Measuring DC offset on a power amp

[monstrous lie]

I recently bought a Quad 306 power amp in v nice condition, and it appears to work fine. However, reading the thread on Rogers JR149 speakers has persuaded me that I should at least check it for potentially harmful DC offset. How do I do this? Is it OK to have this amp switched on and running with no loudspeakers connected? Any help would be appreciated. Cheers.
ML

 

[Barrymagrec]

The amplifier has to have a load connected or you will probably get a false reading. The load can be a resistor or any suitable speaker but if you genuinely think you have a problem it would be best to use a resistor or at least a very cheap disposable speaker.

 

[MJS]

No, the DC offset needs to be checked without a load. Amp on, no input, no speakers. Measure voltage across loudspeaker terminals. That's it.

 

[Julf]

No, the DC offset needs to be checked without a load. Amp on, no input, no speakers. Measure voltage across loudspeaker terminals. That's it.

Doesn't that depend on *why* you measure the offset? If you want to know how much DC there is going into your speaker, you should measure with the load connected.

 

[Tony L]

Is it OK to have this amp switched on and running with no loudspeakers connected?

Solid state amp such as the Quad, yes no issue at all, but definitely do not do this with a valve amp!

PS Capacitor-coupled amps such as the Quad 303 play by different rules, so don’t show a stable DC offset at the outputs. The 306 has no internal adjustment for anything as stability is built into the design somehow. My old recently recapped one, now with Lordsummit, showed no offset beyond a meter rounding error of about 0.04mV. Someone better versed in Quad’s architecture may be able to explain why it is so well behaved!

 

[Arkless Electronics]

Doesn't that depend on *why* you measure the offset? If you want to know how much DC there is going into your speaker, you should measure with the load connected.

And time how long it takes for smoke to appear? Do it with no load connected! There should be no offset beyond say 30mV tops so if you see a few volts or more you have a big problem regardless of whether there's a speaker connected.

 

[Arkless Electronics]

Solid state amp such as the Quad, yes no issue at all, but definitely do not do this with a valve amp!

PS Capacitor-coupled amps such as the Quad 303 play by different rules, so don’t show a stable DC offset at the outputs. The 306 has no internal adjustment for anything as stability is built into the design somehow. My old recently recapped one, now with Lordsummit, showed no offset beyond a meter rounding error of about 0.04mV. Someone better versed in Quad’s architecture may be able to explain why it is so well behaved!

A normal valve power amp can't have DC offset anyway...There's a transformer that stops that. On a 303 and other cap coupled amps the cap stops the DC so it should never be an issue.
The 306 uses a DC servo to keep the offset very low, as do lots of other amps. It can be very low without a separate servo anyway due to the similar "servo" of the amps negative feedback loop. A separate DC servo is most often used to get rid of an electrolytic in the feedback loop. Removing it also removes a good deal of the offset reduction ability and so it would have say 0.5V offset without the separate servo.

Whilst it's not a good idea to run a valve amp with no load it will usually be ok in amps with feedback. The worst would be pentodes or tetrodes with no negative feedback but that would be hopelessly low fi and is never used in hi fi amps.

 

[Julf]

And time how long it takes for smoke to appear? Do it with no load connected! There should be no offset beyond say 30mV tops so if you see a few volts or more you have a big problem regardless of whether there's a speaker connected.

Depends on your amp topology. If the DC offset is high impedance, and drops down to millivolts with a load, the unloaded voltage doesn't matter.

 

[Arkless Electronics]

Depends on your amp topology. If the DC offset is high impedance, and drops down to millivolts with a load, the unloaded voltage doesn't matter.

The only possible case where this could happen would be a capacitor coupled amp with a leaky coupling cap... The 306 is direct coupled.

 

[monstrous lie]

Thanks to all for the guidance. I think I've got enough to be going on with here.
Cheers. ML