monica - where are we now and moving forward

[NicMac]

Hi Yeo and Paul H+N,

I'm happy to understand that you are cooking on a DAC based on ES9018. Uniting your expertises you should be able to come up with something really outstanding. I wanted to try the Mojo as output-stage for the Buffalo some time ago http://diyparadise.com/forum/index.php?topic=880.0 but I never got dow to actually try it.

What about making the I/V stage first? The buffalo II is very flexible and could easily be used for testing purposes.

There might even be a market for "modular" alternatives to the IVY III (and Legato) seeing how the Buffalo II sells out week after week......

Cheers,
Nic

[yeo]

hi nic

definitely will approach it slowly and make it modular. will take some time as i need time to digest the intricacies of getting this part to work.

i am still of the opinion that a good implementation beats all the specs in the world. but i'm open as always.

thx!

yeo

[Paul N]

Hi

This what I started to build, to try with my Buffalo DAC. Paul H's circuit, with the CFT installed. I'll need one per channel. I'm using two Mojo boards, modified to use just one diode/CCS string to power both halves. I'll try listening to just one output per channel, and also try converting the balanced output to single ended with a suitable op amp

The drawing's a mess I'm afraid- hopefully the circuit will be neater :-)

[ian444]

I don't mind drawing up a schematic now and then, hope this helps.

Ian.

[Paul N]

Hi Guys

I've built a working version of the circuit now- see attachments. It sounds superb- much better IMHO than IVY III, and better than Monica with the CFT mod

Note the 220R I/V resistor. This gives about the "right" output level. Larger values lead to distortion due to overload of the stage. The top CCS is now delivering about 50mA which is getting near its safe operating limit, but OK

I'm using the I/V circuit with the op amp based Bal to SE converter (this is just the circuit Twisted Pear use for this). This has proved to be essential. Unfortunately, if I omit this stage, and take a SE output from just the + (or -) out and ground, the signal is highly distorted and accompanied by birdie-like noises (the birdies you can get due to poor FM reception, not the garden variety!).

Paul

[Paul N]

Hi again

The construction is a bit Heath Robinson. I used an IVY III pcb, on which I installed the Bal to SE circuitry only. I built the L and R  I/V circuits on two Mojo boards, which I extensively modified (including sawing bits out of them!) These are mounted over the left and right sections of the IVY III board. I fitted the same long pinned headers to the IVY board that Twisted Pear used (actually, a little longer), so the Buffalo II board can connect in the same way as if using IVY III or the forthcoming Legato. So it's a three layer stack- four, if you count the Paul Hynes Z17 regs on top I use for Vacc

Paul

[kp93300]

Hi Paul,
Many thanks for the picture  and wish i have the skills and knowledge to build one like yours.
I thought the present version of Monica is already very good. There is a proverb in chinese that says that there is always a mountain higher than yours!!
Keep up the good work .

regards
kp93300

[yeo]

hi paul

thx again for the report and the wonderful effort you put in!

this tells me this:

1. a 20 year old dac chip with all the tweaks and mods, implemented to the max that we are capable of, can swing it against the best of the commercial offerings today.
2. even with the state-of-the-art, you got to mod/tweak to the max, to fully realize its capability.


hi guys

this brings me to a dilemma now.

the tda1545a dac chip costs around usd3. the sabre dac chip... usd60!

now being a manufacturer, i got to give warranty for what i sell. a minimum valuation (in order not to go bust) is at least 2x part cost (this isn't practical for a business to survive, let alone strive) and thus just looking at the cost of the dac chip alone, we are looking at minimum usd120. [unless they offer huge quantity discount but that will involve a huge outlay of cash first.]

and there's still pcb, component cost, regulators, engineering effort etc to contend with. and don't forget i need a HUGE outlay just to keep stock of this chip.

thus i'm not so sure should we go the path of this sabre dac.

now this ess company. i don't know what direction are they taking. at the end of the day, i don't know how they justify such a high price for their dac chip. is it any wonder there are so few commercial offerings using this dac chip?

i'm aware that in this industry, cost is seen as "no object" but who set it this way? isn't it us as consumer willingly and abidingly cordone this practice?

i used to laugh at "audiophools" spending hundreds even thousands on wires. i'm less inclined to laugh these days.

i think very soon folks will laugh at us for spending hundreds on just a dac chip!

penny for your thoughts?

yeo

[Paul_Hynes]

Hi Yeo,

I recall that you were reticent about the cost of my regulators when you first started selling them yet you continue to sell good numbers.

I think it is a good point to consider the motives of the DIY enthusiasts you sell to. These will vary but a general consensus could be reached with a site questionnaire. Some will place performance above cost others will consider cost to be part of the decision to buy. Some numbers will help with any decision.

