3
\$\begingroup\$

I have a very large high voltage device that a friend asked me to take a look at. It has a bunch of BIG thyristor packages (Semikron, SKKT 106/18 E, Dual Thyristor Module, 600A 1800V).

I'm testing them on a Fluke 87V multi-meter, and they all read EXACTLY the same.

In Diode mode, they all read 0.018v between gate and cathode in both directions, open circuit on all other pins.

In Resistance mode, they all read 18.1 ohms between gate and cathode in both directions, open circuit on all other pins.

I would assume that this means that the devices have all failed in exactly the same way with a gate to cathode short, but it would be very odd if this was the case.

Does anyone think that these devices have failed with a gate to cathode short, or would you expect this behavior on a multi-meter?

\$\endgroup\$
  • \$\begingroup\$ Can you see any forward voltage if you inect a bit of current into the base? \$\endgroup\$ – winny Oct 29 '18 at 13:47
1
\$\begingroup\$

Try putting 180mA or so into the gate (positive wrt the cathode) and check for forward conduction. Limit gate current with a resistor, do not depend on the current limit in your bench supply (because it probably has a big capacitor on the output that could damage the thyristor gate).

It's certainly possible they include some shunt gate resistance to limit sensitivity to dv/dt and other false triggering. 18 ohms sounds way too high for a gate-cathode short.

According to the datasheet it can take 150mA to trigger the SCR.

Assuming there is some G-K shunt resistance, your SCR readings on a multimeter are as-expected.

\$\endgroup\$
0
\$\begingroup\$

This does look like a gate-cathode short unfortunately

The gate-cathode should appear like a diode so it should block one way and show some impedance the other. If you are seeing very low impedance both directions either the silicon is damaged or there is another short somewhere.

The method I used to check SCR's and their associated gate drivers is use a DMM at the gate-cathode but then forward bias with a bench powersupply [1] with say ... 10V and current limit to say 1A. Then gate it with either another PSU [2] ( current limited, 100mA) or with the driver under test. What should should see is [1] will enter current-limit once [2] is turned on.

\$\endgroup\$
  • 1
    \$\begingroup\$ Most power SCR's read exactly that, 15 to 50 Ohms G-K. The only positive test with a multimeter would be a completely open GK, or shorted KA. The proper test is to test them live with a reasonable gate pulse. \$\endgroup\$ – R Drast Oct 30 '18 at 13:12
0
\$\begingroup\$

I tried biasing the gate with a 9v battery and 50 ohm resistor, and the device seems to be working fine.

Diode test voltage drop from anode to cathode is about 0.7v when the bias is applied, but I doubt the dmm has enough current to keep it latched.

I think Spehro Pefhany is correct and there is a shunt resistor in there.

Thanks everyone!

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.