Generically, I'm referring to the ubiquitous "classic" Philips trio, 7210, 7220 and 1541, because this is where most of my albeit limited experience lies. However, the queries below could effectively apply to any digital-section IC combination (or layout), even NOS.
Specifically, I've noted certain DIY and some "high-end" designs use inter-IC resistors (10 to 150 ohm) and myriad layout techniques (e.g. varying proximity of ICs, and location of each IC's respective PS regulator).
Commercial/mass-market Philips-based models (e.g. Magnavox, Grundig, etc.) did not use -- TTBOMK -- inter-IC series resistors (i.e., on signal lines between BCK, WS, Data pins). Also, inter-IC distance was, in commercial/mass-market units, somewhat arbitrary.
Should these ICs be as close as possible? (The use of inter-IC signal resistors create cubicles, as discussed in Tent's paper on the topic. Not sure whether that's always good thing, tho').
What's your experience with respect to how the following affect sound quality or measurements:
- inter-IC signal signal resistors (series) -- use of and R values
- inter-IC distance (e.g. electron-travel distance between 7210, 7220 and 1541) -- as close as possible, or otherwise?
- proximity of IC's power regulator
P.S. Yes, I've looked at the datasheets, but they (at least those from Philips) are not specific on these points.
Specifically, I've noted certain DIY and some "high-end" designs use inter-IC resistors (10 to 150 ohm) and myriad layout techniques (e.g. varying proximity of ICs, and location of each IC's respective PS regulator).
Commercial/mass-market Philips-based models (e.g. Magnavox, Grundig, etc.) did not use -- TTBOMK -- inter-IC series resistors (i.e., on signal lines between BCK, WS, Data pins). Also, inter-IC distance was, in commercial/mass-market units, somewhat arbitrary.
Should these ICs be as close as possible? (The use of inter-IC signal resistors create cubicles, as discussed in Tent's paper on the topic. Not sure whether that's always good thing, tho').
What's your experience with respect to how the following affect sound quality or measurements:
- inter-IC signal signal resistors (series) -- use of and R values
- inter-IC distance (e.g. electron-travel distance between 7210, 7220 and 1541) -- as close as possible, or otherwise?
- proximity of IC's power regulator
P.S. Yes, I've looked at the datasheets, but they (at least those from Philips) are not specific on these points.
Series resistors are used to damp reactances that may be present and cause over/undershoot and could cause potential damage or instability in the IC affected. The best way to determine this is with a suitably bandwidthed oscilloscope. Of course you need to account for the effect of probe reactances when determining optimum values.
Distance of ICs from each other is not a concern generally in the constraints one normally encounters in common designs for this type of logic IC. Shorter connections are better to preserve rise time of signals. For extremley fast rise times transmission line techniques are used. This will not be an issue with the ICs you talk about.
Same with power regulators. You should concentrate more on RF bypassing and ground returns. If the regulator does have a long trace to the components then look at the possible tank circuit created and try to damp it well.
National Semiconductor had a good application note available on alot of what you are asking.
Distance of ICs from each other is not a concern generally in the constraints one normally encounters in common designs for this type of logic IC. Shorter connections are better to preserve rise time of signals. For extremley fast rise times transmission line techniques are used. This will not be an issue with the ICs you talk about.
Same with power regulators. You should concentrate more on RF bypassing and ground returns. If the regulator does have a long trace to the components then look at the possible tank circuit created and try to damp it well.
National Semiconductor had a good application note available on alot of what you are asking.
Agreed: 'scope-tweaking this is the only way. A lot of datasheets (e.g. PMD100 dig. filter, etc.) show the series R's in their "suggested" ckts; OTOH, major-manufacturer Philips-based CDPs (don't know about the "high-end" stuff that use Philips components, like Rotel) don't -- IME, anyway -- use these R's.CraigBuckingham said:
I’m not sure what sonic correlates can be attributed to tight rise-time control (perhaps “timing” and “pace”? See comments below*).
This image, posted by a member on another forum, shows the 7220 DF and the 1541 DAC I2S pins almost touching:
An externally hosted image should be here but it was not working when we last tested it.
Yeah, I hear you on the bypassing issue. Tho’ I think I’ve got a lot of that covered.
BTW: I use an external (remote) PSU, and bring power in this way for several sections such as analog IV and certain digital ICs. Still haven't nailed the best strategy here for particular sections: raw DC (remote) --> regulator ckt (local) vs. complete regulation (remote). I think the former is better. But I’ve heard a DIYer note something like if one can afford the additional voltage overhead, the latter may be desirable (not sure what this means). Any suggestions/comments here are appreciated.
* Characteristics I, personally, hold paramount (above other characteristics -- perhaps even as "trade offs") include timing, rhythm, pace, and tight bass slam. These are normally strong correlates of better oversampling designs.
hollowman,
I know you asked about sonic performance but I concentrated more on proper engineering.
I am less sure about sonic benefits of all of this, except that keeping RF out of audio circuits seems to me to offer sonic benefits.
I know I went through an exhaustive design and optimisation of an external power supply (PS) for a custom DAC design quite a few years ago and in was unbelievably esoteric only to have some sealed lead acid batteries (SLA) easily knock it off the high mantle it had progressed too. We ended up converting the PS to a battery charger and housing the SLA batteries in it.
I know you asked about sonic performance but I concentrated more on proper engineering.
I am less sure about sonic benefits of all of this, except that keeping RF out of audio circuits seems to me to offer sonic benefits.
I know I went through an exhaustive design and optimisation of an external power supply (PS) for a custom DAC design quite a few years ago and in was unbelievably esoteric only to have some sealed lead acid batteries (SLA) easily knock it off the high mantle it had progressed too. We ended up converting the PS to a battery charger and housing the SLA batteries in it.
- Status
- Not open for further replies.