There is no reason why you cannot do both DACs without holding huge stocks. Just cost the product accordingly. A fully tricked out Monica DAC is considerably cheaper than a fully tricked out Sabre DAC. The big plus for a Monica build is that it can be applied in stages which eases the financial constraints for those on limited cash flow and the end result can hold it’s own against the big players in the market. Also constructors learn a lot along the way. The Sabre can be treated in the same way, as a long-term project, to ease the cash flow. OK it has a higher entry price but that is the nature of the beast.

Further development work will be required to get the Sabre optimised before you can move it into a production unit. It is worth experimenting with one for a while to see where it takes you. Paul N is pushing the development and there will be others joining in as the project progresses. I do think you need to actually listen to a Sabre, with the same engineering treatment that you give to Monica, in your own system. You can then judge if it is worth your effort for your own peace of mind.

I hope this helps.

You owe me a penny. 

Regards
Paul

[weaver]

Another point of view:

as someone who tends to sit back and wait for you good people to do all the hard work, wait for it all to settle down, and then do a known good build - my take on cost is 'what can I re-use and what will I write off'.

Unlike many of the rather skilled folk here I can't be sure that once installed I will be able to extract a given component and re-use it (ie, there's a fair chance I'll trash it while trying to extract it.)

For example: I have a couple of Paul's regs currently on hold as I'd much prefer to install them on a revised Mojo board (if/when it arrives) rather than install them now then try and rip them out for a new project and ruin them in the process.

If there were to be a new output module now which could then be re-used pretty much in it's entirety with the new chip then that would make a lot of sense to me. As Paul H has suggested, being able to do things in steps has a definite appeal.

From a trouble shooting point of view, if I've got a module that I know works, adding an expensive dac chip to it is more palatable than doing the whole build from scratch. Put another way: what I'd be prepared to spend on a chip is proportional to how confident I am that I'll actually be able to get it to work.
If half the build is working already (with a tda) then I'm halfway there.

[ps. did I drop enough hints about a revised mojo sometime soon?]

[ole]

What makes Monica so sexy is simplicity and tweakability. The component count is rather low, and even for the lesser gifted like me it is still possible to grasp the principle well enough to attempt tweaking. The affordable price of the components is another big plus. It is just so much more satisfying to see a thoughtfully worked out circuit made up from outdated standard parts beat the hell out of highly touted top class gear.

The Sabre chip is a lot more complex and so is the required external circuitry if we want it to be better than Monica.

That said, a better DAC chip could still make sense if it comes in a modular package. Most people here would not want to screw up a rather expensive PCB with a non-removable SMD chip, so this board would have to contain everything that should/could not be tweaked, to ensure it's still reuseable when the rest is all hacked up.

Is there really no alternative to the expensive ESS Sabre chip?

Ole

[Paul N]

Hi Yeo

Just to be clear- IMO, when used with similar I/V circuitry, the ESS9018 sounds much better than the TDA1545A!

I’ve no idea really if the Sabre chip is overpriced. It’s a much more sophisticated chip, which , when properly set up, performs its basic tasks to a world leading standard, while offering up to eight channel output, I2C control, supposed jitter elimination, direct I2S and DSD inputs, etc; crucially it offers compatibility with digital sources up to 384kHz/32 bit!

Paul

[zsero]

I am a long time reader of this site, and at the moment I am looking into some nice sounding solution for a FLAC archive (either a WAC player or a USB DAC). I was reading about the QLS audio player and started reading the big thread about that player at DIYaudio, and there someone suggested pairing it with the 9022 from ESS, because it's about $2 per DAC for that chip. I have no understanding about the DAC designs, but don't you think you could make a nice implementation to a $2 DAC from ESS in a pure diyParadise style? If something is 30x times cheaper than the other ESS it might be worth trying it out what can it do. It might be a true follower of Monica (but again, I have no understanding about the inner workings of a DAC).

Here is a post mentioning it.
http://www.diyaudio.com/forums/digital-source/150199-qa-550-sd-card-16-44-1-wav-transport-12.html#post2167975

BTW, I think these times it would make more sense to make I2S-input only PCB, what anyone can use with a modded HiFace or a QLS player or anything depending on their choice. And sell separate add-ons boards for SPDIF or USB if needed.

[yeo]

hi guys

thanks for the comments. got few more pennies to give out though!

yes, i plan to do an updated mojo board. also a better spdif-eiaj convertor, and not to forget an i2s-eiaj convertor.

so that's quite a few things to do! i'll do those first before studying the ess seriously again.

i still feel that the ess chip is over-priced. why heck, if the qls player could give such high quality i2s signal, a dac that takes i2s is all we need. hmm... perhaps i could incorporate the i2s-eiaj to mojo as well?

yeo

[leo]

Did you ever look at the Philips TDA1387  Yeo?   its supposed to be basically a TDA1545 but with I2S input .  Maybe worth looking into if your wanting I2S as an optional input, saves the messing about with I2S to eiaj conversion