advertisement


The GJ general purpose Gyrator / Regulator

Hi, nice little thing this, when will the negative version of the circuit be presented for the public? I'm waiting...
Regards, Staffan
 
13mh13 said:
How much is "a little under the zener's voltage"?
E.g. If one wants 5V out -- say, for powering ICs -- what zener voltage should one select? In Geoff's ckt, what is the sum of losses (voltage "dropouts") to consider?

With respect to shunting, how about using a "precision" zener, like LM329 (6.9V) or LM336-5.0 (5V)?

Nice job, Geoff!
Click to expand...

Voltage losses amount to 0.5V to 0.6V for each transistor used. My reg as described in the Buffalo thread uses an 8.2V zener and two transistors - a darlington pair. Output voltage is 7.2V
 
Do higher-current apps -- say 200 to 300 mA -- require a more robust trans for Q1 ... say BD139? Also, for higher-current use, do any of the other components needed to be "beefed up", such as R's?
 
Yes it is possible to hook it up backwards and join the "positiv" part to gnd at the output if:
you have one rectifier bridge for each winding of the transformer.
I don't know if there is any negative side effects such a setup
Staffan
 
As per recommendation from someone on diyhifi.org, C1 should be connected as follows:

"The +side of C1 should be connected to the base of Q1 and -side of C1 connected to ground."

Also, use several LEDs (in series) -- instead of zeners -- to achieve the desired voltage drop. LEDs are way less noisy.
 
13mh13 said:
"The +side of C1 should be connected to the base of Q1 and -side of C1 connected to ground."
Click to expand...

That's not a good idea. A darlington has very high gain, and Q1 needs a resistor at its base to prevent oscillation. However the resistor used is rather a high value - 1K, whereas around 100 ohms is more usual.

Another way (apart from using LEDs) to reduce noise is to use an RC filter. Connect a 1K resistor between the top of the zener diode and the top of C1. This will start filtering noise above 1.5Hz.

Other good examples of similar regulators:

http://www.tnt-audio.com/clinica/regulators_noise3_e.html

http://www.pedjarogic.com/gc/supplies.htm

Dan
 
Sparts155 said:
That's not a good idea. A darlington has very high gain, and Q1 needs a resistor at its base to prevent oscillation. However the resistor used is rather a high value - 1K, whereas around 100 ohms is more usual.
Click to expand...
I'm not sure I understand...the resistor stays. Perhaps I can clarify via schematic (this is a modified version of Geoff's orig)...
http://www.diyhifi.org/forums/download/file.php?id=3350&mode=view

...and a somewhat similar design, with the C1 cap in the position I noted ... and a few other similar components such as the 1K base resistor:
http://www.diyhifi.org/forums/viewtopic.php?p=34386#p34386
 
led-5v-reg.gif


In this version, you have filtering for the diodes - that's good. But the cap is now directly at the base of Q1 - that's bad for stability. You need a resistor between the cap and the base of Q1. 100 ohms is fine.

Going without this resistor the reg will still work, but you will/may get HF oscillation, which rather defeats the object of building a low noise reg.

Dan
 
Sparts155 said:
In this version, you have filtering for the diodes - that's good. But the cap is now directly at the base of Q1 - that's bad for stability. You need a resistor between the cap and the base of Q1. 100 ohms is fine. ...
Going without this resistor the reg will still work, but you will/may get HF oscillation, which rather defeats the object of building a low noise reg.
Click to expand...
I agree. There was some confusion at "the other place" about the design but I think the following may work (note the new position of R6 and a new 47R base resistor):
led-reg-v3.GIF
 
There's a world of difference between 'the best' and 'far more than good enough'.

Most of thr time the rest of the circuit will introduce far more noise/variability than the LED, and in the case of this circuit - strictly speaking the output isn't even regulated (deliberatley, there's no feedback). As a result the way this circuit's output impedance interacts with the load is several orders of magnitude more relevant than the choice of reference.

Build something and play with it, to find out what matters :)
 
martin clark said:
There's a world of difference between 'the best' and 'far more than good enough'. ...
Most of thr time the rest of the circuit will introduce far more noise/variability than the LED, and in the case of this circuit -- strictly speaking the output isn't even regulated (deliberatley, there's no feedback). As a result the way this circuit's output impedance interacts with the load is several orders of magnitude more relevant than the choice of reference.
Click to expand...
Yeah ... the more I dig into the issue the more I realize this is the case. Anyway...
Build something and play with it, to find out what matters :)
Click to expand...
I built my version of Geoff's ckt. But I get the full Vin at Vout. I.e, Vin = Vout. My Vin = 11.6VDC

I have three LEDs (the same green model as in Flea) dropping about 5.9VDC total. I'm also using E44H11 and J201 (NOT Geoff's suggested J204) -- but I don't think these diffs are important.

If I disconnect the emitter-to-base connection in the Darlington, I obtain Vout = 8.3VDC.

Dunno ... maybe Geoff's orig. was correct after all?? I'm away from home currently so I won't be able to troubleshoot for a day or two.

Any suggestions?
 
Dear Geoff,

please advise would this circuit able to deliver around 10 Amp continuous ? as far as regulation the most my concern to drive and class A amp, (beside noise at higher freq).
or what transistor/configuration should be taken for delivering such huge current with the same quality result.

And is there any negative rail circuit for this ?

thanks a lot..

Sejati
 
You must log in or register to reply here.


advertisement


Back
Top