Secret project (Svart AD/DA converters)

It's just that i feel the audio device world as a whole has resorted to boasting and what boils down to outright lies about things. You hear things like "just like a u87, but for 1/10 the price!" And its nothing like it. Now it's "mastering grade!", when that's just a marketing term.


I don't want to play that game. I want to bring the honesty back, even if it means losing a couple sales.

It's alive!

The white unit works as it should. One less worry. Now on to the next worry.

So I broke the unit, then fixed it.

Also, the ADA4898 parts do not get very warm. I measured 105F during full usage.  If someone else is getting hotter parts, I'm going to have to say that their implementation is wrong.

That's barely above the ambient for the board itself. The regulators get a tad warm at full usage, but are still well within the package temp specs.

More testing to happen today.

And now back on topic. I very appreciate your open discussion of the design decisions you made, svart. You may measure the difference between with and without input caps in practice to be sure there is no lowpass effect in the audio band, and if there is no harm at all i would say go for them. Better safe than a disclaimer, my personal opinion, YMMV. The strict avoidance of caps in the audio path may be a marketing argument people would swallow easily because they consider this "good" from reading other companies' buzzword advertizing claims, but i wouldn't give a damn if i know it is unhearable unless you are a dog. Sooner or later someone *will* probably feed dc into it, accidentally, or due to a failure of a feeding device, disclaimer or not...
Input caps may not be needed under normal circumstances, but failure tolerant solid gear is a very good marketing argument in itself. I consider this a good part of the "pro" in pro-audio...but maybe i am just a bit oldschool...

Got around to looking at the DAC board again today. Found a couple things to change.

Populated a second production board. Came right up. Need to test it out on audio now too.

Need to get back to the coding and work on the front panel design.

Things are a little slow right now since I'm also working on a project for the day job, but still on track.

Some mechanical updates.

Finally got my aluminum plates and the hardened steel drill guides..  So time to make the test chassis drill jig!



There are big drill guides for the rear panel holes, and smaller ones on the bottom for the holes for the press-in standoffs.  I'm still working out how to do the bigger cutouts for the power inlet on the back and the LCD on the front.



And they make this:



So far this makes perfect hole alignments for the bottom and back.

In other news, today I orderd 1200$ worth of parts to add to the 1500$ I already have. Now to finish some other things and then order 1000$ worth of PCBs in the next couple weeks.

Thankfully I also ordered a few reels of parts from ebay to save myself a couple hundred $$ and I found a couple old reels of parts at work that they graciously donated to me.

Which part?

I've stated multiple times (3 or 4 I think by now) in this thread and the other one, that the ADC board inputs have DC blocking capacitor footprints, but will be jumpered for the build.  Some have stated they'd rather see caps populated for safety's sake, and I typically agree with this sentiment, however I've also been asked a few times to keep it DC coupled.

So I'm not sure which way to go just yet, it's just that the board DOES have the places for the caps should myself or anyone else want them.  

I might even add it as an ordering option, although I want to keep those to a minimum.

DC and transformer cores do not mix at all unless they are gapped ala older Neve stuff like the 1272/1073/1084, etc.  When you put DC on a coil of copper wire around an iron (or other ferrous metal) what happens?  You make an electromagnet!  Since transformers work by fluctuating the magnetic flux between the coils through the core, if you create a steady mag field, it's simply going to start distorting more and more as you add DC bias.

As for resistors, I somewhat disagree with some of his theory.  Thin film resistors are a proven part that have very well documented attributes, like linearity, that tell me that they are perfectly usable, and low enough noise for the purpose. 

Besides the material, which has some noise addition (Johnson noise or also called..  resistor noise) is also a well understood phenomenon.  However, what ALSO matters is:

Gain..  Amp gain will bring up the noise of the resistors on the input and feedback.
Voltage through the resistor.. You want lower feedback voltages, so higher values are better, but are noisier.
Current through the resistor.. Lower values of resistor lessen the Johnson noise, but will increase the voltage induced noise.
temperature of the resistor.. Thermal noise is always there, colder is better.

And a load of other things that affect noise.

The truth is that I picked the best SMD resistors for the job and they are already purchased.  The design is set in stone at this point, and there isn't much use in arguing academics, as I mentioned before, which he did not bother to reply to either, yet you jump on my case!

But I digress.  

I appreciate his input, but in the end, his input is but one of many things I have to consider while doing this.

I also understand what he does for a business, and that business is based on selling people the idea that their gear will never be good enough without his upgrades.  I have no problem with this, but I also don't believe in it 100% and I take it with a grain of salt.


Also, if you must know, the resistors for the analog circuit are 25PPM 0.5% thin film parts.

EDIT:  to discuss caps and things like dielectric absorption is tantamount to arguing politics.  It goes nowhere.  DA has never once been shown to actually be the effect that determines the majority of the sound of a capacitor.  Many have set out to prove it, none have had any success with scientifically doing so and typically fall back to the "but it sounds different and that's what counts" argument once they've been defeated by their own attempts.  

I hate to sound harsh here, but I also do not believe that it's a pivotal attribute.  That being said, I also have some film caps to test in the feedback locations simply because it's easier to just do it than try to fix years of programming people have undergone with rational discussion.

A few updates.

Ordered the bulk of the front panel PCBs today.  Still need to do a few tweaks to the DAC board before sending it out.

Also, got the bottom/back drill jig finished up over the weekend.  Proved out the router theory I had and it works like a champ, just had to finish tweaking the setup.  Now I start on the front panel drill jig.  I need to order a few more drill guides and another aluminum plate for that.

So I made a few chassis as tests after the first one turned out perfect.  So far they are pretty perfectly repeatable thanks to the jig.

And a teaser:

I'm always #1...

But seriously, that chassis is the first "production" worthy one.  I have two previous ones that I used as test platforms for various things and are unsuitable for selling.

I'll be slowly getting them ready as I await the bulk of the parts for the ADC boards.  I'll be building those while waiting for the DAC boards, and so on.  Going to try a leapfrogging approach for the build so that I'm not trying to rush building the first ones at the same time as getting the software done and trying to start on the industrial design portion (front panel look, etc).

Thanks!

I'm actually quite surprised at how it's turning out.  I've designed chassis for work before, but I've never actually done the work on one, building the jigs and all that.

One thing I've done is draw it in CAD so I can print out the parts on paper and match them up before I ever touch the metal.  

Once it's time to cut/drill the jig parts, I simply cut the paper out and glue it directly to the metal as my guide.  Then I can just drill and cut on the lines and not worry about doing a bunch of measurements.

Once the jig is made all I have to do is drill and cut using the jig as the guide.

I can drill, cut and install the captured standoffs in about 8 minutes for the main chassis.  I expect that the whole thing will take about 15 minutes per chassis to drill and cut.  Assembly should be around 30 minutes per unit since I need to do other things like wiring and stuff.

Sure beats the hour it takes to stuff a single ADC board by hand, even using the nice tools at work.