Benchmark DAC-1 mods

audioengr · « **on:** 3 Dec 2004, 03:52 am »

I just received a Benchmark to mod. First, my listening impressions:

I can see why so many have been anamored with this DAC. It has the best HF extension and HF dynamics that I have ever heard in a stock DAC. However, the midrange and bass is inconsistent with the HF dynamics. This is just the opposite of what I usually find in stock DAC's. Generally, the HF extension and dynamics is what is lacking and the midrange is very good and the bass okay. The High-frequencies are usually recessed or at least imbalanced. In this case, the midrange and particularly the bass is less dynamic than the highs. With some classical music that contains a lot of HF information, this is hardly noticable, but other pieces that have loud kettle drums or bass viols, it is noticable.

Examination of the circuitry lends some clues as to why this unexpected imbalance is occurring. Everything on the board is surface-mounted, including the large filter capacitors. These should really be through-hole parts. Even the very best SMT parts cannot yet compete with good through-hole caps for this function. Typical bridge rectifiers are used as well. High-frequency decoupling caps are located very close to the op-amps and they are larger than what I usually see. This probably accounts for the excellent HF extension and dynamics. Another thing that I noticed is a slight hum in my system. This is typical of many DAC's that do not galvanically isolate the digital input.

Next, to do some mods.

sys1 · « **Reply #1 on:** 3 Dec 2004, 05:33 am »

audioengr -
I've also tested the DAC-1. I agree with your impressions that upon first listening, it sounds fast and clean, but I found it's soundstaging and dynamics lacking. IMO, it was obviously designed for nearfield studio or headphone listening. It uses NE5532 op-amps and has a high output impedance (1.25k Ohms).

It'll be interesting to see what you can do with it.

audioengr · « **Reply #2 on:** 3 Dec 2004, 06:18 am »

Yes, I see why it is 1K+ ohms output impedance. They added a voltage divider using two 2290 ohm resistors to the unbalanced outputs. This cuts the voltage in half, but results in slightly higher than 1K ohm output impedance. I like to see 100 ohms or less. The output is already fairly high amplitude by audiophile standards, so I'll either have to load the op-amps more or change the gain to keep the amplitude low and lower the output impedance.

Mad DOg · « **Reply #3 on:** 3 Dec 2004, 07:00 am »

Quote from: audioengr

I just received a Benchmark to mod. First, my listening impressions:

I can see why so many have been enamored with this DAC. It has the best HF extension and HF dynamics that I have ever heard in a stock DAC. However, the midrange and bass is inconsistent with the HF dynamics...

steve,

i, too, found this to be true when comparing to my MW lev2 P-3/A...more open sounding in the HF. midrange was harder and leaner. bass wasn't as tight and as controlled. but for $975, it's a dayem good DAC in stock form. then given that it does have decent quality sound when hooked up direct to the amp as a preamp. and has headphone jacks, too! that's alotta features offered for the $...

cjr888 · « **Reply #4 on:** 3 Dec 2004, 12:42 pm »

Probably worthwhile to talk to the Benchmark folks to identify why they made certain choices -- they've responded quite a bit regarding modifications to the unit and explaining why they made the choices they did.

audioengr · « **Reply #5 on:** 3 Dec 2004, 06:00 pm »

Quote from: cjr888

Probably worthwhile to talk to the Benchmark folks to identify why they made certain choices -- they've responded quite a bit regarding modifications to the unit and explaining why they made the choices they did.

I've read many of the posts of the Benchmark engineer. Everyone has their version of the "magic bullet". The problem is that most of these bullets only solve one problem and do not address the full spectrum and full dynamics problem.

audioengr · « **Reply #6 on:** 3 Dec 2004, 11:41 pm »

Update - I now have developed a schematic.

There are 5 op-amps in the signal path from the DAC to output when using the RCA outs and bypassing the volume control. There is one AC-coupler consisting of an electrolytic 22 uFd cap and a small chip ceramic cap in parallel. I would like to try to bypass this and eliminate it. I may have to add components to tune-out the DC-offset. AC couplers will prevent it from ever being as good as the P-3A DAC that I mod. I will probably first improve the post-DAC filtering and the power delivery to the op-amps, the DAC chip and the upsampler chip, before changing the op-amps.

