Is it better to use 68000µF caps before or after regulators of a DAC psu ?
I was thinking of something like 4700µ > reg > 1µ > 10 ohm > 68000µ + 2200µ + 10µ film.
I was thinking of something like 4700µ > reg > 1µ > 10 ohm > 68000µ + 2200µ + 10µ film.
From an scientific point of view, you would need a capacitor before the regulator sized according to the current drawn by the DAC in order to achieve manageable voltage ripple, and another capacitor after the regulator of a type suitable to compensate for the characteristic rising regulator output impedance as frequency increases. There would be no place for paralleling (as it causes resonance and it usually worsens impedance curves) or for such a thing as a 68000uF capacitor.
On the other hand, from a non-scientific point of view you can do whatever you like as there is no objective criteria to determine what performs better.
On the other hand, from a non-scientific point of view you can do whatever you like as there is no objective criteria to determine what performs better.
Hi,
I wonder if you can measure the effect the reg is having in amongst all that? I don't need to think about whether anyone can hear, or not hear, the reg in there.
AndrewT said:
The reg is to get the desired volzage.
I would like to have the ICs decoupled from the reg. with small R and very large C perhaps.
I would try this:
4700µ>10ohm>68000µ>reg>2200u>R>C (Digital)
R>C (Clock)
R>C (Analog R)
R>C (Analog L)
Even better would be two separate supply for analog and digital.
D
4700µ>10ohm>68000µ>reg>2200u>R>C (Digital)
R>C (Clock)
R>C (Analog R)
R>C (Analog L)
Even better would be two separate supply for analog and digital.
D
Dag said:
Analog:
4700µ+1µ_MKP>10ohm>Cx_68000µ+1µ_MKP>reg>4700µ+1µ_MKP>10ohm>Cy_2x4700µ>on each IC: 47µ_MKL+10n_SilverMica ?
Digital:
4700µ+1µ_MKP>10ohm>Cx_68000µ+1µ_MKP>reg>4700µ+1µ_MKP>10ohm>Cy_2x4700µ>on each IC: 220µ_Oscon ?
Or swap Cx <--> Cy ?
I have seen 20V>1000µ>R>1000µ>R>1000µ>R>1000µ>12V to feed some transistor circuits in a R&S L-C meter.
Hi Bernhard,
Why are you calling up Cx & Cy? Other than value what is the difference?
what is an MKL?
Why are you calling up Cx & Cy? Other than value what is the difference?
A 10 ohm resistor anywhere after the regulator will ruin regulation because the average current consumption of the DAC or whatever circuit you are powering will modulate its own supply voltage. You are very deep into metaphysics and very far away from electronics.
Regulators you're refering are circuits with feedback just like amps, an't they ? A lot of capacitance at their output slows down the whole regulation process. The aim would be to have the best voltage regulation with the lowest noise and the best transient behaviour with the current demanded by the intended load. I fear that huge and fancy loads made from different types of capacitors (each with its own parasitic inductive value) could only do harm for both.
Eva said:
If C is very big, the modulation will be very small ?
In Philips players they have 33ohm>47µ>opamp
Btw. most discussion on this board is about metaphysics 😀
forr said:
That's what I want to keep away from my circuits.
The DAC could be with passive I/V and with passive filter.
forr said:
The reg may have a 1s time constant, like a DC servo of a power amp.
AndrewT said:
MKL is a Siemens film cap with a §|&?'#!% dielectric.
If the value is very high + local low ESR Cs , any modulation will be very low ???
Bernhard said:
That's becuase it allows them to get away with a VERY basic regulator (or sometimes even just a cap multiplier) and still avoid most of the mutual interference. They do this not for engineering reasons, but for the shareholders value. I don't think you want to use that as a role model, so don't bring it into the argument.
If you would use all those fancy cap banks etc mentioned above you really can save the money for a fance reg. Just a 78xx would still be overkill.
Don't make the mistake that combining two things that in itself have merit give you a result that is the sum. It isn't, except for the cost 😉 .
Jan Didden
Hi Bernhard,
Regulators with feedback
---That's what I want to keep away from my circuits.
The DAC could be with passive I/V and with passive filter.---
Avoiding regulators was unclear for me in the initial and following posts. I hope you are aware what is the effect of very high capacitance directly connected to the rectifying diodes. This has been discussed in a thread with details, many of them given by Eva, I think.
Regulators with feedback
---That's what I want to keep away from my circuits.
The DAC could be with passive I/V and with passive filter.---
Avoiding regulators was unclear for me in the initial and following posts. I hope you are aware what is the effect of very high capacitance directly connected to the rectifying diodes. This has been discussed in a thread with details, many of them given by Eva, I think.
I tried few different combinations in my DAC and it sounded the best with:
1000µ >R>L>1000µ> reg > 47µ
1000µ >R>L>1000µ> reg > 47µ
forr said:
I never mentioned connecting the cans directly to the rectifiers.
10 ohm limits the current.
Bernhard said:
This makes me feel at home.
The following approach works for me. If a regulator contributes too much of a horrible sonic signature (78xx etc) an inductor and large cap following it is probably a best compromise, still providing a measure of dc stability. If, otoh, you like the sonic effect of the regulator (Jung, various shunt types, feedback free series type), the minimum of additional filtering seems best.
I have never felt the need to use a 68000µF following a regulator.
- Status
- Not open for further replies.