Tonight replaced the stock shielded sony output wiring with vampire tcss wire directly soldered to the rca output jacks, but for now I have retained the plugs and board sockets. Not sure whether this really made a difference or not, but subjectively it does sound slightly more open on the top end, very subtle and possibly not real difference, but no harm either.. 
As I have listened over the last two days I realize that this modification has substantially altered the performance of player - it sounds much better cold than it did, less impression of it "warming up" over a couple of hours of use.
Modding these players is considered controversial in some quarters and not even more so at least over on AA... My mod focus has been to address areas where I think some improvement can be expected from an engineering perspective. There has been substantial progress in audibly improving the sound of audio op-amps in addition to some very real parametric improvements. None of the op-amps in this player were particularly expensive in their heyday although they did represent good choices for the time. They've all been sonically superceded by much more recent designs, and frankly the AD712 in particular is mediocre by today's standards. I've kept the stock topology with only a couple of minor tweaks, and plan to continue with what I think is cost disregarded approach to optimizing the original design approach.
I will possibly have the opportunity to a/b this player against a stock player and I expect the differences will not be that subtle. I was a little worried in fact that I would "ruin" this player's sound by modding it and in fact did the work in such a way that I could quickly undo most of it if I didn't like the result. I won't be undoing it, and in fact will proceed further as time and money approach.
The media server is not getting much use right now...
As I have listened over the last two days I realize that this modification has substantially altered the performance of player - it sounds much better cold than it did, less impression of it "warming up" over a couple of hours of use.
Modding these players is considered controversial in some quarters and not even more so at least over on AA... My mod focus has been to address areas where I think some improvement can be expected from an engineering perspective. There has been substantial progress in audibly improving the sound of audio op-amps in addition to some very real parametric improvements. None of the op-amps in this player were particularly expensive in their heyday although they did represent good choices for the time. They've all been sonically superceded by much more recent designs, and frankly the AD712 in particular is mediocre by today's standards. I've kept the stock topology with only a couple of minor tweaks, and plan to continue with what I think is cost disregarded approach to optimizing the original design approach.
I will possibly have the opportunity to a/b this player against a stock player and I expect the differences will not be that subtle. I was a little worried in fact that I would "ruin" this player's sound by modding it and in fact did the work in such a way that I could quickly undo most of it if I didn't like the result. I won't be undoing it, and in fact will proceed further as time and money approach.
The media server is not getting much use right now...
I've heard a lot about removing the line filter chokes in these players lately, and also about inverting the ac line polarity.
I'm thinking about deleting those chokes shortly and may swap the phases on the board at that time.
I have the easy option of powering the analog and digital power transformers separately with a little rework as I have separate dedicated ac power for the analog and digital portions of my system. (This would require sheet metal work and strikes me as beyond the point of diminishing returns.)
Comments on anything above?
I'm still waiting for the 10K MK-132 resistors to arrive, and have decided to order some other resistors, mlcc caps for the supplies and some additional large BG caps also for supply bypassing.
I have come up with what I hope will be a low noise 5V supply design based on xetex ZXT849 pass transistors, LTC6240HV op-amp and LT1009 reference - seems to sym well, and should be more than an order of magnitude quieter than the lowest noise versions of the 7805 as well as having much lower Z out above dc. (Note gbw is at best 18MHz with this op-amp so performance will start to degrade significantly above a couple of MHz, however this is probably several orders of magnitude better than the 7805 as well. Caps with very good HF performance are a must on the output.) The pre-regulator is a fixed voltage LT1086-2.85 referenced to the output. I'm going to start work on versions for 7V and 9V shortly.
As I commented in an earlier post I have decided to limit the number of mod cycles and am doing mods in blocks rather than individually, unfortunately this does prevent me from determining what mods worked best. I made the decision to do this as quite a lot of disassembly is involved and I am concerned about keeping board handling to the minimum required to reduce the likelihood of cracked connections to FP, soic ics, esd damage, and cracked solder joints at the connectors.
