dCS modification.

[fmak]

If you want to exploit full potential of N caps, always use them in much smaller values than with regular caps. When I insert them in a circuit, the values are usually 3-10 times smaller of the caps I'm replacing.

The Delius might be an interesting choice. Would you mind suggesting a particular model? The Audiogon seem to offer steady selection of those units.

Using a BGN 3 times less than the value of the cap being replaced is ludicrous as this changes the lower roll off frequency by a similar factor. You are comparing airplanes and submarines!

I never said I liked the Delius; I like the 954 and Elgar with upsampling; very analog and musical.

Non OS DACs can have loads of upper frequency harmonics which may make a system sound interesting; just like some valve amps. Both are not for me.

 

[Peter Daniel]

Ludicrous is your comment on something you did not try, in fact quite dissapointing.

To be clear, I was not talking about coupling caps.

 

[Lourens]

Now back to music

I wouldn't mind trying dCS to see how it would compare. Which model would be recommended with the best price/performance ratio?

Hi Peter,

A used Delius seems an obvious choice in terms of p/p ratio. Personally I am not impressed by DSD-upsampling nor SACD, so if you agree, watch out for a pre-Firewire model, in Europe they sell for about € 2000-2500. Insiders told me early software versions could sound harsh (maybe that's the thinness fmak mentioned? something I don't recognize at all). Mine is version 1.13 which sounds perfect to my ears, a little better than a Weiss Medea, almost like the non plus ultra Elgar (Plus) -- at a price )

I do agree with fmak that a dCS upsampler is the better choice if you pursue analogue-like sound. However I do prefer the Delius/Purcell combo over the pro 9xx versions he mentions, and not only because of their looks
For word clock synchronizing a Purcell you need software version 1.25 or later. Units can of course be updated but this can be costly if a hardware upgrade is involved.
Hope this helps.

Lourens
PS if you still have that ML31.5, slave it to a Delius, and be surprised....:

 

[Peter Daniel]

Thanks for the info, I will be looking into that whenever opportunity arises.

My ML31.5 is currently not functional as I'm in a proccess of transport mechanism replacement (the previous one developed certain skipping defect).

 

[Ruach]

PS: The ML360 panel is for decorative purposes only, it covers chassis free NOS DAC that sits behind

 

[fmak]

Ludicrous is your comment on something you did not try, in fact quite dissapointing.

To be clear, I was not talking about coupling caps.

Peter

We WERE discussing coupling caps. Even if you were not, reducing C by 1/3 means that you end up with a loyt more noise at LF, and doubtful improvement at HF.

It's fine if you like the subjective effect. But please don't say that it is generally true.

 

[Peter Daniel]

We WERE discussing coupling caps. Even if you were not, reducing C by 1/3 means that you end up with a loyt more noise at LF, and doubtful improvement at HF.

It's fine if you like the subjective effect. But please don't say that it is generally true.

No, we were discussing BG caps in general, as you mentioned some of them are surrounded by too much hype. I just indicated a way to use them, where hype has no place and results speak for themselves.

Now, with all that subjectivity, your last line is completely out of place. Wasn't it you suggesting at one time resistors directivity and great changes it brings sonically? This is more subjective that whatever I had ever written here.

"reducing C by 1/3 means that you end up with a loyt more noise at LF, and doubtful improvement at HF"

That is complete BS. Reducing cap value in most instances will not bring any noise at LF, and the improvements at HF will be substantial. I tried it many times, I suggest you'll try it as well.

 

[Peter Daniel]

Being involved in a thread like that sometimes has advantages. When I was writing about Bidat, just out of curiosity I checked Audiogon for any listings and apparently there was one for sale: http://cls.audiogon.com/cgi-bin/cls.pl?dgtlconv&1149868288

Well, I just bought it. According to some folks, this is one of the best DACs ever made, will see about it

 

[maxlorenz]

Hi timcatn et alter
I have to comment that BG-N take some days to brake-in. Don't judge them from first impressions, please. They can sound awful at first.

Well, I just bought it. According to some folks, this is one of the best DACs ever made, will see about it

What ugly connectors!

fmak:

There we are; our perceptions are different. I like the much closer resemblance to analog without noise of upsampling both f and bit depth. I have never liked 16/44.1 in any system. No matter how 'good' (expensive) CDs are simply too 1-D and coarse at hf for me. Why do you say the space and greatness are fake. Upsampling lowers the noise floor significantly and pushes all the phase changes at 20 kHz much further up the f range.

OT:
I am a nullity in electronics but I believe I'm a good listener.
I hapen to like both: nonos and upsampled DACs. Of course my humble upsampled digital chain is on an inferior level compared to dCS (or isn't it?).
I recently made a CDP inspired by that Yeo man (very funny guy) with a CDPRO transport and a minimalist NOS DAC (4*TDA1543 paralelled), fed directly through I2S outs. Well, about your impression of CD's sounding 1-D, you should check out this player's depth and soundstage Dynamics and midrange are to die for, also.
Of course, I blame my ineptitude if something sounds bad

(I had a little problem, though, with it connected to UCD amps: probably HF garbage messing with UCD's carrier F?)

