8 Channels ADAT converter

[jdbakker]

However, you won't necessarily have an Apogee-quality clock or analog path. But it is Focusrite, be it Platinum.
Because the Octopre ADAT card is capable of external sync, you could still use a Big Ben or comparable master clock if you are worried about Focusrite's ability to deliver a good internal clock.

[...]

If you're concerned, just run external sync via wordclock.

There is no way that external sync can improve the jitter performance of a halfway decent free running crystal oscillator.

Do the math on your typical PLL, be it analog, digital or hybrid. Its noise floor (predominantly phase detector noise, but the loop filter can add plenty as well) lies higher than you get for a good XO, and that gets modulated onto your VCXO carrier. One of the few ways that external sync can match free-running phase noise is if you have a DPLL driving a really quiet DAC into the XO's frequency control port; flavor-of-the-month would appear to be sigma-delta so you can filter out any discrete frequency jumps due to finite DAC resolution.

JDB.
[now, it may be possible that you prefer the sound of your converter on external WC. That's jitter distortion you're hearing. Like transformer distortion there's no accounting for taste in that area, but if you measure it you will see that converter clock jitter has gone up. Unless, of course, your converter has an LC clock or any other low-Q tank, but not even Behringer use those for internal converter clocks]

 

[Skylar]

Thanks, JDB.

I'm not very knowledgeable on how clocking works from a very technical level, so your explanation here is quite helpful.

I guess what I'm saying is audibly, the Octopre's ADAT out sounds fine and measures fine (well, more than fine).

Somehow the idea that the Octopre ADAT board clock needed to be "fixed" cropped up, but I don't see anything wrong with it.


So, here's a side question:
If you have two pieces of digital gear and you know that the clock in piece A is better than the clock in piece B, you should use piece A as the master clock source, right?
In my case, I'm pretty certain the clock in the 1820M is better (less jitter) than the clock in the Octopre (according to Focusrite specs...I haven't measured it).
So, using the 1820M as the master may introduce more jitter into the Octopre ADAT board's clocking system, but the 1820M's jitter performance will be normal and pretty low.
Or should I use the Octopre ADAT as the master? That would yield better jitter performance from the Octopre but higher jitter from the 1820M; however, the increased jitter from the 1820M may still be much lower than the jitter from the Octopre.
Is any of that accurate?
If so, what's the best option...the one that sounds best? What if they sound the same?



Throughout the course of using the Octopre over the years, I've sporadically switched between ext and int sync from session to session, depending upon whether I had an E-MU sync card installed and/or BNC cable handy. I haven't heard a difference either way.
I'm going to try to measure the performance with RMAA when the Octopre board is externally synced.
I don't suspect there will be a difference.

 

[Whoops]

Hi,
thank you all for this interesting discussion.
It's a really important subject that sometimes is overlooked.

Skylar,
so basically , if I buy an Octopre Adat card, follow your pinout diagrams and do a Suitable PSU + case/XLR/Jacks I have an affordable 8 channel digital converter?
is that right?

So can you inject Line level signals directly to the card ? there's no need to step down the levels to match the A/D chips like Joe did on this page?

http://www.jlmaudio.com/phpbb3/viewtopic.php?f=2&t=22

thanks

 

[jdbakker]

If you have two pieces of digital gear and you know that the clock in piece A is better than the clock in piece B, you should use piece A as the master clock source, right?

As discussed recently on Bruno Putzeys' forum, you'll want the device with the best clock driving the device with the best PLL. If you're not sure which is which, use your best converter as clock master and have it convert your most important tracks (more specifically: those with the most HF content and the fastest transients). And yeah, "whatever sounds best" works well enough for most people.

Throughout the course of using the Octopre over the years, I've sporadically switched between ext and int sync from session to session, depending upon whether I had an E-MU sync card installed and/or BNC cable handy. I haven't heard a difference either way.
I'm going to try to measure the performance with RMAA when the Octopre board is externally synced.
I don't suspect there will be a difference.

I mostly rely on my own homegrown test software so I don't use RMAA that often, not sure how good it is at determining jitter. You could try a procedure like the one I outlined for the ADA8000.

JDB.

 

[blue_luke]

Hi all!  

