About DHT look for the cheap 4P1L or if you can afford for tubes with filaments of thoriated tungsten.
I no doubt quality of these two transformers brand but are very expensive and possible not needed depending final schematic.
I will look at kusy audio.
Also there is the possibility to use a PFGA as digital filter.
So several ways to skin the cat.
I did a similar (but on a smaller scale) experiment during the previous 2-3 years. Completely identical environment, power supply and USB/SPDIF boards, I just changed the PCBs with those vintage DAC chips. PCM56, AD1856, AD1860, PCM1702, AD1865, AD1862 and PCM63. PCM63 is the best for me, you just need to pay attention to a fairly bright sound, and accordingly a suitable I/V stage. I have PCM63P-K and P-Y chips, but they don't sound identical. One complete PCM63 DAC is waiting in a cardboard box for me to build some tube I/V stage, and pack everything in a decent metal box.
@elviukai
Same impresssions here, used CS8414-16 DF1704 PCM1704 Borbely I/V output stage with Borbely regs., Borbely PSU with chokes, teflon PCB. Also used Ian FPGA to use PCM1704 as NOS. Nothing changed the sound signature of PCM1704 only thing I still not tried is a valve output stage but experience said never will have the musicality of PCM63.
Also owned PCM63K and made the parallel trick, but disliked the sound of both.
I will go for DHT for PCM63 because I consider the output stage a very important part of the final sound like the PSU and regs.
Lately my experience with noDac to play DSD confirmed how important are the filters, output stage & PSU+regs. Here I'm using a NP BA-3 with R-Core and Salas BiB, the LPF is a passive 4th order L-C-L-C with teflon Russian FT-3 caps & Jantzen wax coils and yes the quality of components matter a lot because is reflected in the SQ I learned this when I made the 3 way crossover for my DIY loudspeaker.
Same impresssions here, used CS8414-16 DF1704 PCM1704 Borbely I/V output stage with Borbely regs., Borbely PSU with chokes, teflon PCB. Also used Ian FPGA to use PCM1704 as NOS. Nothing changed the sound signature of PCM1704 only thing I still not tried is a valve output stage but experience said never will have the musicality of PCM63.
Also owned PCM63K and made the parallel trick, but disliked the sound of both.
I will go for DHT for PCM63 because I consider the output stage a very important part of the final sound like the PSU and regs.
Lately my experience with noDac to play DSD confirmed how important are the filters, output stage & PSU+regs. Here I'm using a NP BA-3 with R-Core and Salas BiB, the LPF is a passive 4th order L-C-L-C with teflon Russian FT-3 caps & Jantzen wax coils and yes the quality of components matter a lot because is reflected in the SQ I learned this when I made the 3 way crossover for my DIY loudspeaker.
Keybord is fine, its just careless user who typing without even checking both his mind and screen for errors.
See thats the issue, one is bipolar and one is cmos. Very different regarding surrounding circuitry in terms of decoupling, reducing reflections and parazitics. No wonder why you have sound signature, in analog drinking milk with steak.
True, but those old ones are even less capable of dynamic load, much higher zout and psrr is quite bad. Simply worse for multibit dac chip, that requires very low zout in a regulator and highish psrr. It sounding better means you like sound that is on one way or another changed to your taste, and that is fine. But i general view it cannot sound better.
Indeed. There is always doing it proper, make it as best as possible regarding available technology, and in global it will be a better device. Now there is that subjective side, which just suits us as individuals and our pair of ears and little pink/gray mush sitting in between 😅 Best thing about diy is that we can tweak it to our liking.
Yeah i was looking at 4p1l, but lot of cases i heard is it is microphonic, and as an alternative 2p29l triode strapped. But not that much gain. Will have to spend more time regarding that, but many other things are priority atm unfortunatelly.
