How much settle time do you take to review the SQ between the changes.
Is there a chance that you create a discrete output stage or maybe with tubes ![]()
How much settle time do you take to review the SQ between the changes.
Is there a chance that you create a discrete output stage or maybe with tubes ![]()
I just settle until I know what its doing. There is also parts break in time involved, which complicates things. But itās usually pretty quick to assess what its doing though; the changes weāve progressed through are pretty big, so thereās usually a pretty big audible change along with it.
Itās playing now in fact.
Iāve thought about buffering. Unfortunately I know ~ nada about tube circuits. But I do have an excellent, ~ no distortion discrete BJT open loop buffer available in the tool bag (itās in the preamp as well, years of development effort).
But I think if we keep the driving impedance ~ low to the transformer we then have a low (enough) output impedance to do the job properly IMO. What Iām hearing so far at least itās not crying out for a buffer (to me) ![]()
Why are you not using the Edcor transformers?
RE: transformers, I have not double checked the pin out yet but hereās a high nickel core transformer that might drop in and someone might want to try rather than the Edcor M6 steel. And the correspondence with the mfgr (Cine-mag) from 12/10/21.
Turbo
Thank you for your inquiry. The CMOQ-2H will drop into the same PCB holes. We do not apply an external copper foil āshieldā as shown for the Edcor transformer, but for most applications it is superfluous. The CMOQ-2H is $62.04 + shipping. The CMOB-2H is the equivalent of the CMOB-2H ($56.60) but with the windings hooked up for 1:1. āBā = bifilar winding and āQā= quadfilar winding. Please note that, because of the Wuhan19 pandemic, our lead times are running a bit long. Thank you in advance for your patience.
All the best,
David Geren
(818) 993-4644
I miss spoke⦠Is anyone helping with capacitor upgrades (of course for $$)?
Carnhil, eg Sowter, it was tough to get parts from them for me here in the US recently. I had to invoke the nuclear PayPal option to get them to finally ship them, unfortunately.
I used a Sowter mu metal transformer early on in the DS DAC trials. They (Carnhill) donāt have anything that drops in though, so it has to be fly wired.
Personally, my opinion Iām for anything that uses a more transparent sounding core. Otherwise if you just stick with M6 steel then just use the Edcors that drop in.
Turbo
That is very interesting⦠Iāve seriously thought about installing some Souter transformers in mine, but the mounting issues have made me shy away from the idea. But if Cinemag makes some permaloy core transformers which are drop-in compatible, that would be a far more practical mod.
The big thing I would wonder about though is the bandwidth of these cinemag transformers. Ted deliberately chose transformers with high leakage inductance so that they could play a part in the low-pass output filtering. So having such wide bandwidth might be an undesirable characteristic in a DS.
Turbo. May I ask which software package you are using to perform your modeling?
Also, Iām curious why you would refer to a circuit which clearly shows feedback being fed inside the filter itself as being a "passive LPF ". While that is indeed a very interesting circuit (and it might sound wonderful, for all I know), it appears to me that only the first two poles are truly passive.
@wctaudio you did an excellent job with transformers. Are you planning to help us with the capacitor upgrade too? We need help here ![]()
Well the Naim circuit is a passive filter because the filtering is all done with passives. But it has feedback, unity gain buffered. Iām not sure if it could run properly otherwise (unbuffered). So not the best choice of words perhaps. The main point though was, interesting concept for creating a deep null at CD 44K1 freq back in the day. CD was new back then.
I just use LTSpice for simulation.
T
Frankly thereās not much LPF left to do after the OA in this design. I think the stock MFB filter is > 100dB down around the bitstream frequency.
Tedās comments are valid of course. My opinion, for most of us whatever very small ultrasonic is left into and after the M6 or even mu metal or amorphous trafos is ~ of not much concern really. That ultrasonic content is generally way above the BW of even a mu metal trafo.
In other words if you want to hear a clearer trafo, go ahead. You wonāt harm anything. If you find thereās some issue, just put the others back in.
