Replacement for NJM3404AD dual op-amp?

[ChrisLody]

Hi everyone

I'm attempting to repair a Korg Poly 800 synthesizer and i think i have tracked the problem down to a dodgy op-amp which is a 3404AD which doesn't seem to be easily available, if available at all. I've been looking at datasheets for other dual op-amps but i'm self taught in electronics so most of the data goes over my head a bit.

Here is the 3404AD datasheet:

http://www.datasheetcatalog.org/data...3/233709_1.pdf

I'd prefer if the replacement were one of the op-amps listed on Rapid which is where i buy most of my parts. Here's there op-amps:

Integrated Circuits from Rapid

Any advice would be gratefully received.

Great site by the way. I've been getting more and more into electronics so a resource like this could come in very useful

Cheers
Chris.

[jrfrond]

Mouser has 1,506pcs. in stock.

NJM3404AD NJR Op Amps

[Enzo]

He appears not to be in the USA.

If only to see that the rest of the system works, I would suspect most any dial op amp would function there. From that list, I'd try a 5532 myself.

[ChrisLody]

Yeah i'm in the UK, should have probably mentioned that, sorry.

Thanks for the recommendation of the 5532 and the link to mouser as i've never been on there before. I could order from there if i put a big order together but i have a half completed order with Rapid at the moment so i'll add a 5532 to that and see what happens.

Thanks for the help, i'll let you know if i get it working.

Cheers
Chris.

[Steve Conner]

The 5532 is kind of high on bias current for a general purpose op-amp.

Also the datasheet specifically calls this one out as "single supply", which a 5532 certainly isn't. If it were me I would take a LM324 and break off one quarter.

Maybe the synth doesn't take advantage of the chip's single supply capability, though. And other than those extra 2V of common mode range, the chip is very similar to a 741, so you could substitute one of those if you can still get them.

[ChrisLody]

Thanks steve.

Yeah the op-amp does need to be single supply. I didn't know what that meant so i looked it up, the poly 800 supplies the chip with +5v and Gnd so the 5532 is no good. The 741 is no good either as its not a dual op-amp, only a single.

A bit further down on the same page is the LM358 which is a dual op-amp, seems to be single or dual supply and has the same pin out. Do you think that would work? I'm going to solder in a chip holder anyway so i can always swap for another alternative if that one's no good, and also to stop me killing the chip with heat as i put it in.

Is there any difference between an LM358 and an LM358AN? Different manufacturers maybe?

Here's the data for the LM358: http://www.rapidonline.com/netalogue/specs/82-0258.pdf
Cheers
Chris

[doctor]

You can use as a replacement the following chips without any problems

BA728
LM358
M5223
MC34072
MC33172
uPC842
uPC4742

Please pay attention on the chip name’s suffix following after digits – it means package type.

[Enzo]

5532 does draw more current than a TL072 or something, about twice as much if I recall. But that is usually only an issue when you try to replace a mixer full of 4558s with them. The few extra milliamps draw when replacing just one should not tax the system.

As far as I can see, they are not running single supply. I spot one running off +/-5, which I think all of them use.

The difference between LM358 and LM358A would be that the A version is a later revision. ANy difference or improvement in specs would likely be irrelevant to this application. And final letters in the IC designations are usually there to specify the package. That is more important now than it was when your sunth was made. Today we have to chose between DIP and surface mount and other things. Also the chips can be had in plastic or ceramic packages. Each type or shape has its own letter. In your case plastic or ceramic matters not, but you do want DIP packages and not SOIC.

Here is the Poly 800 manual with schematics.

[ChrisLody]

As far as I can see, they are not running single supply. I spot one running off +/-5, which I think all of them use.

My apologies Enzo, i just checked and you are right, it is +/-5v. Thanks for the service manual and the advice on package type too. Atleast now i've got a few chips to try. I can add a few to my order and swap them about until i get the best results.

At the moment i only have oscillator 2 getting to the output so i might swap the other working 3404 (for osc1) for which ever chip works best as a replacement, that way atleast there outputs will match when i'm in 'whole' mode which uses both oscillators for 8 notes of polyphony. Might not be necessay but worth thinking about atleast.

I knew this forum would be helpful. Thanks everyone

[J M Fahey]

*Very* simplified answer:
If it uses a split supply or uses a single one but at least 9V and is biased approximately halfway, you can use a regular dual Op Amp: TL072/82 , LM1458 , RC4558, LF353, etc.
If it uses a single supply, with very low voltage to boot, as in running from a single +5V supply, use an LM358.
What does that particular IC do?
FWIW in an emergency on late Saturday night I have replaced a TL072 with a lowly LM358 and it was ... usable, saved the night.
It wasn't a horrible sounding preamp, just somewhat less bright, but still usable, so ....
I told my customer to come back on Monday for a proper replacement, he never did, said "it works .... "

[ChrisLody]

Thanks for the advice, looks like there's plenty of IC's i can try.