I have pcm63P-K, PCM63P-KY (16psc) and PCM63P-Y(24psc) I never realy compared AB them, but i think diferencies is not so big. I choosed Y ranking based on-
1) user description of sound in 2008 I am audiophile and are sold on great lyrical expressions UNTILL i hear results
2)Burbrown datasheet rankings
In paralel with this 24psc I am working on 2psc PCM63P-KY DAC which is not lsited in burbrown datasheets (each dac 4 pcm63P-KY)- they already finished 90% and working/tested since 2024. each PCM6P-KY is MSB adjusted at factory.
comutation of digital front end is quite complicated and at the same time smple- with disconectng some power suplies and few tricks DAC works(without opening lid) with 3 digital filters ( SM5803, sm5842 and PMD100) and with 2 diferent spdif receivers- CS8412, yamaha 3436(not 3623) all those 5 parts can function in diferent conections which eatch other. I am also keen to repalce cs 8412 with sany or akm spdif receivers- I like sanyo spdif receivers but need micocontroller.
I have feeling that 4x PCM63KY with MSB adustement and active IV in some aeas can outperform 24x pcm63-Y in passive mode (which on top of that will have superior Denon digital filter comapring to ancient pmd100/sm5842) but we will see. my main goal is use fully passive IV. and unlike most sugestions in 2014 that it will not work it works . even 8x pcm56 8mA total have enought voltage for my amps (with 330ohm resitor). with pcm63 I will have 48mA per channel so resitor values will as as low as 5-10ohms- nearly as opamps and lower than discrete stages.there is not much price to pay in this implementation- except dac chips and pcb size price. buts its irevelant in this project.
Attachments
Last edited:
Sorry, I didn't read everything from the beginning, so I don't know what was tried. The small Iout of the PCM63 does not seem to me to be favorable for passive I/V with parallel chip connections. Too many hard-to-find chips are needed for low output impedance and that chip doesn't like any voltage on Iout either. For passive I/V design, chips with higher Iout are better, for example PCM1794A, as Doede did.
As for these 78/79XX regulators, these new improved ON 78/79XX ACTG are good for analog circuits. I don't know if they also work so good for digital. My favorites are the LT1963/3015, the only thing that annoys me is the 20V limit on the input voltage of the LT1963.
As for these 78/79XX regulators, these new improved ON 78/79XX ACTG are good for analog circuits. I don't know if they also work so good for digital. My favorites are the LT1963/3015, the only thing that annoys me is the 20V limit on the input voltage of the LT1963.
Last edited:
All DHT is better to use small silent-blocks to avoid microphony, I use Rod Coleman regs. for filaments + silent blocks = never problems!!!
I made several DHT: 26, 4P1L & 801A.
Attachments
Last edited:
and that where is subjective parts chim in. upon caefull listening testa and comaprision I have remooved ALL LT1936 from all my designs. LT3015 were removed even eralier and replaced with LT1963 when I realized that LT1963 is better regualtor for negative voltage than LT3015.
some brands of some no low drop out reglator SONICALY outperforms LT1963 , Jung and other high end regulators. By saying sonicaly i mean just that nothing more. for me sonicaly is preserved transparenty and dynamycs.
Unless something goes teribly wrong with pcm63 (as i tried passive only on pcm56 dac arrays where i comapred passive R with vishay bulk foils vs sowter transformers vs with variuos opamps and vs variuos discrete IV solutions) 48mA output per chanel is enought for quality, for voltagae output and for driving cababilities of next amplification stage. there really no price to pay.
I liked the sound of the PCM56, I even liked the sound of the identical AD1856 better. You can play parallelization with it because it is available for ridiculous money. For PCM63, I barely got the ones I have.
As far as these LT regulators are concerned, I don't consider them to be the ultimate solution. They are excellent regulators but they have limits. They are certainly better sonically than the 78/79XX or LM317/337.
