Audio Interface Repair (SMD and QFN IC component questions)

[matrx10503]

I'm pulling my GS forum post here as the questions are better suited for the EE community.

I have a Focusrite 2i4 usb audio interface. The Focusrite Scarlett line shares these components I have listed in most of the models they offer.
Background Information/Issues/Symptoms:
The pad and phantom buttons (internal led) DO light up if I push them and the device is plugged into usb port.
The 1-2 input gain knob "halo/ring" shows no lights (should light up when input is detected from mic/instrument pluged in). And midi/usb activity light are very very faint, not bright/fully green/active (should be an obvious bright green)
The device is NEVER seen by windows (or mac) in the DevMan.

TL;DR:
So I suspected it had to be some type of USB controller/board issue.
I found some tear downs of these Scarlett's and found the internal components used.

Quote:
Chips:

XMOS XS1-L01A-TQ128- (GT1244L1 C5 PKH293.09) - Microcontroller (500MIPS, 32BIT, 64KB SRAM).
Cirrus Logic CS4272-CZZ YFAABXNJ1238 - 114 dB, 24-Bit, 192 kHz Stereo Codec.
AKM 4384ET 6R230 - 106dB 192kHz 24-Bit 2ch DAC.
JRC 4565 C096F - DUAL OPERATIONAL AMPLIFIER.
JRC 2122 C109H - ULTRA LOW NOISE DUAL OPERATIONAL AMPLIFIER.
ATMEL AT25DF041A ATMEL1113 - Serial interface flash memory device.
Texas Instruments TI HC4066 2BAQDCK G4 - Quadruple Bilateral Analog Switches.
SMSC USB3343 1511900 - Hi-Speed USB 2.0 Family Featuring RapidCharge.
CN PC900V - Photocoupler.
7519A G2KSZ
C1L95 Y72

After a lot of digging, I found 1 person who attempted to repair their interface.
This person replaced 3 chips. (The xmos controller chip, atmel serial interface flash chip, and the USB3343 transceiver chip.
They claim after replacing these chips....they have their Focusrite device functioning again!!


Now......I don't even have the proper equipment for this repair (bare bones experience).
I don't even have a DMM, I have an old analog multimeter.
So, I'm thinking if I can find a electronics repair shop and order the chips, I may can salvage mines.

Update:
I had found one more person (so now total of only 2 people have attempted this repair¿), But they only replaced the USB3343 chip, which makes sense, since this is the usb interface and is what interacts with the computer.
I do not know how to test the IC's as to know if I need all 3 chips, or just the 1. And from reading, I cant really test them with MM, I can just find a short maybe??

Here are pictures of the interface board and IC's in question:
http://img.photobucket.com/albums/v2...209_200201.jpg
http://img.photobucket.com/albums/v2...209_200211.jpg

Update 2:
Wanna-Be Soon-To-Completely-Fry-Anything-That's-Still-Good Amateur Electrician checking in...
Ok, as I said I don't have a DMM.
So I used my analog MM and set it to ohm mode. I grounded my ground lead to the rca jacks and started touching some of the IC chip legs.
So on the pretty much all the legs of the atmel chip, my meter I think is what you call..open-looping (like when you are touching the - and + leads together)
On the USB3343 chip same thing
Now, on the xmos microcontroller, when I run my + lead across the bottom legs, on 1 leg I can get the midi LED to come on solid green and another leg close by it, if I touch it, I can get the USB light to shine.
So I have to extrapolate for now that the xmos chip is at least good/working?
And I don't understand fully yet what the serial interface flash memory chip does, so I can't rule it out.
The USB3343 transceiver chip would be the only for sure suspect, but I would hate to find out it IS working and something else down in the chain is causing the issue instead.

So basically........what things can I do myself at this point?

[Ian.M]

Chuck it in the trash.  An analog multimeter outputs far too much current and voltage on its ohms or continuity ranges to be safe for probing circuits including nearly all ICs made in the last 50 years, so the odds are you've just totally FUBARed it.

