Let's discuss ZM's Good Gemini shunt

[Sorenm]

Thanks. I see the figure is labeled high current shunt, which may be overkill for a DAC chip requiring less than 50mA. Anyway, will try to enlighten myself.

 

[Zen Mod]

soon to get posted few preamp projects ....... you'll recognize that approach in Tanka preamp, both for +90V and +25V rails

since there was no need for KosherUberMumboJumbo reg for -25V, I did use simple LM337

 

[Sorenm]

Cool.

Made those resistor changes and it now tops out at 14.9v, which is good enough. It runs cooler with the lowered lq and is currently "making music". Thanks again.

😎

 

[Zen Mod]

change to 27K and go skyhigh

 

[Zen Mod]

currently "making music".

No Porn, No Glory

!

 

[Sorenm]

It's a quiet Sunday, so probably need neither of those 😉 But if it gets me in line for a 5v version or helps someone else...here it is.

and with the rest of the ever evolving DAC

 

[Vunce]

Plank DAC!
Love it, lookin' good Sorenm.

 

[Sorenm]

Update: Loving the sound of the Good Gemini! It replaced a Miro PSU2 (LT1963/LT3015) and in this position I slightly favor the GG shunt.

 

[Sorenm]

Quick question. Thinking of using GG shunts for a DCG3 pre-amp. It draws about 170mA/rail @17V. Would programming the Iq to 190mA or 200mA provide sufficient overhead for the shunt to function right? Maybe setting R3/4 and R5/6 at 5.6R and 7.5R.

 

[Zen Mod]

taking in account that load is pure class A...... say that 210mA of programmed current is absolute minimum

reg itself is needing just few mA for function ( error amp, references) but from practical reasons, say that 20% of overhead is OK in this case

take care of dissipation .... as final criteria - use thumb or temp. meter

 

[Sorenm]

Thanks ZM. Will shoot for ~230mA with two 5.6R. Do you think the bottom plate of a 330x350mm galaxy chassis would be sufficient for keeping two GGs cool?

 

[Zen Mod]

proper analysis is half of work

what's intended ( necessary) voltage of secondaries, and then working down ... voltage sag across CCS mosfet etc

for shunt mosfet - it's easy - rail voltage *(programmed current - load current)

so, 17V*(230mA-170mA)= 17*50mA=0W85 ..... relatively small TO220 hsnk is OK



say that your goal is 17Vdc at rails

let's see 18Vac secs

ZM's Thumb of Rule - with heavy loaded, mult. factor is in best scenario *1.25

so , you can expect 18Vac*1.25=22V5 after diode bridge

means 22V5 prior CCS mosfet, 17Vdc after

volt. difference is 22V5-17V=5V5 ..... almost covering or just covering even scenario of -10% mains voltage sag ( just proper engineering praxis to allow that)

5V5 and programmed current of 230mA..... measly 1W265 per series mosfet

even medium sized TO220 htsnk will cover that

meaning .....

bottom plate of a 330x350mm galaxy chassis

is good for, say , 20 complete Good Geminy units, in this context

 

[Sorenm]

Thanks ZM. All makes good sense and I won't need to ask this question again 🙏🙂

 

[schultzsch]

Which is the lowest vdrop that you would use on this ccs ZM? Without including the mains variation.

 

[Zen Mod]

no less than 4V to cover all losses and have some headroom

 

[schultzsch]

Thanks

 

[Sorenm]

Quick question. So, the 18vac trafo is actually 22.5vac with the load of the shunt. Per ZM's logic above, this means something like 11.5v voltage drop and 2.7w heat dissipation for Q1/2. The heat can be managed but is this too much of a voltage drop? Most days, the pre amp runs 12 hours...so would want to make sure it has longevity, even if it means getting two lower voltage trafos.

Cheers, Soren

 

[Zen Mod]

choice is yours

if you manage heatsinking to have max 60C at Q1/2, all good

if that's too much of trouble, change xformer

 

[Sorenm]

Thanks ZM. The heat is manageable. It is in a temporary layout (=mess) for now, while I figure out how to organize the pre amp. Are 10 cm power wires between the shunts and DCG3 OK, or too long?

Cheers -

 

[Zen Mod]

speaking of principle, wires between any reg and (load) stage are always too long