Parasound D/AC-1000 tweaking

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Member
Joined 2002
Paid Member
I just got this old DAC in house, and plan to tweak it a bit.

I already have "K" grade DAC chip's, will implement the mandatory receiver-chip filter update, then my way of tube-rolling; swap op-amps.

Maybe a new clock too...

Anybody with experience here? ideas?

Arne K
 

Attachments

  • para-1000.jpg
    para-1000.jpg
    91.4 KB · Views: 2,038
Cobra2 said:
The digital filter is the older SM5813...

Arne K


Any change to that will require more extensive surgery or the construction of a daughter board. The small size of the DF1704 and the SM5847 should make the construction of a daughter board fairly easy. You could also find out if there really is a difference between the DF1700P and the SM5813, the DF1700P being a rebadged SM5813. Lastly you could become a NOSser:darkside: and ditch the filter.

ray.
 
Re: Adding clock

Cobra2 said:
It seems that the clock is recovered by the CS8412, and then sent to the dig. filter.
Can I just disconnect the output from the CS8412, and "inject" an external clock into the SM5813?

Arne K

What you should be able to do is to place the oscillator at or near the DAC chips (or I guess in your case the digital filter), and feed the same or a compatible division of this master clock to the source.

This way you are slaving the source off from the DAC. The input receiver should unless you are really unlucky find the digital signal from the source and set it up in a synchronized manner.

I believe there was info on the LC Audio site about this sort of connection some years ago. Try both the Danish (www.lcaudio.dk) and the global (www.lcaudio.com) site as I have sometimes been able to get more information out of accessing both sites :)

Feeding a master clock into the input receiver will not work IMVHO since the input receiver can only be slave of one item - and two clocks running separately (one in source, the other in target) will never run close enough over time to enable close synchronization.

Petter
 
Re: Re: Adding clock

Petter said:


Feeding a master clock into the input receiver will not work IMVHO since the input receiver can only be slave of one item - and two clocks running separately (one in source, the other in target) will never run close enough over time to enable close synchronization.

Petter


MCLK on the CS8412 is strictly an output signal whatever mode you choose. What can be accepted as input signals are SCLK and FSYNC/LRCLK. In this mode data is double buffered to allow a degree of slippage between the incoming SPDIF data and the incoming SCLK and FSYNC which now clock data out.

ray
 
True, let me be a bit more precise:

The goal should be to have an accurate master clock at the digital filter and/or DAC chip.

The means I described involve making sure the digital source ("CD Player" for example) is effectively in synch with a master oscillator in the DAC.

Thus, it becomes necessary to disregard (cut the trace) from the regenerated MCLK of the input receiver in an appropriate position (depending on implementation, typically safets at or near DAC/DF). Thus, you are left with regenerated BCLK and L/RCKL and SDATA. Most devices are much less susceptible to jitter on these minor clocks than the master clock for obvious reasons.

... and then, unless you are truly unlucky, you should get automagic synchronization.

Petter
 
The original post was about modifying a DAC not a source. While you suggestion sounds nice in theory, in practice it effectively ties the DAC to the CD player, not the most flexible approach.
The clocks you dismiss as minor are everything to the PCM63 and its ilk, hence the suggestion to derive new ones to drive the CS8412 and the SM5813 rather than use the clocks derived from the SPDIF datastream, bearing they have to pass through the SM5813 before reaching the PCM63.

ray
 
rfbrw said:

The clocks you dismiss as minor are everything to the PCM63 and its ilk, hence the suggestion to derive new ones to drive the CS8412 and the SM5813 rather than use the clocks derived from the SPDIF datastream, bearing they have to pass through the SM5813 before reaching the PCM63.

ray

Agree about flexibility

The "minor" clocks could be double-bufferend and passed with the MCLK to clean up jitter.

I don't think "my" recovered "minor" clocks will be any worse than other simple alternatives.

The CS8412 can as noted by others not be driven by external clock. One will have major issues with synchronization of that were possible. An crystal controlled oscillator or a memory based PLL seems more appropriate for such duty.

Petter
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.