Upgrading opamps in MOTU 1224 - ran into snag

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

dasbin

Well-known member
Joined
Jul 2, 2004
Messages
106
Location
Vancouver, Canada
My latest project has been to upgrade all the opamps in the Main Out stage in my MOTU 1224. I've swapped the filtering and I/V from the original JRC NJM4508's to OPA627's. I've swapped the output opamps from the same JRC's to OPA2134's. I think this should be a good combo, and both are relatively "compatible" chips, running well over a wide range of voltages. The MOTU is spitting out 17 volts each side at them - pretty high, but within limits.


Here's my problem.
The left channel works - sort of. It's full of garbage noise and distortion, but at least there's a bit of a signal coming out. However there's a 2.21K SMD resistor that went missing on that channel during the upgrade process, and I'm guessing that has something to do with it, so I'll reserve judgment on the left channel until I get a new one.
The right channel is being really goofy. It just gives me buzz. The first opamp in line (I'm assuming it handles filtering duties?) is extraordinarily hot to the touch (like, it could boil your skin if you left it on there) and measures some really weird random DC voltages on most of its pins - like 12V, 13V, 4V. I've taken the chip out, double-checked every connection, put it back in again, and checked every connection to the traces and board components. I've also checked every pin for shorts - nothing.
This opamp seems to be passing the voltages on to the next one in line (IV?) as it too measures these weird DC voltages.

I should mention that the other opamps are getting really freakin' hot too, nearly as hot as the one that's acting funny.
Any ideas?
 
Fried Motu?

Was it broke in the first place?

Re: new opamps, Are the pin outs and designations compatible with the original opamps?
 
[quote author="owel"]Fried Motu?

Was it broke in the first place? [/quote]

If you read my post, you'll see that I'm just upgrading the opamps. There was nothing wrong with it before, and I'm sure there's still nothing wrong with it, besides a problem with the new opamps.

Re: new opamps, Are the pin outs and designations compatible with the original opamps?

Yup.
 
Dasbin,

Hi! I was wondering how you were getting on with this project.

I am confused...you say you removed NJM4508. I don't know this number and I don't recall seeing this in my unit. I think you mean to say NJM4580 as you said in a previous thread.

Also, you say you used OPA627. I see that this was discussed in our previous thread on this topic. I did not bother to check this particular chip out at the time as you said:
Problems:
- OPA627 is far too expensive, ruling it out immediately.
- AD825 is a single opamp ONLY, so I need 2 each plus DIP adapter
So just now, I looked...OPA637 is the dual...OPA627 is a single
This is probably the source of the problem.

Please let us know what you find out.
Charlie
 
[quote author="SonsOfThunder"]So just now, I looked...OPA637 is the dual...OPA627 is a single This is probably the source of the problem.[/quote]
Where did you look? According to the datasheet I have, they are BOTH single op. amps.

Best regards,

Mikkel C. Simonsen
 
[quote author="SonsOfThunder"]Dasbin,

Hi! I was wondering how you were getting on with this project.
[/quote]
Slowly. I'm disenchanted about it at the moment. I hope it's fixable.
I've also .probably. got a new PLL chip on the way for the unit, and I'm going to experiment soon with upgrading the input ADC opamps, which will be tough considering they have to be happy with the single-supply 7V the MOTU gives them.

I am confused...you say you removed NJM4508. I don't know this number and I don't recall seeing this in my unit. I think you mean to say NJM4580 as you said in a previous thread.
You're right! Just a typo.

Also, you say you used OPA627. I see that this was discussed in our previous thread on this topic. I did not bother to check this particular chip out at the time as you said:
Problems:
- OPA627 is far too expensive, ruling it out immediately.
- AD825 is a single opamp ONLY, so I need 2 each plus DIP adapter
So just now, I looked...OPA637 is the dual...OPA627 is a single
This is probably the source of the problem.

First of all, regarding my comments about the expense of the chip (something like $47 per pair :shock: ), I managed to score 8 of them for free off Burr-Brown as samples :grin:
Secondly, I'm aware it's a single-channel chip. I'm using the two-single-channel-SOIC-to-dual-DIP adapter available at http://cimarrontechnology.com/
This isn't the problem. I've also double-checked all my connections on the adapter, BTW.
 
MCS,
Ooops, my fault, you are correct on the 637! At any rate, the object of the question is the 627... and the fact that dasbin is using an adaptor.

dasbin,
Now the picture is more clear knowing that there is an adaptor!
Don't get discouraged, but we must take a logical, step by step approach. This usualy involves making only one change at a time. Replacing ADC opamps seemed like a REAL challenge to me as well...that's why I haven't touched my units internals yet.
I see the slew rate on OPA627 is 55... What's the output look like on a scope? I suspect oscillation.

Peace!
Charlie
 
[quote author="SonsOfThunder"]MCS,
Ooops, my fault, you are correct on the 637! At any rate, the object of the question is the 627... and the fact that dasbin is using an adaptor.

dasbin,
Now the picture is more clear knowing that there is an adaptor!
Don't get discouraged, but we must take a logical, step by step approach. This usualy involves making only one change at a time. Replacing ADC opamps seemed like a REAL challenge to me as well...that's why I haven't touched my units internals yet.
I see the slew rate on OPA627 is 55... What's the output look like on a scope? I suspect oscillation.

