Is the SAA7220P/B really that bad ?

Phil/All:
philpoole said:
Yeah, I noticed you've been busy with PMD100 and TDA1541A recently on the other forum. Glad you've got it working.
And it’s just that: “working”. For whatever reasons (and for the most part), I still way prefer my original 7220B-TDA1541A combo. Anyone else out there prefer 7220B-TDA1541A to PMD100-TDA1541A, for non-HDCDs, anyway?
Alas, there are so many variables to tweaking these things, involving a lot of art as well as science. I’m still very much a rookie!
BTW (Phil): Are you using the combo in a CDP or stand-alone D/A processor? What brand/model, or is it all DIY?
Me, I’m working in a Philips CD-650, and sometimes CD-60.
All my DEM reclocking (mis?)adventure has been with SAA7220 (and I think maybe NOS as well).
My initial thoughts, and hence that post, were that it was an improvement, when really it was just different (and not for the better). It seemed brighter, and further listening, especially with headphones, revealed this horrendous hash, and that actually, it was of no improvement at all.

It became apparent to me that actually my DEM reclock circuit was useless and was equivalent to not having anything there at all. This was reinforced by the fact that the hash vanished as soon as I replaced this circuit with the standard 470pF cap. Hohum.

I think these digital artefacts you hear are either due to similar issues I had with DEM reclocking, or perhaps poor grounding (I discovered great improvements in reworking my ground return paths a bit). I don't think it's related to the PMD100, or the TDA1541A selection, as rfbrw suggested. I have 10 TDA1541As and I've not noticed any sounding the lesser in this current setup I have (although I haven't needed to try all of them). The TDA1541A is still working within its rated limits (8x oversampling).
No, the artifacts are related to my implementation of the Guido PMD100-to-1541 module w/o DEM. The artifacts I hear are not due to DEM, because I have not been able to get that to work (with PMD100) period. (Will try DEM again when 7220B is back in.)
With the Guido PMD100 module, when starting the CDP from cold, there is quite a bit of ‘hash’ noise, but that diminishes very quickly as playback begins. During quiet passages, I can certainly hear noise. Also, there’s something “unnatural” about the sound overall: while there's lots of detail (maybe more than 7220), there's also lack of focus, congestion, and glare. All that said, I haven’t played with careful grounding yet -- it’s all still on protoboard! Also, separate PSUs for Vdd1 and Vdd2, as well as isolating the PMD100 from the rest of the logic ICs in that module -- these and others are tweaks yet to be played with.
I think, since I've improved my layout and grounding a bit, I may well return to attempt DEM reclocking. However, I'm in the middle of implementing a new I/V and Bessel filter for it instead.[/B]
Your comments here point to an important consensus: at any given time, one has myriad options to improve SQ. As far as I/V, I’m using Pass/Promethius D1. I’m looking for a way to servo the output so I can ditch that cap. If you’ve tried the Rbrorer-Jocko I/V, for the TDA1541, let me know if it worked.
Continues success!
 
For whatever reasons (and for the most part), I still way prefer my original 7220B-TDA1541A combo. Anyone else out there prefer 7220B-TDA1541A to PMD100-TDA1541A, for non-HDCDs, anyway?
Blimey! Well, that's a shame. I hope you sort it out. I much prefer the PMD100 to the SAA7220.

Are you using the combo in a CDP or stand-alone D/A processor? What brand/model, or is it all DIY?
My system is rather DIY. I have a Philips CD950 (CDM9) stripped down to just a transport. I2S is tapped off, and sent via RS423 to a seperate DAC.
The DAC has a seperate PSU (two toroids), RS423 receiver, PMD100, dual TDA1541A simple opamp IV stage followed by a 3rd order passive LPF (very soon to be replaced by an active equivalent and output buffer, with HDCD gain switching, done with opamps).
There's a photo or two some where in the forums - perhaps the CD650 thread.
Future developments would feature revisiting discrete I/V and DEM reclocking, as well as further improvements to the regulation.

-- it’s all still on protoboard!
Nothing wrong with that. So's mine! And it will probably stay that way.


when starting the CDP from cold, there is quite a bit of ‘hash’ noise, but that diminishes very quickly as playback begins
That sounds like you're not muting the PMD100 when you power up, so its munging the garbage coming out of the decoder. The error signal should sort that out, I think it can be applied to soft mute on the PMD100 (not that mine does at the moment).

