Multimeters

[Mick Bailey]

I use this as a test for marginal instability in Blues Jr amps that needs fixing. They can also oscillate to the point of red-plating if a meter probe is touched on the EL84 screen (pin 9).

[Dave H]

The trouble is, I was hoping to verify that both output valves actually had (nearly) the same bias current. I failed at that because of the instability problem.

You could have tried using the real transformer shunt method (meter on current range) to measure the bias current. It most likely won't oscillate with one half of the OT shorted.

When probing the plates the oscillation can usually be prevented by using a 'stopper' resistor clipped to the end of the meter lead attached to the plate. I can't remember the value I used. It was probably between 10k and 100k.

[Gnobuddy]

I use this as a test for marginal instability in Blues Jr amps that needs fixing. They can also oscillate to the point of red-plating if a meter probe is touched on the EL84 screen (pin 9).

I've only ever worked on one. That one absolutely fits your description of "marginal instability". Simply moving the ribbon cables running to the EL84s a couple of millimetres was enough to trigger the amp into oscillation.

I was given a Super Champ XD that had gone dead, because the digital signal processing board had fried itself. I brought it back to life by replacing solid-state preamp with a 12AX7, but in the process, found out that the power amp section was even closer to the hairy edge of instability than the Blues Junior. The PCB design was one of the worst I have ever seen, with no separation between high and low current grounds, and bizarre layout (such as the pcb track that ran two inches north on top of the board, went through a plated-through-hole via to the opposite side, and then ran south two and a half inches, for absolutely no reason whatsoever. It could just have run half an inch south in the first place.)

I still have the '65 Princeton Reverb reissue. It will go unstable if I place a multiFX pedal on the floor within a couple of feet of the amp, and set it to a fairly high gain blues or rock (not metal, i.e. not bizarrely high gain) patch. It is a nice amp, though, if used the way it would have been used in 1965.

Three amps is not exactly a statistically large sample, but three unstable amps out of three total, leaves me wondering if Fender Corp pays any attention at all to the stability of their amps. When you see pennies saved by not using grid stoppers on output valves, et cetera, you start to get the picture.

-Gnobuddy

[Gnobuddy]

It most likely won't oscillate with one half of the OT shorted.

But if it does oscillate, you have no anode load to limit peak currents. That sounds pretty risky for the health of the output valves, no?

When probing the plates the oscillation can usually be prevented by using a 'stopper' resistor clipped to the end of the meter lead attached to the plate. I can't remember the value I used. It was probably between 10k and 100k.

Thanks for the tip! 100k will cause about a 1% error with the typical 10 meg input resistance of many DMMs. That's too small to matter in an amp full of 5% resistors, so it sounds like a good solution to the oscillation issue!

-Gnobuddy

[Enzo]

Sometimes "pointless" traces are used as shielding and hum abatement. Many have discovered little anti-hum antennas in their Mesa amps. You find a tiny maybe 3/4" piece of bare wire soldered to teh board and sticking straight up. Snip it off, and hum goes up. It is there to pick up a little bit of hum out of phase with the hum in the circuit at that point, thus cancelling. To the unwashed tech it looks like an untrimmed lead, or a temporary test point never removed, but it is there on purpose.

I often see ground traces snaking across a board between two signal traces, and ultimately it just stops at some point. it wasn't going anywhere, but it was acting as a shield between two signal traces to reduce crosstalk.

Your funny trace could easily be something along those lines.

[Enzo]

Quote Originally Posted by Dave H View Post
It most likely won't oscillate with one half of the OT shorted.
But if it does oscillate, you have no anode load to limit peak currents. That sounds pretty risky for the health of the output valves, no?

One doesn't do this while signal is present, there shouldn't be much in the way of peak currents. In the shunt method your ammeter is put parallel the half primary winding. The meter has a very low resistance, so essentially is shorting across the winding.

[Gnobuddy]

Your funny trace could easily be something along those lines.

I think you are a kind soul, who gives credit where it maybe wasn't deserved.

There were a lot of uglies on that board - as though someone ran autorouter software while designing the PCB, and didn't have the faintest clue how grounding in audio circuits should be done.

As for "unwashed tech", I may be unwashed, but I don't qualify as a tech.

-Gnobuddy

[J M Fahey]

But if it does oscillate, you have no anode load to limit peak currents. That sounds pretty risky for the health of the output valves, no?

Tubes are self current limiting by design, so much so that shorting output jacks are are a common and accepted feature, not a design error but quite the contrary-

[Gnobuddy]

Tubes are self current limiting by design, so much so that shorting output jacks are are a common and accepted feature, not a design error but quite the contrary-

Nevertheless, if you replace the anode load of an output valve in a guitar amp with a short-circuit, and drive it hard, you will exceed anode dissipation limits handily. With the full B+ voltage across the valve the entire time, power dissipation is much higher than with the proper load. Extend the Vgk=0 anode current at the "knee" horizontally across the graph until it intersects the vertical line corresponding to B+, and you will find the intersection point far above the maximum anode dissipation parabola...

I do understand that the shorting jack is better than the opposite (open circuit), because it buys you a little more time before your output valves red-plate and your amp fries.

IMO there is a better way than a shorting jack. I usually wire an 8-ohm power resistor to my guitar amp (switching) output jack, wired up in such a way that, if nothing is plugged in, the amp sees an 8 ohm load.

I wire a pair of high-brightness anti-parallel LEDs with a protective series resistor across the 8 ohm resistor, so I also get a visual indication that nothing is plugged in.

I have contemplated wiring a bridge rectifier and a loud piezo alarm across the resistor, to provide an audible warning, but never actually did that.

I should mention that I have no guitar amps with more than 15 watts output power, so it is easy for me to find an 8 ohm power resistor that can cope with the amps output (and therefore protect it) for a reasonable period of time. It would not be so easy with a 200-watt monster.

-Gnobuddy