Where does PWM fit on AC / DC component usage?

Thread Starter

cspwcspw

Joined Nov 8, 2016
78
AC fuses, switches and relays are generally rated higher for AC usage than DC usage. I understand that is because the AC wave crosses zero, which will turn off any DC arc. So DC fuses are often spring-loaded to open a bigger gap to stop the arc. My little relays say they're good for 10A @ 30V DC, or 10A @ 250V AC.

Now I'm driving DC that is PWM chopped (with a guarantee that it will always be zero for part of the period). Would the AC reasoning apply - arcing will stop, so an AC fuse could work, or maybe we can safely pump more chopped DC through a switch (even at high duty, but not 100% duty cycle). I'm looking at 10kHz to 20kHz chopping, and I'm guessing any DC arc should die down in microsecs. (Or does the plasma hang around and the arc might survive a too-short interruption?)

The question is more for insight than deployment right now - I'm just trying to appreciate some of the boundaries and trade-offs.

Thanks
 

DarthVolta

Joined Jan 27, 2015
521
Ok what type of thing uses a coil/relay at that 10-20kHz ? I guess some engines might, but what about just going right to a MOSFET type circuit ?

I want to know all plasma physics, so I'll learn some more of it, but yeah, I've seen video's of plasma's arcing in circuit breakers, does LTSPice simulate that ? It probably can
 

Ian0

Joined Aug 7, 2020
9,668
I don’t think cspw’s switches were opening and closing at that frequency. I think he was feeding pulsed current through a switch-contact that opens and closes fairly infrequently. Is that correct?
(Still don’t know why)

By the way, it is quite common drive relay coils (especially big ones, and contactors) at 10-20kHz PWM to reduce coil power dissipation once the contacts have pulled in, but the contacts don’t open and close at that speed!
 

nsaspook

Joined Aug 27, 2009
13,081
It's never simple as arc physics can be complex. Pulsed DC arcs are just another factor in the mix.

https://www.nature.com/articles/1761019a0.pdf
https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.939.9774&rep=rep1&type=pdf
A. Cathode Root Emission Mechanisms The emission mechanism at a cathode root will vary over the duration of an arc and will affect the arc motion. The cathode root is the source of electrons for the current through the arc column, and is generally considered to dominate the movement of the arc column. There are two main emission mechanisms, thermionic and field emission.
B. Thermionic Emission Thermionic emission occurs when a refractory electrode material is heated to a sufficiently high temperature. Thermionic cathode arc roots on refractory materials have been observed to be stable and stationary with current densities of 10–100 A/mm
C. Field Emission Electrons can be extracted from a cathode material by strong electric fields. Although the average field strength between the arc electrodes in an MCB is not high enough to cause field emission it is thought that there may be a high field strength local to the cathode surface.
 

Tonyr1084

Joined Sep 24, 2015
7,853
My little relays say they're good for 10A @ 30V DC, or 10A @ 250V AC.
So it's good for 10 amps whether you use 30VDC or 250VAC. I think the rating says it all. Since it's good for 10A at 250VAC I would just go ahead and assume the contacts are good for 10 amps at 120VAC. Or 70VAC. I wouldn't go over the 250VAC rating because of potential arcing. Since DC is the hotter current I wouldn't exceed the 30VDC either.
Now I'm driving DC that is PWM chopped (with a guarantee that it will always be zero for part of the period). Would the AC reasoning apply
AC crosses zero twice per cycle. However, assuming you switch exactly at the zero crossing - you're not going to get any arcing. However, there's no guarantee, and it's very unlikely you'll be able to hit the zero crossing every time; or even close to once in a while. Same would be true of PWM. While there is a zero period, switching at that moment depends on how long the zero period is AND if you happen to catch it right at zero. But on the other hand, the same thing is true that there is a period of time with PWM where the voltage is at full pressure. (on). Switching at that moment has to be considered too. So at 30VDC PWM during a HIGH signal, switching through relay contacts at that moment will have the same affect as if you had a constant 30VDC (no PWM).

Just accept the relay's ratings and don't exceed them. It's better to over engineer than to under estimate the capabilities by some possibly faulty reasoning. AC is AC. It has an ever varying voltage. It's neither zero or full on ONLY, it has other voltages depending on what time you throw the switch. Could be at 2V (AC) and climbing (or dropping). Could be at 93V (AC) (going up or down). Could be anything in-between full on and full zero. And lets not forget the negative side of the sine wave. It could be -2V(AC), -70V (AC) or even -120V (AC). AC is AC. It has positive and negative components to the voltage and to the current. So switching at any given moment will all depend on exactly where the AC wave is at that instant. Why mess with it? Just get a relay rated for the load you want to control.
 

