Hotrodding the UCD modules

[classd4sure]

Hi Gertjan,

Excellent work, thanks for sharing.

Do you suppose the Wima perform better and are more stable in frequency due to having two in parallel and thus lower ESR? Also how were they mounted, one on top and one on the bottom?

I'm sorry if I missed it but what was the input signal, or was this done at idle? Might be interesting to step the input/amplitude. Maybe we'd get to see the microphonic effect begin to appear with the stacked film or something?

What 470nF are you going to try?

Thanks,
Chris

[analogspiceman]

Originally posted by "classd4sure" Chris

Do you suppose the Wima perform better and are more stable in frequency due to having two in parallel and thus lower ESR? Also how were they mounted, one on top and one on the bottom?

I'm sorry if I missed it but what was the input signal, or was this done at idle? Might be interesting to step the input/amplitude. Maybe we'd get to see the microphonic effect begin to appear with the stacked film or something?

It takes much care in measurement setup to get good "apples-to-apples" comparisons.  To keep the noise floor out of the picture, the basic measurement must be very clean.  This is probably best done using a 'scope tip jack (facilitates a "zero" length ground lead) soldered in-circuit at the measurement point and a common mode choke installed on the probe cable (wrap it several times through a high perm ferrite toroid).  Also, the fft time window probably must be carefully realigned for each measurement as the UcDs operating frequency will vary slightly with variations in the output capacitor's value and ESR (this could explain some of the subtle differences in the harmonic ratios and "noise" floor prior to the first peak, etc.).

Just a thought. -- analogspiceman

[maxlorenz]

Sorry to interrup this interesting talk.

Quick report about my HF noise on UCD180. I found the time to address it and one of the little 22uF/50V caps was dammaged: it had brownish material on bottom. I changed it for a 4.7uF BG-N that I had on hand and this channel has now good HF performance. The other unmoded channel has tolerable distortion. I could hear the wonderful performance of Tchaikovsky's violin concerto by David Oistrach without getting sick.

No one seem to have suffered from this failure and the other (visible) caps look good so I guess it's bad luck. Anyway, my iron is hot and nervous and I want to swap more caps.
Please help me to identifie the caps wich are good to be changed.
What do you use for AC coupling and low voltage PS decoupling?
I have here some Elna Starget 22uF/50V wich usually make the sound colorful and warm but are not too transparent. Are they OK for those tasks?

[IMG][/IMG]

The one that failed is the red cap on bottom/left, with an arrow and a "1".
Is the bootstrap the lonely, solid blue cap? (on the figure, that is)
This image was imported from the parallel thread and has markings for outboard opamp PS (for the brave among us ).

Many thanks for helping with UCD "nomenklatur"
Mauricio

[classd4sure]

Originally posted by maxlorenz
Sorry to interrup this interesting talk.

Quick report about my HF noise on UCD180. I found the time to address it and one of the little 22uF/50V caps was dammaged: it had brownish material on bottom. I changed it for a 4.7uF BG-N that I had on hand and this channel has now good HF performance. The other unmoded channel has tolerable distortion. I could hear the wonderful performance of Tchaikovsky's violin concerto by David Oistrach without getting sick.

No one seem to have suffered from this failure and the other (visible) caps look good so I guess it's bad luck. Anyway, my iron is hot and nervous and I want to swap more caps.
Please help me to identifie the caps wich are good to be changed.
What do you use for AC coupling and low voltage PS decoupling?
I have here some Elna Starget 22uF/50V wich usually make the sound colorful and warm but are not too transparent. Are they OK for those tasks?

[IMG][/IMG]

The one that failed is the red cap on bottom/left, with an arrow and a "1".
Is the bootstrap the lonely, solid blue cap? (on the figure, that is)
This image was imported from the parallel thread and has markings for outboard opamp PS (for the brave among us ).

Many thanks for helping with UCD "nomenklatur"
Mauricio

Brownish material you say, sure it's not just some left over resin? Rather odd for that cap to have a problem I'd think, it's just the AC coupling cap, hardly put to work at all, but, if you say it helped..

So yeah the two red ones are the coupling caps, get ride of them or upgrade them (you've seen the warnings etc)

Yes the blue cap is the bootstrap cap.

Black ones need to be high ESR, I haven't changed mine.

Blue stripped is the filter cap, upgrade it! Black stripped are the high voltage local supply/decoupling caps, take your best shot with those, they can change the sound a good deal.

[Mike2]

Here is a shot of the Wima's on my v5 boards. My new v6 boards have the red Panasonic caps so I will be able to compare them with the Wima's. Also installed a separate PS and regulators for the Op-amps late last Friday. Only about 50 Hrs. on them so far but the sound is very good. I will post more on this later.

[ghemink]

Originally posted by classd4sure
Hi Gertjan,

Excellent work, thanks for sharing.

Do you suppose the Wima perform better and are more stable in frequency due to having two in parallel and thus lower ESR? Also how were they mounted, one on top and one on the bottom?

I'm sorry if I missed it but what was the input signal, or was this done at idle? Might be interesting to step the input/amplitude. Maybe we'd get to see the microphonic effect begin to appear with the stacked film or something?

What 470nF are you going to try?

Thanks,
Chris

Hi Chris,

The two in parallel may actually be a factor, this should lower the inductance and the caps itself have a lower loss, and thus a lower ESR already.

There was no input signal as I wanted to have the FFT as clean as possible, not modulated by an input signal.

With the 470nF, I mean 470uF for the power supply decoupling caps. Thinking of trying panasonic FC 680uF/100V but also have some other caps lying around.

Best regards

Gertjan

[ghemink]

Originally posted by analogspiceman


It takes much care in measurement setup to get good "apples-to-apples" comparisons.  To keep the noise floor out of the picture, the basic measurement must be very clean.  This is probably best done using a 'scope tip jack (facilitates a "zero" length ground lead) soldered in-circuit at the measurement point and a common mode choke installed on the probe cable (wrap it several times through a high perm ferrite toroid).  Also, the fft time window probably must be carefully realigned for each measurement as the UcDs operating frequency will vary slightly with variations in the output capacitor's value and ESR (this could explain some of the subtle differences in the harmonic ratios and "noise" floor prior to the first peak, etc.).

Just a thought. -- analogspiceman

Hi spiceman,

I know it is easy to make a worthless measurement. I connected about 30cm speaker cable to the UcD module. I used an 8Ohm dummy load and hooked up the oscilloscope probe at the end of that speaker cable. So the GND loop of the probe was quite far away from the UcD module itself. I also used a ferrite core on the scope cable to suppress common mode noise and I connected only one probe, so using only 1 measurement channel. Since the FFT window I used was long (32768 samples at 50Mhz sample rate) I think I do not have to worry too much about the accuracy related to that.

I have actually made a home made BNC cable ending with two twisted wires (for GND and signal) to be soldered directly to the PCB to minimize the loops that can pick up noise. I actually have used that already to measure the signal over the 470uF power supply decoupling caps, using it with a common mode suppressing ferrite core as well. I could nicely measure the waveform over the cap. Want to experiment with different caps later.

Best regards

Gertjan

[ghemink]

Originally posted by Mike2
Here is a shot of the Wima's on my v5 boards. My new v6 boards have the red Panasonic caps so I will be able to compare them with the Wima's. Also installed a separate PS and regulators for the Op-amps late last Friday. Only about 50 Hrs. on them so far but the sound is very good. I will post more on this later.

Are those the same WIMA`s as I`m using?

Gertjan

[Mike2]

Yes, these are the exact sameones you have.

[t.]

Where did you guys buy those Wima's? they look small for a polypropylene cap

Cheers