Digidesign Mbox 2 mods

[Sallivan]

Hi, after about a week of research, I decided to write this post because there aren't many informations on this audio interface.

I decided to open my Mbox 2 and  look inside it to find some possible modifications.
If anyone is interested and would like to make some suggestions it would be great.
I posted the internal pictures of the unit.

The AD converter are Cirrus Logic cs5361 KSZ and the DA is cs4392 KZZ.
Data transfer is cs8427.

The first thing I noticed is that the AD and Da can sample up to 192 Khz.. so why this limitation?
But, from what I have read in the web their are decent converter.

I think the worst part is the analog input and output stage.
The mic line pre's (and also the DI) are decoupled with Jamicon electrolytic capacitors (220 uf 100v, 47 uf 63 and 25 v), and they have a reputation of poor audio quality, and a small and cheap smd capacitors.
The opamp in the preamp stage is the jrc 2122 and ,I think, MAX4477.
In the DI section the only active part that I see is the K170 and one similar transistor but with 2 pins marked as AQ506R W6COO2A B0704.
The board is full of max 4477 opamps.
Near the output section I can see these max 4477 opamps no other.

It would be interesting bypassing the mic line and DI stage and go straight into the converter, but, for me, it is very difficult to trace the signal path.
An improvement that I would like to do is to change all the Jamicon capacitors with high quality ones.
Change the JRC 2122 mic line opamp with another one but I can't find one with the same specifications.
I've read that the MAX4477 is quite good as amplifier, but maybe it can be improved.
And the last mod I would like to do is to build a separate 5v for power the unit, instead of the one that comes from my laptop.

Could you help me by giving me your opinion?

Really thanks!!!

Pre amp Mic line stage

http://imageshack.us/photo/my-images/21/dsc7200g.jpg/

DI stage

http://imageshack.us/photo/my-images/203/dsc7193.jpg/

Converters

http://imageshack.us/photo/my-images/824/dsc7189.jpg/

 

[Rochey]

CS8427 is doing the S/PDIF interface, not the USB.

The limits below 192kHz are likely set due to the USB capability of the unit, not the converters.

 

[Winetree]

Can't help on a tech level, but I have 2 M-box 2s and will follow for any mods.
I know I purchased better op-amps for a mod but never opened it up.
Keep us posted.

 

[Sallivan]

I read the data of the AD and DA converter.
This is the pdf of the Cirrus Logic cs5361

http://www.cirrus.com/en/pubs/proDatasheet/CS5361_F2.pdf

At page 18 there is a circuit that they recommended  for the analog input buffer.

And that one is the data of the cs4392 kzz

http://datasheetz.com/data/Integrated%20Circuits%20(ICs)/Digital%20to%20Analog%20Converters/598-1065-5-datasheetz.html

At page 9 there is the analog circuit for the output ( it's not balanced) of the DAC.
There is only one thing that I don't understand in the DAC : if you see the block diagram at pages 7 and 8 it shows the output A pins 18/19 and mute output pin number 20, the same is for the output B.. but in the analog output scheme there isn't this mute pin.
In the data there is this explanation for the mute pin:
"Mute Control (Output) - The Mute Control pin goes high during power-up initialization, reset, muting, power-down or if the master clock to left/right clock frequency ratio is incorrect."


The two circuit are filtered for the high frequency.

The question is this: can I create one external board  whit both circuits so I can bypass the analog part of the Mbox and go straight into the pin of the ADC and from the DAC to the board?
In this way I can use high quality opamps and components  without touching the Mbox PCB ( smd !!!).
Is this a good concept or I have miss something?

Really thanks for the help!!!!!!

 

[Andy Peters]

The first thing I noticed is that the AD and Da can sample up to 192 Khz.. so why this limitation?

MBox 2 is a Full Speed (12 Mbps) USB device so that impresses a bandwidth limitation on the design.

-a

 

[Andy Peters]

The board is full of max 4477 opamps.
Near the output section I can see these max 4477 opamps no other.
I've read that the MAX4477 is quite good as amplifier, but maybe it can be improved.

Its main feature is that it has rail-to-rail input capability (maximizes input headroom) and rail-to-rail output for moderate loads.  A good substitution might be OPA2365 if Avid used the SO-8 package for the Maxim part.

-a

 

[Rochey]

I read the data of the AD and DA converter.
This is the pdf of the Cirrus Logic cs5361

http://www.cirrus.com/en/pubs/proDatasheet/CS5361_F2.pdf

At page 18 there is a circuit that they recommended  for the analog input buffer.

And that one is the data of the cs4392 kzz

http://datasheetz.com/data/Integrated%20Circuits%20(ICs)/Digital%20to%20Analog%20Converters/598-1065-5-datasheetz.html

At page 9 there is the analog circuit for the output ( it's not balanced) of the DAC.
There is only one thing that I don't understand in the DAC : if you see the block diagram at pages 7 and 8 it shows the output A pins 18/19 and mute output pin number 20, the same is for the output B.. but in the analog output scheme there isn't this mute pin.
In the data there is this explanation for the mute pin:
"Mute Control (Output) - The Mute Control pin goes high during power-up initialization, reset, muting, power-down or if the master clock to left/right clock frequency ratio is incorrect."


