james - shunt reg q

[dave dove]

was talking to nick the other day
and now i'm confused
so

with your bouncy shunt reg design

do i need to shunt 120% of idle current
or
120% of signal current
and if the latter
how do i calculate it

dave

[James D]

For SE the best sound is when you shunt 120% of the idle current... for PP its about 50% of idle current

ciao

James

[dave dove]

)^&(£)&_^_%6

[Andrew]

Hey Dave,

Is it for a phono? Then somewhere between signal current double the signal current should be more than enough. As James said elsewhere, there's usually quite a bit of standing current in a phono by keeping gm high and noise down, no point shunting this - as its a fixed load.

cheers,

-- Andrew

[James D]

But you can start from about 50% of the idle current and work upwards as and when able. Just make sure you have some good capacitance before it to buffer the shunt when it runs out of current on peaks...

As with all music most of the time the amp is cruising at 20% of peak or less and thats where the shunt works best - of course it helps deliver clean peaks and cleans up after the peak really well if you have the current capacity in the shunt but it is still worth doing if you can get to 50% of idle current - or there abouts...

Andrews point is accurate for some/most phono's but ones that run at 2 or 3 mA per stage should have full idle current shunt reg

James

[Andrew]

Hi Dave and James,

Of course my mind is on the the 1k5, you are, as usual, quite correct!

cheers,

-- Andrew

[NickG]

Now I am getting more confused, why is a SE output stage stage best at 120% when there is a big inductor to reduce the variation in current, whereas in a phono SE stage, where there is no inductor and the PSRR is higher because of that, and you have huge gain, especially at low frequencies, less is ok?

I can see the point of 120% if you are actually regulating the DC voltage, but I thought the shunt was only operating at AC in this case, and at DC was just a constant current load,

Confused of Halifax.

[dave dove]

andrew

no
already done the phono
just trying various choke:cap combos

chaps

for a few moments there was light
now i find myself all damp + extinguished
slapped by a wet herring or something.....
that's 560mA current plus headroom for the tx

dave

[James D]

The ac shunt does three things for an SE amp:

1) From the perspective of the output stage signal current loop it provides an active component for the return half of the loop and so replaces/minimises the sonic signature of the WE/ultrapath and psu caps. In theory this requires the shunt to swing the same signal current as the output stage so the shunt needs an idle current of maximum signal current plus enough "dc" current to keep the valve in its linear region when sourcing maximum signal current. this suggest the shunt needs the same idle current as the output valve for the same type of valve and maybe more if its different.

2) From the perspective of the psu the shunt plus output stage looks like a constant current load i.e. make it look like a resistive load ... so it removes all the signal based stresses from the psu and keeps its life easy - no current steps to get all bouncy over :-)

3) The shunt provides a broadband noise filtering action and so makes the amp quieter.

Listening tests have shown that the amplifier sound improves with increasing shunt idle current up to about 120% of output stage shunt current or there abouts. I put this down to two effects a) the signal current swing is not just due to the output stage but due to all the amplifier stages b) more idle current through the shunt than the output stage makes the shunt more linear and you can hear that...

For a phono stage with idle current of 10 to 1000 times the signal current swings then 1) and 2) are minimised so you can get away with much less current in the shunt reg. But I guess the acid test is to try it and see if more current in the shunt makes further improvement to the phono...

Hope that helps

James