The mods are simpler if the volume control, XLR outs are not used and headphone amp is not modded, but it is clear that they can be quite extensive. There is a LOT to mod here.

cjr888 · « **Reply #7 on:** 4 Dec 2004, 12:05 am »

Wouldn't it be better to use the XLR outputs wired for unbalanced so that you end up with the 30 ohm unbalanced output?

audioengr · « **Reply #8 on:** 4 Dec 2004, 03:43 am »

Quote from: cjr888

Wouldn't it be better to use the XLR outputs wired for unbalanced so that you end up with the 30 ohm unbalanced output?

It's the same buffer. One is just an inverting buffer and the other is a non-inverting that goes only to the balanced outputs. The non-inverting is bandwidth limited, so I would rather just replace the output resistors and use the RCA outs driven from the inverting buffer.

audioengr · « **Reply #9 on:** 5 Dec 2004, 04:09 am »

First round of mods:

1) Bridge rectifiers replaced with HEXFRED's
2) Power delivery improved to 8 dual op-amps
3) Some filtering improved after the DAC chip
4) power delivery improved to DAC chip, receiver and upsampler
5) Coupling cap for S/PDIF input improved
6) Analog stage AC-coupler bypassed and DC-offset manually tuned out

After reading how this DAC was reported to be jitter-tolerant by folks that have them, I was surprised to find that replacing the S/PDIF coupling cap with a good Poly cap made a marked improvement in the clarity. Every time someone says that a DAC is totally tolerant to cables and Transports and that jitter doesn't matter, I am really suspicious. This one may be less affected by jitter than most, but this result indicates that jitter still matters.

First listening impressions of these mods:

The dynamics in the midrange and bass are improved. Bass is very strong and focused now. The high-frequencies are still excellent. The entire frequency range is now equally dynamic and focused. Imaging is excellent. The problem is that now I think I am hearing the real character of the NS5532A op-amp. It is a bit "cold" sounding (my wife's observation). Once the power delivery was optimized to this op-amp, it's real character seems to have shown through. I will give the Benchmark DAC-1 a few days of play-time to insure that the Black Gates are fully broken-in. This will probably improve the sweetness a bit.

In order to add sweetness, I may need to change 8 op-amps to my own favorite (it's a secret). This will add significantly to the cost of the mod, because they are around $12 a piece and I have to swap them in, which will take a couple of hours. As it stands, the labor is $650 and the materials are around $50.

audioengr · « **Reply #10 on:** 5 Dec 2004, 07:35 pm »

Holy smokes! - this is sounding like a Turbomod now. I finally changed the output resistors so that it has about 120 ohm output impedance on the RCA outs. With some break-in on the Black Gates it is sounding superb now. Very airy and 3-dimansional. I dont think there is any need to change the op-amps. In fact, the HF response of these op-amps seems cleaner than the OPA-627's used in my P-3A DAC. Still need the wife to make the final call as to which one is better.

This is looking like it may be neck-and-neck with my P-3A. Finally, a world-class balanced-output DAC that is direct-coupled

cjr888 · « **Reply #11 on:** 5 Dec 2004, 09:18 pm »

And thoughts on completely ditching the opamps?

audioengr · « **Reply #12 on:** 5 Dec 2004, 09:41 pm »

Quote from: cjr888

And thoughts on completely ditching the opamps?

I dont believe there is any need to change op-amps. If I did, I would have to change two regultors as well, to lower the voltages to +/-15V. Most op-amps will not handle +/- 18V.

Steve N.

audioengr · « **Reply #13 on:** 6 Dec 2004, 05:43 pm »

Comparisons last night to the P-3A:

The P-3A is more musical. The midrange is more "there". I believe this is a difference in the op-amps.

The Benchmark has pristine, clear, focused highs, even better than the P-3A, but you dont get as much echo off the walls of the venue, so it is not as airy as the P-3A. This is probably due to the 5 op-amps or the fact that several stages are bandwidth limited. This is necessary since it has the volume control pots. These can introduce noise or even oscillation if there is no B/W limiting.

