advertisement


It's Taylor DAC v2.0

Kit,
Kit Taylor wrote:
Does anyone know where to buy the SN75179BP buffer chip? I remember reading about how spdif signal boosting just before CS8412 could substantially improve the sound.[...]
you should consider opening Reichelt's UK branch. :)
I've found a chip with that exact name there, with no hint at what it does and no spec sheets or anything. It's 0.57 EUR ppc, so it might be worth trying.

Best regards,

Oliver
 
Finished one of the DDDAC boards on Tuesday (I'm building two of them, one for a friend). It's quite a 'straight' implementation, the electrolytics are all OsCon SPs with no bypassing, all resistors MF with 1% tolerance (one pair of carbon resistors in the I/V stage).

At first, with Pana FCs for coupling and no fine tuning done, it sounded like an old car radio. :)

Then I started playing around with the pots (VRef and supply voltage for 1543) and changed the caps for my favourite ones, Mundorf MCaps. Whoa! I set the SV to 8.0 Volts and tuned the VRef until I had about 3.9V DC at the output (hint by the board's designer, output DC should be just below half of SV), and the distortion went and the sound opened up considerably.

Apart from the DAC's attributes already mentioned by Kit (good dynamics, dry presentation, good bass) I also find it lacking some sizzling high frequency noise floor that the NAD generates no end; it is also more tuneful.
In fact, it sounds much more like a turntable IMHO. :)

It's 'darker' quality (-3dB @ 20KHz, starting at 5000Hz, apparently a result of the absence of digital filters, or so I've been told) matches the 32.5 very well.

I'll provide a nice case for this board, I have a feeling it's going to stay longer.

Best regards,

Oliver
 
Kit,

I guess I'm not going down the multiple DAC route yet, as the sound is already very solid with lots of slam using one chip (but this is only my current opinion and probably subject to change sometime :)).

It's just playing into a hicapped Nait3, a bit too rich with some recordings, but awesome fun with others. TGIF (free time to try some tweaks).

Best regards,

Oliver
 
It's just playing into a hicapped Nait3, a bit too rich with some recordings, but awesome fun with others

This is a good description of the sound I'm getting at the moment. I think I've finally found the most neutral speaker setup I'm going to get, and the sound is warm, dry and very physical. There's loads going on in the bass, which whilst soft and diffuse has excellent pitch and rhythmn.

Anything sparse and well recorded sounds awesome (eg the tracks with just voice and acoustic guitar on John Martyn's Solid Air), whilst busy rock music is effervesant but murky.

Everything is basically right. I just need to find the mod(s) to snap things into perfect focus without making the sound any brighter.
 
I've replaced the 1M carbon films in the I/V stage with MFs of the same rating; this restored some treble bite that is most welcome with rock music.

It is now connected via the good ol' Chord Crysalis, which always seemed somewhat midrange-focussed and a bit harsh to me; well, it matches best of all the cables I've tried so far.

* TNT DYI Piano 6 gave too much bass and seemed too dark (while it works very well with your average digital source)
* Phonosphie (black) interconnects give a hard-hitting, PA-like presentation, which is nice when you happen to own a CDP with a soft presentation, but sounds way OTT in this case
* I've yet to test DNM/Reson with bullets, but will have to do that with a different amp for obvious reasons (still using Nait3 for now, may switch to 32.5 in the afternoon).
 
I've done all the basic improvements to the current design and I'm not presently in the mood to build the more sophisticated power supply that I suppose is the next logical step.

So I've had a go with the old LCR 1uf coupling I was using before 220uf g-type OSCON.

It sounds good. Sweet, less grainy and generally more melodious, without the stridency it gave with the less tweaked versions of the dac. It has good dynamics but lacks the sheer physical oomph of the OSCON, and the output is noticably lower. I suppose the bigger caps makes for an easier load on the 1543.
 
Is this a more natural sound? I dunno.

I wonder if music is naturally "warm" and fairly soft and fuzzy, and if "clarity" isn't sometimes induced by a kind of distortion that makes things unaturally smooth and shiny.
 
