Apogee Duet firewire IC chip weird behavior HELP

[pedalpower]

Hello to everyone.
I have a Apogee Duet Firewire. It developped wierd issue. All lights exept two on channel 2 are on when connected to Macbook pro. Duet doesn't recognized by Mac. Contacted Apogee they suggested to replace TSB41AB1 IC chip. I replaced it with no difference, still the same problem. Contacted Apogee for shematics, but they refuse to give it up, even the unit was discontinued and not supported by the company. I started to figuir out myself what might be wrong. Ane what I found was: it has another IC next to TSB41AB1, OXFW971-TQAG and when I heat it up with the solder the unit start normally and work with no issues even if I live it on and working for 24 hours. Or I can pluge the unit on, all lights are Up, leave it for 7- 10 menutes then reconnect and it started normaly. I thought the problem is that IC cheap. I replaced it with Brand New one but no luck. I purchased secon Brand new one still no luck. Measurred all voltages on another working Duet and they are different. Soldered back original one heat it up and it works. My question what might be wrong, now I don't think it is a IC. Any Help would be appreciated. If any one has a Schematics of a Duet will be hellpfut also.
Thanks.

 

[BradtheRad]

when I heat it up with the solder the unit start normally and work with no issues even if I live it on and working for 24 hours. Or I can pluge the unit on, all lights are Up, leave it for 7- 10 menutes then reconnect and it started normaly.

The unit start up normally when the IC is heated (either by solder gun or a few minutes of operation)? This sounds like a tiny physical connection is broken when cold. But then it is restored when heated.

Just to try anything that might work... Did you flex the board, press down on the IC's, pull up on them, etc.? Examine copper traces with a magnifier/ microscope? Examine for bad solder joints?

Or the problem might be inside the IC. A simple thing such as flexing the pins might cure an intermittent contact. Of course it is not easy to flex soldered pins, but anything is worth a try.

Also look for a connection which disrupts operation, caused only by cold temperature. This type of problem may be easier to spot visually.

Also check with your voltmeter at various points, both when the unit is cold and when it is warm. Write down volt levels. Look for any differences.

 

[pedalpower]

The unit start up normally when the IC is heated (either by solder gun or a few minutes of operation)? This sounds like a tiny physical connection is broken when cold. But then it is restored when heated.

Just to try anything that might work... Did you flex the board, press down on the IC's, pull up on them, etc.? Examine copper traces with a magnifier/ microscope? Examine for bad solder joints?

Or the problem might be inside the IC. A simple thing such as flexing the pins might cure an intermittent contact. Of course, it is not easy to flex soldered pins, but anything is worth a try.

Also look for a connection which disrupts operation, caused only by cold temperature. This type of problem may be easier to spot visually.

Also check with your voltmeter at various points, both when the unit is cold and when it is warm. Write down volt levels. Look for any differences.

I've tried everything possible what I came up with but nothing worked.
First I thought the chip itself was a problem, but after replacing it the problem still there. I checked all traces with the microscope and everything is fine. Looked at the voltage difference in working and not working unit. There is a difference in reading. Both brand new IC chips I soldered in didn't fix the problem but has the same reading, wich is different from working unit. I took the reading from the board without that IC chip and find out the are no voltages on some output legs until I head up that chip, then I restart the unit and all voltages are back and unit start normally. Replacing the chip twice did nothing. I did not try to head up the brand new chip. Only a difference is with new chip only three indicator lights are on but very dim. Measured the voltage on those LED's and found 3.2V on positive and 1.2v on a negative side of LED. on working unit there is no voltage on a negative side and 2.1v on a positive side. Maybe it will tell you something I have no knowledge of electronics but can perform very complex repairing task.
Thanks.

 

[BradtheRad]

until I head up that chip, then I restart the unit and all voltages are back and unit start normally.

If you are able to get normal operation at all, then you may have to be content with that, even if you need to heat it up each time you turn it on. You've done very well to remove and replace those devices, and still have a unit that works. It's risky to attempt to repair a commercial unit. One small mistake can ruin it.

Firewire operates within a certain voltage range. Do you read the correct voltage on your board?

Are there times when your Mac recognizes the Apogee? Open 'About this Mac' and look at Firewire parameters. See if your device is listed, and write down parameters. Compare these with parameters you see at other times, when the unit is working or not working, etc.

