Simple 555 PWM - disappointing performance

I've tried to cancel my other response, that I dashed off before even bothering to read the rest of the thread, so if it got through anyway, please ignore it.

(I've left the one posted where I suggest the slide pots for soft braking.)

Thanks, Rich

The existing circuit has been working reliably for nearly a year.

A small, robust geared screwdriver motor is surely unlikely to suffer from problems of heating and vibration?

What switch are you referring to in "...when the switch is off, the motor keeps turning..."?

The output torque must be closer to ~ (duty)^2 and not ~(duty), so that at duty=50% you will get somewhat less than torque equivalent to

Base is driven from 12V, not 5V, as shown in screen shot. Hence plenty of base drive across 100 or 50 ohm base resistor. (I wonder if that's why others have also mis-identified low base drive as a problem? Maybe I should have added '12V' to the schematic as well as showing the waveform.)

I don't follow you on that one. I'd *like* to get the collector voltage swinging down to 0V! But please see my reply to Bob Eldred, including my corrected 'scope shots. I set the scale wrongly on that earlier one, adding unnecessary confusion.

Bottom line though is that Ed and Bob Eldred correctly identified the cause: power supply. See my reply to them, and latest screen shots.

Correct.

No, diode polarity is OK. (Now a 1N4006 BTW instead of the puny

Ed and Bob Eldred correctly identified the power supply as cause. See my reply to them, with latest screen shots.

Bob, Ed: Thanks - you're right, the cause of the problem was my bench power supply! It was not delivering the voltage it was supposed to under load. And the set of NiCads I used yesterday also proved to be suspect. However, I'm pleased to say that after re-charging them overnight they sustained 4.0 V throughout today's tests. I'm now getting the sort of result I expected.

Nothing wrong with the base drive, BTW. I'm still puzzled why everyone except Bob Monsen thought it inadequate?

I also carelessly set my PC 'scope's scale incorrectly, so that earlier screenshot was a bit misleading. And adding to ambiguities, one of my probes was ~ 100 mV adrift from zero when shorted, and I've now nulled that out. Here is the latest result:

So...apart from my screenshot carelessness, it was all down to a poor supply. I'll make that my first check next time .

The only remaining 'disappointment' is that, even with the circuit properly powered, torque at 50% duty cycle is poor. Low enough for me to be able to stop the motor easily by hand. That's almost impossible with the directly driven motor. Even with the original screwdriver (2 x 1.2V C-type Nicads), it's very hard to stop it. I'd assumed that PWM would reduce speed, but virtually maintain torque? IOW, this add-on circuit would let my curtains close at half their present speed, but still do so reliably, overcoming the considerable friction. Now I don't think that will be the case. So what *is* the effect of PWM on torque?

You can talk about the 555 sinking/sourcing 200mA all day long, BUT you still have to deal with the fact that the 3055 isn't being driven hard enough to work as a switch for you. You are asking a very poor amplifier to do the job of a saturated switch. If you want it to work you need to do the following : Use a FET or at least an NPN darlington. Change the 4148 out for something like a 1n4004, or better yet something rated for 3A or more like an SB560. Make sure you have good decoupling on your 555 circuit.

Jim

Me too, but I still stand by my observations. Just because it is working now does not mean I would rely on it to continue far into the future. Parts age, and it is best to think in worse case mode.

Jim

The voltage is proportional to the speed of the motor. The current is proportional to the torque.

vic

Guilty as charged - I thought you were running the 555 from between 4 & 5 volts. Max base I would be < 50 ma ...

Glad you got it working. I think you may get better torque with a higher frequency.

Here's another possible approach: run it direct (no PWM) and move the limit switch actuator to shut the motor off sooner. That will give you full speed. Alternatively, you could go with two limit switches like Rich (?) mentioned - the first one slows it down (via the PWM circuit or just a resistor) and the second one stops it.

Ed

I think that's also been adequately explained by now, so not to worry. :-)

This is interesting - I'm nexting through the thread, and the last (most recent) post said TP should use a lower frequency. This could be a whole topic of its own. My last successful (in the smoke test sense - no commercial potential, because the "customer" turned out to be a swindler) PWM ran at about 100 Hz. It drove a 12V, 1.5A PM motor, which is probably comparable to a screwdriver motor.

