Is the SAA7220P/B really that bad ?

[JonHarrison]

Andy,

Has anyone thought about producing a daughter board to combine an external clock, divider, SAA7220 etc ?

Yes, all with seperate regulation. Kind of a NOTNOS upgrade. I've socketed my chips for just such a purpose !

I've also thought about replacing the SAA7220 with an FPGA / DSP so that I can experiment with different filter co-efficients and reconstitution / interpolation algorithms.

Jon

 

[Fin]

Good idea for a thread!
2) Simply stop putting the clock signal through the filter. Use a separate 11.288Mhz clock with D-flipflop divider* to feed 5.66Mhz to the 1541, and 11.288Mhz to the 7220 clock input and to the preceeding 7310 decoder. That is - take all responsibility for clock generation and distribution away from the 7220. Problem solved!

Hi Martin

Do you think that is is worthwhile trying to reclock WS and DATA in addition to what you recommend above?

 

[martin clark]

Dunno, I've not tried it.

I have tried RC filtering of these signals though, as recommended in the monster TDA1541 thread of last year to keep HF currents out of teh dac substrate. I liked the effect, so perhaps the precise timing of these two isn't so significant. It is for LE though!

 

[philpoole]

Martin,
Did you try the voltage level adjustment as well as LPF for the TDA1541?
Cheers,
Phil

 

[martin clark]

Not yet!

 

[zygibajt]

Halo,

Considering SAA7220 and NOS mode.

For many years I have been modifying CD players with TDA1541A and SAA7220 on the board. I have pleyed with virtually any old model of Philips or Marantz with this old chips (so far my favourite is Philips CD 960/Marantz CD 94).

1. Nos sound sucks. I'm really annoyed to hear all this voices how great imrovement it brought to the sound. Omitting SAA7220 really sucks. And every nos dac I tried (TDA1541 or TDA1543) realy sucks to me. If one has trully reveling set up (really hi-end audio) one can clearly hear no oversampling really does make it sound worse.

2. All the above comments apply to highly modified machines with seaparate clocks, separate supply lines for the dac and digital filter and separate supply for analog stage and additional separate analog stage itself, and many many other tweaks and up grades.

3. SAA7220P/B sounds better that P/A version of this chip.

Bartek

 

[Fin]

Hi Martin

.....the key tweak is to use RC decoupling in the signal lines between these two chips to reduce HF noise injection into the dacs substrate (1K/10pF was suggested)*. Your Arcam already has some resistors in place in these lines, so definitely add small caps (10-22pF) to digital ground at the 1541 input pins. It helps a surprising amount.

*this is because the 7220 swings 5v pk-pk on its outputs, and the 1541 only needs a tiny current centred c1.4v to switch - it's a current-routing logic; the excess voltage swing just pushes HF currents into the dac.

The resistors in the Alpha 5 are 390R for WS and DATA, and 120R for BCK. In the Delta 70.2 they are 390R for all lines. Should these values be altered?

I have tried RC filtering of these signals though, as recommended in the monster TDA1541 thread of last year to keep HF currents out of teh dac substrate. I liked the effect, so perhaps the precise timing of these two isn't so significant. It is for LE though!

Sorry for my ignorance but can you elaborate on "LE"?

 

[philpoole]

Nos sound sucks. I'm really annoyed to hear all this voices how great imrovement it brought to the sound. Omitting SAA7220 really sucks. And every nos dac I tried (TDA1541 or TDA1543) realy sucks to me

So you're not a fan then
Fair enough, I think most people are with one or the other.
I'm hoping to get hold of an old player off of ebay in the next few days with an SAA7220 on board, so I can form more of an opinion.

Fin, Martin,
I'm not sure what LE is either, but too timid to ask.
WRT the resistors, on a different thread (can't remember which one) I read of a player that 200ohm resistors on the I2S lines prior to the TDA1541.
Prompted by this thread, I found some 240ohm resistors and did the same to my NOS 1541 over the weekend, and I think it made an improvement.
My DAC is driven by an RS423 receiver chip, so might have similar output characteristics to the SAA7220, so thats probably why it made a difference to the 1541.
I am interested in filtering and voltage level conversion at some point (especially if I get somewhere with oversampling).
Cheers,
Phil

 

[rfbrw]

I've also thought about replacing the SAA7220 with an FPGA / DSP so that I can experiment with different filter co-efficients and reconstitution / interpolation algorithms.