The two circuit are filtered for the high frequency.

The question is this: can I create one external board  whit both circuits so I can bypass the analog part of the Mbox and go straight into the pin of the ADC and from the DAC to the board?
In this way I can use high quality opamps and components  without touching the Mbox PCB ( smd !!!).
Is this a good concept or I have miss something?

Really thanks for the help!!!!!!

That is a global mute pin that you can use with an external mute circuit to avoid pops and clicks that you might get during those erronious moments.

To use external opamps etc, you'll need to desolder the old ones and solder wires out to your new board. Not impossible, but a major pain in the ass.

 

[Sallivan]

So the mute pin can be floating.
I found another data of the cdb4392 evaluation board.

http://www.cirrus.com/en/pubs/rdDatasheet/cs4392eb-1.pdf

Very interesting circuits from page 10 to 13. There are 2 balanced output sections and also a mute circuit.

 

[Sallivan]

This morning I looked more deeply the buffer stage in the front of the AD input pins.
I took photos of both analog parts (ad / da).
The opamp buffer before the converter has the feedback 619 ohm and the output 91 ohm resistors soldered under the pcb, as you can see in the picture. The capacitors are over the PCB near the opamp.
I drew the schematic ( hope that there are not errors!) and it is very similar to the circuit that the data of the cs5361 suggest.

The only 2 difference are:
1. In the original schematic of the datasheet there is a voltage (VQ from pin 22 of the AD chip) that pass trough two 10 kohm resistors and go to both + of the dual max4477. But in the Mbox there is a 15 Kohm between both + of the dual opamp, and there is no contact with VQ pin of the AD converter.

2. There are 2 resistor of 4k9 in the feedback of the dual opamp that there isn't in the suggestion of the datasheet.

Also  there is a 2.45 volt in every point of the buffer circuit... when I measured the voltage I had not connected the two small pcbs of the leds and the in/out connectors.

The output section seems to be very simple, because the output of DA chip go to 2 capacitors and than to the output buffer circuit.


I attached the pictures with the signal path (red) and the schematic for the  left channel.

I would like to try to desoldering the feedback resistors under the PCB so as to isolate all the analog part and connect the AD L and R to my board.
The same for the output, I would like to desoldering the 2 capacitors an go to the output circuit of my board.

Any suggest?

Input analog buffer and ad converter path (the r25 and r24 are the input 2k resistors). Note that I drew a 15k resistor which is located under the PCB

http://imageshack.us/photo/my-images/832/26022013132.jpg/

Bottom of the PCB (feedback resistors)

http://imageshack.us/photo/my-images/23/26022013134.jpg/

Output section from DA

http://imageshack.us/photo/my-images/27/26022013131.jpg/

 

[barclaycon]

Black Lion Audio used to do mods to Digidesign converters.
I think they modded all of the Mbox range at one time.
Might be worth asking what they did?

 

[Sallivan]

Yesterday I built the input buffer circuit with the OPA 2604 and High quality capacitors, just for one channel.
The schematic is the one in the datasheet of the AD chip, the only thing is that I've used 10K input impedance instead of 100k and a 4700pF instead 2700pF.
I desoldered the feedback resistors and the filter capacitors of the Max4477 opamp and connected the new protoboard to the input of the CS5361.

For the test I have used a pair of Shenneiser 414 (both very close) and 2-610 Uad preamplifier with the acoustic guitar.
At the same Pre level the board have approximately 6/8 dbs more then the Mbox line in ( if the pot is turn to minimum level).
So I made 2 recording:
-one whit the two different level,than I matched the volume in the daw and compare the two result
-one with the new board level down (of about 7 db) so the levels recording was the same.

It is incredible how the sound is changed, more details and more clean, it seem that the original Mbox have a "opaque veil" between the source and the converter.
Very impressive.

Next week I'll modify the channel number 2 input, and also I'll try to build the output stage with the OPA 2604 because I really like this opamp.
After that I will build a separate  power supply to avoid using the USB cable voltage supply.
In that way I think I'll have less noise and more headroom .

 

[Andy Peters]

Yesterday I built the input buffer circuit with the OPA 2604

again, note that OPA2604 doesn't have rail-to-rail input and output, so you sacrifice headroom on the input and max output level vs something like OPA2365 (which is also quieter and faster).

After that I will build a separate  power supply to avoid using the USB cable voltage supply.
In that way I think I'll have less noise and more headroom .

It may or may not be quieter, and it won't give you more headroom.  (How can it?)

You also have to manage separating the USB supply (which has to go to the micro) from the rest of the circuit. The design probably does have some kind of power switch which turns on the analog stuff after enumeration.