Does your computer have Firewire 400 or 800 capability?
Is your Apogee designed for Firewire 400 or 800?

Another thing that might be possible, a firmware update. Is this available for your unit?

 

[pedalpower]

If you are able to get the normal operation at all, then you may have to be content with that, even if you need to heat it up each time you turn it on. You've done very well to remove and replace those devices, and still have a unit that works. It's risky to attempt to repair a commercial unit. One small mistake can ruin it.

Firewire operates within a certain voltage range. Do you read the correct voltage on your board?

Are there times when your Mac recognizes the Apogee? Open 'About this Mac' and look at Firewire parameters. See if your device is listed, and write down parameters. Compare these with parameters you see at other times, when the unit is working or not working, etc.

Does your computer have Firewire 400 or 800 capability?
Is your Apogee designed for Firewire 400 or 800?

Another thing that might be possible, a firmware update. Is this available for your unit?

Of corse, I reinstall and install the Duet software before I performed the IC's swap. The Mac System info shows, Unknown device when the Duet doesn't start up normally, after heating up the IC the System info changed to " Duet"
and everything works. I compared the System info with my second working Duet and everything is identical except GUID: FireWire Bus:

Maximum Speed: Up to 800 Mb/sec

Duet:

Manufacturer: Apogee Electronics
Model: 0x1DDDD
GUID: 0x3DB0A000025FF
Maximum Speed: Up to 400 Mb/sec
Connection Speed: Up to 400 Mb/sec
Sub-units:
Duet Unit:
Unit Software Version: 0x10001
Unit Spec ID: 0xA02D
and :
FireWire Bus:

Maximum Speed: Up to 800 Mb/sec

Duet:

Manufacturer: Apogee Electronics
Model: 0x1DDDD
GUID: 0x3DB0A00008A55
Maximum Speed: Up to 400 Mb/sec
Connection Speed: Up to 400 Mb/sec
Sub-units:
Duet Unit:
Unit Software Version: 0x10001
Unit Spec ID: 0xA02D
On the faulty unit, I've found resistance on suspected IC chip legs 58 ohms to ground on working unit about 7 khom.
What do you think where to look to find the problem.
Thanks

 

[BradtheRad]

On the faulty unit, I've found resistance on suspected IC chip legs 58 ohms to ground on working unit about 7 khom.

Yes, that could be important. It implies that the pin has developed internal low resistance to ground (when it is cold, that is).
Perhaps this low resistance path has opened to other places as well.
It may also create a low resistance path to supply +.
Perhaps it affect voltages as you've seen on led's, etc.

Are you determined to try to make it work flawlessly? It's unlikely you can fix a problem inside the IC. Perhaps the best you can do is to mix in a bit of supply+ (through a high-ohm resistor to pin 58). Just enough to lift its voltage a tenth of a volt or two (for a start).

First as a safety measure, check resistance across the suspect pin, to the pin that gets supply+. If it is 58 ohms too, then this avenue may be too risky. It means 100 mA is going through the IC, which sounds high.

So be surprised if this cures anything. I'm only suggesting this because you mentioned the difference between the two IC's. If it does work then it is proof that your determination is strong enough to succeed.

 

[pedalpower]

Yes, that could be important. It implies that the pin has developed internal low resistance to ground (when it is cold, that is).
Perhaps this low resistance path has opened to other places as well.
It may also create a low resistance path to supply +.
Perhaps it affect voltages as you've seen on led's, etc.

Are you determined to try to make it work flawlessly? It's unlikely you can fix a problem inside the IC. Perhaps the best you can do is to mix in a bit of supply+ (through a high-ohm resistor to pin 58). Just enough to lift its voltage a tenth of a volt or two (for a start).
enough to make the IC
First as a safety measure, check resistance across the suspect pin, to the pin that gets supply+. If it is 58 ohms too, then this avenue may be too risky. It means 100 mA is going through the IC, which sounds high.

So be surprised if this cures anything. I'm only suggesting this because you mentioned the difference between the two IC's. If it does work then it is proof that your determination is strong enough to succeed.