The player is now getting at least 10hrs of use in an average week, and the mechanism has not missed a beat since the "gear fix" of some weeks ago, and ultimately gear replacement a few days ago. Incidentally all gears ordered have been received. If you have a player with a white gear in the gear "c" position, the new gear has more hub clearance and is black in color.
I'm thinking about deleting those chokes shortly and may swap the phases on the board at that time.
I have the easy option of powering the analog and digital power transformers separately with a little rework as I have separate dedicated ac power for the analog and digital portions of my system. (This would require sheet metal work and strikes me as beyond the point of diminishing returns.)
Comments on anything above?
I'm still waiting for the 10K MK-132 resistors to arrive, and have decided to order some other resistors, mlcc caps for the supplies and some additional large BG caps also for supply bypassing.
I have come up with what I hope will be a low noise 5V supply design based on xetex ZXT849 pass transistors, LTC6240HV op-amp and LT1009 reference - seems to sym well, and should be more than an order of magnitude quieter than the lowest noise versions of the 7805 as well as having much lower Z out above dc. (Note gbw is at best 18MHz with this op-amp so performance will start to degrade significantly above a couple of MHz, however this is probably several orders of magnitude better than the 7805 as well. Caps with very good HF performance are a must on the output.) The pre-regulator is a fixed voltage LT1086-2.85 referenced to the output. I'm going to start work on versions for 7V and 9V shortly.
As I commented in an earlier post I have decided to limit the number of mod cycles and am doing mods in blocks rather than individually, unfortunately this does prevent me from determining what mods worked best. I made the decision to do this as quite a lot of disassembly is involved and I am concerned about keeping board handling to the minimum required to reduce the likelihood of cracked connections to FP, soic ics, esd damage, and cracked solder joints at the connectors.
The player is now getting at least 10hrs of use in an average week, and the mechanism has not missed a beat since the "gear fix" of some weeks ago, and ultimately gear replacement a few days ago. Incidentally all gears ordered have been received. If you have a player with a white gear in the gear "c" position, the new gear has more hub clearance and is black in color.
Without a control, this, like all mods, is done in a vacuum, ...
...thus basicaly unquantifiable in terms of objective (or even subjective) quantification of results, for better or worse.
not that that has stopped people before, but simply relying on 'audio memory' as to before and after results is nearly useless IMNSHO. One really needs a control variable (e.g. a stock player) to even consider trying to quantify results. It will be interesting nonetheless, I suppose.
-Ed
...thus basicaly unquantifiable in terms of objective (or even subjective) quantification of results, for better or worse.
not that that has stopped people before, but simply relying on 'audio memory' as to before and after results is nearly useless IMNSHO. One really needs a control variable (e.g. a stock player) to even consider trying to quantify results. It will be interesting nonetheless, I suppose.
-Ed
Hi Ed,
In general I would agree with what you are saying. Auditory memory is quite unreliable, and I try to rely more on impressions described in careful detail before and after to assess whether anything I have done has merit. The best way of course is to compare a modified player to a stock one, the caveat being that the two need to have sounded more or less identical at the start for this to be meaningful. Sony interestingly made some running changes to the op-amp line up in these players, so it is not necessarily the case that any two factory players sound exactly the same unfortunately, and may be more likely to diverge over time due to component aging, particularly if one player is used on a more frequent basis than another. (Sony apparently did believe these op-amps all had specific sonic attributes otherwise they would have specified the same minimally electrically adequate device for all locations to keep bom costs down and they didn't - and neither did I in fact) Unfortunately it is not the case that most of us have more than one player, particularly in the case where the player in question is still relatively expensive.
In many cases you may design and build a device and spend months voicing it, you will depend heavily on notes taken based on subjective criteria along the way as well as measurements, as a hobbyist you probably will not have more than one device to compare to, and commercially I didn't either. I think general trends however are observable with care, and I try not to go to far in regards to what I think I hear.