Well, It was I who suggested something like this, but actually, since I like doing stuff with physical hardware, I had suggested an arrangement of a couple of slidepots, which would give a very dramatic (theatrically) slowing effect, such that it comes to a complete stop just as it actuates the limit switch. There are those who would call it a klooge, but "those people" are probably of the same ilk as those who misspell "klooge", to make it look like it rhymes with "sludge." ;-)

Cheers! Rich

Yes, screwdrivers aren't noted for coasting a lot! :-)

Of course, I like my own, which the determining factor is a constant- current driver - really quite simple, just the standard emitter resistor and voltage-regulated base drive, but it gave full torque down to zero RPM.

It ran at about 100 Hz. The guy whose design I was one-upping at the time had evidently used a freq. somewhere in the Khz range, because one of the complaints from the customer was that it "whined".

I've since misplaced the schematic, but it's trivial. A two- transistor astable, where the two base resistors are replaced by a pot with about 1K at each end, and the wiper to +V. Standard emitter follower, a couple of 1N4148s, and a TIP33C current sink. I picked the TIP33C for its robustness, and MPSA06's for the little trannies, for the same reason.

But I had 20-40V of headroom, and needed the robustness because it was supposed to be in a MIG welder spool gun.

Wanna buy one? You've seen the pic of part^H^H^H^Hmost of one of the prototypes:

The TIP33C mounts to the heatsink, of the same footprint, but with about 3/4" fins - if anybody asks, I'll go dig out a whole one.

Ah, heck. Before hitting "send", I always like to take a little break, such that if I make a complete spectacle of myself, I at least know I'm being foolish. Well, I went out to the RV to refresh my tipple, and they just jumped out at me. There's a couple more pix there, which I've gone ahead and uploaded just because they're there. You should be able to figure out which is which:

Yeah, they all exhibit crappy workmanship. Sorry, but I was underpaid, and I've always been terribly, terribly lazy. If I got paid accordingly, I do have the capacity to really dress stuff up. (of course, they always say, "dress up first!" Screw'em. I'm too old for that crap.

And I don't know if there's anybody with significant money who'd want to pay me enough to cut off my gorgeous hair. ;-P

BTW, the pot isn't shown - it mounts off-board, and solders to the three turret terminals on the right-hand side. The DC input is on the left. The motor negative lead goes to the central turret terminal where the collector of the TIP33C is. I don't remember what the other turret terminal is for. :-Z

Cheers! Rich

From Motorola's data.

Saturation is measured at Ic/Ib = 10

@ 3 A Vce sat = 0.5V ; Vbe = 1.25V

That may help explain the 'high' Vbe commented on by others.

Graham

Ah, that explains it, thanks. Intuitively I'd have expected a linear relationship, with torque proportional to (average) current. And therefore I'd hoped that I could use a supply of say 4.8V @ 50% duty to get same torque as 2.4V direct, but at half speed. From that ^2 relationship, I'd need to go to 10V, which might be a tad *too* high for a motor originally using two Nicads!

Is that ^2 proportionality easily derived?

Thanks, but see Fred's message and my reply.

Thanks for all the follow-ups. Appreciate the advice.

Ed: Potential snag with 'earlier switch off' is that curtains are then unlikely to reach their limits. In particular, when opening, the last few inches requires the greatest torque, because of 'bunching'.

James: May well try a MOSFET to see if there's any noticeable improvement.

Didn't realise you were running the 555 off 12 or 15 V.

Also haven't used a 555 in ages - decades actually ( well actually I just did recently - but just driving a tiny load and wasn't interested in the load current specs ).

Have you tried stopping the motor when driven with a DC voltage of 1.2V ? Is it comparable to the 50% duty cycle ?

Note that when the 2N3055 is off - the motor continues to rotate but acts as a motor-generator and isn't it loaded by the 1N4004 ( as it is now ) . That is likely to reduce both torque and power.

Half bridge would eliminate this effect.

Graham

No as this means the power supply delivers current to both rotor and stator- but it is the only other explanation given that your measurements and circuit are shown correctly.

I just measured a small screwdriver type motor. It came out at 6mH. If yours is similar (or worse) then at 300Hz the inductive current will never rise to it's optimum value. Maybe only gets to 60% of where you think it should be, which makes the apparent torque even worse than square law. Chop the motor at a much lower 50-100Hz. regards john