Somewhere on the net is a Japanese implementation of an improved, I assume, version of the SAA7220 using a FPGA.

LE is Latch Enable. Used by mono dacs and the TDA1541A in simultaneous mode.

 

[tubee]

Nos sound sucks. I'm really annoyed to hear all this voices how great imrovement it brought to the sound. Omitting SAA7220 really sucks. And every nos dac I tried (TDA1541 or TDA1543) realy sucks to me. If one has trully reveling set up (really hi-end audio) one can clearly hear no oversampling really does make it sound worse.

SAA7220P/B sounds better that P/A version of this chip.

Bartek [/B]

Yes i can confirm too the TDA1541 is not the best dac to run nonos. PCM56 is a lot better then, more transparent as the TDA, figured that out. But i can hear some of the same weird nonos facts with the PCM's as with the TDA. But the TDA has some qualities: certain recordings can sound more like a "Q-sound" recording, with pcm it is not that obvious anymore, more "flat" sounding.

Maybe i should try a highly rated PMD100 instead of a 7220. Didn't check that possibility in the datasheets yet. (input-output formats)
Have a 7220B lying in the attic, should power it up first.

 

[tubee]

I've also thought about replacing the SAA7220 with an FPGA / DSP so that I can experiment with different filter co-efficients and reconstitution / interpolation algorithms.
Jon [/B]

Imo a decent DSP with better algorithms instead of the 7220's FIR is the only way to go?? Or buy a Marantz CD7....

 

[tubee]

Somewhere on the net is a Japanese implementation of an improved, I assume, version of the SAA7220 using a FPGA.

This one?: http://homepage2.nifty.com/electrart/digital_amp/digitalamp.htm

 

[JonHarrison]

Imo a decent DSP with better algorithms instead of the 7220's FIR is the only way to go?? Or buy a Marantz CD7....

Ooh a CD7 that would be good. However, coming back to earth ...

... I did some quick calculations earlier and think a C55x running at 100MHz should do the job. FIR in an FPGA may be simplier but I want to use the flexibility of a DSP.

 

[guido]

Maybe i should try a highly rated PMD100 instead of a 7220. Didn't check that possibility in the datasheets yet. (input-output formats)
Have a 7220B lying in the attic, should power it up first.

PMD-100: recommended!

Possible; yes.

Gluelogic: quite a bit

 

[rfbrw]

Ooh a CD7 that would be good. However, coming back to earth ...

... I did some quick calculations earlier and think a C55x running at 100MHz should do the job. FIR in an FPGA may be simplier but I want to use the flexibility of a DSP.

Ummm, c55x flexible......

 

[Fin]

LE is Latch Enable. Used by mono dacs and the TDA1541A in simultaneous mode.

So - what Martin is saying is that in "Time multiplexed mode", the precise timing of WS might not be so significant and reclocking is not necessary. However, in "simultaneous mode", WS is now LE and the precise timing of this LE is significant and reclocking of this signal might be worthwhile.

 

[rfbrw]

So - what Martin is saying is that in "Time multiplexed mode", the precise timing of WS might not be so significant and reclocking is not necessary. However, in "simultaneous mode", WS is now LE and the precise timing of this LE is significant and reclocking of this signal might be worthwhile.

I see things the other way around.

 

[Fin]

I see things the other way around.

You would relcock WS if possible?

If so - would you also apply the suggested RC filtering of the signals between SAA7220 and TDA1541?

 

[rfbrw]

You would relcock WS if possible?

If so - would you also apply the suggested RC filtering of the signals between SAA7220 and TDA1541?

Can't say I've given it much thought but given the mode associated with LE and simultaneous data input, I don't see much point in reclocling DATA and BCK.
The data input mode shared by the SAA7220 and TDA1541,OTOH, would probably benefit from reclocking all three lines. As for RC filtering, you would have to be prepared to spend time optimising the values.

 

[Fin]

The data input mode shared by the SAA7220 and TDA1541,OTOH, would probably benefit from reclocking all three lines. As for RC filtering, you would have to be prepared to spend time optimising the values.

Thank you for your comments.

I will attempt the reclocking of all three lines as the layout in this CD player allows for it to be done without significantly increasing the size of the circuit.

For the RC filtering the options would probably be:

1. Use the previously suggested value of 1K and 10-20pF.

2. Retain the existing values of 390R (and possibly add 20-50pF to acheive the same overall value as above).