I have chosen technically uncontroversial changes in a few key areas to see what effect they might have if any, on the player's performance. My focus has been largely on upgrading op-amps with devices I know from personal listening (even some a/b listening tests) and bench experience to strongly outperform the stock devices used in these players. I also spend my days designing high performance analog electronics so I do have a fundamental grasp of the tradeoffs made in this design (and they are there) as well as in my own work. Those are the only sorts of issues I'm trying to address, I'm not in pursuit of sonic nirvana..
The one other thing that enters the equation is that I am a relative cheap skate, I don't mind paying what a heap of money for this machine, but I won't spend 5X what it cost me to acquire to modify it, that's just crazy.
Hopefully next meet I can persuade someone to bring along an unmodded SCD- 777ES or SCD-1. It would be very interesting to hear what if any differences exist outside of my mind..
In general I would agree with what you are saying. Auditory memory is quite unreliable, and I try to rely more on impressions described in careful detail before and after to assess whether anything I have done has merit. The best way of course is to compare a modified player to a stock one, the caveat being that the two need to have sounded more or less identical at the start for this to be meaningful. Sony interestingly made some running changes to the op-amp line up in these players, so it is not necessarily the case that any two factory players sound exactly the same unfortunately, and may be more likely to diverge over time due to component aging, particularly if one player is used on a more frequent basis than another. (Sony apparently did believe these op-amps all had specific sonic attributes otherwise they would have specified the same minimally electrically adequate device for all locations to keep bom costs down and they didn't - and neither did I in fact) Unfortunately it is not the case that most of us have more than one player, particularly in the case where the player in question is still relatively expensive.
In many cases you may design and build a device and spend months voicing it, you will depend heavily on notes taken based on subjective criteria along the way as well as measurements, as a hobbyist you probably will not have more than one device to compare to, and commercially I didn't either. I think general trends however are observable with care, and I try not to go to far in regards to what I think I hear.
I have chosen technically uncontroversial changes in a few key areas to see what effect they might have if any, on the player's performance. My focus has been largely on upgrading op-amps with devices I know from personal listening (even some a/b listening tests) and bench experience to strongly outperform the stock devices used in these players. I also spend my days designing high performance analog electronics so I do have a fundamental grasp of the tradeoffs made in this design (and they are there) as well as in my own work. Those are the only sorts of issues I'm trying to address, I'm not in pursuit of sonic nirvana..
The one other thing that enters the equation is that I am a relative cheap skate, I don't mind paying what a heap of money for this machine, but I won't spend 5X what it cost me to acquire to modify it, that's just crazy.
Hopefully next meet I can persuade someone to bring along an unmodded SCD- 777ES or SCD-1. It would be very interesting to hear what if any differences exist outside of my mind..
I received the 10K resistors for the "mixing" diff amps last night, and had the opportunity to install them tonight. I very carefully matched them and ended up with matched pairs which matched to better than 0.074% theoretically giving a cmrr of about -62.6dBr which with the 1% resistors previously installed is close to a 30dB improvement with worst case resistors - they weren't matched, but they were better than 1% pair to pair.
Did I hear a large improvement? I was able to complete the mod in about 30 minutes, and the changes were fairly subtle particularly considering what these resistors cost so this part of the mod should probably be skipped. The Riken resistors probably sound a little sweeter and perhaps more neutral than most would realize. I think the previous mix was probably optimal, although I won't be reverting.
The player sounds very good, but until I can compare it to another one I am not sure how significantly it differs from the stock player. I do think it sounds more open, smoother and more detailed, but the power of suggestion is a strong thing particularly where the wallet is involved, and I am relying on auditory memory which is not particularly reliable..
