Sowter/Pioneer AX-10 Mods

[rljones]

I recently acquired a Pioneer AX-10 and found the sound exactly as described in a recent Stereophile review: very good DVD-audio, but SACD and CD is so, so.

Opening up the AX-10 (to the horror of my family) I found a about 12 OPA2134 serving as I/V, filters, and output buffers for PCM1704s. Main channels have a balanced pair of these DACs with one apiece for each surround.

My initial modification was to change all op amps to AD825 (15V rails precluded my favorite AD8610/20) using BrownDog adapters, bypassing all output electrolytic and adding 100 nF decoupling caps for the AD825s. The improvement was noticeable, but there was still some mid freq difference between this unit and a BOW CD player I've used as a reference. At this point, the modified AX-10 played SACD slightly better than an Sony 333ES with a Modwright change.

I then turned to Sowter transformers. I tried various models (the load is a solid state preamp, so their DAC-tube transformer wouldn't work for me): 4935, 8230 and 3603. I tried all 3 types taking the signal off of DAC directly, using the transformer as the only load, using a resistor in parallel with transformer primary, and finally, taking the signal off of the AD825 used as an I/V per the PCM1704 datasheets (they used an OPA627 as I recall). The latter method worked best, and sounded best with the 3603 used as a balanced output transformer. (One warning is not to use the 8230 off of the op amp I/V: its impedance is too low and made the AD825 rather warm.)

So now it's the PCM1704 to AD825 I/V to Sowter 3603 to preamp. I've eliminated many op amps and the soundstage and dynamics are improved. I don't yet have the BOW available for comparison, but the SACD now sounds as good as the DVD-audio.

Regards, Robert

 

[fmak]

So now it's the PCM1704 to AD825 I/V to Sowter 3603 to preamp. I've eliminated many op amps and the soundstage and dynamics are improved. I don't yet have the BOW available for comparison, but the SACD now sounds as good as the DVD-audio.

Regards, Robert [/B][/QUOTE]
--------------------------------------------------------

Robert

I did this with a SCDBX940 using the Sowter 9360 600/2.4k unit, and with the AD825 as I/V.Two points:

1. The Sowter adds distortion and picks up (inaudible but measurable) hum. Performance is akind to good analogue preamp and provides a tube like, liquid sound. However, I have concluded that it isn't necessarily better after many comparison trials. The sound depends a lot on any filters (first order) connected to the secondary.

2. Current loading the AD825 to at least 2mA with a cascode circuit connected to the -ve rail brings much better benefit in both the treble, and surprisingly, the bass.

The waveform from the I/V stage looks clean and if one uses a valve preamp, offers the chance to delete most of the following circuitry and perhaps balanced connection with passive lf filtering.

There is a major issue with switching noise with coupling methods that bypass the relays. Modders be careful!

 

[rljones]

fmak,

I'm not using any filters on the secondary side. Is this a mistake? I thougth that since the 24-bit DAC is sampling at around 1MHz (per data sheet), that the natural roll-off of the transformer was sufficient, allowing one to thereby bypass the 2nd order Butterworth that are so commonly used.

I'm running the signal from the transformer into my preamp, which uses only a PGA2310 and likes a 600 ohm load (such as the 3603 provides). Do you have other transformer recommendations, considering the output impedance should be around 600 ohms? (Space is also an issue and the Sowter is reasonably small.)

Current loading the (-) rail may be a little difficult (possibly on the adapter?). How did you do this, with a jfet, any bypass caps?

Thanks, Robert

 

[dorkus]

weird

i find it interesting that your analog mods improved SACD reproduction to where it is comparable to DVD-A now. i would think that the mods would improve both DVD-A and SACD equally, so the margin between them would still exist. i also thought the AX10's limiting factor with SACD was the conversion of DSD to PCM for the PCM1704 DAC, which no amount of analog tweaking would be able to solve.

also, my exposure to DVD-A is very limited but from what i can tell it sounds very different from SACD. not better or worse (that is a subjective call), but different. but then, if DSD is being converted to PCM internally anyway then i guess it would make them sound similar.

 

[rljones]

dorkus,

You're correct, all signals, DVD-SACD-CD, in this unit go through the PCM1704s. I don't know where the DSD conversion section is (I don't know digital circuitry and wouldn't know what I was looking at if I saw it).