Good luck with your mods, timcatn
M

 

[fmak]

"reducing C by 1/3 means that you end up with a loyt more noise at LF, and doubtful improvement at HF"

That is complete BS. Reducing cap value in most instances will not bring any noise at LF, and the improvements at HF will be substantial. I tried it many times, I suggest you'll try it as well.

Whis is this BS? Reducing capacitance by 1/3 means less noise reduction at LF. Read up a textbook!

You actually agreed at the time that treistors are directional.

 

[Peter Daniel]

I still don't oppose resistor directivity, I feel in some cases it is percivable.

Capacitance values are usually much higher than necessary, both in raw PS filtering and after regulation.

After experimenting with much lower values, I'm coming to conclusion that values lower than usually chosen by manufacturers are much better for speed and high frequency extention and don't affect low frequency perfomance or noise susceptibility. Don't take my word for that though, but try it yourself. The feedback I'm getting from other people, greatly confirms my findings.

 

[Peter Daniel]

Somebody e-mailed me about the mods I did to ML37 transport. Here short description:

Normally the transport was very uninvolving and annoying, with lifeless hypnotizing sound signature. Switching from ML31.5 I couldn’t listen to it for more than half an hour.

The initial mods were as follows:

The main filter caps (in transport section) have been changed for 6,800 Panasonic Audio grade (previously 10,000u Chemi-Con); in digital section I used BG N 1000/50 (previously Chemi-Con 6800/35).

Bridges have been replaced by discreet diodes, RGB 15B (the same as used in ML31.5); they work better than MSR860 and need cap bypasses.

I didn’t touch regulators yet, as I feel pretty good about the sound as it is. The caps after two main regulators have been changed to BG N 47/50 (previously Nichicon 330/25).

The caps after 3 additional regulators have been changed to BG N 33/16 (previously Nichicon 100/25). I also removed all ceramic bypasses, as well some of tantalum bypasses (BG N work better without those additional caps).

All other places on board got BG N 10/16 or BG N 4.7/16 caps (previously 100/25).

The output coupling caps have been replaced to Vishay ROE 0.001. I also simplified the output signal path and removed ferrites and ceramic filters.

All caps have been placed in special sockets and many different values/types tried until optimum sound was achieved. I also tested STD, FK, different Nichicons, Rubycons and Oscons. The BG N were the best, but the sonic signature greatly depended on actual cap value. If I placed the originally used values, the sound would be unlistenable.

One cap was completely removed.

Such modified transport is better than ML31.5. I still consider modifying ML31.5, but the usage of cascaded regulators make it a bit too laid back for may taste and I’m not completely convinced such use of regulators is really neccessary. While ML31.5 does not benefit at all from air suspension table (it actually sounds worse), the ML37 is greatly improved when using it.

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

 

[jan.didden]

[snip]"reducing C by 1/3 means that you end up with a loyt more noise at LF, and doubtful improvement at HF"

That is complete BS. Reducing cap value in most instances will not bring any noise at LF, and the improvements at HF will be substantial. I tried it many times, I suggest you'll try it as well.

Petr,

I think what he means is that when you decrease the coupling cap, you decrease the lf response appreciably. THAT is true, no way around it. Coupling caps are sized to give the required lf roll-off frequency.

For instance, with a 1uF coupling cap into a 10k load, your lf roll-off (-3dB) starts at around 15Hz (f=1/(2.pie.R.C). So, if you change the cap to say 0.33uF, you lf response now starts to drop at around 50Hz, which you clearly wouldn't want.

Jan Didden

 

[Peter Daniel]

That is true, but usually the value of those caps is not critical either. Many times one can see 200uF coupling caps where 10uF is still sufficient for most common applications (as the filtering action of a coupling caps at the output is greatly influenced by input impedance of the next stage). I'm presently using 0.1uF in my DAC as the next stage still provides enough input impedance for filtering not affecting the sound.

However, the above quote was specifically in response to cap application in PS filtering.

 

[jan.didden]

That is true, but usually the value of those caps is not critical either. Many times one can see 200uF coupling caps where 10uF is still sufficient for most common applications (as the filtering action of a coupling caps at the output is greatly influenced by input impedance of the next stage). I'm presently using 0.1uF in my DAC as the next stage still provides enough input impedance for filtering not affecting the sound.

However, the above quote was specifically in response to cap application in PS filtering.

0.1uF, really? That's real lean. Even with 100k load that gives you already 45 degrees phase shift at 15Hz. What input stage does that thing after the DAC have? Tube?

Jan Didden

 

[Peter Daniel]

TVC (S&B TX102). I was a bit surprised myself with bass performance, but I'm actuall getting more bass than previously with output buffer and resistor devider (6R buffer output imp, 20k attenuator). I'm not completely sure if it's better though

When caps cost $200 per 1uF one tries to save every penny

 

[Peter Daniel]

Apparently TVC stage represents quite high input impedance at low frequencies: http://www.stevens-billington.co.uk/page102.htm I'm usually set at -20dB attenuation.

I tried it with an output buffer after the low value coupling caps, but I prefer it much better without buffer.

 

[jplch]

mods for elgar plus ?

Hello all, I folowed with interest your exchanges about modifying the delius. I have an elgar plus, and would like to know if any of you have tried anything on this unit.