Voodoochile
Supafly & The Funky Pimps
- Joined
- Sep 13, 2002
- Posts
- 4,883
- Likes
- 14
Figured it's time to post up some details on the light mods to my Marantz DV7600. This is the same source I had at the NYC meet recently; it has PSU, coupling, and chassis damping mods at that time. This weekend I reclocked it at long last.
Short history: Used to own a NAD C541i; got a Toshi 3960 to try out some modding on; picked up the Marantz DV7600 universal player; acquired an Electrocompaniet ECD1 DAC; sold the 3600; sold the C541i; reluctantly sold the ECD1. Footnote: I loved the ECD1, did not wish to part with it, miss it, etc. It was not the end-all DAC, but it sounded beautiful. Hated to see it go. I used to run the DV7600 directly to my amp for SACD, and fed the SPDIF to the ECD1 for redbook.
Now I've had the DV7600 since new, and while it's always been a nice sounding piece, and certainly very flexible WRT formats, the sound was a bit on the heavy side. Nice bass, clean highs, but the highs were a little overshadowed by the plush low end. Not a function of the Senns, as it was consistent with Grado and AKG as well. The DV7600 shares much componentry with the ECD1, which is good. They both use the same DAC, they both utilize a discrete transistor output stage using the same transistors. The details between the two are what separates them, though. The ECD1 has a much better PSU and better coupling, amongst other things- that drive the cost of the ECD1 up, along with it's performance.
The DV7600 uses the Cirrus CS4398 DAC, CS494003 DSP, "HDAM" discrete transistor output stage. The PSU, while switchmode, uses hexfreds before the transformer and high speed soft recovery rectifiers in the audio output PSU. It has a separate PSU channel for audio, transport, and video boards, and includes line filtering of the mains. The discrete output stage is built with Toshiba 2SK369 jfets for the first stage, and complimentary output with Toshiba 2SK170 and 2SJ74 pairs... times six channels. If that rings a bell, look at the Gilmore Dynamic design. It's not the same, but it's very nice for a commercial design. Yes, there are also muting transistors, no there are no integrated op-amps. FWIW. It is certainly interesting to see so many discrete transistors in a contemporary consumer product.
Pics are terrible, but I was afraid if I waited until I got better ones, it would be another month. Overall picture of the DV7600, with some areas highlighted for reference:
Section A is the PSU, B is the DAC and audio board, C is the new clock board. In between A and B is the digital output board, below B is the video output board. It's a pretty full chassis. Most of the chassis has been damped with dynamat. It's heavier near the PSU, and the transport frame is also damped. It should be noted that Marantz added foam blocks to the transport that couple it against the cover, and also includes a dense foam block between the SMPS transformer and the side of the chassis. It's a minor effort on their part, but better than nothing. I added 400 square inches total- about 2/3 of that is on the top and bottom chassis plates.
Section 1 of the psu:
Replaced the generic main filter cap with a Black Gate VK series cap. I'll admit I did this because I had the cap on hand, and the value and pitch were correct for the unit. Note the mains filter Marantz included on the board.
Sections 2 and 3 of the psu:
In section 2 you see the pre-reg filter caps for the audio segment of the PSU. Replaced the stock generic 330uF caps with a pair of 470uF Cerafines. Although it's hard to see behind the larger rectifiers, I also bypassed the smaller rectifiers for this segment with .01uF polypropylene caps (small green boxes). I was going to replace the rectifiers, but those from Marantz are already very fast and quiet, so no need.
In section 3, the two caps to the rear are the post-reg filter caps. I replaced the stock 47uF generics with 33uF Black Gates, bypassed with non-polar Black Gate 0.1uF NX HiQ caps. The audio PSU ripple and noise was low already, but it was reduced by almost 40% versus stock.
Section 4 and 5 of the output board:
Section 5 is simply where I tapped the +/-8vdc for the add-on clock.
Section 4 shows the output coupling caps. Stock had 330uF generic caps. The rearmost pair are for the L/R channels. I tried several different caps and combinations here, including various polypro caps, nonpolar electros, polystyrene, PIO, and various combinations thereof. Cerafine 300uF caps were a clear improvement, and when bypassed with 1uF Auricaps, sounded better than anything else. I had high hopes for no electrolytics, or at least non-polars. The bass suffers without big electrolytics in this circuit, and there is too much offset to go without anything. My amp has no blocking caps.