When I said the SACD section sounded better, I was comparing the same SACDs pre- and post-mods in the AX10 to a Sony 333ES. I didn't have the same recording on both SACD and DVD to make even a rough comparison. Since the AX10 converts all to PCM, even if I do such a comparison, it will still not tell me whether PCM or DSD sounds better. Rather, with the extra stage in the DSD decoding in the AX10, I cannot imagine it could sound better than a top DSD unit. I believe I can say, however, that the modified unit sounds better than before it was modified, and better than the semi-modified Sony (which is not a top-of-the-line unit).

I seem to have read that many currently available SACDs are re-released PCM (CD) recordings, so this may contribute to many of those that I've heard that are only so-so. With so many potential variables at this stage, it's difficult to say one is better than the other. I didn't mean to imply that. All formats were improved with the mods, but I cannot quantitate the amount of improvement for each format. (How many LPs do I have that are only so-so, and variable in terms of recording techniques, artists, venues, etc--yet I still prefer good vinyl.)

 

[dorkus]

agreed. comparing formats is pretty tricky because of all the variables in the recording chain. i also agree with you about preferring one format in general despite variables... i have a few crappy-sounding SACDs but i still like SACD better than CD, even if a good CD sounds better than a bad SACD. i think even with the variable recording quality, it is still possible to get a handle on the "sound" of a format if you listen to enough recordings. PCM and DSD sound very different to me. vinyl is great too but i'm a child of the digital age and don't have the patience to deal with it tonearms and needles and brushes.

 

[fmak]

agreed. comparing formats is pretty tricky because of all the variables in the recording chain. i also agree with you about preferring one format in general despite variables... i have a few crappy-sounding SACDs but i still like SACD better than CD, even if a good CD sounds better than a bad SACD. i think even with the variable recording quality, it is still possible to get a handle on the "sound" of a format if you listen to enough recordings. PCM and DSD sound very different to me. vinyl is great too but i'm a child of the digital age and don't have the patience to deal with it tonearms and needles and brushes.

--------------------------------------------
It is difficult, if not impossible to compare DVDA and SACD on commercially available software. There are decent SACD discs but very few hirez DVDAs. The 192 tracks sound very good but most of the 24/48 discs are ordinary and no better than DAT. This is why DVDA hasn't btaken off - I for one won't buy new DVDAs until the situation improves.

On the other hand, DVDV 24/96 discs do sound very good through a good DAC and comparable in some ways to SACD in terms of improvement over 16/44. Rationally, 24/96 and 24/192 in a video format would have sufficed if there were no need for surround.

 

[rljones]

Listening Follow-up

I was able this weekend to compare the modified AX10 as described above, to a Bow Technologies ZZ-8. Prior to modification, the Bow was decidedly better with regular CDs and even more so when playing HDCDs (which the Pioneer cannot decode).

After the mods, the AX10 played regular CDs better than the Bow with better depth and air. (I had a friend over who has professionally played violin and another who plays clarinet; both concurred). The HDCD discs on the Bow were slightly better as compared to when the same discs were played on the AX10 without HDCD.

Another test was using two copies of 'Nature's Realm' by Water Lily with the modified AX10. One is on DVD-audio, the other, SACD--both of the same recording. Listeners preferred the DVD-audio, but this was very close. Slightly more air about the instruments was noted with the DVD-audio version. However, as mentioned previously, this unit sends the DSD signal throught PCM processing, so it is not a truly fair comparison of the formats, except on this machine.

Regards, Robert

 

[bmwrob]

Hmm, did you measure 15V rail? If you did you scare me a bit, as I didn?t measure (maybe a big mistake) but trusted the service manual which states its 12 V. I have already replaced the 2134 with the AD8610. I?m very pleased with the result. If the rail is 15V I guess I?m frying the OP-amps sooner then later.

Where did you buy the AX10? I thought all American sold AX10 had the AD DACs and not the BB 1704 as the Japanese version. Can you also run your AX10 as an standalone DAC?

 

[jam]

Robert,

How about following the AD825 with a buffer? I believe Jensen might make a better transformer.

Jam

 

[rljones]

bmwrob,

I bought the ax10 in the US from a dealer. I did not have access to a service manual; I studied the pcb and drew out a schematic and then proceeded with mods. I have not tried it as a stand alone DAC.

Are service manuals easy to get; I'd like to buy one.

Jam,

I did not wish to use another op amp; I like well made transformers.

 

[Ric Schultz]

No buffer!