[highlux]

They are $100 used.  Any electronics shop will have that as a minimum.   It's gooched.   No longer economically viable.

[matrx10503]

This is not a dire attempt to not buy another interface.
Its just for basic base knowledge. Regardless if I can fix things, I like to understand what the issue is to the furthest extent that I can figure out....as you can see.. to a fault.
Its costing me nothing but time right now....besides the fact that my brain knows 2 people in this world have repaired theirs and that makes me uneasy for some reason    (although I probably SHOULD be being productive and mixing right now!    )
At the most the chips will cost me maybe around $10 total (a piece) if I wanted to get them...I've looked up domestic and the usual CN sources.
The only real issue would be finding someone to do the job..aka solder the USB3343 QFN's....Which if I really want something done bad enough, i'm pretty good at figuring out how to make it happen.

Like one thing that is intriguing me...since I have selective knowledge.
I'm guessing since the analog meter is battery operated, its powering through the leads....
ok.. So when I touch the leg of the XMOS micro controller chip and get the USB indicator light to shine, and same with the midi light....
I understand that these lights would normally come on once the USB3343 transceiver chip probably sends "word" and confirms that's its connected to the computer....therefore sending the command to the XMOS chip telling it to turn on the led light on the board..which is the USB activity indicator.
So....am I bypassing the usb chip (or some other chip) and sending "word" manually to the xmos chip with the analog mm?

[Ian.M]

Effectively, all you've done is connected a battery and series resistor to the LED.  The only thing you have proved is the LED isn't blown, and the odds are you have damaged the XMOS chip in the process.   Most Microcontrollers are custom programmed with the code they run, so you need one from the original equipment manufacturer or salvaged from an identical board with a totally unrelated fault, and cant just get one from a component dealer or Chinese broker.

If you had taken a chance on replacing the USB3343, *BEFORE* probing with an inappropriate continuity tester, you might have got lucky, but unless you knew someone who did hot air SMD rework and had all the right equipment + owed you a favour so would do it off the books for beer money or coffee and donuts, it wouldn't have been economic.

Rough rule of thumb: if you can replace it for twice the parts and labour do so.  If you can 100% repair it for less than a third of the replacement cost, do that.

[matrx10503]

Effectively, all you've done is connected a battery and series resistor to the LED.  The only thing you have proved is the LED isn't blown, and the odds are you have damaged the XMOS chip in the process.   Most Microcontrollers are custom programmed with the code they run, so you need one from the original equipment manufacturer or salvaged from an identical board with a totally unrelated fault, and cant just get one from a component dealer or Chinese broker.

If you had taken a chance on replacing the USB3343, *BEFORE* probing with an inappropriate continuity tester, you might have got lucky, but unless you knew someone who did hot air SMD rework and had all the right equipment + owed you a favour so would do it off the books for beer money or coffee and donuts, it wouldn't have been economic.

Rough rule of thumb: if you can replace it for twice the parts and labour do so.  If you can 100% repair it for less than a third of the replacement cost, do that.

I concur (my fault of knowledge quest is not greater than economics). Thanks to all that replied.
Really, if I would have waited to try and get a response from the 2 people who repaired theirs.....i'm sure they could have answered most if not all of my questions.
I did jump the start a bit...I really should have at least had a DMM, which I knew.
So no harm really.
I've done the impossible a few times....a la Tom Cruise.  So i'm content with this learned-lose.

My last question, how would a technician properly test the IC's if they were troubleshooting an interface?
Is it commom-rule to replace the chips normally to determine if the previous chip had faults most of the time? Are there quick ways (with common store-buyable tools vs technical trained ways (with special measuring tools)?
Edit: I read a little on ICs and I see what you mean about programmers and such to test the MCU IC's. So manufacturer companies would have programmers to "test" the atmel flash chip and xmos type chips. IC resistors are the simpler beasts.
Good information to know.