Pass Shunt Circuit

[Bengali]

Hi,

for Mr. Pass's shunt circuit. Is this how you calculate for output voltage?

7 x 9.1zener = 63.7V minus drop across fet = 60V

so varying factor is voltage drop across the fet?
how do I determine wha the fet voltage drop will be
from fet to fet?

thanks

[janneman]

Errr... This is not a shunt. It is a source follower. And yes, the output voltage varies a bit from FET to FET, but that should not be a problem.

jd

[Bengali]

Errr... This is not a shunt

yeah, what do I know

so what would be the cause for it not to regulate.
I'm did a sim for neg. rail. -10VDC going in and I'm
trying to achieve -3VDC on the output. using a 7.5V zener in
the zener string.

what I'm seeing is the output voltage will drop a lot depending
on the load current.

does this psu circuit require a circuit amount of load current
to regulate properly?

thx

[janneman]

Originally posted by Bengali
Errr... This is not a shunt

yeah, what do I know

so what would be the cause for it not to regulate.
I'm did a sim for neg. rail. -10VDC going in and I'm
trying to achieve -3VDC on the output. using a 7.5V zener in
the zener string.

what I'm seeing is the output voltage will drop a lot depending
on the load current.

does this psu circuit require a circuit amount of load current
to regulate properly?

thx

Well, it's not a regulator either, sorry I don't want to be pedantic, but it doesn't regulate anything. It tries to replicate the gate voltage with a few volts of loss. As you draw more current you lose more volts so the Vout falls. How much load current are you drawing?

Do you have a circuit drawing for that neg regulator?

jd

[Bengali]

thanks for your reply. I thought it did some kind of regulation
and noise reduction.

so how would this source follower circuit differ from using an LM337 to regulate it to exactly -3VDC? it's to drive the jboz.

where I'm at right now, I can't shrink file size for the screen capture. my circuit is same as the neg. rail. just replaced the zener string with one 7.5V zener and input volt of 10VDC.

thx

[janneman]

Originally posted by Bengali
thanks for your reply. I thought it did some kind of regulation
and noise reduction.

so how would this source follower circuit differ from using an LM337 to regulate it to exactly -3VDC? it's to drive the jboz.

where I'm at right now, I can't shrink file size for the screen capture. my circuit is same as the neg. rail. just replaced the zener string with one 7.5V zener and input volt of 10VDC.

thx

An LM337 does actively regulate because the output voltage is fed back into the regulator, and if the Vout starts to fall, it compensates very quickly so that Vout doesn't fall (almost...). The speed of reaction is not infinite so you will see that with increasing load current frequency, the Vout is dropping more and more.

This dropping is normally spec'd by an equivalent 'Zout'. So, if say Zout of the LM337 is 10 milliohms and you draw 1A the Vout drop is 10mV.

In the FET source follower you posted, there is no feedback to allow it to regulate. The FET itself has a spec called the transconductance, in volts per ampere. That tells you how many volts Vgate-source should rise for a certain load current. Alternatively, if you don't regulate Vgate-source, as in your case. it will tell you how much Vout will drop with a certain current.
Say your FET has a transconductance of 1 amp per volt, and with no-load Vout is 3V. Then if you draw 1 amp out of the thing, Vout will drop 1V to 2V.

I think for something like 3V out you are better served with something like the LM337 than a source follower. What's the expected load current?

jd

[woody]

For a 3v output this may be a poor choice of a circuite. The
gate turn on voltage may varry from fet to fet by a large value
but assuming a 4v gate turn on and 10v input you will need
a 3v zener.

[Bengali]

In the FET source follower you posted, there is no feedback to allow it to regulate. The FET itself has a spec called the transconductance, in volts per ampere. That tells you how many volts Vgate-source should rise for a certain load current. Alternatively, if you don't regulate Vgate-source, as in your case. it will tell you how much Vout will drop with a certain current.

that's make more sense now. so this type of circuit is
typically for higher voltage requirement?

my load is about 60-80mA. yes, lm337 would be much easier.

originally I was searching for a low noise -3VDC shunt circuit
since it to drive the 2sk170. something simple and clean. I did
not find any negative shunt circuits. Since Mr. Pass uses
this in his preamps, I would think it would work very well.


thx

[janneman]

Originally posted by Bengali


that's make more sense now. so this type of circuit is
typically for higher voltage requirement?

my load is about 60-80mA. yes, lm337 would be much easier.

originally I was searching for a low noise -3VDC shunt circuit
since it to drive the 2sk170. something simple and clean. I did
not find any negative shunt circuits. Since Mr. Pass uses
this in his preamps, I would think it would work very well.


thx

Well if you find a good pos shunt it should be easy to convert it to negative.

I think in the Pass case where you have say 60V output it doesn't matter too much if it drops a volt or so, but in your case it is not acceptable of course. Also maybe (but am not sure) there is a supply cap after the source follower to take the varying load current so the source follower only needs to deliver the average current into the cap which is normally pretty low in audio.

jd

[lineup]

Hi!
You could use the similar circuit for lower voltage. Like 3 Volt.
If you replace transistors with BJT, Bipolar.

This will give a voltage drop from B (=zener) to E (output) like -0.7 Volt only.
With very small changes depending on current drawn.

Depending on how much current you need you can select suitable devices.
For <500 mA then BD139 and BD140 would work well.
For higher currents we can use MJE15030/MJE15031

I would recommend some smaller cap (say 2 - 10uF) from base to ground on both transistors.
This would reduce ripple further och make clean output.