If you are chasing a subjectively better sound, I would say that the best way is passive I/V then tube gain stage or transformer I/V then tube gain stage or just tube I/V stage. In some future period, I am planning to try those variants. The problem is that there are too many of those tube I/V options and many good tubes. I'll need another year or two of those tests for sure. 2xPCM56 in parallel equals Iout of one PCM63, and I will experiment with that instead of the expensive and hard to find PCM63.
As far as these LT regulators are concerned, I don't consider them to be the ultimate solution. They are excellent regulators but they have limits. They are certainly better sonically than the 78/79XX or LM317/337.
If you are chasing a subjectively better sound, I would say that the best way is passive I/V then tube gain stage or transformer I/V then tube gain stage or just tube I/V stage. In some future period, I am planning to try those variants. The problem is that there are too many of those tube I/V options and many good tubes. I'll need another year or two of those tests for sure. 2xPCM56 in parallel equals Iout of one PCM63, and I will experiment with that instead of the expensive and hard to find PCM63.
Last edited:
You could try out the unexpected: use a very recent sigma delta chip (or very recent sigma delta DAC module) followed by a tube stage.
I heard that recently too. I was helping a friend with the digital inputs (USB/SPDIF/TOSLINK) of an AK4493 with DHT as output stage. The AK4493 itself is an average DAC, it doesn't stink or smell. However, the DHT output stage is a real game changer (with Rod Colleman regulators). Just an amazing experience. Only that particular one was too big and too expensive for me. And I can't lift easily those machines heavier than 15kg. 😁
I had the Eversolo Z8 DAC as a guest, for about a week. I've listened to everything, and it's a solid economy class DAC, but not boring like many others I've listened to. It has some weird output section with a bunch of OPA1612s. Who knows, maybe there is hope for sigma-delta. 😁
I had the Eversolo Z8 DAC as a guest, for about a week. I've listened to everything, and it's a solid economy class DAC, but not boring like many others I've listened to. It has some weird output section with a bunch of OPA1612s. Who knows, maybe there is hope for sigma-delta. 😁
Last edited:
No problem here, i have over 50pc pcm63
We cant comapre Analog Devices LT 1963 regulators with "the 78/79xx"because there is no such part or manufacturer. Only phyhical/exicting l things can be comapred . You need to name exact models, like my mentioned Signetic 5534 with Texas Instruments 5534.
I tested and did not liked any of solutions for pcm63 except OPA and carefully selected passive parts. It took many year and many brken dreams to understand this. My tube era ended in 2008,my siemens halske Cca and other tubes happily resting in drawers. My sowter trasnformers doing the same- I do not want to smear sound of pcm63 with those devices, actualy i want oposite.
Well, Signetics/Phillips is the only real 5532/5534. TI and other versions are not at that level. As for the 78/79 regulators, I only use it for auxiliary circuits, and I don't look at the manufacturer.
An amazing experience even. That is good to hear. You see, it is not just about the DAC chip or its topology but also what tube stage you connect after it to modify sound character to your likings. It makes matters easier with regards to cost and availability too.
Last edited:
I don't quite agree. Because both have an effect on the final sound. I confirmed this with tests as I have already mentioned. The DAC and the one before the DAC and the one behind the DAC and the power supply have a cumulative effect. Not to mention the wires now, or PCB design, or box, or PC/streamer etc.
Let's say the sound from Tidal and the same file played directly from the PC HDD are different sonically. Both are digital and have the same content. Those zeros and ones don't sound the same.
Let's say the sound from Tidal and the same file played directly from the PC HDD are different sonically. Both are digital and have the same content. Those zeros and ones don't sound the same.
Last edited:
Which one? And not the same as in worse/better or just different? It would be OK to test a few just like with all the old ones and mention specifics like is done when it is about the old ones.
Maybe you used the wrong tube 😉 With regards to getting the most agreeable sound character the devil is in the details it seems.
Maybe you used the wrong tube 😉 With regards to getting the most agreeable sound character the devil is in the details it seems.
Last edited: