Parts connexion have these for $500us incl power supply which is
far from cheap but at the same time not crazy money....
What did the total cost of Andrea's clock + PS + doubler + squarer etc end
up at?
TCD
Thanks Terry, I did not really count, but my project with all clocks and doublers, the Clock switch stuff I built and the power supply and Chassis, was getting close to 2k Euro - Eduard is not far off indeed !
to be honest, a lot of money, but it kept me nicely busy in the absence of going on vacations (Covid....)
I'm not sure if I understand you correctly. You mean the primary winding of the transformer needs a pair, don't you? I tried to draw a draft.
Attachments
That won’t work on this dac. Although it is desirable to have a low DCR for a current output dac that is being used to convert current to voltage using resistors only, the DDDAC 1794 is not operating in that mode. It has been specially designed and optimized by Doede to use the specified load resistors.
Using low DCR transformers alone will not work and could even damage the DAC.
I suggest you have a look at the dddac 1794 documentation where the unique operating mode is explained in detail.
Hi Ivan,
no doubt, basically this is what you do with a balanced tube amplifier. This will technically work, as long as the load is correct. Point is that the secondary load is critical for the result.
also the "load" the current source sees is the load x transfer function of the transformer. this will cause the signal at the output to swing around zero. I have no clue to be honest if the PCM1794 is liking this (in the long run)
The reason I am using Ra Rb as loads is to have a clean load for the current source and from there do voltage transfer for the transformer. It is a more straight forward case. Plus you can go out with capacitors.
Now if any one wants to experiment with a transformer where the primary winding is the only load for the current source, go ahead, but no guarantee from my side for the DDDAC.
just image a primary of 4H and an open secondary, than the load is the primary winding with a 1kHz signal the current source will see roughly 25kOhm. The current source would like to put 12mA signal in that load - this would clip dramatically. So you need to be very clear what you put as load on the secondary. Also this will be very transformer type dependent. So hard to make any general recommendation
At the end you could as well go back to using Ra and Rb?
I fully admit, I did not think this through in the very detail, just airing my thirst train of thoughts.
no doubt, basically this is what you do with a balanced tube amplifier. This will technically work, as long as the load is correct. Point is that the secondary load is critical for the result.
also the "load" the current source sees is the load x transfer function of the transformer. this will cause the signal at the output to swing around zero. I have no clue to be honest if the PCM1794 is liking this (in the long run)
The reason I am using Ra Rb as loads is to have a clean load for the current source and from there do voltage transfer for the transformer. It is a more straight forward case. Plus you can go out with capacitors.
Now if any one wants to experiment with a transformer where the primary winding is the only load for the current source, go ahead, but no guarantee from my side for the DDDAC.
just image a primary of 4H and an open secondary, than the load is the primary winding with a 1kHz signal the current source will see roughly 25kOhm. The current source would like to put 12mA signal in that load - this would clip dramatically. So you need to be very clear what you put as load on the secondary. Also this will be very transformer type dependent. So hard to make any general recommendation
At the end you could as well go back to using Ra and Rb?
I fully admit, I did not think this through in the very detail, just airing my thirst train of thoughts.
Hi Doede! Your doubts are pretty understandable. Moreover, I can say that I have tried this approach more than 3 years ago, but due to it was on PCM1792A (not PCM1794), I have waited 3 years for somebody who will take the (so-called) risk. So just imagine how patient can I be in terms of somebody's safety 
Well, let me try to answer your questions, partially at least.
All I can say is that the primary coils should be absolutely identical in terms of DCRs and reactive impedance (inductance). There should be no difference how large DCRs are, but SQ wasn't satisfied me at ~100 ohms DCRs (1+1:2). So I have moved to current trafos (1:18), so ~1 Ohm is great
Regarding loading resistor across the secondary. Well, it should adequately be transformed into primary coils by simple rule of square dependence. 1+1:2 - then 266 ohms across the secondary, 415 for 1+1:5,
2.66k for 1+1:20, and so on... to be at the same outut level.
Sure, I am a simple man too
Doede, I have offer. Do you want to try my transformers to ensure that you really do not missing a very good (suitable) approach? I can offer a needed pair for free. Just a honest feedback as a payback will be enough.
Well, let me try to answer your questions, partially at least.
Sure, the load seen by DAC itself is critical for the SQ, no doubts.
I suppose it is simple. Any current output type DAC (any current source) works best on to zero load. But... We still wants to grab the signal somehow, so absolute zero is impossible (in terms of practical transformers usage).
No doubts - it is pretty straight approach. No, really, it is clear technically. 4 boards gives you 36mARMS of the signal, it is a heaven in terms of signal power compared to single PCM1704 for example, which in 50! times is weaker. But trafos are still enjoybale in the audiophile games...
I suppose any DAC is limited by its internal impedance (circuit/PSU), so no way to get 300V from 12mA across 25k. But sure, better to ensure by self.