Probably just as unreliable, but I have always preferred the back panel "normal" filter setting to "custom" - just the reverse of what I expected in fact. It sounds a hair "glassier" to me in the custom setting, and the mods have not changed this at all. It could be that something in my system does not like the additional ultrasonic energy in the output, or I could be imagining it.
I've included a picture of the interior of the player, not a very good one so no real detail is visible unfortunately. If you look very closely you can just see the sockets I installed for the LM4562 I/V converters which are in dip packages.
Did I hear a large improvement? I was able to complete the mod in about 30 minutes, and the changes were fairly subtle particularly considering what these resistors cost so this part of the mod should probably be skipped. The Riken resistors probably sound a little sweeter and perhaps more neutral than most would realize. I think the previous mix was probably optimal, although I won't be reverting.
The player sounds very good, but until I can compare it to another one I am not sure how significantly it differs from the stock player. I do think it sounds more open, smoother and more detailed, but the power of suggestion is a strong thing particularly where the wallet is involved, and I am relying on auditory memory which is not particularly reliable..
Probably just as unreliable, but I have always preferred the back panel "normal" filter setting to "custom" - just the reverse of what I expected in fact. It sounds a hair "glassier" to me in the custom setting, and the mods have not changed this at all. It could be that something in my system does not like the additional ultrasonic energy in the output, or I could be imagining it.
I've included a picture of the interior of the player, not a very good one so no real detail is visible unfortunately. If you look very closely you can just see the sockets I installed for the LM4562 I/V converters which are in dip packages.
Attachments
I've added a couple of pages to my site about the SCD-777ES/SCD-1 mods I've tried so far. Not too different than some of the things I've posted here - which in some cases form the basis for further notes and comments I made on my site.
Link here: www.kta-hifi.net/projects/SCD1/scd1.html
Link here: www.kta-hifi.net/projects/SCD1/scd1.html
I have done some additional mods, this time focused on the ac inlet board. I measured the ac leakage current and determined that reversing the transformer connections resulted in slightly lower leakage so I reversed the transformer connectors on the board.
I also removed the common mode choke for the analog supply transformer and jumpered it.
I noticed an odd thing as well - and that is that the fuse is only on the primary of the analog supply transformer. This is correct as indicated in the schematics, but I was surprised - I'm assuming that the unfused transformer has either a sufficiently high primary impedance that a short or partial short on one of the secondaries is not an issue, or they thought that using regulators with electronic protection was sufficient - not sure which.
I replaced the common mode cap on the analog side with a RIFA 0.1uF film with VDE and other safety approvals. I also added one across the inlet and from neutral to ground. I left the others in place as I figured they weren't hurting, and probably work better at high frequencies than these films do.
Improvements are very subtle, slightly more dynamic, tighter bottom end, slightly increased sound stage.. Or it could all be in my imagination. Cheap and easy to do, hence worthwhile. Others have claimed huge gains from similar changes..
I also removed the common mode choke for the analog supply transformer and jumpered it.
I noticed an odd thing as well - and that is that the fuse is only on the primary of the analog supply transformer. This is correct as indicated in the schematics, but I was surprised - I'm assuming that the unfused transformer has either a sufficiently high primary impedance that a short or partial short on one of the secondaries is not an issue, or they thought that using regulators with electronic protection was sufficient - not sure which.
I replaced the common mode cap on the analog side with a RIFA 0.1uF film with VDE and other safety approvals. I also added one across the inlet and from neutral to ground. I left the others in place as I figured they weren't hurting, and probably work better at high frequencies than these films do.
Improvements are very subtle, slightly more dynamic, tighter bottom end, slightly increased sound stage.. Or it could all be in my imagination. Cheap and easy to do, hence worthwhile. Others have claimed huge gains from similar changes..
I did recently the mod described here: http://www.audioasylum.com/forums/hirez/messages/93363.html
changing I/V conversion stage resitors and caps.
I ordered the same resistors as described in a mod: Caddock CC1512FC 10 ohm (they are available at $1/pc with 50pcs minimum order)
changing I/V conversion stage resitors and caps.
