Toshiba SD-3990: audio out bypass -- help needed with mods

[headfone]

Miniaturization* and circuit/topology changes (since the release of models 3950 --> 80; I have NO experience with any prev. model so I can't make concrete statements) -- plus the unavailability of a service manual for this model** -- has made upgrading/tweaking the SD-3990 a challenge.

For starters, I would like to perform the Swenson mod (http://johnswenson1.home.comcast.net...3960_mods.html ; also outlined in this excellent article: http://www.spectrum.ieee.org/jun05/2651) which bypasses the onboard output stage.


Uncertain assumptions I am making for the SD-3990:

- The two caps that need to be bypassed will be located adjacent to each other.

- The two caps that need to be bypassed will have values identical to those used in previous models. (E.g. c910 and c912 at 47Uf/25(?)V, in the 3950).

- The big Zoran processor is doing the all DACing (no separate DAC, like the BB/TI used on the 3950).

- The caps in question will be coming off the DAC.

Questions:

On the SD-3990, where exactly are the two caps that need to be bypassed?

On the SD-3950, the Swenson/Vinnie mods list c910 and c912 as the ones that need bypassing (both their values are 47uF).

As noted above, a lot of components -- caps included -- have been ultra-miniaturized (and are surface-level) on the 3990 main audio/video board. However, methinks that since the analog output stage is usually beefier by design, these caps are still standard sized (??)

The photos below show the main board of the SD-3990 (the caps with the value of 47uF are denoted with a red X):

Reference threads:

http://www5.head-fi.org/forums/showt...7&page=1&pp=25
http://www5.head-fi.org/forums/showt...7&page=2&pp=20
http://www6.head-fi.org/forums/showthread.php?t=156289
http://www6.head-fi.org/forums/showthread.php?p=1129145

* This also means more surface-level components (i.e. on the board's top layer; look at thge middle image above (bottom view of PCB) to see what I mean).

** The official Toshiba service center informed me it's cheaper for Toshiba to replace the product than to offer a service manual; hence, none will be made avail. for the 3990.

 

[Nerull]

I have the exact same player, and was wondering the very same things aforementioed above. Perhaps I'll trace out the signal path tonight or something, I have time

.

I also wonder of the RC filter values so I can replace the coupling capacitors (where ever they may be hiding) with some foil ones I have. The opamp too, but lifting a pad would suck (could always bypass though).

~Tom

 

[headfone]

Quote:

Originally Posted by Nerull I also wonder of the RC filter values so I can replace the coupling capacitors (where ever they may be hiding) with some foil ones I have. The opamp too, but lifting a pad would suck (could always bypass though).

I'm also stuck as far as where the audio op-amp(s) is/are. Is it, perhaps, the one closest to analog out jacks (see image in orig mesage)? In a brief search, I could find no references for it ("N6EZBC" ?? and "UTC4558E" ?? can barely read it!) so I have no clue.

 

[Nerull]

It's definately that 4558, that's a dual, cheap-as-dirt opamp that sounds like crap. It's the 8-SOIC located nearest to the RCA outs. Another thing are those SOT23-3 packages (look up a picture of the package, you'll see them then); which I believe are the muting transistors (seeing as there aren't anymore transistors in the local vincinity). Removing those will also give an improvement in sound.

I still have to trace out the signal path though (gotta find those coupling capacitors and god knows what else in the signal path...)

~Tom

 

[aphex944]

I had one of these players, because I thought it had the exact same audio section as the 3950. I replaced the opamp with an LM833, and while the audio improved significantly(no more sibilance), it still fell way short of a high quality component. I've since returned mine to Best Buy, and bought a 3950 on eBay. Pretty disappointing though, after hearing this player right out of the box I thought it had great potential. From what I understand from my electronics friend who did the work, that opamp is used as a buffer. There are no output coupling caps.

 

[headfone]

Quote:

Originally Posted by aphex944 I had one of these players, because I thought it had the exact same audio section as the 3950. I replaced the opamp with an LM833, and while the audio improved significantly(no more sibilance), it still fell way short of a high quality component. I've since returned mine to Best Buy, and bought a 3950 on eBay. Pretty disappointing though, after hearing this player right out of the box I thought it had great potential. From what I understand from my electronics friend who did the work, that opamp is used as a buffer. There are no output coupling caps.