Unfortunately, I can't find any info for OXFW971-TQAG IC, all that I've found out it is a 24bit AD converter with Firewire link also. I took all measurement from working and not working IC. Also, I took the measurement of the PC board without OXFW971-TQAG IC chip and I've found everything looks OK to me, all supply voltages looks OK. Unfortunately, I did not take a resistance measurement of the PCB with no OXFW971-TQAG on the board, to compare it .
As far as I understand I have to connect, let say 5 kohm resistor to pin with 58-ohm resistance and ground to elevate a voltage on that pin little bit, just enough to make the IC start.

 

[BradtheRad]

OXFW971-TQAG

Does OX stand for Oxford? I purchased an external firewire hd which I later found had an Oxford chip on the board. It lost all data when I accidentally disconnected power. (I got it back using a disk recovery program.) I did some online research and found out the Oxford chip was a major source of problems. A firmware upgrade was recommended.

As far as I understand I have to connect, let say 5 kohm resistor to pin with 58-ohm resistance and ground to elevate a voltage on that pin little bit, just enough to make the IC start.

That is what I would try, if it were me. It will be a big surprise if it works, of course. Remember it's a borderline problem. Don't apply a bigger change than you have to. This is the sort of experiment that can ruin a device. A 400 or 500 ohm resistor may carry too much current into that pin.

Measure volt level frequently, before a change, after a change, before it heats up, after it heats up, etc. You may discover that the unit works more reliably, therefore you want to know exactly what you did to cause it.

Of course 'I am not responsible' if it were to become more problematic.;-)

 

[pedalpower]

Does OX stand for Oxford? I purchased an external firewire hd which I later found had an Oxford chip on the board. It lost all data when I accidentally disconnected power. (I got it back using a disk recovery program.) I did some online research and found out the Oxford chip was a major source of problems. A firmware upgrade was recommended.



That is what I would try, if it were me. It will be a big surprise if it works, of course. Remember it's a borderline problem. Don't apply a bigger change than you have to. This is the sort of experiment that can ruin a device. A 400 or 500 ohm resistor may carry too much current into that pin.

Measure volt level frequently, before a change, after a change, before it heats up, after it heats up, etc. You may discover that the unit works more reliably, therefore you want to know exactly what you did to cause it.

Of course 'I am not responsible' if it were to become more problematic.;-)

Of course, you are not responsible for everything I will do to my unit. I still think if the chip is faulty why the problem persist even I've replaced it with Brand new ones two times. Very hard without schematics and IC pinout to guess, but today I discovered pins number: 3,22,30,46,50,71,83 and 97 connected together and received +3.276V The resistance of those pins and ground is about 42 ohms when the unit is not plugged in. As soon as I connect the Firewire cable resistance climbed to about 6-7 khom and stays until I disconnect the firewire cable. Then resistance starts to drop slowly to 42 ohms. It takes a while. Looks like some electrolytic capacitor discharges. Also, pin 100 has a resistance around 45 ohms but stay the same even if I force the unit to start with the heat. Pin number 100 dos not connected to those pins I've mentioned before.

 

[BradtheRad]

pins number: 3,22,30,46,50,71,83 and 97 connected together and received +3.276V The resistance of those pins and ground is about 42 ohms when the unit is not plugged in. As soon as I connect the Firewire cable resistance climbed to about 6-7 khom and stays until I disconnect the firewire cable. Then resistance starts to drop slowly to 42 ohms. It takes a while. Looks like some electrolytic capacitor discharges. Also, pin 100 has a resistance around 45 ohms but stay the same even if I force the unit to start with the heat. Pin number 100 dos not connected to those pins I've mentioned before.

I was going to say that an IC's pins usually have high resistance between them, and when I have measured low ohms between several pins, it's a sign the innards are fused. However you describe large swings in ohm readings. So perhaps 42 ohms is okay.

To say whether this behavior is normal, it would help to have detailed knowledge about both (a) the IC, and (b) Firewire protocol. It's not easy to get all the information you need. (Not from the manufacturer anyway.) To get an idea what's going on you'll need to take a lot of readings, and perform a lot of experiments. So far you've been lucky. You've been careful about soldering and de-soldering the IC. However the repeated heating could change its operating characteristics, and ruin it.

There is still the primary clue you have observed... that hot/cool causes internal changes inside the IC. A neighboring component might be involved too. Electrolytic capacitors are notorious for going bad unexpectedly. Many technicians just go ahead and replace them on a board, rather than test them.