No more tweaking necessary, this thing will get a proper box now (Farnell #939766). Implementation details:

PSU
I settled with a Panasonic LC type valve regulated lead-acid battery; the 'blackness' and 'naturalness' it introduces have to be heard to be believed. My guess is that this level can only be achieved with SRegs when using normal (mains based) PSUs.

There is some amount of decoupling capacity near the DAC board (2*1.000uF Pana FC, 1*4.7uF Pana FC, 1*0.01 uF Wima MKP10), which delivers enough current for demanding dynamics, and filters some HF the battery produces (which it does, but obviously not as bad as a mains based PSU).

Board and Parts
The board itself can be seen here [dddac.de]. I settled with 1% MF resistors and 22uF OS-CON SPs, good quality trim pots, WIMA caps, etc.; there is no real hifi boutique stuff to be found here. That comes in the next section. :)

Coupling cap & internal cabling
Lucky me I still had a couple of metres of DNM shielded SC cable, which seems to work well in this context. I decided to go for a really good coupling cap and tried Mundorf MCaps (0.47uF, as recommended by the board's designer), but found both the sonic signature and the amount of capacity lacking in this context. So I went for MCapZNs, as these are said to have a very low loss and 'quick' and lively sound. I tried 4.7uF and was immediately pleased. They inserted a certain crispness into the sound that made rock music fun again. At the same time the bass gained a bouncy, well-defined character. They fit so well that I cannot be bothered to try other caps now. :)

Comparison to internal DAC
This is the fun part. My NAD CD540 had its share of critical acclaim by the hifi press, yet I was never really happy with it. It sounds bland, has a weak bass and is overall too bright. When you open the box to look for the possible problem, you get to see a comparatively large toroid, lots of regulators, nice caps and all; yet the DAC chip is a PCM1716E, which is an "all-in-one"-DAC. I've heard several opinions about this chip, all of them pretty negative.

The new DAC sounds a tad quieter (passive I/V, 1543 running at 8V), so I have to turn up the volume slightly. It is still louder than the other sources (tuner, deck) though.
I got hooked on its 'naturalness' immediately; the total absence of any audible digital artefacts makes this the most analogue-sounding digital unit I've ever come across. In direct comparison the NAD produces lots of background sizzle, which makes me feel uncomfortable.

Then comes... the bass. The NOS1543 produces oodles of it, well-defined and well-timed. It sinks NAD here (honestly, the NAD is no match); it will reproduce everything, no matter what you throw at it. I discovered groovy bass lines in grungy music I had not heard before. Amazing.

Another aspect is the sheer presence of music. With the 1543 the music is _very_ present, images are stunningly solid.

On the negative side, the NOS1543 with passive I/V puts out an inverted signal. Some claim to be able to hear absolute phase, other don't. I don't mind really, as I've been told that CDs vary in terms of absolute phase, too.

The treble resolution could be somewhat better, but can successfully tackled by using Reson/DNM interconnects. This may not be enough for some, I guess, but in the context of my system it is sufficient.

Does it make music?
Yessir :)

On the whole, a DIY project worth doing. Thanks to everybody contributing his knowledge to this topic.

Best regards,

Oliver, happy camper
 
Great write up Oliver. I think I'll have a go with the DDDAC sometime. Battery sound does seem to have something to it, even if it might be technically not so good.

I've tried some more tweaks.

The loopfilter resistor has gone from 499R to 384R. More definition and impact, bit stiffer sounding. 100R the mathematically correct value to use with a 4.7uf cap.

I've also changed the decoupling caps on the CS8412 regs to the 4.7uf SMD ceramics used on the chips, soldered directly onto the pins right next to the reg body. Big, lush and dynamic, very present, with a kind of "dark brown" coloration. Sounds quite unrestrained though, perhaps a bit out of control.
 
As there's about a metre of cable between my regulated wallwart dc source and the filter cap feeding the three fixed regs on my dac, I was thinking of putting one of Avondale's dual LM317 regs in between.