I have disapeared somewhat from the discussion, but I am lurking and find the discussion going very well and extremely interesting.
I Have note shelved the project, but my second grandson was born last March 27th.
He is no problem whatsoever, but he requires quite a bit of attention from his mommy.
Meanwhile, the big one with all of his 2 1/2 years old requires quite a bit of attention from papy and mammy.

I just love being with him, but then, projects get side tracked!!  

But I am reading and learning....

Luc

 

[Svart]

If you have two pieces of digital gear and you know that the clock in piece A is better than the clock in piece B, you should use piece A as the master clock source, right?

Theoretically yes but realistically its not so easy to say yes or no, ESPECIALLY if it means needing external cabling.  Two words: Transmission line.

You can have the best clock on the planet but if your transmission lines aren't designed properly then you will get trash on the other end.  There are folks who do nothing else but design high speed transmission lines and make a good living doing it.  That's why JDB mentioned using differential clocking for ADC and DAC ICs. 

Skylar,  I design test and measurement equipment, specifically analog tuners for handheld testing units for the broadband over cable industry.  I don't directly design the sampling stack but my work intermingles with the digital guys so much that we both know what we are each doing.  We use high precision 14bit A/D ICs and we work in the mV ranges which means we can resolve single digit mV changes through some fun and interesting DSP stuff.  This means super clean designs are an absolute must have or the box doesn't work. 

Since I design PLL clocking circuits and PLL tuning circuits I can speak more to that.  We measure our jitter in single digit pS ranges.  I couldn't get an affordable XO for driving a PLL with a tight enough tolerance so I switched to driving a DDS with a differential LVDS TCXO.  The DDS is actually more steady than any XO that i've tested.. 

You could theoretically do something like this to get a "perfect" clock for real testing. 

Anyway, JDB has forgotten more than I'll probably ever know and I've discussed things like this with him before. 

 

[Skylar]

Hi,
thank you all for this interesting discussion.
It's a really important subject that sometimes is overlooked.

Skylar,
so basically , if I buy an Octopre Adat card, follow your pinout diagrams and do a Suitable PSU + case/XLR/Jacks I have an affordable 8 channel digital converter?
is that right?

So can you inject Line level signals directly to the card ? there's no need to step down the levels to match the A/D chips like Joe did on this page?

http://www.jlmaudio.com/phpbb3/viewtopic.php?f=2&t=22

thanks

I'm pretty sure it will work, but there is still a lot of testing that could be done.
So far, I've been able to successfully pass a line-level signal through channel one at 48kHz and 96kHz as well as emulate the dither control/LED circuit.
The RMAA results look good, but I have yet to do any REAL recording with this setup.
The sine tone sounds clean and the SNR/freq response looks clean, but, as I said, more investigation is needed.

I think I jumped the gun and got overly excited earlier, so I don't want to give any false information based on my assumptions at this stage.
You could get the Octopre ADAT card and use my pinouts as a good starting point, but I cannot make any promises!
Do so at your own risk!


AFTER-THOUGHT [as I'm writing this post]:
Listening to Gustav Holst's "The Planets" played into the Octopre ADAT card at line level, back into the 1820M via ADAT.
Sounds crystal clear—but only in mono right now—I only have one channel hooked up.
A/B'ing between playback straight from 1820M and the loop through the Octopre...I don't hear any problems, distortion, artifacts, noise, etc.



By the way, thanks Svart and JDB for sharing your knowledge.
I have some learning + reading to do.
And thanks for the links!

 

[Svart]

No problem. 

I suppose I sounded a little harsh and I apologize. 

Around the forum we get a lot of pseudo-science, you know the usual suspects: caps in audio chain always hurt your signal, must use XXX type of resistor or you can't make YYYY project sound good and so forth.  While I don't mean to single out your project/information, I felt that there would be a lot of people running out to buy some kind of ADAT converter card and either the project would be much too involved for them to handle or the real problem here would be that they would end up with a converter box that actually costs more to make into a competitive unit then if they just bought something off the shelf. 

Everyone who has designed/built something ends up investing more than just money into a product.  They invest their pride as well and I'll be the first to admit that my pride has gotten the best of me in situations where I would subconsciously ignore problems in something that I designed.  That's why we have peer reviews in the design world, so that someone else who has nothing invested can take a good honest look into your work and see all of it's warts and point them out.