I ordered the same resistors as described in a mod: Caddock CC1512FC 10 ohm (they are available at $1/pc with 50pcs minimum order)
Attachments
I didn't like those Caddocks at all and after few hours of listening I removed them and replaced with Rikens which produce much better result (comparing to Caddocks).
Overall, Caddocks were certainly not an improvement to the original MELF resistors, contrary, it was quite a degradation. As to Rikens, they might sound better, but if so, only slightly better and I don't recommend messing with those resistors at all.
As to caps, they are quite influential (0.068uF). The originals are not bad and probably good match for MELFs. I tried best grade Panasonic film, ERO 1830 and 1837, as well V-Caps, none of them are really good and actually the sound is most natural without caps at all, but tends to be a bit too bright. I presently switch between V-Caps ( which are not that good here as one could expect) and ERO 1837 which although not as smooth as V-Caps sound more natural (but with certain edge).
You can also notice I changed feedback ressitors in op amp stage (I presently have LM4562 in there).
The Caddocks are available at $10 for set of 8 shipped, contact me if you are interested.
Overall, Caddocks were certainly not an improvement to the original MELF resistors, contrary, it was quite a degradation. As to Rikens, they might sound better, but if so, only slightly better and I don't recommend messing with those resistors at all.
As to caps, they are quite influential (0.068uF). The originals are not bad and probably good match for MELFs. I tried best grade Panasonic film, ERO 1830 and 1837, as well V-Caps, none of them are really good and actually the sound is most natural without caps at all, but tends to be a bit too bright. I presently switch between V-Caps ( which are not that good here as one could expect) and ERO 1837 which although not as smooth as V-Caps sound more natural (but with certain edge).
You can also notice I changed feedback ressitors in op amp stage (I presently have LM4562 in there).
The Caddocks are available at $10 for set of 8 shipped, contact me if you are interested.
Attachments
Hi Peter,
I didn't change out the stock 10 ohm melfs you mentioned, and based on your comments I'm glad I didn't.
I've been using Caddock MK-132 which do seem to be more transparent than the Riken through hole resistors they replaced.
I see you installed the resistors for the I/V directly on the op-amp pins, I noted the large loop - area present in the pcb and thought this would be a good idea, but still ended up using the stock locations.
I like the idea of going directly out as you are doing, but I don't have the ability to handle balanced signals in my system at this time. Internal transformers are a possibility at some point down the road.
I did replace the AD712 in the current sources with AD8512 and replaced the 510 ohm resistors with 402 and 110 ohm caddocks with a 100uF BF FK to ground. I suspect this is one of the best things I've done so far besides the change to LM4562 and AD8512 in the signal path.
High on my list is a change to the dc servo integrators to make them 2nd order.
I'm also still working on low voltage versions of the Jung super regulators, although they are now diverging from the Jung in that they have become ldo types. Output noise should be about 20dB better than the 78XX/79XX series regs they are meant to replace, and output impedance will be much lower as well. This is all in theory at the moment as I have yet to build and test them. (Simulations look good, but won't tell me whether they will start up into a load properly and won't give me accurate noise performance data at all.)
Overall the changes so far have made for a much more involving sound, the Sony at this point would have to rank as my best source.
I didn't change out the stock 10 ohm melfs you mentioned, and based on your comments I'm glad I didn't.
I've been using Caddock MK-132 which do seem to be more transparent than the Riken through hole resistors they replaced.
I see you installed the resistors for the I/V directly on the op-amp pins, I noted the large loop - area present in the pcb and thought this would be a good idea, but still ended up using the stock locations.
I like the idea of going directly out as you are doing, but I don't have the ability to handle balanced signals in my system at this time. Internal transformers are a possibility at some point down the road.
I did replace the AD712 in the current sources with AD8512 and replaced the 510 ohm resistors with 402 and 110 ohm caddocks with a 100uF BF FK to ground. I suspect this is one of the best things I've done so far besides the change to LM4562 and AD8512 in the signal path.
