diyAudio (
-   Solid State (
-   -   Simple shunt reg (

Onvinyl 19th October 2007 12:59 AM

I remade the perf-board, placed 15pf between base and collector of Q1 (the driver of the shunt-Q (wich is a mosfet now) *and* put a cascode on top of the error amp. It's now stable throughout. Now I have a problem with my sope, it seems to pick up something at the highest input gain from anywhere...

The sim indicates a output impedance of 2.8 mOhms at 1Mhz (12µthrough the audio band. Isn't that far to optimistic?

Another point is, if you really can call it 'open loop'. My understanding was, since there is no dedicated feedback path (as in a series pass reg), but the feedback goes over the V+ line, I thought about it as a kind of local feedback. What do you think it is?


AndrewT 19th October 2007 07:49 AM

an open loop (= no feedback) design it is not.
Imagine a Zener fed Emitter Follower. It's open loop.

Your circuit monitors the output voltage across the resistor string and compares that to the fixed voltage across the LED string. It then adjusts the LTP output to control the severity of the shunt. Very much a closed loop feedback circuit.

Please post the latest schematic.

Onvinyl 19th October 2007 09:27 AM

2 Attachment(s)
Hi Andrew,
ok, it's not simple and it's not open loop, but it's a shunt... :)

It surely has feedback, but I think it is *in a way* local. Since the input and the output voltage are the same in a shunt-reg (in respect to the error amp), you don't have the output devices working in both the load and the feedback path, only in the load.

Here is the updated schematic. The cfp-input pair is meant to be an replacement for those now hard to get Toshiba FETs in respect to linearity. It additionally lowers output impedance as indicated above.

Please note, that for the moment I don't have a 2sc1775 so I used a BC550 as driver. That rises output impedance by a factor of 5 according to simulation.

Later, I will cascode the simple reg shown in the first post as well and compare them soundwise, perhaps against a super regulator as well.

Onvinyl 19th October 2007 09:53 AM

...oops. C6 shall go to ground, of course...

AndrewT 19th October 2007 09:53 AM

can you measure the Vdrop of the CCS?
at a couple of output currents, 20mA and 100mA might be suitable.

What is the lowest input voltage and still get 24Vdc on the output at each of these currents?

AndrewT 19th October 2007 09:56 AM


Originally posted by Onvinyl
...C6 shall go to ground...
but C3 and R11 don't??

Onvinyl 19th October 2007 10:22 AM

2 Attachment(s)
Wrong! I updated the schematic, see below.

The ccs itself drops very little, a few 100mV up to 150mA. The lower the Vreg in respect to Vunreg, the more.

I can't check very well in reality, because my standard r-c-r-c psu filter looses to much voltage. No real bench supply here at the moment, sadly.


Onvinyl 19th October 2007 10:29 AM

It is convenient to make R8 adjustable since there seems to be an optimum current share ratio between Q4/Q6 and Q5/Q8 in respect to regulation. It *seems* to be best to: Q4:Q6 = 1:2 -- 1:3 but I don't know why. (It regulates best in reality and has lowest output impedance in sim)

Onvinyl 31st October 2007 03:56 PM

As shown, the circuit does not work properly.

aparatusonitus 31st October 2007 04:14 PM

You have cascode Qs 3 and 7
drown up-side down in your schematics.

All times are GMT. The time now is 11:06 PM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 17.65%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Copyright ©1999-2019 diyAudio