That's how I feel about this project.  I'll help as much as I can with answering questions.  I'll be glad to review data and such as well but I can't offer a lot of design time due to my normal work commitments.

In closing, if it sounds good then it's a winner regardless of specs.  Not all of the best looking devices on paper sound good.  Just look at Class A preamps.  Hot, heavy, lots of power wasted, lots of harmonics but they generally sound "better".  On paper these things are a mess but in the euphonic world they are gold.  That's the subjective nature of audio devices.  I generally find the analytically designed devices to be rather cold and sterile.  In this way, sometimes bad things like a little jitter CAN help.  Strange eh?

 

[RAM]

What would the D/A side of this card be like then? Would you need to pass the signal through any kind of amp or balancing circuit to get balanced line level outputs?

Awesome work by the way Skylar! I'm a bit lost in the world of digital electronics but I'll learn slowly with you crazy guys around....

Rob

 

[Skylar]

No problem. 

I suppose I sounded a little harsh and I apologize. 

Around the forum we get a lot of pseudo-science, you know the usual suspects: caps in audio chain always hurt your signal, must use XXX type of resistor or you can't make YYYY project sound good and so forth.  While I don't mean to single out your project/information, I felt that there would be a lot of people running out to buy some kind of ADAT converter card and either the project would be much too involved for them to handle or the real problem here would be that they would end up with a converter box that actually costs more to make into a competitive unit then if they just bought something off the shelf. 

Everyone who has designed/built something ends up investing more than just money into a product.  They invest their pride as well and I'll be the first to admit that my pride has gotten the best of me in situations where I would subconsciously ignore problems in something that I designed.  That's why we have peer reviews in the design world, so that someone else who has nothing invested can take a good honest look into your work and see all of it's warts and point them out.

That's how I feel about this project.  I'll help as much as I can with answering questions.  I'll be glad to review data and such as well but I can't offer a lot of design time due to my normal work commitments.

In closing, if it sounds good then it's a winner regardless of specs.  Not all of the best looking devices on paper sound good.  Just look at Class A preamps.  Hot, heavy, lots of power wasted, lots of harmonics but they generally sound "better".  On paper these things are a mess but in the euphonic world they are gold.  That's the subjective nature of audio devices.  I generally find the analytically designed devices to be rather cold and sterile.  In this way, sometimes bad things like a little jitter CAN help.  Strange eh?

Great post, Svart!

What would the D/A side of this card be like then? Would you need to pass the signal through any kind of amp or balancing circuit to get balanced line level outputs?

Rob,
This card does not have D/A.
It's A/D only.

 

[stickjam]

Skylar,

I don't see links or images of any of the schematics.  Are they undergoing revision since you posted them? 

I'm eagerly looking forward to duplicating and refining on your success with the FR AD card.

You can email me directly using my forum name at comcast dot net.

Thanks

-Bob

Hi,
thank you all for this interesting discussion.
It's a really important subject that sometimes is overlooked.

Skylar,
so basically , if I buy an Octopre Adat card, follow your pinout diagrams and do a Suitable PSU + case/XLR/Jacks I have an affordable 8 channel digital converter?
is that right?

So can you inject Line level signals directly to the card ? there's no need to step down the levels to match the A/D chips like Joe did on this page?

http://www.jlmaudio.com/phpbb3/viewtopic.php?f=2&t=22

thanks

I'm pretty sure it will work, but there is still a lot of testing that could be done.
So far, I've been able to successfully pass a line-level signal through channel one at 48kHz and 96kHz as well as emulate the dither control/LED circuit.
The RMAA results look good, but I have yet to do any REAL recording with this setup.
The sine tone sounds clean and the SNR/freq response looks clean, but, as I said, more investigation is needed.

I think I jumped the gun and got overly excited earlier, so I don't want to give any false information based on my assumptions at this stage.
You could get the Octopre ADAT card and use my pinouts as a good starting point, but I cannot make any promises!
Do so at your own risk!


AFTER-THOUGHT [as I'm writing this post]:
Listening to Gustav Holst's "The Planets" played into the Octopre ADAT card at line level, back into the 1820M via ADAT.
Sounds crystal clear—but only in mono right now—I only have one channel hooked up.
A/B'ing between playback straight from 1820M and the loop through the Octopre...I don't hear any problems, distortion, artifacts, noise, etc.



