Roland Cube 60D Humming

[Renate]

As I explained above, there is no "+/- 30V supply".
There is a +/- 40V supply and the minus leg is being dragged down to 30V.
That's because you're supplying 60 Watts of power, 1/2 being dissipated in the amplifier IC, half in the speaker.

100 ohms is fine by me.
You'll see the minus supply go to the correct -40VDC.
Then you can try the headphones.
The headphones are not going to work correctly until you fix the -40V which is screwing up the mute.

[paul_g_787]

As I explained above, there is no "+/- 30V supply".
There is a +/- 40V supply and the minus leg is being dragged down to 30V.
That's because you're supplying 60 Watts of power, 1/2 being dissipated in the amplifier IC, half in the speaker.

100 ohms is fine by me.
You'll see the minus supply go to the correct -40VDC.
Then you can try the headphones.
The headphones are not going to work correctly until you fix the -40V which is screwing up the mute.

That makes more sense! Thanks. Will try that and post back.

According to the datasheet. the LM3886 (Power amp) is has these features:
50W cont. avg. output power into 8Ω at V CC = ± 35V
|V+| + |V-| max of 84V / min of 24V
Output protection from a short to ground or to the supplies via internal current limiting circuitry
Output over-voltage protection against transients from inductive loads
Supply under-voltage protection, not allowing internal biasing to occur when |V EE | + |V CC | ≤ 12V

[paul_g_787]

Right some progress! Please see attached v3 diagram with the new voltages.

With the light load connected, the voltages are much more stable on the power rails and on the mute lines which now appear to be working fine.

I tried the headphones and this is what happens measuing the voltage on the MUTE line and listening to the headphones.

• At power on it is immediately 11.6V
• The headphones hiss slightly for approx 1.5 seconds
• Then the headphones start to hum at 50Hz. Immediately the hum starts to gt louder as the voltage starts to drop.
• Once the voltage drops to -4.4, the hum immediately stops and the headphones are silent.

[Renate]

So, it's all non-muted now.
Have you tried to play guitar to the headphones?

[paul_g_787]

Yes all non muted.
Although when the amp is turned on, but before the mute turns off, there is 35VDC across the input jack. Also there is 9vdc across the tuner output and 76vdc across the line-out.
Once the amp un-mutes these all go to 0V.

Then I tried a guitar input, I have a sound from headphones although the sound is very distorted even on clean with all effects off.
The tuner and line out outputs sound fine, perfectly clear output and the effects all work beautifully.

[Renate]

Well, as long as the line out is clean we can skip the headphones for now.
(Have you used the headphones previously and did it work fine?)

With the 100 ohm still in place, what's the voltage on the hot end of it? (I'd guess -20VDC or so)?
Find the junction of R15, C11 and pin 8 of IC6.
Check the voltage, it should be minus a couple of volts.
Ground that point and see if the voltage on the hot end of the 100 ohm goes to zero.
Remove the ground.

Measure carefully and exactly the voltage on pin 9 & 10 of IC6.

[paul_g_787]

Well, as long as the line out is clean we can skip the headphones for now.
(Have you used the headphones previously and did it work fine?)

With the 100 ohm still in place, what's the voltage on the hot end of it? (I'd guess -20VDC or so)?
Find the junction of R15, C11 and pin 8 of IC6.
Check the voltage, it should be minus a couple of volts.
Ground that point and see if the voltage on the hot end of the 100 ohm goes to zero.
Remove the ground.

Measure carefully and exactly the voltage on pin 9 & 10 of IC6.

No never tried the headphones before.

Will try as you suggested tomorrow. Thanks for all the help so far

[paul_g_787]

Well, as long as the line out is clean we can skip the headphones for now.
(Have you used the headphones previously and did it work fine?)

With the 100 ohm still in place, what's the voltage on the hot end of it? (I'd guess -20VDC or so)?
Find the junction of R15, C11 and pin 8 of IC6.
Check the voltage, it should be minus a couple of volts.
Ground that point and see if the voltage on the hot end of the 100 ohm goes to zero.
Remove the ground.

Measure carefully and exactly the voltage on pin 9 & 10 of IC6.

OK just done the voltage measurements.

All voltages below are DC
Voltage across output load = -42.2V 
voltage from IC6 pin 8 to ground = -3.1V 
Voltage across output load with IC6 pin 8 connected to ground = -41.5V 
Voltage from IC6 pin 9 to ground = -9.6V 
Voltage from IC6 pin 10 to ground = 0V

[Renate]

Yup, it looks like the low transistor in the output IC is shorted.
Grounding the mute pin switches out the input stage and uses a separate stage whose only purpose is to DC balance the amplifier to zero volt output.