The AX10 is perfect for mods since the DSD info is now PCMed so that there is no DSD noise shaping to deal with.....so, why use any buffer after the I-V converter? I have been using I-V converters (both op amp and discrete) with no output filtering with pure PCM DACs for over 10 years. Sounds fantastic. If you must, then use a large value of cap across the feedback resistor for one pole of filtering(set it to start rolling off around 16K). You should use at least 10 picofarads across the feedback resistor for stablization of the op amp. Simply run a 75 ohm resistor from the output of the op amp to the output jacks. You MUST float the output jacks from the chassis and run a ground wire for each channels ground from the ground point of the op amp (probably pin 3, ---you might try other ground points near the DAC also). No need for a transformer, unless you like the sound it adds.

If you are more adventuresome, you could put a resistor to ground and add a tube output stage.......but I would not. A very good I-V stage straight out is pretty killer.

Ric Schultz

 

[Ric Schultz]

Forgot to add

Oops, Did not see that it had two DACs on the output. What I am suggesting is throwing away the use of one of the DACs on the output......Anotherwords you are using one of the two DACs per channel. Now, if you want pure balanced, then you can use both phases (DACS and I-V converters) and mount an XLR on the output. A super transformer might be better than running single ended out, but I kind of doubt it, as I just tried it with a silver transformer and I like the single ended (direct from the I-V stage) out better.

Another option is to have two sets of RCA jacks on the back of the chassis (one set would be non-inverted and the other inverted). This way you can invert phase in the analog domain without any loss of fidelity. Most Sony SACDs and many other discs are inverted phase. If you have a preamp with two inputs, you could just go between the two inputs for the different phase. Me, I have to unplug and plug each time, as I run my source directly into the power amp (built-in attenuator).

Have fun!

Ric Schultz

 

[bmwrob]

Robert, I have attached the L/R output stage schematic. I'm at the moment scanning in my manual (very time consuming). If you have an ftp server I can send it to you. Otherwise I can send you a CD-R.

Ric, the AX10 already has XLR outputs (see link). What I was thinking of was if I could simplify the balanced outputs from IC341/IC391 onwards. There is a lot of stuff in the signal path before the actual XLR connectors. As I always run my machine in balanced mode, I don't need a switch to turn off the balance outputs. Any suggestions for improvement?

Robert as seen on the pictures this is the Japanese version of the AX10. It has digital inputs which I believe the U.S version doesn't have.


schematic

AX10 Rear view

 

[Ric Schultz]

AX10 mods

Rob,
Just do as I suggested. Take a 75 ohm resistor directly from the I-V ICs (IC 310, 331 for one channel) and connect them directly to the output XLR jacks. Throw all the other circuitry away. You might want to hook up the output directly at the XLR to the output relay to get rid of any turn-on or turn-off surges. Try it! You can adjust the gain via the feedback resistor across the op amp.

Ric Schultz

 

[rljones]

Ric,

As I recall, I tried the resistor straight out and could not get a signal out. I'm not working from memory, which is a dangerous thing, but I seem to remember that these BB DACs need an I/V stage, unlike AD DACs. The transformer that I use, going direct to the XLR, bypassing the output XLR switch (as you mention, bmwrob), keeping everything fully balanced.

bmwrob,

I'd really like to get the manual from you. I'd be happy to send you a check for your efforts; just email me separately at [email protected]

regards, and happpy new year to all,

robert

 

[jam]

Ric,

I suggested a buffer because it would help drive the transformer better and isolate the i/v stage from the transformer. A good buffer is pretty transparent.

Jam

 

[Ric Schultz]

Robert,
Please reread what I wrote. I did not say to run a resistor directly from the DAC, but from the output of the I-V converter ICs. It will definitely work, and work tremendously. Try it.

Again, what I am suggesting is to leave the first op amp in place (the I-V converter) and then run a resistor from its output pin directly to the output jack. And if you want balanced then do it with the other phase as well. One 75 ohm resistor to pin 2 and one 75 ohm resistor to pin 3.

With this output stage you only have one resistor and one op amp in series with the signal with each phase. It is seriously transparent. Of course, you could always use a discrete I-V converter for even better results (if done properly).

Ric Schultz

 

[J Chen]

I/V opamp

Hi To All & Merry Christmas & Happy New Year.
Ric, You post an interesting suggestion.
Can you suggest a good opamp for I/V used &
the proper way to implement it for maximum
performance in this application.

 

[bmwrob]

Ric, just to clarify, do you mean on pin 1 on IC311 (IC331, IC361, IC381) or pin 7? Any drawbacks by completely bypassing IC341 and the other corresponding OP-amps?