I donāt know, maybe Ted might know of some content due to his fw scheme that might be a factor down lower.
Turbo
I now have roughly 100 hours of run time on the XS4400. The DirectStream is paired with a RSA A-10 Thunderbolt II preamplifier / electrostatic headphone amplifier. The volume appears to have increased some as I have been able to lower the volume by two clicks. I am able to listen through my Stax SR-009S back to 1:30. This pairing is awesome. I am glad I did the upgrade even with the volume drop. Everything is so clear I can actually go much higher.
This quieter perceived volume has been my experience when noise is reduced in the system too.
Me too. It can be considered a measure of a successful change.
Less fatigue.
Hi Adriaan,
Which alternative did you find for the on/off switch?
Is it a pick and place type?
Well, this iteration of CLC filter is the winner I think. No more iterations needed. It has just a fraction of a dB rolloff at 20KHz. Flat to 20 Hz as well. Square wave looks ~ perfect. I used large capacity, low DCR shielded inductors this time as well.
PS- understand, this passive filter IS the D/A converter here in this re-design.
Itās got 2 OA/channel now. Also took some time and finally put in 3M material on the major digital chips. Included the (4) OA too, because now they are full digital as well.
I used solid pure silver wiring for all of it except the output side of the transformers (Audience pure Cu). Now⦠when do we try the Bypee purifiers I bought for it on the trafo output side? ![]()
Itās currently rendering āGrown Jazzā (Tidal mix) really, really well
Such clarity and accuracy, I love it.
In the next day or so weāll make some more checks and then get it buttoned up for tour duty. Weāre planning a gathering of some of the local gang at https://www.bluemarbleaudio.net/ for early/mid next week. He also needs me to fix one of his CDT ![]()
Should be interesting! I drug out the Terminator Plus again this week as well. I had added (56) local capacitors at each of the switches in the ladders (whew). Itās sounding really nice too, no question. Weāll bring it along on the tour session as well.
Turbo
So today I put the Bybee Cu bullets in the output to the XLRās. Iāve never tried these things before. Just replace a ~ 2" length of very good Cu wire and⦠Wow, I wasnāt mentally prepared for this amount of change to the sound.
I was hoping they might reduce the effective noise floor slightly. They seem to be doing what is claimed, and not slightly. It sounds much more āpresentā/fwd (but not harsh). It also perceives as ālouderā sounding, aligned with some above comments I think. The DAC went from politely sitting near the middle/back to straight up in front (near the Tplus).
Iāll let them break in awhile (do they break in?) and re-assess, try and see that they are not maybe too much of a good thing. But very interesting and unexpected, the level of change to the sound. As far as I know these are just a special metal oxide. Stick this in the learning bank.
Otherwise the new filter (D/A converter) setup x2/channel seems to check out fine. The L/R channel balance is near exact. Itās nearly flat to 20KHz. Thereās no sine wave (or square) peaking out to 40+KHz, using the actual preamp load (XLR). Donāt think the output transformer needs termination per se. All good.
Also today. took a plunge and ordered a ~ well done PCM1794 DAC (NOS mode). Weāll see about that, where it winds up in the pecking order vs TDA1541 (Abbas 2.4se arriving soon too), etc. But I do have high hopes for it. And likely some changes in mind (gears turning) ![]()
Turbo
A sample of it playing after Bybee install: The Griot- first tune with Bybees in place-WhatsApp Video 2022-05-30 at 7.20.34 PM.mp4 - Google Drive
What is a Bybee Cu bullet?
A kind of resistor?
Turbo, I would be hesitant to change the transfomers more than once, and would be very concerned about doing it three times. My experience with modding over the years has been that when you replace parts multiple times, that the traces begin to fall off of most boards no matter how careful you are with your desoldering work. So I may well give this a try eventually. But I will need to read some reports about what sort of sonic changes to expect with these over the better Edcors that I have already upgraded to (and absolutely LOVE the sound of) before Iām going to attempt yet another upgrade to my OTs.