Page 1 of 3


Posted: Mon Mar 26, 2007 8:16 pm
by Joe Malone
8 channel transformer to ADAT interface. The 8 audio transformers step down the +18dBM full scale inputs to match and feed the A/D chips directly so power needed is only one 5v rail. 24bit 44.1k and 48k with word clock in and out.


Posted: Tue Mar 27, 2007 5:49 am
by Lowfreq
Joe. You are my Hero :mrgreen:

Posted: Tue Mar 27, 2007 7:05 am
by Holger
Joe, do you see any market potential for
a) D/A conversion
b) doing the same with AES/EBU and with that
c) give 96kHz capabilities

Posted: Tue Mar 27, 2007 11:36 am
by astrovic
Bye bye ADA8000! Good one Joe :D

Posted: Tue Apr 03, 2007 5:08 am
by usekgb
Looks pretty sweet! Now we just need one that does A/D and D/A.


Posted: Fri Apr 06, 2007 2:22 am
by Moogus
Apparently these sound really good. Ive not used one to record yet myself, but the few people that have say theyre quite amazing. Simple idea isnt it...

Unfortunately D-A is another kettle of fish. You cant "cheat" like you can on the way in by simply padding the input signal down to match the few volts the A-D chip can swing. If you want good output drive up to and above +24 you need a powerful DOA per channel and possibly an output transformer aswell. Then of course for 8ch of that you need a much more elaborate PSU. Suddenly the box has gone from 1u and mostly empty to 2u and completely full and from ~$1000 to ~$5000.

Wed love to make one of course - it would be interesting to make a 'colored' D-A for analog mixing.

Also, doing an A-D with AES in and 96k capability is as simple as finding a card to do this. The analog side is all worked out.


Posted: Sat Apr 07, 2007 12:06 am
by LapTop006
Moogus wrote:Wed love to make one of course - it would be interesting to make a 'colored' D-A for analog mixing.
I *like* the idea of that just in stereo form which should be a lot cheaper.

Posted: Tue Apr 10, 2007 7:03 pm
by greenmanhumming
does anyone else think its odd to crank the level way up with a preamp, involving a transformer etc, and then knock it down again with another transformer?

if the A/D doesnt need high levels why do we bother with preamps with line level outs?

as far as I can see the A/D should aim for transparent and shouldnt have transformers in it. comments?

I'd also be interested in a 2 channel A/D

Posted: Tue Apr 10, 2007 8:39 pm
by Moogus

The reason we still do is partly a hangover from the days of analog recording and partly to do with wanting to design gear that can be used in any interfacing situation without the user having to think too hard about what theyre doing.

You want to be able to patch your gear together in any way the muse takes you and not have to worry about impedance matching or noise or other issues. Our industry serves artists after all. This isnt broadcast or telecom. Most of the gear we like the sound/feel of is completely redundant from a purely functional POV. In fact theres no functional reason to use any analog gear any more, with the possible exception of microphones and speakers.

So yes, its entirely possible to build a mic pre with a digital output that eliminates a lot of analog circuitry. BUT then what do you do if you want to plug that into a compressor or eq? Well unless you run very short and well shielded cables between the two devices youll need to boost the output level to get further from the noise floor and probably balance it to further reduce noise pickup.

The design reasons for using transformers A-D box are twofold - Firstly to make an A-D with *some* color (not a lot but some) based on the idea that tape machines sound good partly because theyre full of transformers. This is especially useful if youre mixing in the box, as everyone who has so far bought one of these boxes is. Remember theyre all being used with 002s.

Secondly, A-D chips are quite low Z in and need to be driven with something that can provide lots of current at low volts (only about 3v max PP). This is exactly what a stepdown transformer is best at, with no extra noise because theres no active circuitry. They also roll the top end just a hint which eases the load on the anti aliasing filters in the A-D chip.
Oh, and they also tend to have a 'soft' limiting effect when pushed close to the A-D's clip point, reducing bandwidth slightly and smoothing over the top end as you push them harder.

So really, a transformer in A-D has a lot going for it from a design POV.

Remember, the alternative is going thru an opamp that has anywhere up to 100 transistors doing the same work. If you think about it that way it doesnt seem quite as clear cut.

Also remember that this box is about $1000, so its not meant to compete with super expensive A-Ds. We have had one person say he likes it much better than his Apogee Rosetta though...


Posted: Tue Apr 10, 2007 9:33 pm
by chrisp
Feel like giving away the trade secrets? Like what values do you have for the pad resistors, and the load resisters/capacitors?

I understand if the answer is NO. But I ask because (1) I'm tempted to design a board to mount the transformers and get rid of that interface kludge; and (2) if I do it, it may be a 16 channel affair based on DB25 connectors.

Like I said, though, Joe had to spend time getting the load and level matching right, so I really would understand if you don't want to give that away.

Posted: Tue Apr 10, 2007 11:41 pm
by Moogus
You mean you cant tell from the photo?? :P

Were actually going to make just such a board, as the amount of wiring is silly for such a simple box. The pad/load values are easy enough to figure out but are partly done to taste - we did play with them for a while to get it to sound the way it does. See if Joe pipes up, otherwise Id assume the values are 'secret' for now.

What we really need is a different digital card to use, that has AES out and goes to 96k...and is cheap! Any suggestions let us know.


Posted: Wed May 23, 2007 11:58 am
by objectx
this looks like it uses the SM Pro Audio ADAT Lightpipe Output for PR8MKII card? I'd like to sign up as a beta tester for this one! s/pdif and AES option would be good.

Posted: Wed May 23, 2007 2:22 pm
by chrisp
Objectx, you're right with the card, but I strongly suspect that any magic in this comes from the right impedance and level matching undertaken by the tranformers and loadings, not from the card.

And from what I've heard, this is far past the beta stage - you can order this today from Joe if you want.

Posted: Thu May 24, 2007 8:08 pm
by Landmark
I'm sitting here with an unused Lynx AES16. So... If there was an aes/ebu option I'd order the a2d at once, probably together with a couple of BAs :)

A d2a with 2 channels or more with aes/ebu would also be very much appreciated.

Posted: Thu Aug 09, 2007 1:22 am
by living sounds
Is it possible to use the JLM14 instead of the Opamps with this DIY-Converter:

Or maybe the TI PCM4204EVM: