Shunt method still useful??

[diydidi]

Hi all!
I use to use this method on valve guitar amps, but read some where that its not very accurate with modern amps anymore due to the way they construct OP transformers nowadays. Is this true? I just used an ordinary DMM.
Can someone shed some light here??

[R.G.]

The "shunt method" uses the resistance of the halves of the output transformer to cause a DC voltage drop, which is what you measure. The problem is that you know only the voltage generated, not the resistance that generates it, unless you separately measure the resistance and correct for any differences between the two halves of the OT primary.

At one time, OTs were designed and manufactured with the resistances of each half being closely matched. If not much care is given to matching the halves, then the resistances (among other things) will be un-matched, so any DC voltage drop you measure will also be un-matched for identical currents.

The shunt method is not terribly accurate. It's quick and (can be) dirty. 1 ohm sampling resistors are more accurately matched. Putting them in the cathodes means the screen currents are in the measurement too, but this is less inaccurate than relying on the transformer windings. You could put them in series with the plates as well. I prefer them at ground level (i.e. in the cathodes) for safety reasons.

[g1]

R.G. I believe he is referring to measuring plate current by "shunting" the OT primary half with DMM in current mode (rather than measuring the voltage across the OT primary half).
The problem arises when some modern OT's have low DC resistances on the primary, and some modern DMM's can have "shunt" values of up to 10 ohms depending on the range.
To be accurate, the shunt method requires that the meter have a very small resistance (in current mode) compared to the resistance of the OT primary.
Otherwise there is still current flowing through the OT and we get an inaccurate reading (we want our "shunt" to completely short out the OT for our measurement).

[R.G.]

Ah. Could be. The comment about output transformers throwing it off sent me down that resistance path.

If that's what you meant, OP, then what g-one said applies. There is inaccuracy from the (unknown) internal shunt resistance of the meter interacting with the unknown and probably unequal resistances of the windings. To be fair, if the resistance of the meter is much, much less than the resistance of the winding, the variability in the winding doesn't matter much. And a measurement with a different meter will give a different result if its internal resistance is different from the first one.

I personally still prefer the inserted sampling resistors on the cathode.

[hasserl]

The "shunt method" uses the resistance of the halves of the output transformer to cause a DC voltage drop, which is what you measure. The problem is that you know only the voltage generated, not the resistance that generates it, unless you separately measure the resistance and correct for any differences between the two halves of the OT primary.

At one time, OTs were designed and manufactured with the resistances of each half being closely matched. If not much care is given to matching the halves, then the resistances (among other things) will be un-matched, so any DC voltage drop you measure will also be un-matched for identical currents.

The shunt method is not terribly accurate. It's quick and (can be) dirty. 1 ohm sampling resistors are more accurately matched. Putting them in the cathodes means the screen currents are in the measurement too, but this is less inaccurate than relying on the transformer windings. You could put them in series with the plates as well. I prefer them at ground level (i.e. in the cathodes) for safety reasons.

I usually measure & record the resistance of both halves, then measure the voltage drop of each half and make calculations for each. IT is quick and it is easy, and if you use clip on probes you reduce the risk of slipping with the probe and shorting the plate to something else.