C16J Thyratron vs SCR switching speed and possible damage to motor insulation

Some have pointed out that DC and AC motor controller switching transients and inductive voltage spikes can damage the insulation in electric motors (and other wound devices like transformers). Historically, some inverters and motor controllers (both AC and DC) are less than kind in their delivery of switching transients. For example, so called 'modified sine wave' AC motor controllers typically use MOSFET, IGBT, or bipolar transistor on-off switches to deliver voltage wave forms that are only step wise approximations to a smooth sine wave. The output of these controllers look a lot more like square waves than sin waves. "Inverter rated" motors have extra good insulation on the windings and are designed to resist the negative effect of these less than perfect controller electronics.

So the worry about SCR switching speeds in a solid state thyratron replacement is logical and justified.

I've done some tests to see what the differences in switching speed are between SCR solid state switches and the original C16J tubes in a modular 10ee lathe drive. My test drive has a 5hp Louis Allis DC motor. Ideally I would have liked to measure the voltage between armature windings directly, but as the motor armature is obviously rotating I don't see a way to do that. Instead, I took measurements across the rotating motor armature, the motor field windings, the modular drive main transformer (T5) secondary windings, and directly across the thyratrons.

A suggestion was made to do measurements with the motor loaded at half power in order to produce the most negative transient conditions. Since, on my test stand, I have only my hands on the motor pulley as a brake and have neither the skin nor the strength to make that very effective, I did tests under no-load conditions. Loading lengthens the thyratron on time, but doesn't seem to have much effect on the character of the switching transitions.

I did the tests at a base speed (mid position on the speed pot) with the armature voltage at maximum and the field unweakened.

The first photo below shows about one cycle of the voltage across the armature (red) and the armature current (yellow) with the original C16J tubes. The ramp portion of the top trace is the period that the tube is conducting. At the start of conduction there is a sharp voltage transition, and then another at the end of conduction.

The second photo shows a close up of the start transition. The rise time is about 10 us, and there is a slight undershoot. The third photo shows a close up of the end transition. This is much slower and the rise time is more like 80 us.

The photos below show the armature voltage and current with the SCR solid state thyratron replacements. For these, the start rise time is again about 10 us. The end rise time is a bit shorter at maybe 60 or 70 us.

What has surprized me about these traces is that the SCRs don't seem to switch any quicker than the tubes. I would even say that the SCR start transitions are a little gentler than for the tubes and also with very slighly less undershoot.

Since, under ideal conditions, I do think the SCRs do switch much faster than the tubes, this is leading me to think that the dynamic performance of the drive and its switching behaviour is much more dependent on the magnetics and composition of the whole circuit than on the particulars of the switches themselves.

The photos below show the voltage across one of the thyratrons. The first two photos are one cycle and a close up of the conduction start for the C16J tube. The second two photos are for an SCR replacement.

In the first photo of each pair, the tube is conducting during the flat portion of the trace that starts in the center. During this conduction period, you can see the voltage across the tube is pegged at zero. You can see the parasitic effect of the opposite tube conducting as the zig zag shape on the subsequent rising part of the trace.

Once again, there is nothing spectacular happening at either transition, and both the original tube and the SCR version are performing in an almost identical fashion.

As an (hopefully talented) aeronautical engineer with a deficient ability to understand EE stuff, thank you very much for this work.

I do not understand why you are looking at switching speeds as it is totally irrelevant in a motor controller. What is relevant is cut off current of the device, tube or SCR. These devices turn off only when current through the device drops below a designed threshold current. This area is typically where device failure occurs. Further, the use of an "O" scope is difficult to decipher because the scope will only recognize voltage not current. Additionally, your scope trace will be further confused by the back EMF from the collapsing fields of the motor windings, which of course is the source of these transients because the back EMF rule is that the induced voltage of the collapsing magnetic field of the winding will be opposite in voltage polarity and twice the amplitude of the driving voltage. In point of fact these transients are necessary for the correct operation of the circuit. So, I don't understand the point the OP is trying to make.

Ok, how many burnt up EE drive motors have shown up on this forum..... How many burnt to a crisp from using any aftermarket controller?

Some may have pointed this shit out, because they don't know how to run the machine, so they stand around it thinking this crap up.

One of the better aftermarket drives in the past was the Sabina, done by smart engineers, only problem was they had no experience actually running the machine, so it failed.

donie said:

Ok, how many burnt up EE drive motors have shown up on this forum..... How many burnt to a crisp from using any aftermarket controller?

Some may have pointed this shit out, because they don't know how to run the machine, so they stand around it thinking this crap up.

One of the better aftermarket drives in the past was the Sabina, done by smart engineers, only problem was they had no experience actually running the machine, so it failed.

Click to expand...

Donie,
Tube Thyratrons and SCRs pretty much work the same way. Back EMF will stress the wire insulation used on the motor windings because the flyback voltage doubles, but the motor failures you are referencing are motors made when they only had varnish and shellac for insulation and that stuff does break down over time. Those motors are old. I'm pretty sure the insulation failures were not caused by the controllers, but most likely simply age.

steve-l said:

What is relevant is cut off current of the device, tube or SCR. These devices turn off only when current through the device drops below a designed threshold current. This area is typically where device failure occurs.

Click to expand...

Steve, I agree with you somewhat. One reason I think there is no problem in the drive (and no danger to motor windings from SCRs or glass tubes) is that when the thyratrons turn on there is no current flowing and when they turn off (as you point out) there is also no current flowing. The thyratrons are never switching off significant amounts of current which is the condition that causes inductive elements to fly away at high voltage (which can obviously damage insulation).

What matters for insulation integrity is conductor to conductor voltage. You can certainly measure that with an oscilloscope. Someone raised the issue that abrupt switching can cause an inductive spike and I think looking at the switching edges is a good way to see if anything like that is actually happening. It doesn't.

As a point of interest, I know from hard experience that in the case of solid state SCRs the thing that kills them in the 10ee drive is the very high voltages (due to back emf) that occur during speed changes when the motor is trying to adjust to a new equilibrium. These voltages can hit several thousand volts (I know because I've measured them). These are gentle events that persist for milliseconds and not inductive spike events.

Thanks to Timothy for the tests. You can definitely generate thousand volt spikes with this sort of equipment. I don't recall the exact circuit, but I was doing something with a magamp feeding a three phase full wave bridge made of rectifiers used to power the traction motors in Diesel electric locomotives. Fortunately, I didn't have it hooked to my lathe. I got into some sort spike generation mode and took out the whole mishpucha. Six expensive diodes gone just like that. I learned to put MOVs across them in the future. This sort of damage is not an old wives tale, but very real.

Bill