Hockey Puck SCR drive testing

[vator]

I am currently testing hockey puck SCR’s on an elevator drive that smoked. These are still compressed between the heat sinks, all wires are disconnected and the assembly removed from the unit. I am using a continuity checker that consist o 2 AAA batteries in a tube with a probe at one end (+) and a lamp at the other end, with an alligator clip lead attached. With the alligator clip mounted on the cathode, a jumper on the probe, I’ve placed the probe to the anode, and then gated the SCR with the jumper. The lamp lights and stays lit with gate jumper removed. I have found no shorts between cathode and anode. However, on a few, when I have attached the clip to the cathode and placed the jumper from the probe to the gate, without even touching the anode, the lamp has lit, although dimly. Is this normal or am I getting some type of bleed through? Should this SCR be replaced? Also on each heat sink there is mounted a large canister capacitor along with an MOV and a metal resistor. Everything checks well with a DMM. What is the purpose of this, current or surge suppression or to commutate the SCR? Any comments would help. Thanks Vator

 

[jbartos]

Suggestion. Since the devices are linked to the elevator, it will probably be better, if each SCR has its characterics checked. The circuits attached to SCRs could be snubbers.
Visit

http://www.eupec.com/p/editorials/Apppress.htm

(Technical Explanations "Thyristors for Phase Control", total Download 1,0 MB)
etc. for more info

 

[rawelk]

The SCR stacks where the lamp was dim even w/o gating the SCR are suspect - leaky SCR, failing RC network, MOV, or some combination thereof.

What I often do is to use a variable-voltage megohmmeter
(something like an AEMC model 1000) that offers, 50, 100, 200, 500, and a 1000 volts range. This allows stress testing at a point near the SCRs operating voltage (that is, if the SCR is in a 460V bridge then use a test voltage of 500 VDC or less). DON'T use substantially higher than line voltage, or the SCR junction can be damaged.

Disconnect one side of the "snubber" network (this is the RC network with paralleled MOV) wehich is paralleled with the SCR. The MOV clips voltage transients, and RC network helps to keep the dV/dt within SCR specs.

With the snubber network disconnected the SCR itself can be tested. If the SCR is not shorted then the resistance across it should be fairly high (usually, several megohms) regardless of megohmmeter polarity.

Then, connect the megohmmeter across the snubber, and repeat. There should be an initial 'kick' while the capacitor charges, then build up to a higher resistance. If the resistance doesn't build up then the capacitor is probably shorted or leaking badly, or the MOV is breaking down (split up the circuit further to figure out which is which).

Note: discharge the snubber before handling it (for instance, using a 50 ohm, 50W resistor), because the cap will pack quite a punch.

Then (like you are already doing) check that gating occurs, and the SCR will pass current (although you might want to put together a simple, unregulated DC supply, and provide a bigger - say, 500W - load, in case the SCR is OK at very low currents, but flaky when passing higher amounts of current).