The IsDAC

Decided to give this its own thread.

I currently have in mind a project to design and build a DAC based on the AKM AK4395VF DAC chip, which is the higher spec version of the one used in the SuperDAC.

In order to make this viable, I need other people to be willing to build one or buy a built one. I can't offer any performance guarantees, but with a few things up my sleeve I hope to be able to better the modified SuperDAC for a similar outlay (ie similar to buying a superDAC and modding it).

What do you get? I'm planning on designing and building PCBs. In terms of BOM cost, you can decide what you want to put on it in terms of op-amps and caps and regs, etc. The only fixed parts will be the DAC chip (AK4395), DIR (probably going to be an AK4117 for cost and availability reasons over the CS84xx chips) and the logic on the control side. The minimum "kit" that I will sell to people will be PCB, DAC chip, DIR chip and programmed PIC.

The question for those interested is what do you want? Do you want an AES/EBU input? How many coaxes? How many optical? Do you want XLR balanced outputs, or just single ended?

In self-build, there is no reason that you can't leave the components for balanced output out, saving the cost of op-amps, caps, sockets, etc, if you're not going to use it. It's just a question of what I lay out on the board. Same goes for the inputs, although I will probably only offer one logic set that will assume a full set of inputs.

From the self-builders, I need to know what you're prepared to deal with. It is my intention to make the DAC entirely surface mount except for electrolytic decoupling capacitors. Are people happy with that (Imo it gives access to a wider choice of quality op-amps, eg AD8065, AD825, various BB, etc) or would people prefer DIL?

So, before I embark on this, who would buy one and at what price? I believe that a full-featured DAC could be built for £250ish, dropping to around £200 if you leave out balanced connections.

If I can have 10 people interested then the project could easily go ahead. It could probably work with 5.

 

Iaasc, I put this on pink fish as well, they are cost concious bunch this sort of thing may well appeal to them .

 

Tony - I think you meant I'd put this on PFM aswell . Isaac, if you do so, you might want to mention BlueMax's little Yahoo group about the SuperDAC - that went down like a house on fire

Michael.

 

I'm not registered on PFM, and have quite enough forum registrations to my name...

I'm not too keen on getting millions of people as yet because timescales are flexible... ie I'm busy and it must fit around other stuff, so could be a few months...

 

So I never won any prizes for grammar !!!!,
Hopefully Iaasc will remove all rectification and regulation from the board, while he's drawing out the circuit board & do a fully isolated psu suppiles, this make a world of difference

 

Sounds interesting Isaac, I'm presuming the built ones would come with a box. Could be interested if I can sell some of my old gear

 

Yes. As tone suggested (and he knew I'd do it this way), it would be a two box affair. I favour cast aluminium as they're reasonably priced and offer reasonable shielding and good solidity. Both boxes would, however, be compact.

 

Isaac,
I may well be interested. I'm OK with SMD stuff aswell.
I dont use balanced as yet, but would like the opertunity in the future.

 

Well you never know I may even do a Quantum regulated never conneted psu for those after 'A 1062 quotaint', have some nice boxes too

 

I would possibly interested in this Isaac. Unfortunately smd is too much for me at my soldering level. So for me it would be a single ended pre-built affair if available. Btw - how good is the superdac? - just wondering what sort of sound could be expected from something like this.

 

The superDAC is good, sure enough. It responds very well to modification as the subject of much writing on this forum. Essentially it is constrained by cost, however, with cheap op-amps, basic caps, noisy regulators, etc. Modding only goes so far, and a clean-sheet design allows more room for growth. An all-SMD design will bring various benefits in lower noise and shorter signal paths.

Would you be interested in having the board made up for you, but do your own cutting/socketry on the casework?

 

Casework would not be a problem for me.

 

I want, but can't afford it definitely sounds an interesting project, and bettering the superdac would be excellent - not that the superdac is crap, IMHO the one you did for me isaac was damned good, but better is always good.

However the need to buy another cd transport and sort out/replace this trichord block means I can't give a solid assurance I'd buy the finished article.

If I was to buy one though, in terms of inputs it should ideally have 1 optical, 1 or 2 co-axial, AES/ABU, and both balanced and unbalanced outputs.

 

How do other people feel on inputs?

 

Im not that surprised that you would opt for SMD. Most of the latest and greatest chips of all sorts are increasingly only available as surface mount (as you mentioned earlier with op-amps). I dont mind drilling and assembling. For preference though i would probably like to avoid any soldering. But i might consider some if necessary.

EDIT: input for me would be co-ax.

 

How many? My current thinking is 1 AES/EBU, 2 coax, 1 TOSlink.

 

Your thinking on that is good. Thats a pretty comprehensive set of inputs.

 

why not I2S as well the 8414 will accept this input

 

Because no one will use it. The other problem is, from experience, I2S doesn't travel well. If it did, they wouldn't have bothered with SPDIF.

 

You may yet be surprised young Luke, feel the force