Zener triggered SCR

[wn1fju]

I was fixing a Tek 2230 scope which was stuck in a startup-shutdown-startup loop.  There is an overvoltage circuit (nominally 42.8 VDC)
consisting of a 51 V Zener firing the gate of an MCR72-4 SCR.  When I isolated the circuit and supplied my own voltage, the SCR tripped
pretty consistently at about 37 V.  So I pulled the SCR and it at least tested OK on a component tester.  Figuring maybe the Zener was
leaking, I replaced the Zener with a new one and put back the original SCR.  The scope then worked fine.

But then I hooked up the "defective" Zener to my DC calibrator (set to 2 ma maximum) through an ammeter and cranked up the voltage. 
I got zero current until the calibrator reached 51 V.  So it my mind, it looks like the original Zener was good also.  So then, why didn't the
scope work with those two original parts?

Any ideas where I went wrong?

[T3sl4co1l]

There should be a resistor G-K to absorb leakage current and induced dV/dt (via diode capacitance).

Maybe the SCR is actually what needs to be replaced, by having too sensitive a gate?

When you say "zero" current, you don't actually mean zero current; what was the least digit in that measurement?

Tim

[wn1fju]

Yes, there is a 120 ohm resistor from the gate to ground (K is also connected to ground). 

At least in my crude test setup, I have the DC voltage simply connected through the current meter directly to the 51V Zener on the bench.
At voltage of 50V, I get 0.005e-6, 0.056e-7 and 0.03e-6 on a Fluke 8502A, HP 4140B and Keithley 414A, and I get 0.52e-3, 0.534e-3 and
0.52e-3 at 51V.  At 52V, I simply get the current limit of my DC source.  Admittedly, without proper shielded cables, my microamp current
measurements are somewhat suspect, but in all cases, I wouldn't think the current is enough to fire the SCR at the nominal 42.8 volts in
the scope.

[T3sl4co1l]

That sounds good to me!

How about capacitance, does the sensed rail have unexpectedly high ripple?  (How does the diode capacitance compare between the original and replacement?)

Other paranoia, could be solder or flux residues, maybe it was heat-treated by soldering, who knows?

Tim

[Armadillo]

Yes, there is a 120 ohm resistor from the gate to ground (K is also connected to ground). 

At least in my crude test setup, I have the DC voltage simply connected through the current meter directly to the 51V Zener on the bench.
At voltage of 50V, I get 0.005e-6, 0.056e-7 and 0.03e-6 on a Fluke 8502A, HP 4140B and Keithley 414A, and I get 0.52e-3, 0.534e-3 and
0.52e-3 at 51V.  At 52V, I simply get the current limit of my DC source.  Admittedly, without proper shielded cables, my microamp current
measurements are somewhat suspect, but in all cases, I wouldn't think the current is enough to fire the SCR at the nominal 42.8 volts in
the scope.

Can you share your HP4140B measurement setup? thanks.

[wn1fju]

"Can you share your HP4140B measurement setup? thanks."

Not much of a setup - just a quickie measurement.  The HP 4140B was set to current measurement, auto range, long integration time.  The triaxial
input connector goes through a triax-to-BNC cable and then to a set of clips.  The DC source voltage positive lead goes to the positive clip, the
negative clip goes to the Zener cathode, and the Zener anode goes back to the DC source voltage negative lead.  Then just read the number on
the 4140B.  Crude, and certainly not very effective at the picoamp levels that the 4140B is capable of measuring, but good enough to see if there
is sufficient Zener current to fire the SCR.

My other thought on the subject is perhaps there was really nothing wrong with the Zener and it was the SCR that was triggering too low.  Maybe
I damaged the SCR (or cured it) when I removed and reinstalled it.  For all I know, it opened up and that's why everything works, regardless of
a new or old Zener.  I'll eventually order a new SCR just to be sure.

[David Hess]

The SCR is a C106 sensitive gate type with a 120 ohm gate to cathode resistor and a 3 ohm series resistor to limit the maximum current.  Besides the 51 volt zener diode, it can also be triggered by a 0.51 ohm ground referred current sense resistor through a diode to raise the threshold to 1.2 volts or whatever.  Later models replaced the SCR with a larger and higher current TO-220 packaged part.  There is a big 1000 microfarad filtering capacitor on the 42.8 volt supply and it is a primary suspect for wearing out since it handles a high ripple current.

As far as the actual failure, could the 120 ohm resistor have been open?

If I found the SCR tripping without reason, I would be inclined to replace the SCR and zener together.  Zener diodes seem to have an annoying tendency to become leaky with age.

[wn1fju]

I should have checked the 120 ohm resistor, although a failure of that part would have been my last choice.  The 1000 uF capacitor had been replaced
by the previous owner at some time, and since the SCR was firing with my externally supplied pure 42.8 volts, it probably isn't the culprit either.

I think, in retrospect, that it was an over-current fault, as supplied by the 0.51 ohm sense resistor on the inverter transformer, rather than an over-
voltage fault.  With my external 42.8 supply, every time I cranked it up, it would fire the SCR at just a little bit higher voltage, e.g., 35, 36, 37 volts...
I have seen many repair instances where the electrolytics needed a little time to reform after decades of non-use.  I think that may have been the case
here.  The electrolytics on the the various secondaries of the supply just needed a little time to come up to speed.  Once everything settled in and the
current dropped to normal, the SCR and Zener operated properly since there was nothing wrong with both of them to begin with.

At least that's my guess (until such time as the scope breaks!).

[David Hess]

The 2230 is a great oscilloscope.  The 2232 is 2 better of course but does not have as much Tek.