What does that particular IC do?

There's a pair of them, one for each oscillator. The Poly 800 has octave divider on each oscillator giving 4 seperate octaves per oscillator. These can also be mixed at equal levels or at decreasing volumes giving a sawtooth like waveform. It looks like the dual op-amps buffer the output of each of the dividers mixing styles and passes them on to an IC which i guess is some sort of selector which can send one of the mixing types on to the filter. So with only one buffer working only oscillator 2 reaches the output. The output of the dividers is identical up until this point.

Fingers crossed i've worked it out right, i'm not sure what else it could be.

[J M Fahey]

Then by your description I guess you are talking about IC4 and IC5.
If so, an LM358 will work very well there, because
1) The maximum gain it needs is (4 x 82/33)=10, *easy*
2) You are mixing squarewaves to generate staircases, no place to worry about a (very) little crossover distortion.
3) It will probably will work better than a regular Op Amp which *might* have trouble with signal amplitude, both input and output wise.
My rule of thumb for regular op amps is: "do not count on the last 2 volts close to Rails"

Now a last question: are you sure the op amp itself is bad?
Maybe you have cracked solder or tracks and signal or power does no reach where it should.
Do you have right power on pins 4 and 8?
Does signal reach resistor arrays Br5 to Br8?
What do you scope on pins 1 and 7 at the Op Amps?
Thanks God one of them works, so you can compare what you find.
Good luck.

By the way, I have fond memories of those little Poly 800's.

[ChrisLody]

Then by your description I guess you are talking about IC4 and IC5.

Yeah, it would have been much easier to describe if i'd just said that wouldn't it?

My rule of thumb for regular op amps is: "do not count on the last 2 volts close to Rails"

So what's likely to happen with the last 2 volts? Distortion?

I'm pretty sure its a bad IC. I haven't got an oscilloscope but i've been testing the signal using a small battery powered guitar practice amp, and a multimeter where appropriate of course. The PCB tracks are in good condition and i've tested them all with a multimeter plus bridged them all temporarily to the appropriate places with no change, I have the right supply voltages to both chips at 4 and 8, I can hear the right octaves at all the inputs to the resistor arrays, the resistors surrounding the opamps are all working and connected properly. The only difference is that the healthy op-amp has a nice loud output coming out of 1 and 7 where as the one i suspect is dodgy has a weak sounding output from 1 and 7, even quieter than can be heard at the input to the resistor arrays. Also on the healthy op-amp i hear nothing when i monitor pins 2 and 6 (which i assume is it doing its job correctly as its a working component) where as on the dodgy op-amp i can hear the output of the resistor arrays clearly.

Glad to see you've had a play around with a Poly800 in the past. Mine's fun even in its current state, i'm just hoping this can restore it to it former glory. It was in this state when i found it. I suspect that the damage was done to whoever had opened it up before me. It has a mod which alter the battery compartment to hold a pp3 for memory backup and there was also some random looking extra connection the the solder side of the PCB which i removed without altering the synth at all. Poor thing.

[J M Fahey]

Also on the healthy op-amp i hear nothing when i monitor pins 2 and 6 (which i assume is it doing its job correctly as its a working component) where as on the dodgy op-amp i can hear the output of the resistor arrays clearly

Good deductive skills !
That's a "virtual earth" so you should hear practically nothing there ... with a good Op Amp.
The fact that it does not happen, plus lack of sound on pins 1 and 7 indicates a dead op amp.
It was also a good idea to use a signal tracer, the "poor man's scope" , to check functioning there.
I'd slap an LM358 there and call it a day.
I always have a few of them available, for general purpose "non quality audio" service, such as servos, full wave rectifiers, current or temperature sensing, you name it.
That's also why in some emergencies I had to press one of them into Audio duty, it's what I had available at that moment !!

[ChrisLody]

Good deductive skills !

Thanks

I was hoping to come on here tonight to tell how i'd got on, but my parts haven't turned up yet. I'm itching to get stuck in and hopefully fix this thing. I've also ordered the parts to do the 'moog slayer'and 'FM800' mods for it which i thought i might as well do while i've got the thing to bits. It'll be the closest thing to a proper analog synth in my keyboard arsenal.