Cheers,
Phil
 
Yes I am.
Pretty much everything has its own regulator, except the dual TDA1541As share their regulation at the moment (the second DAC is appended to the first's circuit board for now). The PSU is also regulated, so everything is dual regulated. For the PMD I have two 7805s feeding each supply pin.
Every circuit has a regulator, and at least a decoupling capacitor. A lot of them have an inductor in series prior to that, forming a CLC pi filter with the main PSU's capacitors.
 
philpoole said:
Yes I am.
Pretty much everything has its own regulator, except the dual TDA1541As share their regulation at the moment (the second DAC is appended to the first's circuit board for now). The PSU is also regulated, so everything is dual regulated. For the PMD I have two 7805s feeding each supply pin.
Every circuit has a regulator, and at least a decoupling capacitor. A lot of them have an inductor in series prior to that, forming a CLC pi filter with the main PSU's capacitors.
Your PSU scheme is way better than mine, which may be part of the reason my 7220 sounds better. My PMD setup is similar to Guido's module, shown here.
Did you start out this way, or did you improve/upgrade incrementally?
This project was, for me, first a proof-of-concept: that two unlikely pairs -- 7220 and 1541 -- could be made to work together. That homework was done by Guido and others! So in that way, the project is a success.
IAC, my 7220-1541 setup had two lines reclocked (Data and WS); and BCK was fresh-squeezed (and /2), not imported from the dirty OS/DF. If one adds all the pluses of the old 7220-1541 system together, I think the performance sum is greater than the PMD100-1541 system I currently have.
I still honestly believe the PMD100 is way better than 7220, and a PMD100-1541 system can, all else being equal, way outperform the earlier. One huge obstacle for me is that my CD-650 is a bi*ch to mod (for one thing, it's one of the first-year models with an MAB daughter board). So I have not experimented as much as I like. I'm prepping my CD-60, which is a lot easier to mod. First on the list is DEM, then PMD100, so stay tuned ...
 
In-series resistors for dogotal lines

philpoole said:
For the PMD ...
Forgot to ask if you installed low-value in-line R's in the logic ckts? These were not noted in Guido's design schema. There are quite a few loops in that schema, so I'm not exactly sure where these R's would help/hurt? At least after the decoder and after DF, tho'.
 
I want to mod my Rotel RCD951 with PMD100 and PCM69 to TDA1541A. not much help in this thread Isn't it decode the HDCD automatically, then output to dac? What is the min bit of the dac needed to have HDCD function? May be someone can help here, thank you.

Well, in response to the thread you refer to, its obviously possible to do PMD100 to TDA1541A. Hollowman refers to a link to Guido's implementation a few posts ago. This is what I have used, and it works.

HDCD works with a 16 bit DAC (perhaps not optimally). It can either work in 16 bits (but normal CDs are reproduced within 15 bits) or an extra pin is used to switch the gain the output stage to be 0dB or +6dB, which sort of gives you an extra bit. Admittedly, not the full 20 (or is it 22?) bits that HDCD can support, but its an improvement. Oversampling can sort of be regarded as increasing the bit depth as well (is it 0.5bits per 2x oversampling, so in this case an extra 4 bits? - I could well be wrong there) - so more resolution than a 16 bit NOS anyway.
HDCD or not, you'll probably see from discussion in this thread (which is rather off topic now 🙄 ) that, in general, PMD100 is often regarded a better digital filter to the SAA7220.
 
The artifacts I hear are not due to DEM, because I have not been able to get that to work (with PMD100) period

I've just remembered another thing I did, which hindsight has proven rather dumb, was keeping the two GNDs on the PMD100 seperate. I had lots of artefacts when it was like that. Tying them together fixed that for me. It was one of my grounding problems.
 
HDCD gain

philpoole said:
HDCD works with a 16 bit DAC (perhaps not optimally). It can either work in 16 bits (but normal CDs are reproduced within 15 bits) or an extra pin is used to switch the gain the output stage to be 0dB or +6dB, which sort of gives you an extra bit.
TTBOMK, the +6dB gain is best done in the analog domain: SCAL = high. GAIN pin is 'on' for those specific HDCD discs that contain +6dB. A bit complex, here. Can't folks "simply" crank up their vol. control a bit?
 
Hi Hollowman,

My current output stage is going to use the SCAL pin. I was of the opinion that it was to some extent dynamic, so that it would switch the gain on quiet parts of tracks. So its not necessarilly turning the volume up for HDCD.
However, that'san assumption. I haven't studied the behaviour of this pin yet.

Cheers,
Phil