[PSP]

Hi Hugh,
Please forgive me for asking a question you have answered before but after several attempts searching the archives (and my foot-high stack of  AKSA print-outs and e-mails from you...) I'll ask again:

One of my 55w AKSAs has a DC offset around 30mv.  I want to get that as close to zero as possible, certainly under 10mv.  From memory (I'm at the office now, 30 miles from my AKSA) I think I want to adjust R3, but by how much and in which direction?  Would there be a sonic penalty to putting in a good (Bourne cermet??) trimmer pot here?

Many thanks, and hope all is well at your end.
Peter

[AKSA]

Hi Peter,

No problem at all Sir!  This is the sharp end of the AKSA business, and I enjoy it....

Offset adjustment on the AKSA works by trimming the value of R3.  There is little point in using a trimmer to do this, as the offset has moderate tolerance anyway and every time the mains changes voltage, which is frequent, the offset changes by a few millivolts and thus you finish up chasing rainbows.....

However, you can adjust it so that the median variation is 0mV.

The rule is this:

1.  Increase R3 to drive the offset HIGHER (that is, positive wrt zero)
2.  Decrease R3 to pull the offset LOWER (that is, negative wrt zero).

Generally, you shouldn't have to go more than about 4K above or below the nominal value of R3 to achieve this.  As a rule, 1K of alteration corresponds with 5-10mV, but it depends on other factors too, such as the Vbe of the input devices, and their beta match.

Hope this helps!

Cheers,

Hugh

[PSP]

Thanks, Hugh.  There's enough keywords in this post that I should be able to find this ten years from now after way too much wine!

I am dying of curiosity and excitement about your Nirvana+ release for the AKSA 55.... be sure to give me a shout when all is ready.

Peter

[andyr]

... as the offset has moderate tolerance anyway and every time the mains changes voltage, which is frequent, the offset changes by a few millivolts and thus you finish up chasing rainbows.....

However, you can adjust it so that the median variation is 0mV ....

Hugh,

Thanks for this information.  Could you possibly finish it off by specifying what happens to offset when the mains voltage drops?

IE. if you had offset set to 0v when the mains (in Oz) happened to be at the nominal 240v, what would happen to it when the mains was down at 230v?

Thanks,

Andy

[AKSA]

Andy,

Sure!  Offset will rise wrt ground as mains voltage falls.

A drop of 10Vac would give around 10mV rise in offset.

Until you reach about 100mV, there is almost no effect on the driver.  At this point, 100mV, the standing current in the voice coil is at 15mA - around 25% of quiescent - and while dissipation is trivial at 1.5milliwatts (where most voice coils will dissipate 50W continuously without problem) it will slightly push the voice coil off its rest position, leading to slightly increased second harmonic distortion.

Generally you have to get up to several volts DC across the voice coil before heat dissipation becomes a real issue.  For example, at 2V, the voice coil is running 300mA, a dissipation of 0.6W, but this considerably displaces the cone, causing early clipping at one end of the waveform and thus reducing undistorted power output.  The problem with high offsets is usually that it indicates something seriously amiss, and the chances are that the output offset will soon go ballistic and hit the rails.  THEN you really have a problem;  with 36V across a driver, you have a voice coil current of 5.29 amps and a dissipation of 190 Watts.  Very few drivers will withstand this sort of punishment for more than half a second, even large ones.   Of course, when this happens, the output devices have become elegantly fused wire......  

Cheers,

Hugh

[andyr]

Thanks, Hugh.

I'm probably a bit more worried about output offset than most ... coz my Maggie ribbons are about as thin as a butterfly's wings!!

Regards,

Andy