This is the point the unit was at when it went to the NYC meet. The coupling caps made the most notable/audible difference, cleaning up the high end nicely while maintaining a good low end presence. The damping also seemed to tighten things up somewhat. It's been a pretty smooth running machine anyway, but it's really dead quiet now.
The final tweak (final for now anyway) was to reclock it. I've heard that this is the single best tweak for most players. In a CDP with opamp output, swapping the cheap op-amp with a nice one is a very good tweak, and is very inexpensive to change, so in that realm, I'd say that the OPA swap would be the best first tweak. This machine already had a very nice discrete output stage. There are a lot of choices out there, including Guido Tent's clock, the LC clock, the Hagclock, and a few others. I've built the Hagerman Cornet 2 phonostage already, and have a lot of respect for Jim Hagerman and his design philosophy. He's a pretty no-nonsense guy from an engineering standpoint and the gear stands up very well to scrutiny. Just look at his Bugle phono stage for example. It's difficult to top at 5 times the price. And the Trumpet stage is evidently stellar- I will never know, though! I love the Cornet 2. Finally, the Hagclock was the least expensive of the lot. The output form looks clean and stable, and it includes on-board super regulators.
Section 6 shows where the crystal and clock in/out caps were removed and the original from the Hagerman oscillator feeds in. It's a twisted pair line, which is fine for such a short run of less than 3".
Board C is the hagclock itself, mounted under the audio output board, right off the end of the video processor board.
I had to re-solder the oscillator feed into the hole where the original crystal's leg was mounted, rather than off the smd cap's pad, then it came right up. This player uses a somewhat odd 12.288MHz clock, and Jim was nice enough to build it with that custom frequency. Typically I have bought the board only from him, but in this case, the cost of parts+board was nearly the same as buying it built- so that was the simple way to go.
I checked the output on my frequency counter, and it was right on the money out beyond four decimal places. Looked at the output on the scope and it was beautiful. Considerably more stable than the original oscillator provided, on both axis. There appeared to be a fair amount of jitter, and some ringing on the original clock, which is no surprise. This and almost all CD players use a Pierce oscillator, which works- but that's about it. It's not especially precise, and while everything will function with jitter, in this application jitter can cause all sorts of issues. Manufacturers implement the Pierce oscillator because it uses only four or five cheap parts. Big surprise there, I know. I'm not going to dive deep into a discussion of jitter here, but suffice it to say that jitter is not your friend, and most players have it.
Here is a final picture of the overall output stage, just because it find it interesting to see:
Output caps and muting networks on the left, then the 2SK170/2SJ74 output transistors, then the 2SK369 input transistors and their bypass caps, on the far right the three chips are the DACs. It's a lot of JFETS for sure. Marantz added a nice metal shield below the PCB over the entire output stage, shielding it from the video board below. That was a task to remove to gain access to the legs of the original output caps. I was sure to replace it when finished.
I did also try taking the output from prior to the muting transistors, but the noise when operating any of the transport controls was not subtle, and the SQ improvement was not apparent. So those are still in-circuit. I may upgrade the output jacks if I decide to leave the coupling as it is now.
I'll submit a separate thread on more detailed listening impressions; this one was to focus more on the tweaking and modding. But it is a lot nicer sounding. The recoupling made a very nice improvement, and reclocking it really seems to have cleaned it up even further. The bass is not stronger, but it gives you the impression that it is because it is tighter and somewhat faster now. Some tracks show nice evidence of this. The high end is what was most improved. The placement of instruments is clearer now, and more subtle details are apparent. It seems to have a cleaner sound, but not analytical per se. It's still very lively and involving. I'm looking forward to further listening, and will report back after that time.
I have heard that this machine is available for pretty low cash these days, and I encourage anyone to pick one up. It's really very nice out of the box, playing both CD and SACD. It's got a great output stage, and with a few caps and an inexpensive clock upgrade makes a pretty formidable 1-box source. It also makes a very solid transport with an outboard DAC, but it's a waste of a pretty sweet discrete output stage.