High on my list is a change to the dc servo integrators to make them 2nd order.
I'm also still working on low voltage versions of the Jung super regulators, although they are now diverging from the Jung in that they have become ldo types. Output noise should be about 20dB better than the 78XX/79XX series regs they are meant to replace, and output impedance will be much lower as well. This is all in theory at the moment as I have yet to build and test them. (Simulations look good, but won't tell me whether they will start up into a load properly and won't give me accurate noise performance data at all.)
Overall the changes so far have made for a much more involving sound, the Sony at this point would have to rank as my best source.
In terms of those 0.068uF I have reluctantly come to the conclusion that I would probably be leaving them alone for some time, one area where the larger film caps really fall down is in terms of their ESL, and even the longer leads required contribute to this - given the impedances involved at this circuit node this cap effects ultrasonic behavior strongly (not audio band) and so needs to be well behaved probably out to beyond the dsd sampling frequency in addition to needing to meet all the usual audiophile attributes.
I suggest taking an analog output FAR earlier in the signal path than the I/V opamps!
These are completely redundant and only add typical opamp colourations to the signal.
Look at the signal between the VC24 chip and the S-TACT, it is pure analog and a simple passive low-pass will turn it into beautiful audio. WE use 18dB/oct at 95kHz.
We have done over 500 Upgrades on these machines now using this concept with award winning results.
Regards, Allen
These are completely redundant and only add typical opamp colourations to the signal.
Look at the signal between the VC24 chip and the S-TACT, it is pure analog and a simple passive low-pass will turn it into beautiful audio. WE use 18dB/oct at 95kHz.
We have done over 500 Upgrades on these machines now using this concept with award winning results.
Regards, Allen
Thanks for suggestion, I would had certainly tried it, if I could figure out the connections from VC24 chip.
But, the output level here is probably much lower and I would need to use some active stage anyway as my line/amp combo offers 18dB only.
BTW, interesting discussion on AA: http://db.audioasylum.com/cgi/m.mpl?forum=hirez&n=228665&highlight=sony+allen+wright
But, the output level here is probably much lower and I would need to use some active stage anyway as my line/amp combo offers 18dB only.
BTW, interesting discussion on AA: http://db.audioasylum.com/cgi/m.mpl?forum=hirez&n=228665&highlight=sony+allen+wright
Attachments
BlkNotes said:
As far as I can tell the stock units sound identical. (I've heard several) The issue is that there are minor variations in the audio board op-amp line up and decoupling for the I/V converters and these may end up randomly in either player - and they do sound different.
The board set is identical with the exception that the scd-1 has one additional board to do the balanced output, which incidentally looks almost like an afterthought. The unbalanced output is designed for higher performance. (Hybrid transistor op-amp output vs two straight op-amps.) The transport is identical. There is one service manual for both. Almost everything is common to both players.
The power transformers in the scd-1 are encapsulated, but are otherwise interchangeable with the ones in the scd-777es.
The sheet metal is a little different, that is perhaps the most significant difference.
I am REAL pleased with my scd-777es, with the mods I've done so far it is the best digital source I have ever owned. Let the naysayers say what they will.
Re: differences
Hi Ed,
I quite forgot the 5th foot, and IIRC they are a little different on the scd-1 providing some additional isolation it is claimed.
Do you know whether the potted trannies also include additional shielding?
I will say the scd-777es is making everything else including the turntable sound a bit broken at the moment.. And I'm working on fixing that..
Ed Sawyer said:
Hi Ed,
I quite forgot the 5th foot, and IIRC they are a little different on the scd-1 providing some additional isolation it is claimed.
Do you know whether the potted trannies also include additional shielding?
I will say the scd-777es is making everything else including the turntable sound a bit broken at the moment.. And I'm working on fixing that..
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.