters (http://www.selmaware.com/)
Southern Illinois University Carbondale, Electronic Systems Technologies (http://www.siu.edu/~isat/est)

crgwbr
02-18-2007, 06:04 AM
I would normally suggest a FET for a motor driver, but with such high voltages I'm not sure anything solid state could take the pressure. So you'll probable have to do it with the fastest acting relay you can find and ultra low freq PWM. What king of Amperage does this motor pull. I'm sure you know how the AC-DC conversion would work. Just some BIG bridge rectifyers and capacitors. Just make sure everthing is rated for the voltage. I like these caps for really high voltage stuff:

http://theelectrostore.com/shopsite_sc/store/html/high-voltage-electrolytic-capacitors-400v-3900uf.html

Hope this helps some,
crgwbr

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
NerdMaster
For
Life

Mike Green
02-18-2007, 06:37 AM
Martin,
The IRF620 datasheet says it can switch up to 6A at 200V with a peak current of 24A. I'm sure there are other similar MOSFETs. It seems to need a gate voltage of about 6-8V to switch higher currents so you'll need to use a driver to drive the gate properly, either another MOSFET or junction transistor.
Mike

crgwbr
02-18-2007, 06:39 AM
Wow, 200V. I always thought FETs topped out at 'round 50V.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
NerdMaster
For
Life

Mike Green
02-18-2007, 06:54 AM
It's a bit pricier than the one I mentioned, but Jameco has a MOSFET for $11 that will switch 8A at up to 900V. I've seen some listed that'll switch over 1KV at much higher currents.

BTX
02-18-2007, 07:09 AM
Martin.
Like Mike said I think you could use some of this FET, IRF620, IRF634=250V, or IRF840 = 500V 8A 0.85ohm Rds(ON), but you could choose some IGBT too.
Don't forget to opto-isolate inputs and outputs, the transistor of the opto could serves you as driver for the FET (choice the opto for the outputs acordly with the minimun width of pulse, that you'll have in them), and take in care a protection..to avoid some FET driving continuosly some output...(If you'll use a full bridge to change direction).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Regards.

Alberto.

Martin Hebel
02-18-2007, 07:35 AM
Thanks for the input all, I was hoping to find some dandy little unit that I could interface too. I'm not sure I'm I have the time to do much hardware design work on this. but it might make an interetsing, if not dangerous class project :)

-Martin

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
StampPlot - GUI and Plotting, and XBee Wireless Adapters (http://www.selmaware.com/)
Southern Illinois University Carbondale, Electronic Systems Technologies (http://www.siu.edu/~isat/est)

Phil Pilgrim (PhiPi)
02-19-2007, 11:26 AM
If peak performance isn't an issue, you could do what I did for a friend's painting machine, which uses a 90VDC motor: Use a triac with an optoisolated triac driver (e.g. MOC3080). An output pin from the Propeller would drive the triac driver. You'll also need an isolated 60Hz input from the power mains. A transformer or optoisolator will work for that. Then, for a given amount of drive, all you have to do is wait for the appropriate zero-crossing (one way for CW; the other, for CCW), delay up to 8 ms (more delay = less drive), then trigger the triac with a short pulse. The motor will then be driven for the remainder of that AC half-cycle. It's not as smooth as rectified and filtered DC, but it works, and it would be an instructive circuit for your students. (If you need to close the velocity loop, you'll need some kind of shaft encoder to provide the feedback.)

-Phil

Martin Hebel
02-19-2007, 11:48 AM
Hi Phil, I was considering something similar if I have to build from scratch. The controller we smoked used a rectified AC and SCR's I beleive, where speed was a function of the SCR's on-time. For the closed loop we have a flow sensor to monitor oil flow with a pulse output that will feed into the Propeller nicely. In this class we study SCR/TRIACs/Zero-crossing opto-isolators, along with other power control devices, but being sophmores, having them work with 120V makes a tad nervous :)

How do you recommend sensing the zero-crossing? Using a zero-crossing isolator as in input to the Propeller performing a WAITPEQ?

Also, if I use the MOC3080 to trigger the triac, won't it wait to actually trigger until it crosses zero again? Wouldn't I need a non-zero crossing isolator to trigger it while in an alternation?


Much appreciated!
-Martin

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
StampPlot - GUI and Plotting, and XBee Wireless Adapters (http://www.selmaware.com/)
Southern Illinois University Carbondale, Electronic Systems Technologies (http://www.siu.edu/~isat/est)

Phil Pilgrim (PhiPi)
02-20-2007, 01:49 AM
Martin,

Hmm, you're right about the '3080 being a zero-crossing device. I was going from an apparently faulty memory on the part number. The MOC3020 would be the better choice, I think, since it doesn't have the zero-crossing circuitry.

For a zero-crossing input to the uC, I used a low-voltage transformer with a comparator and a little bit of RC filtering for line noise. I've tried optoisolators with large series resistors, but dropping 100+ volts at even a few mA requires at least 1/2W resistors. For safety and reliability, if you go that route, I'd recommend a couple 1-watters in series. And don't forget the reverse diode across the LED. Otherwise you'll exceed the LED's reverse voltage limit.

Then use WAITPEQ, followed by a WAITCNT, to do the actual speed and direciton control. In my system, the noise filtering created a slight delay in the zero-crossing detection. This had to be accounted for in the control timing.

Sophomores playing with line voltage? If you have an isolation transformer (not a Variac) that will handle the motor current, it will provide a measure of safety until the system can be buttoned up. But even then, make sure they keep one hand behind their backs!

-Phil

Martin Hebel
02-20-2007, 01:58 AM
Thanks Phil, sounds good, and yes, we have plenty of isolation transformers for when they work with 120V on their breadboards so we don't destroy too many scopes :)· Though in recent years, after having blown holes in a few breadboard, I tend to build myself or simply rip open a dimmer switch for measurement·to demonstrate and discuss safety issues measuring.
Thanks again!
-Martin