The second plot has the high secondary resistance, I used two 12VA power transformers with one as the choke.
I don't know what difference the diodes would make. PSUDII can be asked to ignore the ratings.
I think the 5A constant current load derailed the subsequent discussion - given that the target was just a lower power bias supply, and a constant current load messes with the brain when interpreting circuit operation.
A HT supply was shown a few posts before, not sure what the issue is, sorry?
WRT constant current I use this kind of supply with class a amplifiers and decoupling is part of the strategy so constant current is what I'm expecting.
WRT constant current I use this kind of supply with class a amplifiers and decoupling is part of the strategy so constant current is what I'm expecting.
Perhaps it's my interpretation of your post 100. It looked to me like your left hand side simulation had a few concerns. A 36V winding with a 0.34 ohm DCR that was asked to supply a 5A load. The use of 1N4007. The addition of a 1 milliohm series resistor (perhaps that was used as a current sense, as it is otherwise inconsequential). A 330uF cap that would have to be pretty special to withstand 4Arms. And then I thought Andrew was commenting on that simulation.
The transformer was a 300VA toroid, 18+18. DCR was measured with a meter so was a bit of a guess. The resistance was put there to test its damping potential, reduced to .001 to take it out of circuit. I didn't bother to change the diodes, but I haven't noticed that the type plays much of a role in PSUDII, I could be wrong.
I used this for a few weeks at 2.5A, then briefly at 3.75 and 5A. I adjusted the first capacitor on test each time and used low ESR parts designed for switched-mode supplies. The higher voltage parts seemed to have higher ripple ratings. The capacitor was bypassed with a 1uF Rifa PHE426 low L MKP pulse capacitor.
I used this for a few weeks at 2.5A, then briefly at 3.75 and 5A. I adjusted the first capacitor on test each time and used low ESR parts designed for switched-mode supplies. The higher voltage parts seemed to have higher ripple ratings. The capacitor was bypassed with a 1uF Rifa PHE426 low L MKP pulse capacitor.
Yes PSUD2 just gives an alert if the diode is a bit underdone, but I think will still apply the resistive characteristic in the model. I had quickly looked up a few 330uF 100V electro's and 1.8-2A ripple was common max rating, and the frequency multiplier can sometimes not go your way. As such, most designs would not operate a first-stage electrolytic anywhere near that extreme level of voltage and current ripple.
Yes, I just dug out the datasheets myself and I did exceed the ratings, I'm glad this was pointed out. FWIW I was using 200-250V parts. F=100Hz.
In any event this was converted to a 2A bipolar supply with a pair of chokes. Using PSUDII to find that C input kicked in at 4000uF, I used a pair of 8000uF as the first capacitors, bypassed and with the same HF precautions and layout as the thread topic circuit. I've been using this for a few months now.
In any event this was converted to a 2A bipolar supply with a pair of chokes. Using PSUDII to find that C input kicked in at 4000uF, I used a pair of 8000uF as the first capacitors, bypassed and with the same HF precautions and layout as the thread topic circuit. I've been using this for a few months now.
+1 now I have a life again a choke input PS will be an upgrade for a JLH 1969 amp I've got in my possession. Mostly because "it can't be done"
Last night and this morning swapped !N4003 for UA4007 namely D5,6,7,8, 9,10 in my 2 Quad 306 power amps, 24 diodes changed, basically best money ever spent.
I did one channel last night comparing to 1N4003 , the UA4007 channel appeared
slightly louder imaging slightly further back, and was easy to hear to be much
nicer in presentation compared to the standard 1N4003 channel.
This morning I changed the other channel and what a nice overall improvement.
many thanks !
I did one channel last night comparing to 1N4003 , the UA4007 channel appeared
slightly louder imaging slightly further back, and was easy to hear to be much
nicer in presentation compared to the standard 1N4003 channel.
This morning I changed the other channel and what a nice overall improvement.
many thanks !
Diode noise
I have started to replace all my 1N4007 diodes into UF 4007 about 10 years ago. Always so happy that I did the change. Often I was lazy to desolder 4 diodes, but amazingly, even with often 2 level of cascading choke, I still can hear the differences.
But from the diodes spec, I could not find the measurements of diodes noise. Only the leak reverse current, and the recovery time. So has any body tried to measure the noise using a scope of spectrum analyser?
Last night I replaced the SR 306 Schocttky in my blue ray players 5 v output, with a STTH802. Which has a recovery time of 17ns, and reverse leakage of 6ua. The result is a major major improvement in clarity.
I wonder why an ultra fast recovery cud sound better then the schottky...
I have started to replace all my 1N4007 diodes into UF 4007 about 10 years ago. Always so happy that I did the change. Often I was lazy to desolder 4 diodes, but amazingly, even with often 2 level of cascading choke, I still can hear the differences.
But from the diodes spec, I could not find the measurements of diodes noise. Only the leak reverse current, and the recovery time. So has any body tried to measure the noise using a scope of spectrum analyser?
Last night I replaced the SR 306 Schocttky in my blue ray players 5 v output, with a STTH802. Which has a recovery time of 17ns, and reverse leakage of 6ua. The result is a major major improvement in clarity.
I wonder why an ultra fast recovery cud sound better then the schottky...
Yes, usually the differences are caused by the electromagnetically coupling of the forward current-pulses, as well as the reverse-recovery current pulses.
In a digital player, this may show up easily, because the DAC or the opamps become the 'victim' of the coupled pulses.
Important differences in the diodes:
The PN has an ON-voltage (VF) of about 1.0V compared to the 0.7V of the Schottky, and at low voltages, and bridge connexion, this can reduce the peak current, and in turn, the magnitude of the B field.
The PN also has a capacitance much lower than the Schottky (60pF vs 300pF) and this can affect the size and spectrum of the reverse recovery pulses.
If you want to experiment, measure the average/dc current load, and choose a Schottky which is only just enough to manage the current (derated by 20% to allow for non-resistive load). This should allow for smaller capacitance. Adding a small series resistor may give improvement too.
Or, you can construct a separate enclosure for the trafo/rectifier/caps, to keep the B field away from the DAC.
In a digital player, this may show up easily, because the DAC or the opamps become the 'victim' of the coupled pulses.
Important differences in the diodes:
The PN has an ON-voltage (VF) of about 1.0V compared to the 0.7V of the Schottky, and at low voltages, and bridge connexion, this can reduce the peak current, and in turn, the magnitude of the B field.
The PN also has a capacitance much lower than the Schottky (60pF vs 300pF) and this can affect the size and spectrum of the reverse recovery pulses.
If you want to experiment, measure the average/dc current load, and choose a Schottky which is only just enough to manage the current (derated by 20% to allow for non-resistive load). This should allow for smaller capacitance. Adding a small series resistor may give improvement too.
Or, you can construct a separate enclosure for the trafo/rectifier/caps, to keep the B field away from the DAC.
- Status
- Not open for further replies.