For reference, here are some posts on the aforementioned Toshiba 3960 and NAD C541i machines and their tweaking.
Short history: Used to own a NAD C541i; got a Toshi 3960 to try out some modding on; picked up the Marantz DV7600 universal player; acquired an Electrocompaniet ECD1 DAC; sold the 3600; sold the C541i; reluctantly sold the ECD1. Footnote: I loved the ECD1, did not wish to part with it, miss it, etc. It was not the end-all DAC, but it sounded beautiful. Hated to see it go. I used to run the DV7600 directly to my amp for SACD, and fed the SPDIF to the ECD1 for redbook.
Now I've had the DV7600 since new, and while it's always been a nice sounding piece, and certainly very flexible WRT formats, the sound was a bit on the heavy side. Nice bass, clean highs, but the highs were a little overshadowed by the plush low end. Not a function of the Senns, as it was consistent with Grado and AKG as well. The DV7600 shares much componentry with the ECD1, which is good. They both use the same DAC, they both utilize a discrete transistor output stage using the same transistors. The details between the two are what separates them, though. The ECD1 has a much better PSU and better coupling, amongst other things- that drive the cost of the ECD1 up, along with it's performance.
The DV7600 uses the Cirrus CS4398 DAC, CS494003 DSP, "HDAM" discrete transistor output stage. The PSU, while switchmode, uses hexfreds before the transformer and high speed soft recovery rectifiers in the audio output PSU. It has a separate PSU channel for audio, transport, and video boards, and includes line filtering of the mains. The discrete output stage is built with Toshiba 2SK369 jfets for the first stage, and complimentary output with Toshiba 2SK170 and 2SJ74 pairs... times six channels. If that rings a bell, look at the Gilmore Dynamic design. It's not the same, but it's very nice for a commercial design. Yes, there are also muting transistors, no there are no integrated op-amps. FWIW. It is certainly interesting to see so many discrete transistors in a contemporary consumer product.
Pics are terrible, but I was afraid if I waited until I got better ones, it would be another month. Overall picture of the DV7600, with some areas highlighted for reference:
Section A is the PSU, B is the DAC and audio board, C is the new clock board. In between A and B is the digital output board, below B is the video output board. It's a pretty full chassis. Most of the chassis has been damped with dynamat. It's heavier near the PSU, and the transport frame is also damped. It should be noted that Marantz added foam blocks to the transport that couple it against the cover, and also includes a dense foam block between the SMPS transformer and the side of the chassis. It's a minor effort on their part, but better than nothing. I added 400 square inches total- about 2/3 of that is on the top and bottom chassis plates.
Section 1 of the psu:
Replaced the generic main filter cap with a Black Gate VK series cap. I'll admit I did this because I had the cap on hand, and the value and pitch were correct for the unit. Note the mains filter Marantz included on the board.
Sections 2 and 3 of the psu:
In section 2 you see the pre-reg filter caps for the audio segment of the PSU. Replaced the stock generic 330uF caps with a pair of 470uF Cerafines. Although it's hard to see behind the larger rectifiers, I also bypassed the smaller rectifiers for this segment with .01uF polypropylene caps (small green boxes). I was going to replace the rectifiers, but those from Marantz are already very fast and quiet, so no need.
In section 3, the two caps to the rear are the post-reg filter caps. I replaced the stock 47uF generics with 33uF Black Gates, bypassed with non-polar Black Gate 0.1uF NX HiQ caps. The audio PSU ripple and noise was low already, but it was reduced by almost 40% versus stock.
Section 4 and 5 of the output board:
Section 5 is simply where I tapped the +/-8vdc for the add-on clock.
Section 4 shows the output coupling caps. Stock had 330uF generic caps. The rearmost pair are for the L/R channels. I tried several different caps and combinations here, including various polypro caps, nonpolar electros, polystyrene, PIO, and various combinations thereof. Cerafine 300uF caps were a clear improvement, and when bypassed with 1uF Auricaps, sounded better than anything else. I had high hopes for no electrolytics, or at least non-polars. The bass suffers without big electrolytics in this circuit, and there is too much offset to go without anything. My amp has no blocking caps.
