Almost off topic electronics question!

[Jaakko Fagerlund]

As far as the chip driving the bases of the driver transistors, the base current will not increase beyond normal because the emitter voltage also rises as the transistor conducts. The emitter will actually rise to a point where the transistor ceases to conduct when the gate capacitances have charged up to the voltage available from the output of the regulator. Base current in Q1 will drop to 0 even though there may be 12v on the base. The same thing happens with Q2, but in reverse voltage as the drive voltage from the chip swings to 0.

Ah, this seems to explain it The transistor basically sees a peak current going through the base and the collector. Gotcha now.

For some reason I would still use a MOSFET driver IC instead of transistor pair, but in some cases it probably can add 1 dollar to the price.

[Barrington]

I haven't done the math, but I think Barringtons figures are good to use. Otherwise I'd go with using ratios that would put about the same voltage on the pins as the original circuit would have with its lower supply voltage.

Hopefully that amounts to the same thing !!!


As regards C3 I've got a nagging feeling that things may be a little more subtle than they might look at first sight. Note the feedback isn't simple DC...

As a rule it's usually better to change one thing at a time, and I would be tempted get the supply mod working first, then worry about C3 as a secondary issue.

Cheers

.

[The Artful Bodger]

I put it all together and it works! Of course when I came to look in my supply of bits I had to vary the resistors somewhat and wound up with 68k in place of both the 20K and the 18K resistors, but it works and I seem to have the expected speed range.

I have not put C3 in.

Thanks for all the help.

[Barrington]

wound up with 68k in place of both the 20K and the 18K resistors, but it works and I seem to have the expected speed range.

But do you have the regulation ???

It's a rather over-complex circuit for a poorly regulated controller

Cheers

.

[J Tiers]

For some reason I would still use a MOSFET driver IC instead of transistor pair, but in some cases it probably can add 1 dollar to the price.

if you are concerned about drive capability, that can only be assessed by referring to the total gate charge specification, and the desired switching transition time. Then you know what current is required, and can choose a gate resistor value.

it is common to be able to use a considerably larger gate resistor than one might think, in which case even a few hundred milliamp drive capability may be more than adequate. Most controller ICs can supply that, and many can provide an amp or more peak.

The TL494 has an output setup that is not ideally set up for mosfets or IGBT. It is better for bipolar transistors, since it drives either high or low, but not both, relying on a load resistor for the opposite drive.

But it may actually be able to do a 'totem pole" configuration... the output transistors may be used either as common emitter or common collector, and are driven out of phase in one mode. No apps info shows it (there may be a very good reason), but they might be able to be 'stacked" to provide both high and low side drive. I didn't see info that prohibited it, but I have not searched thoroughly. There seems to be a maximum duty of 45%, with a 5% dead time, so "shoot through" at the driver transistors seems not to be an issue.

Note the feedback isn't simple DC...

Did I miss something? it looks as if the feedback is returning the average voltage of the motor "-" terminal, which is compared to the supply voltage (and the speed command) to stabilize the motor net voltage. The error amp is a heavily averaged integrator, which should take care of the pulses.

The polarities are odd-looking, but in this chip the ramp comparator is inverted from many other chips.... zero duty is a high voltage at the FB pin.

I agree that it might be almost as good to simply rely on the percent modulation set, open loop, if you know the DC voltage is reasonably steady.

[The Artful Bodger]

But do you have the regulation ???

It's a rather over-complex circuit for a poorly regulated controller

Cheers

.

It is a bit difficult to measure the regulation but it does appear to be working as I hoped for. It will turn really slowly with significant, but not infinite, torque and I can vary the load at higher speeds without obviously changing the motor speed. I suppose I could put a 'scope on and see if the wave form changes under load do you think that would be a useful test?

John

[darryl]

I'm wondering about the regulation as well. Offhand it would seem that it should be pretty good, but it would be nice to know what happens with variable loads.

One thing I like about the circuit is that it's fairly simple and small. I've always figured that there's a lot of overkill these days with cpu's, etc- why go complex when a simple circuit will do the job just fine. I like the elegance of simple circuitry if it can do the job.

The 555 timer IC is one which I have more experience with than the TL494, and there are ways of using it which aren't usually shown in literature- ways which can add functionality without adding a lot of extra parts. In many cases, it might be just a few resistors. As a PWM motor controller for example, it can have feedback added to give better speed regulation under variable loading conditions- not something you might normally do with a 555. The 494 is probably a better chip to do this with, though. It's one of those circuits which should probably be designed onto a PC board, along with inputs and outputs, then become a sort of generic circuit. You might make up several and have them on hand for any time you need to control a motor. I've made up 555 circuits like this- reach for one, possibly change a few parts, then it's ready to go.