FWIW, after mechanically damping the 3990, I found it to sound better than several "higher-end" transport/DAC separates. Haven't really touched the power supply yet, but some parts for it are coming in this week so I'll report back.

As far as the output coupling caps go...can you ask your friend how he confirmed this (e.g. resistance test between opamp output and output jack -- I'd do the test myself but I'm traveling right now so I don't have access to player)? If there is some ESR (or other source resistance from components, like the SOT23 transistors Nerull mentioned) revealed by this test, then there may still be some coupling caps, but they might now be in the ultra-miniature surface-mount variety, like the opamp and so many other components on this model (see orig message).

Not sure about how an LM833 sounds compared to, say, an opa627 or AD843 or other well-rated audio opamps.

 

[headfone]

Quote:

Originally Posted by Nerull ... which I believe are the muting transistors (seeing as there aren't anymore transistors in the local vincinity). Removing those will also give an improvement in sound.

Tom/anyone:

Are you saying one can simply remove (i.e. desolder and extract with tweezers) these muting transistors -- i.e. no need to shunt the contact pads (where the muting transistors used to be) w/wire?

 

[headfone]

Quote:

Originally Posted by aphex944 I replaced the opamp with an LM833, and while the audio improved significantly(no more sibilance), it still fell way short of a high quality component.

Another trick would be to replace the dual lo-fi 4558 w/a Brown Dog configuration using two single OPA627AU's.

 

[Nerull]

Yes, I believe all you do need to do is remove them and not have to do anything. This is what I did to my old CD player, and it worked just fine afterwards. Again, I'll have to check the tracing and whatnot (just to make sure it is infact a muting transistor); I've been a very busy Tom as of late

.

~Tom

 

[headfone]

Quote:

Originally Posted by Nerull Yes, I believe all you do need to do is remove them and not have to do anything. This is what I did to my old CD player, and it worked just fine afterwards. Again, I'll have to check the tracing and whatnot (just to make sure it is infact a muting transistor); I've been a very busy Tom as of late .

Thx for your reply and don't hurry for my sake

Here's a higher-rez view of the analog output section of the 3990:

For an even clearer (uncompressed) image, click here.

 

[headfone]

Quote:

Originally Posted by headfone Another trick would be to replace the dual lo-fi 4558 w/a Brown Dog configuration using two single OPA627AU's.

Probably old news to some but I just learned that a popular modifying outfit is using a Burr-Brown/TI BUF-634 opamp to replace the stock 4558.

BTW, anyone know where the DAC (and, specifically, the DAC chip's output) is on the 3990? Is it fed (connected) directly to the input of the opamp or are there some intervening components?

 

[headfone]

Quote:

Originally Posted by Nerull Yes, I believe all you do need to do is remove them and not have to do anything. This is what I did to my old CD player, and it worked just fine afterwards. Again, I'll have to check the tracing and whatnot (just to make sure it is infact a muting transistor); I've been a very busy Tom as of late . ~Tom

I did remove Q317 and Q318 today (see image in last msg) which yielded a modest but noticible improvement in sound fidelity (a bit more "focus", perhaps).

 

[aphex944]

Quote:

Originally Posted by headfone BTW, anyone know where the DAC (and, specifically, the DAC chip's output) is on the 3990? Is it fed (connected) directly to the input of the opamp or are there some intervening components?

The DAC is built into the Zoran Vaddis chip.

 

[headfone]

Quote:

Originally Posted by aphex944 The DAC is built into the Zoran Vaddis chip.

I was gonna say "I was afraid of that" but I don't honestly know good the Zoran DAC is. The last model, 3980, also had a (the same??) Zoran chip but, IIRC, ithe 3980 used an another (additional) DAC. It would be interesting to hear from anyone that has experience with both the 3980 and 3990.

 

[headfone]

Quote:

Originally Posted by headfone I did remove Q317 and Q318 today (see image in last msg) which yielded a modest but noticible improvement in sound fidelity (a bit more "focus", perhaps).

Just curious -- when exactly are "muting" or "mute" transistors used in a CDP or DAC's analog output?