This question has been both asked and discussed here, for example here in 2003 and here in 2014. Why not do it again in 2024? I'm looking for a bit of an education on the relevant considerations, not a 'vote-by-SMS' contest to see what solution gets the most votes.
The situation is pretty straightforward. I have little room in the box with my audio PCB and consider placing the (toroidal) transformer in a separate box. This begs the question: where go the bridge rectifier and the smoothing cap? In the box with the transformer or in the box with the audio circuit? More importantly: why? Regulators are in the box with the audio circuit, this seems the easy part.
Strictly in terms of space, my preference would be that they sit in the box with the transformer. But EMI-wise and otherwise - would that be wise? I suppose the choice is between:
1. rectifying and smoothing in the transformer's box and carrying the DC through relatively long cabling (approximately 30-50cm of cable between the boxes) to the audio circuit box; and
2. sending AC through the cables and rectifying and smoothing close the audio circuit.
If choosing option 1 - is there any benefit or detriment in adding a second smoothing cap close to the audio circuit? Are there good reasons for not placing the smoothing caps a long distance from the audio circuit?
As stated in the question, the practical consideration is space. But to what extent does EMI come in to play to affect the turntable's signal? If yes, it would be a 50Hz (100Hz after rectification) problem, so not straightforward to shield against.
On a side note. Yesterday I was playing around with the (high level) design of a tie-in inverter. The outcome of some measurements made me want to know if the grid frequency at that moment was close to 50Hz, where it should be. Turns out that it was. (Using my multimeter that has a frequency counter - I don't own a scope.) It also turned out that when I took the probes out of the wall outlet (or rather, a socket in a powerstrip on my desk), the multimeter kept showing 49.99Hz. This seemed odd. After having briefly investigated, the wires of my multimeter probes turned out to be inductively picking up the frequency from the powerstrip. After moving the probe wires about 1 meter away (bit less actually, but not much) from the power strip, the meter fell to zero. Moving it back just a little bit saw the meter moving up to 50Hz again. Amazing.
To those who made it this far: my apologies for this long anecdote. For those who object on grounds of relevance, I add to my question: is this phenomenon undesirable inside the box that has the audio circuit? Intuitively I'd say yes. However, I have phono preamps that have a transformer in the box without there being an audible 50Hz hum.
On the other hand: if the rectifier is in the audio box, clearly there is a cable carrying AC current that is happily radiating its 50Hz stuff inside the same box. (Not) a problem?
I wish everyone an otherwise happy Easter.
The situation is pretty straightforward. I have little room in the box with my audio PCB and consider placing the (toroidal) transformer in a separate box. This begs the question: where go the bridge rectifier and the smoothing cap? In the box with the transformer or in the box with the audio circuit? More importantly: why? Regulators are in the box with the audio circuit, this seems the easy part.
Strictly in terms of space, my preference would be that they sit in the box with the transformer. But EMI-wise and otherwise - would that be wise? I suppose the choice is between:
1. rectifying and smoothing in the transformer's box and carrying the DC through relatively long cabling (approximately 30-50cm of cable between the boxes) to the audio circuit box; and
2. sending AC through the cables and rectifying and smoothing close the audio circuit.
If choosing option 1 - is there any benefit or detriment in adding a second smoothing cap close to the audio circuit? Are there good reasons for not placing the smoothing caps a long distance from the audio circuit?
As stated in the question, the practical consideration is space. But to what extent does EMI come in to play to affect the turntable's signal? If yes, it would be a 50Hz (100Hz after rectification) problem, so not straightforward to shield against.
On a side note. Yesterday I was playing around with the (high level) design of a tie-in inverter. The outcome of some measurements made me want to know if the grid frequency at that moment was close to 50Hz, where it should be. Turns out that it was. (Using my multimeter that has a frequency counter - I don't own a scope.) It also turned out that when I took the probes out of the wall outlet (or rather, a socket in a powerstrip on my desk), the multimeter kept showing 49.99Hz. This seemed odd. After having briefly investigated, the wires of my multimeter probes turned out to be inductively picking up the frequency from the powerstrip. After moving the probe wires about 1 meter away (bit less actually, but not much) from the power strip, the meter fell to zero. Moving it back just a little bit saw the meter moving up to 50Hz again. Amazing.
To those who made it this far: my apologies for this long anecdote. For those who object on grounds of relevance, I add to my question: is this phenomenon undesirable inside the box that has the audio circuit? Intuitively I'd say yes. However, I have phono preamps that have a transformer in the box without there being an audible 50Hz hum.
On the other hand: if the rectifier is in the audio box, clearly there is a cable carrying AC current that is happily radiating its 50Hz stuff inside the same box. (Not) a problem?
I wish everyone an otherwise happy Easter.
Place the rectifiers and main filter capacitors in your power supply box. This will minimize the ripple in your cabling to your phono stage. Ripple is AC and will contribute to the radiated noise.
In your phono stage box, place decoupling capacitors as close as possible to your active stages. This will decrease any noise from the power supply or picked up by the power cable. Even better, include a small value resistor before the decoupling capacitor to form a RC filter.
In your phono stage box, place decoupling capacitors as close as possible to your active stages. This will decrease any noise from the power supply or picked up by the power cable. Even better, include a small value resistor before the decoupling capacitor to form a RC filter.
This is how Threshold built my FET10-Pe phono stage, except the capacitors in the main phono stage were close to the DC input to the enclosure along with the regulators.
Salas recommends a braided sleeving over the DC cables from the power supply to the FSP phono preamp. The braided sleeving is electrically connected to the umbilical connector of the power supply and then no connection at the phono preamp connection.
Okay. Even as this has somewhat taken on the character of a vote, thanks everyone for weighing in. The case for DC power output appears overwhelming and this is very helpful.
In the last version of my design I added 1000uF at the input of the regulators in the box with the audio circuit. In one of the earlier threads that I linked to in the original post Jan Didden advises against using heavy gauge cables, saying it’s good to have a bit of R or L between the smoothing caps in the transformer’s box and the decoupling at the load. That’s tantamount I suppose to adding a small resistor as per Rod Elliot’s (and pinholer’s) suggestion above.
Thank you also for this:
The other day the designer of some American commercial audio kit suggested that for DC cabling shielding or no shielding is equivalent. I had considered using cabling with a braided shield and may actually do just that.
The same guy suggested to use star quad cabling for DC because it has the lowest self inductance of any cable geometry. That flies in the face of Jan’s recommendation of over 20 years ago…
I’d much appreciate it if anyone finds the time to weigh in on this too.
In the last version of my design I added 1000uF at the input of the regulators in the box with the audio circuit. In one of the earlier threads that I linked to in the original post Jan Didden advises against using heavy gauge cables, saying it’s good to have a bit of R or L between the smoothing caps in the transformer’s box and the decoupling at the load. That’s tantamount I suppose to adding a small resistor as per Rod Elliot’s (and pinholer’s) suggestion above.
Thank you also for this:
The other day the designer of some American commercial audio kit suggested that for DC cabling shielding or no shielding is equivalent. I had considered using cabling with a braided shield and may actually do just that.
The same guy suggested to use star quad cabling for DC because it has the lowest self inductance of any cable geometry. That flies in the face of Jan’s recommendation of over 20 years ago…
I’d much appreciate it if anyone finds the time to weigh in on this too.