Thread Starter

cspwcspw

Joined Nov 8, 2016
78
Thanks for the useful responses. My question was a bit vague, my apologies. I have no AC, just PWM DC driving a fixed-magnet brushed golf cart motor, with PWM chopping to control torque, presently at 12kHz PWM. No relays. I put about 33 amps (average) through the motor on the bench recently for a short burst (maybe about 70% PWM duty cycle). My fuse is a DC 30 Amp car fuse which did not blow, so it looks like the fuse might handle a bit more than 30 Amps average provided it has some brief off-time to recover in the PWM cycle. But that got me wondering whether an AC fuse would work instead, or whether, if I just want a big red "kill" switch, do I need to size it for 30A DC or 30A AC. My voltage is modest, two 12 volt lead-acid batteries (so probably about 26V-27V when they're charged and in series).
 

Ian0

Joined Aug 7, 2020
9,668
You really should have mentioned that it is an inductive load. The current from a PWM motor speed controller will most certainly be DC, with no interruptions. When the output MOSFET is off the current continues to flow through the diode.

It takes at least 1000 seconds at 40A to blow a 30A automotive fuse. A 10x38 industrial fuse takes 10000 seconds at 55A (see page 16)
https://ep-fr.mersen.com/sites/mers...rule-Fuse-Links-10x38-gG-400-To-690VAC-EN.pdf
The big red kill switch does require the driver to have the presence of mind to press it before he jumps off when his golf-cart is starting to catch fire!
There's something about the 10x38 sand-filled fuse that inspires more confidence than the piece of plastic that is today's automotive fuse. Having said that, you don't hear of many car fuseboxes catching fire.
Do you have overcurrent protection in your motor speed controller? Otherwise what stops the motor burning out when someone is trying to drive up rather too large a kerb?
 
Last edited:

Thread Starter

cspwcspw

Joined Nov 8, 2016
78
Thanks Ian0 - very helpful. I'm a retired hobbyist, so this is a kind of "interest" project that took me way out of my comfort zone. The cart, not working, was given to a disabled friend in the hope that it could be repaired and used as a mobility scooter. It went through a few hands, each of which seemed to do more damage. It has a Curtis 1228-2730 controller that arrived opened, with a few blown power Mosfets, some really bad repair soldering, and the accelerator and front-panel feedback mechanism in bad shape. I think I've got the controller sorted out now, but I don't have a way to get into the software or settings, and last time we tried it on the motor there was a lot of smoke (the Controller survived). It turned out someone had "tested" the motor by directly putting 24 volts across it while the clutch/brake was engaged. So we pulled the motor apart, managed to unsieze the clutch / brake mechanism, cleaned the motor up, and it is now running reasonably under no-load on my desk and my chopper.

I built my controller purely for bench testing the motor under zero load, watching things on the scope, trying to understand the power Mosfets, the snubbing, the flyback, etc. And watching lots of videos on Mosfet circuits and motor circuits, PID control systems, etc. My build is not an H-bridge, not even a half-bridge - just a single Mosfet chopping the DC, and some monitoring of current and temperature. So it can't do reverse, regenerative braking, battery charge management, etc. But nowhere near the power that the real controller can allegedly deliver.

So the next move is to see if the real controller and motor can play nicely together on the bench. If we get over that hump then there is still quite a long way to go - put the transaxle back together, fix the front-panel stuff on the cart, get the flat tyres sorted out, etc. But that sounds like "on the home stretch" kind of stuff.
 

Ian0

Joined Aug 7, 2020
9,668
I know the Curtis controllers. They earlier (non programmable) ones were a lot easier to deal with! Do you know of 4QD? https://www.4qd.co.uk/
Don't leave out the diode! (The one between MOSFET drain and V+) otherwise you will have a very short lived MOSFET as soon as you put any power through it. It should be a 40V Schottky, with the same current rating as the motor.
Some of the more basic Curtis products use exactly that circuit = PWM+MOSFET and DIODE. Not everyone needs to go backwards!
Aren't you glad it wasn't a three-phase brushless DC motor with hall-effect position sensors and all that stuff.
 
Top