I'm wondering if this would be overkill. Some say too many solid state devices make things grainy. I could remove the wallwart reg to compensate. Would the pre-reg have to be really close to the locals to be worth while?

Might it be better to place the local regs with Lm317? This would be a bit of a pain, as the board with solder pads I use is a bit fragile to desolder without damaging it.

TIA
 
Kit,
Kit Taylor wrote:
Oliver, how have you set up your battery to recharge etc?
quite straightforward really; the box will have two connectors, one for the battery, the other for the charger (sorry, German only). The charger has some built-in electronics to prevent overloading, etc.

They are connected to the DAC board via a nifty little switch I found at Reichelt's that is able to switch 4 poles. In that configuration the DAC and the charger are cleanly separated from each other, there are two states: 'charging' and 'converting'. ;)

The board has an additional part for an adjustable battery checking circuit, which allows to connect up to three LEDs to display the battery's remaining charge. I went for a dual LED (green/red) to display two states: 'normal' and 'to be charged soon'.

As I use only one 1543 chip, my 2.2Ah battery lasts for about 6hs (conservative estimation). Recharging from ~9V takes about 1.5hs. As these lead-acid batteries have no problems with memory effects (from what I've read they don't like being fully discharged and being stored half-charged for some time) handling is quite easy IMHO. :)

Best regards,

Oliver

Edit: how many typos can one person produce?
 
I've built and installed the Avondale dual LM317 pre reg between wallwart and dac. Its ouput is set to 9.68v, as curiously it wouldn't go any higher than this.

More fluid, softer and more controlled, subjectively less dynamic as a result. A voice's natural rattle and rasp sounds like part of the sound rather piggybacking noise. You may say that the bombast has been calmed and the cracks neatly filled in.

The reg is fitted with ordinary metal film resistors, and bypassed with 1uf and 10nf Wima film caps. I might change these, as sometimes they make things sound a bit "ringy."

The input is bypassed with the fine Rubycon ZL 3300uf 16v, a tall skinny cap that fits into tight spaces and sounds splendid. Too bad the superior ZA isn't available in this type.

I definitely want to to build a dac with snazzy regs now. I was thinking a single Avondale TPR feeding three ALW superregs. How much are these each? Would it be wise to put an inductor after each or serve them straight?

TIA
 
The reg is fitted with ordinary metal film resistors, and bypassed with 1uf and 10nf Wima film caps. I might change these, as sometimes they make things sound a bit "ringy."
Maybe try 1ohm in series with the cap, to damp the resulting LC tank a bit (i.e give the cap an ESR of 1ohm). Then make sure you have a nice big cap bypassing the 317s adjust pin - essential.

All serial regs have an effective output inductance that varies with output voltage; this interacts with low-ESR caps to produce a peak in the noise output, often in the midband... hence your 'ringing'. It's been mentioned before on this forum; there's a nice write-up here and a more technical Usenet post by Winfield Hill here . The latter is really good stuff.
 
It connects between the adjust pin and ground, across the lower voltage-set resistor. 10uF is the usual recommendation, bigger is fine - it should reduce LF noise even further :) A good quality electrolytic is the thing to use.
 
there's a nice write-up here and a more technical Usenet post by Winfield Hill here

That Winfield Hill quote is fantastic and should be etched in stone somewhere for all to see.

He's absolutely on the ball.

Andy.
 
I don't quite understand the Lm317 circuit as implemented by Avondale, but muddling through I've changed the output cap to a single Rubycon ZL 1000uf, and the adjust caps are Rubycon ZA 10uf.

This is a bit more like it :). Really thunderous and a little quieter subjectively, which I suppose means less distortion. The rig's main problem has always been raggedy reproduction of dirty metallic sounds like brass, slap bass and metal percussion, and whilst not flawless this is probably the safest sound yet.

I guess there's only so much that can be squeezed out of a dinky speaker like the Royd RR2, even by mighty Joe Ackroyd.
 


advertisement


Back
Top