Modification to M-Audio Transit
May 27, 2004 at 2:32 PM Thread Starter Post #1 of 14

Garbz

Headphoneus Supremus
Joined
May 19, 2004
Posts
3,573
Likes
14
Hi all

This is copied form my headwize. I'm trying to reach a large audience.

I'm thinking of modding the M-Audio Transit. It's a beautiful little soundcard, and for $70USD it blew me away for quality. Out of curiosity I opened it up and saw 2 cheap and nasty electrolytic caps on the output. When i posted my review here we started to toss around the idea of modifications. As i don't want to run into power supply problems i've decided not to mess with the opamp. Afterall it gives outstanding line-out performance running from less than 5v.

For thoes people who are unsure what i'm talking about first a few background pics / info. This is the baby and an explicitly erotic picture too.
trans-1.jpg

trans-2.jpg


Basically it uses an AKM4584 audio codec which performs remarkably and is used in a few professional cards. THen it runs through an LMV722 amplifier and then into the output.

trans-3.jpg

trans-4.jpg


I'm not sure if this trace is correct. It appears to be running in unity gain from a single ended source, through an output cap pictured above.

The only problem might be as expressed here:
trans-5.jpg

There's not a lot of lead spaceing to substitute capacitors which may limit choice. I think that film caps may not fit.

Anyway to the details. I'm tossing around the idea of replacing the caps with something a little more linear and less abusive to the audio signal. Ideally this would be non-electrolytic but like i mentioned above i'm limited by the 10uf 25v requiremnets and the size of the leads.

However I'm sure that replcement ELNA or Black Gates will fit. Since I may be ordering a case from THL Audio in a few months Black Gates, ELNAs, and other niceities are within my reach without paying for postage
biggrin.gif


Black Gates are the first that popped into my mind because of firstly reputation and secondly CD player. It's a Harman Kardon HD7500 and although it's OLD (i.e. All discreet DAC old) It sounds wonderful with the exception of slighly inacuracies in the bass (but it was free so I can't complain). Anyway the 7500 uses Panasonic caps for all power requirements and Black Gates for anything that is sitting in the audio path.

What are you ideas on the mod. What caps can I choose from. Does the move from electrolytics to non-polarised matter? There is a SMD cap already linked to ground and a few resistors around there. Will they effect the way I can impliment this?
 
May 28, 2004 at 9:02 AM Post #3 of 14
use AD823 instead of LMV722 and place 2u2 or 4u7 N series Black Gates instead of those stock caps..
 
May 28, 2004 at 10:46 AM Post #4 of 14
that page is exactly where i got the idea from, we discussed it a while ago on headwize but I wasn't to psyched about the modificaiton back then.

I'm not in any mood to attempt to solder SMD components. I don't have a soldering iron small enough, and i'm not in the mood to wreck this yet.
rolleyes.gif


Would using a low value like 2u2 or 4u7 raise the low-pass level ? Headwizers recomended using a 22uf or something similar and dropping the voltage to compensate for size. Low values could create problems when driving low impednace loads.
 
May 28, 2004 at 12:47 PM Post #5 of 14
you use it to drive headphones directly? bad idea.. otherwise 2u2 is enough for line level signal..
 
May 28, 2004 at 1:17 PM Post #6 of 14
no not used to drive the headphones. But doesn't it depend on the amp that it's driving?

On the other forums an idea was to simply jumper the connection if i use an amp with an input capacitor. That would allow me to use a small 0.1uf input cap and a 200K resistor to ground after it without bass problems. (this would go in a gilmore)

Also to prevent caps in the signal path being true to the design of the gilmore I suppose that i could use multiple inputs and have only one set with the resistor / capacitor combination. This would theoretically allow me to completely remove electrolytics from the signal path right?
 
May 28, 2004 at 2:49 PM Post #7 of 14
you probably can't solve this without any cap in signal way, so either in Transit or inside the Gilmore there must be DC blocking cap.. in the amp there's more space, so that you can use some polypropylene caps which are almost undetectable soundwise..
 
May 28, 2004 at 3:23 PM Post #8 of 14
Quote:

Originally Posted by Garbz
no not used to drive the headphones. But doesn't it depend on the amp that it's driving?

On the other forums an idea was to simply jumper the connection if i use an amp with an input capacitor. That would allow me to use a small 0.1uf input cap and a 200K resistor to ground after it without bass problems. (this would go in a gilmore)

Also to prevent caps in the signal path being true to the design of the gilmore I suppose that i could use multiple inputs and have only one set with the resistor / capacitor combination. This would theoretically allow me to completely remove electrolytics from the signal path right?



You're correct on all three points. Glassman's statement that 2.2uF is good enough assumes that the amp impedance is greater than 10k, which is a safe assumption to make. I'd probably go a bit higher, but 2.2uF is fine. If the op-amp is running unity-gain, you can choose a cap rated as low as 6.3v. There is plenty of space for Black Gates, at least the smaller varieties anyway. You could probably fit the good N/NX/NH series in there.
 
May 28, 2004 at 11:54 PM Post #9 of 14
hmmm the gilmore option is out because of future upgrades. I inted to use the transit with an Electrocompaniet Integrated amp in the future.

So are quality electrolytics the only option left? Is it feasable to construct a box or mayb even a cable for the specific purpose of blocking DC?

I.e. Transit -> Box (polyproperline cap inside + 200K resistor to gnd) -> amp?

I can see the issue where extra interconnects will be used in this case. Now the tradeoff becomes Highquality Interconnects / Sockets or an electrolytic cap in the signal.

It doesn't worry me that a cap is in the signal it worries me that a cheap nasty and above all electrolytic cap is in the signal.
 
May 29, 2004 at 12:22 AM Post #10 of 14
Quote:

Originally Posted by Garbz
It doesn't worry me that a cap is in the signal it worries me that a cheap nasty and above all electrolytic cap is in the signal.


The stock caps in the Sonica were indeed cheap and nasty, and I'd guess the Transit would be the same, but BlackGate N/NH/NX series are not at all cheap and nasty. Their performance is similar to an average film cap. The presence of a cheap op-amp in the signal path will have more impact on the treble quality than a quality BlackGate.
 
May 29, 2004 at 6:28 AM Post #11 of 14
my words Wodgy..

to sum it up a bit:

6.3V Black Gate N/NX/NH series, 2u2 - 4u7 or more depending on space and availability
AD823 dual low voltage opamp, yet pretty good sounding one
 
May 29, 2004 at 11:18 PM Post #12 of 14
ok thanks guys i'll start with the caps and work my way up to an opamp when i get an iron capable of doing work that small.
 
Apr 30, 2005 at 9:53 AM Post #13 of 14
Ok, just rebumping this thread..

What values are the power supply caps? Maybe you should try replacing them too?
 
Apr 30, 2005 at 3:00 PM Post #14 of 14
Well the cap values are low and yes it would probably help filtering out the noise from the computer psu, however I doubt much more can be gained. I don't think the actual main chip has a SNR greater then ~98dB. For me it's no longer a problem, my DAC will be finished shortly.
 

Users who are viewing this thread

Back
Top