This is the point the unit was at when it went to the NYC meet. The coupling caps made the most notable/audible difference, cleaning up the high end nicely while maintaining a good low end presence. The damping also seemed to tighten things up somewhat. It's been a pretty smooth running machine anyway, but it's really dead quiet now.
The final tweak (final for now anyway) was to reclock it. I've heard that this is the single best tweak for most players. In a CDP with opamp output, swapping the cheap op-amp with a nice one is a very good tweak, and is very inexpensive to change, so in that realm, I'd say that the OPA swap would be the best first tweak. This machine already had a very nice discrete output stage. There are a lot of choices out there, including Guido Tent's clock, the LC clock, the Hagclock, and a few others. I've built the Hagerman Cornet 2 phonostage already, and have a lot of respect for Jim Hagerman and his design philosophy. He's a pretty no-nonsense guy from an engineering standpoint and the gear stands up very well to scrutiny. Just look at his Bugle phono stage for example. It's difficult to top at 5 times the price. And the Trumpet stage is evidently stellar- I will never know, though! I love the Cornet 2. Finally, the Hagclock was the least expensive of the lot. The output form looks clean and stable, and it includes on-board super regulators.
Section 6 shows where the crystal and clock in/out caps were removed and the original from the Hagerman oscillator feeds in. It's a twisted pair line, which is fine for such a short run of less than 3".
Board C is the hagclock itself, mounted under the audio output board, right off the end of the video processor board.
I had to re-solder the oscillator feed into the hole where the original crystal's leg was mounted, rather than off the smd cap's pad, then it came right up. This player uses a somewhat odd 12.288MHz clock, and Jim was nice enough to build it with that custom frequency. Typically I have bought the board only from him, but in this case, the cost of parts+board was nearly the same as buying it built- so that was the simple way to go.
I checked the output on my frequency counter, and it was right on the money out beyond four decimal places. Looked at the output on the scope and it was beautiful. Considerably more stable than the original oscillator provided, on both axis. There appeared to be a fair amount of jitter, and some ringing on the original clock, which is no surprise. This and almost all CD players use a Pierce oscillator, which works- but that's about it. It's not especially precise, and while everything will function with jitter, in this application jitter can cause all sorts of issues. Manufacturers implement the Pierce oscillator because it uses only four or five cheap parts. Big surprise there, I know. I'm not going to dive deep into a discussion of jitter here, but suffice it to say that jitter is not your friend, and most players have it.
Here is a final picture of the overall output stage, just because it find it interesting to see:
Output caps and muting networks on the left, then the 2SK170/2SJ74 output transistors, then the 2SK369 input transistors and their bypass caps, on the far right the three chips are the DACs. It's a lot of JFETS for sure. Marantz added a nice metal shield below the PCB over the entire output stage, shielding it from the video board below. That was a task to remove to gain access to the legs of the original output caps. I was sure to replace it when finished.
I did also try taking the output from prior to the muting transistors, but the noise when operating any of the transport controls was not subtle, and the SQ improvement was not apparent. So those are still in-circuit. I may upgrade the output jacks if I decide to leave the coupling as it is now.
I'll submit a separate thread on more detailed listening impressions; this one was to focus more on the tweaking and modding. But it is a lot nicer sounding. The recoupling made a very nice improvement, and reclocking it really seems to have cleaned it up even further. The bass is not stronger, but it gives you the impression that it is because it is tighter and somewhat faster now. Some tracks show nice evidence of this. The high end is what was most improved. The placement of instruments is clearer now, and more subtle details are apparent. It seems to have a cleaner sound, but not analytical per se. It's still very lively and involving. I'm looking forward to further listening, and will report back after that time.
I have heard that this machine is available for pretty low cash these days, and I encourage anyone to pick one up. It's really very nice out of the box, playing both CD and SACD. It's got a great output stage, and with a few caps and an inexpensive clock upgrade makes a pretty formidable 1-box source. It also makes a very solid transport with an outboard DAC, but it's a waste of a pretty sweet discrete output stage.
For reference, here are some posts on the aforementioned Toshiba 3960 and NAD C541i machines and their tweaking.