Next, I will replace more of the filtering caps to see what affect this has. Another possibility is to replace the input pulse transformer, but this requires dissassembly of the entire back panel.

audioengr · « **Reply #14 on:** 7 Dec 2004, 03:16 am »

I didn't give up on the Benchmark yet. More changes today:

Bulk Decoupling added - large Black Gates
Replaced the regulator filters - large Nichicons
More improvements to the filter section
Removal of output filter caps

Now this puppy is really singing. We will do more extensive tests this evening, but it is sounding a LOT like my reference P-3A DAC now but with more clear, focused highs. Definitely Turbomod material.

eico1 · « **Reply #15 on:** 7 Dec 2004, 03:39 am »

Quote from: audioengr

After reading how this DAC was reported to be jitter-tolerant by folks that have them, I was surprised to find that replacing the S/PDIF coupling cap with a good Poly cap made a marked improvement in the clarity. Every time someone says that a DAC is totally tolerant to cables and Transports and that jitter doesn't matter, I am really suspicious. This one may be less affected by jitter than most, but this result indicates that jitter still matters ...

Do you think there is actually a change in jitter when changing this cap? What about the Benchmark site showing graphs of no measurable jitter? Just wondering what's going on...

steve

audioengr · « **Reply #16 on:** 7 Dec 2004, 06:00 pm »

Quote from: eico1

Do you think there is actually a change in jitter when changing this cap? What about the Benchmark site showing graphs of no measurable jitter? Just wondering what's going on...

steve

Yes, there is. Not measurable? Depends on the equipment that is doing the measuring...

audioengr · « **Reply #17 on:** 7 Dec 2004, 06:14 pm »

Final report:

I have done everything I can to power-delivery wise to make these op-amps happy. It is evident that there are power delivery problems on-die with them that prevents the midrange from having the slam that the P-3A has. Either that or the Analog Devices DAC chip just does not deliver the goods. More likely the op-amps are at fault.

The Benchmark op-amps do excel at high-frequency dynamics, clarity and extension. This is what makes this DAC so popular. However, it just does not match the midrange bloom and dynamics of the OPA627. The Benchmark DAC is still marvelous to listen to. Very detailed and musical, but on really dynamic passages, it just does not compare to the excitement created by the P-3A. Still highly recommended. I would have to rate it between the modded Electrocompaniet ECD-1 and the modded dAck!. All of the DAC's in the ratings list are excellent, so dont get the impression that any of them sound bad or unmusical. If this is the case, they dont make the list.

This is a case of one strong point. The Benchmark excels at HF extension and dynamics, which are really the weakpoints of the P-3A. I'm trying to fix that in the P-3A today. The P-3A (and the MSB) are great in the midrange and have the best in slam. Dont get me wrong, the Benchmark does not sound analytical to me. It still has good dynamics from top to bottom. It's those orchestral crescendos that are just more impactful with the P-3A. The two DAC's actually have very different characters, but both are great to listen to.

audioengr · « **Reply #18 on:** 23 Dec 2004, 01:38 am »

Another final report:

I decided to put new regulators in it because this improved the P-3A so much. Todays testing shows that it is very similar sounding to the P-3A, but a tad brighter. Must be the op-amps. The midrange and bass seems to have come up in amplitude and dynamics, but the highs are still a tad brighter or more forward. Very detailed and focused now, with some lushness that was missing before. It is so good now, I have to give it a Turbomod rating. It falls between the MSB and the P-3A now, or at least neck-and-neck with the MSB. The mod stands at $1000.

I may offer options for this one:
1) change 10-12 op-amps - $300
2) redo the digital input - new transformer and termination - $150

jeffx · « **Reply #19 on:** 8 Jan 2005, 05:52 pm »

I have been having a lot of fun and enjoyable listening sessions since I received the Benchmark back from Steve earlier this week. Wow! sums it up best.

Improvements seem to be across the board, especially with the midrange and lower frequencies. Originally, the midrange seemed slightly recessed, flat and non-involving. This has been corrected. While bass with the unmodded unit was better than what I was accustomed to, it was not well defined. This too has been corrected. These corrections have helped the dynamics to be consistent through all frequencies. Instuments sound more like real instruments and voices now have body.

I never really experienced the problems with the high frequencies and brightness that others have commented on. Maybe it is because of the tube preamp I use.

In addition to my EA modded S7700 and ZR1600, the Benchmark gives me the most musical system I have had. Thanks Steve!