Hi all,
When powering a couple of ldo regulators like for example the adm7150 giving 3.3v to the vdd of a DAC chip and an oscillator, what's the ideal supply for that regulator? A free flowing unregulated dc supply just with the spikes filtered out? Or a clean filtered dc from another regulator? And if there's the voltage headroom, why not have a regulated power supply feeding a second low noise regulator, with the output from that feeding the adm7150 3.3v reg?
Is there a reason why too many successive layers of regulation would cause a problem? Or is it just unnecessary?
I ask because it seems that for analogue circuits with shunt regulators, it seems that they can sound really special if they're fed with a very low impedance unregulated raw supply which can react to an ever changing musical requirement, but I suspect that digital components like oscillators have more of a steady, consistent requirement and may benefit most from a power supply with different characteristics.
Regards,
James
When powering a couple of ldo regulators like for example the adm7150 giving 3.3v to the vdd of a DAC chip and an oscillator, what's the ideal supply for that regulator? A free flowing unregulated dc supply just with the spikes filtered out? Or a clean filtered dc from another regulator? And if there's the voltage headroom, why not have a regulated power supply feeding a second low noise regulator, with the output from that feeding the adm7150 3.3v reg?
Is there a reason why too many successive layers of regulation would cause a problem? Or is it just unnecessary?
I ask because it seems that for analogue circuits with shunt regulators, it seems that they can sound really special if they're fed with a very low impedance unregulated raw supply which can react to an ever changing musical requirement, but I suspect that digital components like oscillators have more of a steady, consistent requirement and may benefit most from a power supply with different characteristics.
Regards,
James
At 1/3 and 2/3 of rail is the switching point, so a little noise makes no difference. An analogue well designed amplifier using a dual rail supply is not fussed about noise either. Fitting extra smoothing and super ripple filters is just not necessary unless you are running the amplifier at full power. At full power, dips in the rail voltage will cause sag on the output to the speaker and introduce distortion but only on its limits. As long as the power supply has an impedance lower than the load, there is no issue. Take a look at Crown amplifiers. They are the most reliable and best power producers in the commercial industry. Quad use a small tank capacitor in the power supply, (2200uF X 2) because any larger is not required and puts a strain on the mains transformer.
Does this LDO chip produce a sufisant low impedance ? As good as a shunt reg ?
Any interest to have a low impedance also for digital ?
BTW, (sorry off topic) any oem chip as good as complicate shunt circuits ?
Nice car JSE...
Does the ripple capability of this cap stay important ? Inductance (pitch of the cap leads ?) and ESR in the power supply you talk about ?
Any interest to have a low impedance also for digital ?
BTW, (sorry off topic) any oem chip as good as complicate shunt circuits ?
Nice car JSE...
Does the ripple capability of this cap stay important ? Inductance (pitch of the cap leads ?) and ESR in the power supply you talk about ?
I'm asking to me the same questions as you James, and I've heard some ones telling that PSU play an important role in digital audio. The link between SQ and PSU invoked was mostly jitter, and that low impedance is what matters most.
Sooo, I've upgraded, just to verify, in my full digital amplifier (DDX320 from hifimediy), replacing a dc-dc converter with a Belleson one (3.3v) ---> big, obvious improvement. This experiment was suggested reading "soundcheck" diyaudio member's positive experience upgrading digital board's PSU.
Now I'll replace de SMPS on this same unit by a linear one ... who knows ? At "Computer audiophile" website, there are some positives results with linear PSU compared to SMPS.
All the best.
Alain
Sooo, I've upgraded, just to verify, in my full digital amplifier (DDX320 from hifimediy), replacing a dc-dc converter with a Belleson one (3.3v) ---> big, obvious improvement. This experiment was suggested reading "soundcheck" diyaudio member's positive experience upgrading digital board's PSU.
Now I'll replace de SMPS on this same unit by a linear one ... who knows ? At "Computer audiophile" website, there are some positives results with linear PSU compared to SMPS.
All the best.
Alain
If you seek a cost-is-no-object, brute force, and massive overkill design, a cascade of three stages might just be what you want.
Stage 1 is an RCRC lowpass filter whose only job is to attenuate high frequency (RF) noise.
Stage 2 is a linear IC voltage regulator. LDO or not, your choice.
Stage 3 is a discrete shunt regulator, current source + follower, using a low noise reference.
A 3-stage cascade like this will voraciously eat up voltage headroom. It will wastefully dissipate a lot of power. It will occupy lots more PCB real estate than a single IC regulator. It will cost more to purchase the parts. It will also give you that glowing feeling of "leaving no stone unturned," that you did everything humanly possible to reduce noise, reduce ripple, and reduce output impedance.
Stage 1 is an RCRC lowpass filter whose only job is to attenuate high frequency (RF) noise.
Stage 2 is a linear IC voltage regulator. LDO or not, your choice.
Stage 3 is a discrete shunt regulator, current source + follower, using a low noise reference.
A 3-stage cascade like this will voraciously eat up voltage headroom. It will wastefully dissipate a lot of power. It will occupy lots more PCB real estate than a single IC regulator. It will cost more to purchase the parts. It will also give you that glowing feeling of "leaving no stone unturned," that you did everything humanly possible to reduce noise, reduce ripple, and reduce output impedance.
I'm asking to me the same questions as you James, and I've heard some ones telling that PSU play an important role in digital audio. The link between SQ and PSU invoked was mostly jitter, and that low impedance is what matters most.
Sooo, I've upgraded, just to verify, in my full digital amplifier (DDX320 from hifimediy), replacing a dc-dc converter with a Belleson one (3.3v) ---> big, obvious improvement. This experiment was suggested reading "soundcheck" diyaudio member's positive experience upgrading digital board's PSU.
Now I'll replace de SMPS on this same unit by a linear one ... who knows ? At "Computer audiophile" website, there are some positives results with linear PSU compared to SMPS.
All the best.
Keep us informed please. I bought this amp after the good first review of Soundcheck but was very disapointed by the stock sound (a crap !) though not to complain too much in relation to the good price : QP stay good even stock with no tweaks around 🙂
Take care to high voltage involved with the SMPS, please.
Alain
I love Mark's way, and I was hesitating going so far ... no more now 
Before that, my next step will be a LC(LC) with a mains transformer nicely snubbed (thank you Mark for Quasimodo 😉 ).
About shunt reg., I was thinking about Paul Hynes, Salas. About Paul's, some designs have a bi-directional current capability
Is it needed / useful for digital circuits ?
@ Eldam : QP = quality price ?
Alain
Before that, my next step will be a LC(LC) with a mains transformer nicely snubbed (thank you Mark for Quasimodo 😉 ).
About shunt reg., I was thinking about Paul Hynes, Salas. About Paul's, some designs have a bi-directional current capability
Is it needed / useful for digital circuits ?@ Eldam : QP = quality price ?
Alain
- Status
- Not open for further replies.