I think so. I've used them with 120/220 AC, and also 24VDC. I have several of them with me, and are extremely simple to use.

I wonder if you removed a ground loop when you installed the optos and that is why it works better.

 

I wonder if you removed a ground loop when you installed the optos and that is why it works better.

That's probably one of the reasons. But I'm sure it wasn't the only one, because the circuit still misbehaved even with the slow, old optos.

My conclusion is that there were at least three things affecting it.

• The ugly ribbon cable that I used for connecting the PCB to an oscillator and a 5VDC source that I had improvised on a protoboard.
• The lack of an RC filter on the motor's side
• The abrupt inrush current to the motor every time the mosfet switched on, producing a strong EMI

The first problem was solved by simply moving away that ribbon cable from the 120VAC source cable, and the cable going to the motor. Besides, when I finally install this thing, that ribbon cable is going to be only 1" long. It's used to connect an adjacent board to this one.
The second one didn't require much brain power, but some trial an error, until I got it right.
And the last one was solved by using a high impedance resistor (1k) at the fet's gate, as you suggested, so as to open it gradually, instead of abruptly. Of course, the risk of that was that the fet would stay in the linear region for too long and it could overheat and go *poof* on me. But so far it hasn't shown any of those symptoms, and besides, the duty cycle of the motor is only about 50%. That is, it works for about 10 seconds, and then it rests for another 10. Anyway, when I installed that fast diode at the gate to facilitate its quick discharge, the amount of time that it was spending in the linear region per cycle was cut in half.

 

I have used Honeywell for some time, they have quite a wide range.
https://sensing.honeywell.com/sensors/current-sensors
Max.

 

I have used Honeywell for some time, they have quite a wide range.
https://sensing.honeywell.com/sensors/current-sensors
Max.

Yeah, I'm familiar with the brand (it's huge!), though I didn't know about their current sensors.

Question, if these are called "open loop" sensors, then how do the "closed loop" sensors work? ... assuming they're called that way, btw.

 

I've finally put together the RC-diode snubbers for the motors, according to this diagram:

​

They don't look less of a kludge than the previous assembly, but at least they're more compact and have the proper value components, which are: 3.3uf caps@400V, a 330 ohm resistor @ 5W, and a couple of MR758 diodes.

​

 

Won't those diodes prevent the correct snubber action? Once the caps become charged they can no longer provide a low-impedance path for transient currents.

 

Won't those diodes prevent the correct snubber action? Once the caps become charged they can no longer provide a low-impedance path for transient currents.

The diodes are there to prevent the electrolytic caps from charging with the wrong polarity. And according to LTspice they work... also, they do make a difference in the working circuit. The alternative was using 15 0.22uF polyester caps @250V connected in parallel. But perhaps you have a better suggestion? (it's an honest question, I know for a fact that you're far more versed than I am in these sort of things)

​

 

The diodes are there to prevent the electrolytic caps from charging with the wrong polarity

So why don´t you use for example a foil cap?

 

So why don´t you use for example a foil cap?

I thought he already wears one?

 

The diodes are there to prevent the electrolytic caps from charging with the wrong polarity. And according to LTspice they work

I realise the need for diodes if you use polarised caps but, as your sim shows, the caps pass significant current during the first few mS only. Thereafter they do very little. IMO you'd get far better suppression by using one non-polarised cap (a 1uF plastic-film cap rated 440V AC / 1kV DC isn't too expensive) and omitting the diodes.
Here's a comparison sim. I rest my case.

 

I realise the need for diodes if you use polarised caps but, as your sim shows, the caps pass significant current during the first few mS only. Thereafter they do very little. IMO you'd get far better suppression by using one non-polarised cap (a 1uF plastic-film cap rated 440V AC / 1kV DC isn't too expensive) and omitting the diodes.
Here's a comparison sim. I rest my case.
View attachment 122596

Point taken... I'll start looking around, see if I can find a good (and affordable) high-voltage cap supplier around here... either way, I also placed a 180V TVS on the PCB, just below the motor's connectors, and another one below the power supply's connectors.

Thanks again!

 

There are some spikes you just cannot predict.
Max.

 

I thought he already wears one?

I'm wearing my favorite weekend t-shirt....

 

If you want to use electros, put them in series, back to back, with the diodes across each cap, in the direction to prevent reverse volts.

That should work ok I think.
You will need to ensure the diodes and caps are rated for the high ripple currents.

 

I'm wearing my favorite weekend t-shirt.... View attachment 122601

While it isn't a picture of me, he kind of looks like me. Beard and every thing.

 

If you want to use electros, put them in series, back to back, with the diodes across each cap, in the direction to prevent reverse volts.
View attachment 122604
That should work ok I think.
You will need to ensure the diodes and caps are rated for the high ripple currents.

Just tested it in LTspice with 3.3uF caps and a 330 ohm resistor... and the thing's telling me that I should use a 50W resistor for that!!!!

 

If I were to use a 10k resistor instead, it would have to be rated at only 2W ... but would it "snub" properly?

 

The most bizarre of things is happening...

This driver is used to lower and raise a DeWalt router that is being used as a drill. The router is set at its lowest RPM's, and when the 90VDC motor is activated by the driver, the router's RPMs plunge down... it's almost as if the voltage reaching it goes down considerably. Both the driver and router are connected to the same 120VAC source... But I know for a fact that that source has more than enough capacity to feed them both simultaneously.

What's going on?

 

Are they from a transformer? It may be close to the current rating, particularly on start.
The starting current of the motors can be very high, a lot more than the running current.
Have you a clip on ammeter? Very handy for measuring the current. Also vave you measured the supply volts to the motors to see how it changes?
A fairly small change of the voltage will give a noticeable speed change of the motor.

 

It may be close to the current rating, particularly on start.

Like I said, the source is more than capable of delivering the appropriate current.

The starting current of the motors can be very high, a lot more than the running current

This phenomena presents itself throughout the whole cycle (about 4 seconds) of the DC motor, and not just during startup.

Have you a clip on ammeter? Very handy for measuring the current. Also vave you measured the supply volts to the motors to see how it changes?

Haven't done that. Now I will. It's the sensible thing to do first, as you've suggested.

A fairly small change of the voltage will give a noticeable speed change of the motor.

Quite true. I'll be back tomorrow with more observations, if I can. Thanks for your help.