All I can say is that the primary coils should be absolutely identical in terms of DCRs and reactive impedance (inductance). There should be no difference how large DCRs are, but SQ wasn't satisfied me at ~100 ohms DCRs (1+1:2). So I have moved to current trafos (1:18), so ~1 Ohm is great
Regarding loading resistor across the secondary. Well, it should adequately be transformed into primary coils by simple rule of square dependence. 1+1:2 - then 266 ohms across the secondary, 415 for 1+1:5,
2.66k for 1+1:20, and so on... to be at the same outut level.
Well, I have no DDDAC boards by myself, but I can claim based on a lot of many other current-type-output DACs, that yes, better to use i/v resistors (Ra/Rb) across the primaries. Such a low level valued resistors (6-7 Ohms for 1+1:20), are really material/quality dependeced ones. Z-foils/niobium/tantalum/etc... really makes a significant enough changes in sound signature. So I decided to use my own manganin resistors. My humble opininon is not changeble over years, sorry: any current output-type DAC should works on to as small impedance as it possible (within the desireble at the end signal swing, sure).
Sure, I am a simple man too
Doede, I have offer. Do you want to try my transformers to ensure that you really do not missing a very good (suitable) approach? I can offer a needed pair for free. Just a honest feedback as a payback will be enough.
Last edited:
You just got to love this forum! I look forward to the feedback. I am very happy with my "vanilla" build though!
Doede, may I ask you where you have your TVC remote relay board from (Arduino based)? Is it open design? I am looking for a 23 relay remote board for my Silk transformers, currently operated by a Blore Edwards rotary switch. I prefer a proven solution over my own easyEDA brewings
Doede, may I ask you where you have your TVC remote relay board from (Arduino based)? Is it open design? I am looking for a 23 relay remote board for my Silk transformers, currently operated by a Blore Edwards rotary switch. I prefer a proven solution over my own easyEDA brewings
I have a couple PCM1794A chips which I always wanted to build a DAC with, but never got around to that. They were state of the art chips back then, but that was many years ago and things have changed. Now ESS and AK chips took the crown, seemingly. My question is to those of you who have listened to many different DACs: are PCM1794 still viable today? When properly implemented in a DAC (as in the complete device, not just the chip), how do they compare to modern ESS and AK chips, also properly implemented?
Does it make sense to spend time and effort on a 1794 design nowadays?
Does it make sense to spend time and effort on a 1794 design nowadays?
Really? A shortcut? That reminds me of Nelson Pass: "An ideal constant current source will emit a small lightning bolt which will travel until it connects to something." Good luck for your DAC chips...
You just got to love this forum! I look forward to the feedback. I am very happy with my "vanilla" build though!
Doede, may I ask you where you have your TVC remote relay board from (Arduino based)? Is it open design? I am looking for a 23 relay remote board for my Silk transformers, currently operated by a Blore Edwards rotary switch. I prefer a proven solution over my own easyEDA brewings
Hi Robin,
It was designed by me, as well written the code. It is very specific for my use, but if you are fit with Arduino code and a board without to much of a manual, I can get you the information, just email me
I might be biased, but the answer is a clear yes…. Comparing with ESS, which I did, is that the dddac implementation sounds quite analog whereas the ess is more analytic. It might be a matter of taste also what you like best. The far more than thousand dddac users worldwide typically like the natural analog sound print from the dddac.
VERY important, the above applies only when the chip is used in mono / non-digital filter mode ! All other implementations sound like a CD player.
Why do you think ESS and AK took the crown?
My experience is quite different. ESS never made it tick for me, the dddac rules over ESS with quite a margin.
And I made quite an effort trying to get the best out of ES9038Q2M in dual mono configuration and in synchronous clock mode (trying to bypass the horrendous internal PLL)
As I hated that there was nothing better than my simple dddac
In the end I ditched the ESS and kept the dddac untill finally I found a treasure of new old stock TDA1541A S1 dac chips.
In my opinion older is better......
Thank you for sharing your experiences!
I haven't listened to many DACs, certainly no well-made "audiophile" DACs. But some people whom I somewhat trust have a very strong opinion that new high-grade chips beats old multi-bit and delta-sigma stuff, when properly implemented. They praised AK 4497 and ESS9038 Pro (not so much the Q2M). Which is not to say they're right and everything else are wrong, but I respect these people for making thoughtful arguments and having an open mind, so I took note of their opinion.
P. S. You certainly have very interesting design decisions in the DDDAC, I don't doubt its sound is special!
I haven't listened to many DACs, certainly no well-made "audiophile" DACs. But some people whom I somewhat trust have a very strong opinion that new high-grade chips beats old multi-bit and delta-sigma stuff, when properly implemented. They praised AK 4497 and ESS9038 Pro (not so much the Q2M). Which is not to say they're right and everything else are wrong, but I respect these people for making thoughtful arguments and having an open mind, so I took note of their opinion.
P. S. You certainly have very interesting design decisions in the DDDAC, I don't doubt its sound is special!
Last edited: