X-24K DAC Mods

[Deleted]

matt7941
The 100nF capacitor leads should be both as short as possible. Longer leads may result in oscillation of devices like the AD8066, which normally specify 100nF ceramic bypasses, usually surface mount types. I am not sure that you will be able to plug the Adaptor board with the AD8066 into the DIL socket. The pins on the adaptor socket MAY be too thick. In any case, it is always good practice to solder devices like the AD8066 into the PCB, particularly when mounted on adaptor PCBs, due to the extra lead length,because of their very wide frequency response and risk of instability. The LM4562 is much more forgiving, because of it's lower speed, and should be O.K. plugged in, provided that the 100nF capacitor lead length is minimised.
Alex

[matt7941]

Cheers Alex.

The adaptor boards I've ordered for the for the AD8066 don't come with the pins attached, they're separate. I have seen some SIL turned pins that should fit the sockets should the ones supplied not do so.

With my limited skills in this department I've attached the caps to ground at the nearest point I could without drilling through the double sided board so as not to damage any circuits. With the congested nature of the board it really doesn't seem to have lengthened the route by any excessive distance. However, I note your comments and if the AD8066s start to oscillate I'll stick to the OPA2134/OPA2604/LM4562s.

Cheers.

Matt.

[matt7941]

Good evening one and all, it's been a while.

Just a quick update on the DAC front. I have tweaked the thing out of all recognition and it now sounds fantastic.

The large Jamicons have been replaced with Panasonic FCs, power smoothing being handled by 47uf 16v OSCONs. All the dipped poly caps have been replaced with WIMAs and the opamps changed to OPA2134 buffer and LM4562 outputs, supplied by Rubycon ZAs and bypassed with a 1uf and 0.1uf poly combo. The RCA output sockets have been changed to CMC solid copper silver plated and rewired with solid silver teflon insulated wire from the circuit board. There are now no tantalums in sight and the low value (2x 68pf and 2x 22pf) ceramics have been replaced with silver micas of the same value. The output caps are now 100uf 16v non polar Capxons bypassed with the 1uf/0.1uf polys.

The last thing I have yet to tweak is the clock, and this is where I have a problem. The crystal was originally a 6.144mhz, but I have replaced it with a 12.288mhz and it appears to work fine, if not better. The dac chip is the Burr Brown PCM1728e which suggests this crystal should be fine, but the frequency sampling lights on the front of the unit now no longer work. Not a problem to me as it sounds fine but can anyone suggest why this should be? The replacement crystal was an experiment to see if it still operated ok as I can only find a clock upgrade with this frequency and not 6.144mhz. Incidentally 6.144mhz is not a frequency listed on the spec sheet for the PCM1728e as being effective yet it obviously works. I am a bit of a novice at this despite my endeavours with the soldering iron and if someone could help shed some light on the clock situation I would be most obliged as I've grown attached to the DAC and want to wring out that last ounce of performance which it obviously has.

Thanks

[traf]

Hi Matt

You seem to be having fun This thread is bringing back a few memories as the X-24K was the first piece of kit I hacked around a couple of years ago.

You mention that you've upgraded your psu. If you are still chucking an AC supply into the dac, one mod you HAVE to do is upgrade the rectifier diodes (8 of them I think) to schottkys. I tend to use MBR1100 or 11DQ10. You'll find it brings a bit more clarity and focus to the party....and its cheap

I really liked the X-24K. It responded really well to 'word length' changes using a Meridian 518.

[matt7941]

Hello Traf,

Sorry, I forgot to mention that I replaced the diodes from the original 1n400x series to UF400x series which is partway toward the Schottky effect I suppose.

With all the mods it really does sound soooooooo much better in every department. I've similarly modified my MF A3 amp with new internal cabling to match.

Did you attempt to upgrade the clock and if so how did you fare?

Cheers

[matt7941]

Oh one more thing, I also whipped out the muting transistors which could well be one of the simplest and best upgrades yet.

[traf]

[quote author=matt

Did you attempt to upgrade the clock and if so how did you fare?

[/quote]

Matt

I never quite got around to a clock upgrade, unless you count the Meridian 518 which was a rather good de-jitterer. I employed a similar approach using a WM8840 board in an A5 DAB tuner to add a standalone DAC capability (equivalent to XPSUv3+XDACv3+X10v3). It has a very good de-jitter spec at 1/8th cost of the 518.

Cheers

Simon

[dejanm]

Matt,

Where did you bought your clock replacement ? I have difficulties to find two pin oscillator (16, 9344 MHz) replacement for installed SPK ...

[matt7941]

Hello Dejanm.

I was also struggling to find a 16.944mhz 2 pin as I thought I might give it a whirl so to speak. Unfortunately I had no success in locating one. I found my 6.144mhz 2 pin from RS Components and the 12.288mhz was from a seller on eBay (UK) of all places.

Sorry I couldn't be of more help regarding the 16.944mhz.

Cheers

[dejanm]

Matt,

Thanks anyway. It is interesting that nobody is selling two-pin oscillators anymore. I contacted several producers and shops without even getting any reply.

[dejanm]

Sept 25, 2008 3:07:14 GMT mrarroyo said:

Try: www.tubesandmore.com/ Seach for oscillator.

I contacted them and they answered politely that they do not have it.

I am giving up ....

[Deleted]

dejanm

Try Digikey , or in the U.K. ,
www.jabdog.com/crystals.htm
You will most likely find others too. Just Google "16.9344 crystal".

SandyK

[leo]

Why not just use a 4 pin oscillator and feed it a low noise regulated supply, run the oscillators output to the XI where the original 2 pin went

[matt7941]

Good evening folks.

Just a thought, I've virtually finished with the soldering iron now and have just one more question if someone could graciously help me out?

Can I dispense with the final output caps just prior to the RCA outputs as suggested by Alex when modifiying the X-Dac v3? Would it work in the X-24K? They are originally 47uf 16v NP Jamicons (ugh!) and I have replaced them with 100uf 16v NP Capxons, but am toying with the idea of removing them completely or using Ampohm 1uf 630v caps which seem to have a good press. I seem to remember someone once saying the best cap is no cap.

I am using opa2604s as the buffer opamps and LM4562s as the output opamps.

Help would be warmly appreciated.

Thanks in anticipation,

Matt.

[Deleted]

Matt
You would have to read the voltage between the I.C. side of the capacitors and earth. IF the reading is less than 20mV,(preferably only a couple of mV) then you could short them out . Using 1uF 630V capacitors would completely unbalance the sound,and severely attenuate the low end frequency response.
Alex

[matt7941]

Thanks Alex, I'll get the multimeter out.

If I were to replace the caps what would you suggest, something like a Nichicon Muse non-polar of say 47uf and bypassed with a Wima or similar 0.1uf - 1uf polyester or polypropylene, or can I use one of the larger (in size) "audio" caps of the like of the Ampohm or Mundorf but at a different capacitance?

Thanks

Matt.

[Deleted]

Matt
Your choices seem fine to me, but I would try without the Wima bypass cap first.
Mike may have further suggestions in respect to suitability and availability.
Alex

[matt7941]

Thanks for the help Alex, most appreciated.

Just one thing re the voltage reading, I presume this should be done whilst powered up but with no signal playing through. Forgive the naivety.

Cheers,

Matt.

[Deleted]

Matt
Correct. If you are using a Digital Meter, measure on the mV range.
Alex

[matt7941]

Thanks Alex, just tried it and the reading seems to fluctuate from 0.3 mv to approximately 2.4mv so it sounds as if it should be ok to short.

Thanks again.

Cheers,

Matt.

[matt7941]

Hello all.

A quick question, for the output caps on the X-24K immediately prior to the RCA outputs can I use straightforward polarised electrolytics or do they have to be non-polar? In the original design they are 47uf 16v nonpolar, however I have been using some 470uf 16v AM series Panasonic's which are polarised with the positive side towards the ICs and the negative toward the output but bypassed with a 1uf 400v Mundorf MKP nonpolar. It seems that some say this is fine and does not degrade the sound whilst others swear by nonpolars. In this application replacing the the original 47uf caps with the 470uf has extended the bass response somewhat and it sounds fine, I wouldn't want to go back to 47uf caps. However, if replacing the output caps with 470uf nonpolars would sound even better it has to be worth investigating.

Thanks

Matt

[Deleted]

Matt
Didn't you intend bypassing those output capacitors due to the very low DC out ?
You wouldn't normally use polarised electrolytics where there is no reasonable level of DC across them.
Alex

[matt7941]

Good evening Alex.

I did intend to but then got slightly cold feet as it were, I didn't want to risk damaging anything further down the chain. The problem being that I took a second set of readings from the test meter and it read from +9mv to -9mv and fluctuating all points between. That put me off a little. Having said that I suppose the A3 amp will have an input cap and would that not negate the need for the output cap?

Matt

[Deleted]

Matt
No problems at all with those kind of readings, even without an input cap at the other end. Your new readings also illustrate why a polarised electro isn't suitable at that location.
Alex

[matt7941]

Good evening all.

Just thought I'd now add that I've finished playing around with the X-24K (almost). Alex, I finally settled on Nichicon Muse ES 220uf 16V bipolars with a Mundorf 1uf MKT bypass as the output caps. They work really well together; sweet, open and deep. Now with the modifications I've carried out on the DAC, the A3 and my Epos M15.2s I'm really happy. Many thanks to all who have helped along the way.

I just wondered whether anyone could advise re my next step. I am considering reclocking the DAC and transport using one of the low jitter TCXO based modules out there. The Micromega Stage 1 I'm using as a transport is clocked at 33.8688mhz and the DAC is 6.144mhz. I'm going to struggle with space constraints in the DAC but I'll work round that. My question is which component should I reclock first? Is it better to reclock the transport so it outputs a more accurate signal to the DAC or is there more gains to be had by doing the DAC first? I can only afford to do one at the moment when I take in to account the cost of a separate power supply for the new clock to give it clean power. The reported gains of reclocking seem to be worthwhile experimenting with (bang-for-the-buck-wise).

I hope someone can assist as I bow before the vast knowledge of many of the members.

Cheers

Matt