Peace!
Charlie[/quote]

Thanks for your help!
I don't have a scope unfortunately. Is there any other way to check for oscillation?
Also do you think this would account for the extreme heat and DC voltages encountered? And why would it only happen on one channel? I thought oscillation was caused by low voltages, and the voltage it's getting is very high.
 
dasbin,

"Got 'scope?" :shock:

If not, can you remove the 627s and adaptor and replace with 2134? Also on the channel with the missing resistor...wouldn't expect much outta that channel until you get it back in place.

I had to cut the pins on the old chips...
You have just learned a good lesson about modding...one channel at a time...don't pull anything you can't replace unless you are sure about what you are putting in there! Cutting the pins isn't such a bad practice. A lot of thru-hole PCBs don't desolder easily (@RT products come to mind!) either. But you can buy some extra of the stock chips when you order the (hopeful) replacements. That way, if something goes wrong...its easier to go backward.

Sorry that the 627s seem to be giving you a fit. If I had known you were actually considering them, I might have actually looked at them, though I probably would have attempted to talk you out of those, knowing that you wanted "plug-and-play" as you indicated in the other thread.

HTH!
Charlie
 
dasbin, look for an inexpensive scope on eBay. I've seen some decent ones on there for not much bread.

You are a braver man than I, messing around with the MOTU. I hope you don't have any sessions coming up! :)
 
You can mount a leaded resistor inplace of an SMD if you are carefull.

You realy need a scope. I have not looked up the datasheets but I would guess the replacement opamps are higher current draw and have a bigger bandwith. This can/will cause problems with the power supply and osc due to PCB layout. Alot of times you just can't drop a "better" IC in a PCB because of power and layout.
 
For the sake of reference, this is what the "working" channel sounds like. If anyone knows what this kind of distortion is, and/or what it's caused by, I'd be happy to know.

http://www.appletonmusic.com/songs/opamps.mp3


I realized something else: the 627's current draw is way, way higher per amplifier than either the 2134 or the original JRC's. I guess this could have a lot to do with it. I'll try swapping them to 2134's soon.
 
It sounds like clipping.

Several probable causes...

1. Not enough voltage supplied to the chip, thus causing the early clipping... i.e. no headroom

or

2. In one mod that I did, (it sounded exactly like this during my initial power up), I found out there was oscillation going on. Using a bypass capacitor near the IC opamp fixed the problem.

Note: Like what others said, you'll need an oscilloscope. This is very important for this type of work. That way, you can "see" what you're hearing. It's also useful for signal tracing.... to figure out where in the circuit the signal went bad, or disappeared.

In my oscillation problem, I tried different capacitors and some cap values fixed the oscillation according to my ears but looking at the oscilloscope at high magnification showed there were still some going on. I finally nailed it down by picking the right capacitor value and the oscilloscope showing a clean output.
 
the 627's current draw is way, way higher per amplifier than either the 2134 or the original JRC's. I guess this could have a lot to do with it.
If you had swapped every opamp in the box for 627s, I might agree with the thought...though it's a possiblity, its rather unlikely that this extra draw on four opamps would take the whole power supply over the limit.

Let us know what the result of swapping to the 2134 is.

Peace!
Charlie
 
[quote author="dasbin"]
I realized something else: the 627's current draw is way, way higher per amplifier than either the 2134 or the original JRC's. I guess this could have a lot to do with it. I'll try swapping them to 2134's soon.[/quote]

sasbin,

As Gus also pointed out..I would suspect this to be your problem as well..;-)

Kind regards

Peter
 
IIRC, there was little or no decoupling across this PCB. Am I correct, dasbin? I would suspect this first and secondly, the adapter board (if it has no bypass/decoupling caps). Fast opamps like decoupling!

There's a giant pile of opamps in this box and I can't imagine that the power supply was designed so close to the limit that upgrading just a few of the opamps to a higher draw part (OPA627) would put you over. I could be wrong but just thought I'd clarify my thought process here. BTW, the quiescent current for OPA2134 is 5mA vs 6mA for the NJM4580, so changing those out should actually give you a little "breathing room".

I just checked the spec sheet for the 627 again...and it says Iq=7.5mA so I fail to understand how this part has a "current draw that is way, way higher".

HTH!
Charlie
 
[quote author="dasbin"]If anyone knows what this kind of distortion is, and/or what it's caused by, I'd be happy to know.[/quote]
Sounds like (in band) rail-to-rail oscillation, with the audio intermodulated. Yes, there's an element of clipping, but it's not that simple. There's a distortion even during the quiet bits. though you didn't give us a lengthy sample of 'silence' for reference, there are quiet nits that have full-level audi in them... Interestingly enough, the 'note' appears to be 120Hz...

Start with local decoupling. If that doesn't cure it, check for bridges/splashes. Scope the rails if you can. 120Hz in the USA means there's either too much juice being drawn, or there's a bad reservoir cap, but that second suspect is too much of a coincidence to be likely.

Either the chips are oscillating out-of band and pulling the average current drain up -thus loading the rails, or you've done something very unpleasant to one, other or both rails... With the complex intermodulation, it could very well be the former. A scope will almost certainly confirm that you've got silly ripple, I bet an hour's wages on it. -Anything getting hot?

I always counsel people to think twice before re-chipping. I hope you get the thing working again, but I reckon that 50% of all the 'improvements' heard from a number of the re-chipping jobs that I've seen have been self-induced imaginings... However, 100% of the unpleasant aspects have been attributed to the chips.

Keith
 
Back
Top