By the way, thanks Svart and JDB for sharing your knowledge.
I have some learning + reading to do.
And thanks for the links!

 

[Whoops]

Skylar,

I don't see links or images of any of the schematics.  Are they undergoing revision since you posted them? 

I'm eagerly looking forward to duplicating and refining on your success with the FR AD card.

You can email me directly using my forum name at comcast dot net.

Thanks

-Bob

Bob there's no schematics from  Skylar, it's a Pinout diagram, that should be enough:

http://s73.photobucket.com/albums/i222/skylargray/octopre/OctoPre_pinout_20090428.gif


hope this helps

 

[Skylar]

Make sure you use the latest pinout though!

The one above is incorrect.


Here's the latest:
http://s73.photobucket.com/albums/i222/skylargray/octopre/OctoPre_pinout_20090501.gif


EDIT: THIS should quell any confusion!
Also, I still haven't tested J3 yet...so no promises there.

 

[stickjam]

My problem could be that my ISP is blocking photobucket.    Any chance you could post it on the equinox host?  Thanks!

Make sure you use the latest pinout though!

That one is incorrect.


Here's the latest:
http://s73.photobucket.com/albums/i222/skylargray/octopre/OctoPre_pinout_20090501.gif

 

[Skylar]

Wow, that sucks!

I'll just attach it to this post.

 

[stickjam]

Wow, that sucks!

I'll just attach it to this post.

Cool!  Octo1290AD here we come!  ;D

 

[Skylar]

Extra info:

It seems the caps are indeed de-bounce caps.
The cap I used was a 2.2uF, and it de-bounced the switch fine...in fact that value is probably too high because if you tap the switch REALLY fast, the µController that handles the dither switching doesn't receive the switch press, as the LEDs do not change. A liiiiittle too much de-bounce.

So, a better value may be 1uF, but I recommend experimentation.
2.2uF works fine though.

 

[ruairioflaherty]

Throughout the course of using the Octopre over the years, I've sporadically switched between ext and int sync from session to session, depending upon whether I had an E-MU sync card installed and/or BNC cable handy. I haven't heard a difference either way.
I'm going to try to measure the performance with RMAA when the Octopre board is externally synced.
I don't suspect there will be a difference.

Perhaps I have a faulty unit but I've never had any luck clocking my Octopre card using the word clock input.  I get clicks when I try to clock from my Digidesign 192 io.  I've always used the internal clock as the master.

As for the quality of the unit's converters I have no real opinion as I have never used them separately to the Octopre pre amps which are decidedly average.  In general the Octopre is used when I am short channels for things like toms or hats etc and when I do I usually use the line outs into my 192 io.  I do this to avoid clocking the 192 io to the Octopre.

I have zero time in the next 2 weeks to help but if after that point I can do anything to help further the project along just let me know.

Cheers,
Ruairi

 

[walter]

Here is yet another option, the Behringer ADT1616. I hope the image link works. I did a quick search of each part for this list. This seems to be a simple circuit and may help myself as well as others understand how this AD/DA conversion works. With a 50-pin connector it allows 16 channels in and out I am assuming analog signal for AD/DA conversion to two sets of ADAT lightpipe. Alesis or Wavefront convertors. I'm curious as to what is on the PIC IC. Made as an add on expansion card for the DDX3216, a 16 buss digital board with two expansion slots. Too bad they didn't make a firewire card, that's another project.

2)  AL1402 ADAT optical decoders, 2) AL1401 ADAT encoders, 2) HC959 8-bit serial-input/serial or parallel output shift registers, 1)74HC174 positive edge trigger,  1) 74HCT04 Hex inverter, 8) LVT244 octal buffer/line drivers, 1) PIC16C544 14BIT programmable device, 1) raltron 4.000-20 crystal
http://good-times.webshots.com/photo/2234412930094727889sHFhsW

 

[bommaren]

This might seem a bit out of topic but I hope Skylar will give a quick answer with his
history at E-MU.

Can a e-mu 1616 be used as a adat client to a 1616m?
If so This would for me be a resonable way to solve my problem.

This type of info is very hard to find out from E-MU.

- Bommaren