Tube DAC - opinions sought

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

I'm toying with building the Tube DAC detailed at and as featured on the home page of this site ( under "Latest Sites" - "Marc Heijligers DAC pages".

Has anyone either built or heard this design - it looks cool and has strong associations with the acclaimed Sheldon Stokes design - but I'm after some opinion on it before going ahead.



[Edited by peted on 05-10-2001 at 08:15 PM]
While I have not built or heard the Heijligers DAC you refer to, I did manage to read through the considerably lengthy design notes just recently and may be able to offer some opinion from my own research on DACs.

The Heijligers DAC embraces many of the best design principles and there is no doubt the digital section is second to none. Extreme care has been taken to ensure integrity of the digital signal to the d/a conversion, both in component choice and board layout. If you were fortune enough to be able to locate the HiFi News article of a DAC Ben Duncan designed some years ago, within is a detailed discussion of the various digital topologies used in design and their sonic signatures. Least to say, maintaining a perfect digital signal to the d/a conversion proved to be as important and as difficult as it's analog counterpart.

The analog section of the Heijligers DAC has very close parallels to the AudioNote. Although the AudioNote is 0x oversampling, much of what applies to the AudioNote analog section equally applies to this DAC. In fact Heijligers makes reference to having based this design upon his love of the AudioNote DACs. There is little doubt resistive i/v conversion and passive filtering offers a significant improvement in transparency and ambience over the more commonly used active i/v conversion. CLC configuration - as used in the Heijligers DAC - being superior to LCL configuration in this respect, probably due to the CLCs better immunity to RFI. However, the AudioNote DACs, passive filtered oversampling DACs I have heard, and I suspect equally the Heijligers DAC do not come without some significant trade offs.

* Bright and gritty recordings, and dull and lifeless recordings remain just that. ECM jazz recordings are rendered unlistenable. They really do border on the threshold of pain.

* Highly compressed recordings - techno, dance, some hip-hop, highly commercialised pop - appear to stress the DACs ability to cope and this is clearly audible. Yet those passive filtered DACs I have heard seem comfortable with highly dynamic recordings.

* Most important is that the AudioNote which uses a similar configured tubed output stage absolutely demands a high impedence load. 20k minimum, but many claim 40-50k more suited. Low impedence amps/preamps are most definitely out of the question.

There have been a few reviews of the AudioNote DACs and the Hawk DAC on the Web. These may help but remember the AudioNotes are 0x oversampling and so the Heijligers DAC will not provide the same dimensional soundstaging and reproduction of vocals for which the AudioNote is famous, and the Hawk uses a fifth-order passive filter from memory and those absolutely horrid Crystal d/a chips.

If you are unwilling or unable to take the plunge on the Heijligers DAC unheard, then I would strongly suggest you first listen to a commercial passively filtered DAC before making any decision to proceed. I and many others will tell you how absolutely magical AudioNotes and oversampling passively filtered DACs sound, but with caution I would add that their sound is not for everyone. Others may disagree with me, but depending upon the type of music you listen to and the quality of your recordings, you'll either love them or absolutely hate them
Hi plp, many thanks for your detailed reply - well I confess to being left a little confused by some of it, but can relate to one of your points - garbage in, garbage out. I own quite analytical Gallo Solo spheres driven by a Bride of Zen & Zen Revised - still in construction :( - so anything that emphasises even more an average recording doesn't sound like good sense. This tends to mean the majority of one's collection in practice. Perhaps this is being too harsh?

My knowledge of the subtleties of the subtleties of CLC Vs LCL is limited despite an old electronics degree. I'm intrigued by your comment "but remember the AudioNotes are 0x oversampling and so the Heijligers DAC will not provide the same dimensional soundstaging and reproduction of vocals for which the AudioNote is famous" - in no position to challenge, just trying to understand what you are saying if you will.

Anyway, I've a budget problem just now, so all this is hypothetical for the future.

Also, are we talking at such a high level of performance that either way, this is still a red hot DAC?

Is your own research on DACs professional or private?


Nice to see plp's philosophy based on our schematics. Let me give me some of my observations.

1. Non-oversampling
I have some objection against 'zero oversampling' as a word, and prefer non-oversampling instead. I haven't heard the newest Audio Note designs, so I can't comment on the sound of that. There is one owner of our design, who did a non-oversampling experiment, and reports superiour results. He designed a PCB which can be used to replace the SM5842, together with some minor scratching and SMD replacement.

2. CLC vs LCL
One of the main problems with the Ls is to find good Ls which have a wide operating frequency range (from 20Hz to about 5 MHz), and are not sensitive to RF (most are). Because the PCM is a current source, I would see no problems applying a C on its output directly (if it was a voltage source, I would think of it differently). I've done a LCL experiment, but the effects were hard to compare, because in my opinion fthe difference in filter components makes more difference than the filter topology.

3. Tube output stage
Audio Note uses SRPP designs extensively. In our opinion, an SRPP takes away a part of the dynamics, and sound dull compared to a common cathode stage. An advantage of anSRPP is a lower output impedance. W.r.t. the sound, a lot depends on the tube used. Currently, we're using a 6N1P, a kind of replacement for an E88CC (a change in the operating point should be applied for optimal results).

4. Listening to our DAC
If people want to listen to our DAC, they're always welcome to visit me if they are around the Netherlands. Just send me an e-mail, and we can make an appointment.

5. Preference
Although a lot of technical backup can be given, all our design decisions are preference based. In our experiments we've encountered some strange observations, where the side effects seemed to be stronger than the primary effect we were looking at (e.g. filter components vs filter topology). That means that copying the schematics, and trying to build this thing yourself with minor changes, will result in a different sounding design (which might be 'better' sounding).

Hope this gives you some usefull background info.

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.