LM329 is probably the best low noise and low impedance high stability voltage reference you can still source in through hole package rather easily.
I have the Salas up and running. Hooked up to a Touch Squeezebox. Couple of questions, does the wire used have any affect on the sound of the piece of equipment it is hooked up to? Also I am doing some playing around with caps, are there any that are considered positive in the way it affects the sound?
Thanks much,
Mark
Thanks much,
Mark
I like the reg. It is still settling in and doing some improving. It is ver ynice. My ultimate desire is to do a shootout with another supply that I had made that uses a Music Rail and see how they compare. I plan on the first test today or tomorrow.
As I saw in your picture when we were looking in the heat thing you already got Wima MKS4, not sure about your main filter electrolytic, is it a Nichicon? Can't see the print on it.
You also got quality TO-220 diodes. As long as you use all 4 wires to the contact point at Squeezbox, you got a very decently built and in spec reg. There could be some Teflon bypass arrangement for the MKS caps. Smallest Teflon you could find for size, even 10nF value is OK. Or Polysterene, or Silver Mica. Use them under PCB at MKS pins. That ups resolution a bit, each one has a slightly different tonality in applications that their influence can be picked up at all. But give it 48 hours on the play counter first, the electrolytic tends to settle in that time.
You also got quality TO-220 diodes. As long as you use all 4 wires to the contact point at Squeezbox, you got a very decently built and in spec reg. There could be some Teflon bypass arrangement for the MKS caps. Smallest Teflon you could find for size, even 10nF value is OK. Or Polysterene, or Silver Mica. Use them under PCB at MKS pins. That ups resolution a bit, each one has a slightly different tonality in applications that their influence can be picked up at all. But give it 48 hours on the play counter first, the electrolytic tends to settle in that time.
Thanks, your suggestion got me digging and I did put the wrong one in. Just switched it out with a KG which is what I meant to put in. And the rough edge is gone.
Yes, I use them in my poweramp, the ones with the "golden feet", type III. They are somewhat bigger than the other types. Don't know if it's really better but the specs say yes.
KG are a little bit dark but pleasant all the way. Now I work more with KZ, 4700UF KZ (5x 1000uF) sound more alive and with better image than KG. I need less optional parts(teflon/mks for highs and acoustic ) to get the sound I want. KG in poweramp + preamp + phonoamp is too much, shut in dark sound. too much of anything is never good : you control the spices !
KG are a little bit dark but pleasant all the way. Now I work more with KZ, 4700UF KZ (5x 1000uF) sound more alive and with better image than KG. I need less optional parts(teflon/mks for highs and acoustic ) to get the sound I want. KG in poweramp + preamp + phonoamp is too much, shut in dark sound. too much of anything is never good : you control the spices !
Problems with my build - help!!
Here's my build - pretty standard, 18V is my target. I'm using a 24V transformer, but I only get 3.8V at the output, and adjusting the trimmer doesn't have any effect. All the LEDs light up, and all the right parts are in the right places. (The trimmer is in backwards, but it doesn't fit the other way and it shouldn't make any difference anyway, as long as the wiper is in the middle.) There's about 293mA through R101 (10 ohm), about what I was aiming for. However there's a 25V drop over Q101 (R101 to output) and only 0.1mA through R103, neither of which seems right to me, though I don't know what they should be. Q101 gets pretty warm after a bit but Q306 doesn't. Other voltages: 0V across R102, 115mV across R108, 3.8V over C102, 0V over Q103, 0V over R106. 24.8V over Q102. Can anyone help?
Here's my build - pretty standard, 18V is my target. I'm using a 24V transformer, but I only get 3.8V at the output, and adjusting the trimmer doesn't have any effect. All the LEDs light up, and all the right parts are in the right places. (The trimmer is in backwards, but it doesn't fit the other way and it shouldn't make any difference anyway, as long as the wiper is in the middle.) There's about 293mA through R101 (10 ohm), about what I was aiming for. However there's a 25V drop over Q101 (R101 to output) and only 0.1mA through R103, neither of which seems right to me, though I don't know what they should be. Q101 gets pretty warm after a bit but Q306 doesn't. Other voltages: 0V across R102, 115mV across R108, 3.8V over C102, 0V over Q103, 0V over R106. 24.8V over Q102. Can anyone help?
Attachments
Fabelous discussion about the dynamic resistance, Salas, Andrew, merlin and others!
Is the adviced 220uF for C101 (in case of zeners, or dig clocks) based on the adviced minimum impedance on 150-220 ohm in the Vref chain??
Is 220 an optimal value, or could it be as big value as possible??
If yes, could the minimum impedance be lower for the Vref chain, if you used say 1500uF?
Peter
Yes, I had run noise sims with 220uF and 200R. It can be a bigger cap for lower impedance depending how effective you wanna be for below 1/f noise knee. Keep in mind that an non damped big cap can stretch semis during on/off there will be a practical threshold for each Vo level they could take.
So double C and half R (like 440uF and 100 ohm) would give the same frequency knee ??
And higher C, and same R - for better 1/f noise knee
2)
Do you mean that the big cap will unload charge through the Vref-chain, and give a high peak-current, Io? I tried to analyze it, and i cant see how - since basically the Current at on/off is the same as the normal working current, and is limited by R103:
in my understanding the size of the peak-current isnt determined by C. It is I = Io*exp(-t / RC). So C only determines the length of the peak.
Io is limited by the resistor, so Io = Vr / R103
Now asuming the chain is R + Zener:
- then Vr (V over R103) = Vtotal - V(zener)
- which basically gives the same Io as normal working current, determined by Idss of K117.
So if you make R103 less and C bigger, - Io stays the same, because V(zener) has to be bigger.