Help - techie question - DC on amp

[farflungstar]

Ok so I've been listening to my Hiraga (bought from Jake) amp and enjoying it but today I decided to measure for DC at the outputs - I wish I hadn't.

I set the multimeter as in the pic, disconnected speakers and switched the amp on. Measuring both channels fluctuates around 190mv and higher - can this be right? Am I doing something wrong? Am I frying my Tannoys?

https://uploads.tapatalk-cdn.com/20171025/e5fa2d402205d1e29de2d9544558f74b.jpg

[struth]

Should be ok at that. Much depends on the speaker etc. will probably equate to less than a tenth of a watt. think 100mv is classed as low.

[struth]

Think the calc is volts squared, divided by impedance = power (ie ohms law)

[farflungstar]

I read that any DC on the outputs was bad though 0-15mv was acceptable. Hence my worry.

[struth]

The maximum dc offset tolerated by a speaker is a percentage of the linear excursion of the driver that is acceptable, usually 10 to 15%, or so i,m told. The formula used is:

Vdc(max) = % excursion x excursion x driver impedance/[Force Factor x Suspension Compliance]

All these parameters are found in the manufacturer's spec sheets.

You could easily get fitted, dc protection circuit for the speakers

[JimC]

I wouldn't worry too much about that.
What does it measure with Speakers connected? That'd be a more accurate test for me.

Jim.

[Firebottle]

Was just about to suggest repeating the measurement with speakers connected, good one James. Will be a more useful reading.

:)

[farflungstar]

Am at work now but will try wth speakers connected in morning and retest. I don't like the thought of why DC going into my speakers.

[Radford Revival]

The amount of power this will deliver into the voice coil, assuming a 5 ohm DC resistance (and it is probably more), can be calculated by 0.22 / 5, ie V2 / R, which is a whopping 8mW (0.008W)

AKA not much to worry about assuming there is no other problem!

[Jazid]

Surely it is not difficult to adjust the DC offset anyway, aren't there a couple of screwpots under the lid that achieve this function? Should be able to get it down to low millivolt levels though beware as it may drift as the amp warms up. This was a DIY amp wasn't it?

[farflungstar]

Yes it's the Hiraga 20w class A. There are no adjustable components on the circuit boards.
https://uploads.tapatalk-cdn.com/20171025/ea453a723eebb1e61f542b7300897550.jpg

[Pharos]

Connecting the speakers will make very little difference to the measured voltage because power amps are designed to be as near a zero O/P impedance as possible.

I empathise with your concern because 15 years ago I bought an ATC SIA150, which had about 180mV DC on its O/Ps also, but the calculation gives a very low DC dissipation wattage, of (0.18 X 0.18)/8, about 4mW.
I doubt that ATC would mess up their reputation with a faulty design. (Though usually, say with Quad, it is a couple of mV.)

[farflungstar]

Right now I have the speakers disconnected and am reluctant to hook them back up grrrrr

[Pharos]

I think you will be OK for the reasons above, and 4mW is tiny compared with the programme material, the equivalent of running a 90dB/W sensitivity speaker producing about 66dB spl of programme.

[Radford Revival]

If worried, connect a couple of 10,000uF electrolytic caps back to back (i.e their polarities in opposite directions) in series with the speaker. This will remove almost all DC apart from a tiny leakage current inherent to the caps (in the context of a loudspeaker, for all intents and purposes no current).

If you get capacitors rated to the rail voltage of the amp then you will also get some level of protection against the amp going DC (this isn't 100% bulletproof protection in the long term but can't hurt)

But in reality this level of DC (in the non fault condition) will dissipate a tiny amount of power in the voice coil, actually listening to music will produce more.

[walpurgis]

If worried, connect a couple of 10,000uF electrolytic caps back to back (i.e their polarities in opposite directions) in series with the speaker. This will remove almost all DC apart from a tiny leakage current inherent to the caps (in the context of a loudspeaker, for all intents and purposes no current).

How would an amp respond to this. Are there any negative issues?

[Radford Revival]

How would an amp respond to this. Are there any negative issues?

As far as I'm aware with a conventional-ish solid state design this shouldn't cause any issues. Lots of single rail amps use a large electrolytic on their output as part of the design, and in many cases they are even inside the the amp's feedback loop. The capacitor will have a small ESR (10s of milliOhms) but not enough to worry about.

Weirder topologies like circlotron amplifiers probably woudn't work too well but they're pretty uncommon. They depend on a DC path through the loudspeaker to bias their output devices, and the opposing bias current from each half cancels out.

[farflungstar]

Update - I realised that there is an adjustable pot on each circuit board so tried adjusting them - short lived success - got them both down to 45mv even after 15 minutes warming. Switched off and switched on again - now 170, 110. Adjusted again, meter reading dipping into - 25 so too much and adjusted again to around 60mv. Same thing after switching off and back on again around 180, 210 - how can they drift (both) so much and so quickly. They were measuring 300mv before I started messing!

[Firebottle]

It sounds like it isn't the best thermal design.
The transistor characteristics will change from cold to operating temperature.

It is perfectly OK to have a small negative offset, it doesn't have to be positive, just small in actual value.
Probably better to adjust to say -25mV at operating temp, you will probably find that it will be less than 100mV (positive) when cold.

[farflungstar]

Thanks. I'm really unhappy and gutted that I've touched it. I think the pots are a bit flaky but then they're over 30 years old! If the readings would stay consistent (both when warm and cold) between switch on's I would be happier. I'm not a techie and should have known it wouldn't be so simple as using a multimeter and adjusting a bloody plot.

[Pharos]

If the pots are unreliable/intermittent, you should replace them to protect the circuit, and they are quite cheap.
Then leave it on running at typical sound levels and let it stabilise, and adjust for null. (zero) as near as possible.

[farflungstar]

I have a techie friend who will take a look and replace the pots when he has time. This is a strange problem. Not to mention no ground at mains.