There's a 1% chance that the lower clamping diode D21 is shorted.

With the 100 ohm disconnected, measure:
Between the output & the +40
Between the output & the -40

Check with the "ohms" setting (one way) and the "diode" settings (both ways).

[paul_g_787]

Yup, it looks like the low transistor in the output IC is shorted.
Grounding the mute pin switches out the input stage and uses a separate stage whose only purpose is to DC balance the amplifier to zero volt output.

There's a 1% chance that the lower clamping diode D21 is shorted.

With the 100 ohm disconnected, measure:
Between the output & the +40
Between the output & the -40

Check with the "ohms" setting (one way) and the "diode" settings (both ways).

I tested across D21 in both directions. Dead short. So I removed D21 and out of circuit it seems fine and measures 601.
Also tested D20. In and out of circuit it measures 603 so I'd say it is good also. 

With D21 still out of circuit I checked between Pin3 of IC6 and the - rail of BD1 bridge rectifier. Dead short both ways. 

Tested the output + cable and there is a short to -40V with 16Ω resistance. Tested this with the DMM leads both ways. 
Tested output + with +40V and there is no connection and infinite resistance. Also with leads both ways. 

So the LM3886 is duff then. But the question now is why did it fail?

[SilverSolder]

It could have failed due to a manufacturing flaw when it was made - something that can take years to surface.

I have seen them stand up to incredible abuse (120 watts continuous for hours on end with no quibble),  but I have also seen some fail under less extreme conditions.  Sh!t happens.

They are really good chips, basically.  Solder another one in, and let'er rip!  :-)

[paul_g_787]

It could have failed due to a manufacturing flaw when it was made - something that can take years to surface.

I have seen them stand up to incredible abuse (120 watts continuous for hours on end with no quibble),  but I have also seen some fail under less extreme conditions.  Sh!t happens.

They are really good chips, basically.  Solder another one in, and let'er rip!  :-)

Before replacing it I would like to hear what Renate has to say about it. I want to make sure I don't blow up the new IC straight away in case there is another fault which caused the IC to blow.

Although, just looking at the schematic, what if previous owner connected their mixer to the speaker instead of the line out jack? The jacks are right next to each other and are very hard to tell apart.
Surely this would blow the power amp IC?

When I got the amp given to me (for free so I hope I can get it going) I was told "[they] lent it to a friend for a gig and it came back like this".

[Renate]

SilverSolder is right, stuff happens.
Most of the IC amps today are pretty bombproof.
Even a short to ground doesn't annoy them too much.

When removing multiple pin devices, I always like to destroy them.
I cut the leads off as near the body as I can.
Sometimes I even use a Dremel tool.
Then I can unsolder the leads individually.
This saves wear and tear on the PCB.

You can look at the headphone driver now if you like.
The obvious thing would be if the op amps were missing one leg of power.
They could be blown too.

If somebody had plugged a cable from ext speaker out to headphones out....

[paul_g_787]

SilverSolder is right, stuff happens.
Most of the IC amps today are pretty bombproof.
Even a short to ground doesn't annoy them too much.

When removing multiple pin devices, I always like to destroy them.
I cut the leads off as near the body as I can.
Sometimes I even use a Dremel tool.
Then I can unsolder the leads individually.
This saves wear and tear on the PCB.

You can look at the headphone driver now if you like.
The obvious thing would be if the op amps were missing one leg of power.
They could be blown too.

If somebody had plugged a cable from ext speaker out to headphones out....

Yes I tend to cut off legs on faulty components too. I usually dob them with some red paint as well, I have a box of broken electronic bits that might make some good artwork one day haha.

Moving on...

Is it safe to run the amp without the IC to fault find further? Or shall I replace it first?

As for the op amps, they are all on the 12/-12V regulated rail. Do you think they could be affected? I guess I could test for direct shorts on them. What say you?

What about the speaker. It still measures 8 ohms on the DMM  but could it be damaged? It doesn't push in and out very much but i can't feel the coil rubbing either.

And headphone circuit. What to test first?


The last few DIY repairs I have done all tend to be shorted silicon components. Are modern silicon components just rubbish or something? 

[Renate]

It's mostly a question of how quickly you can get a replacement IC and your choice.
I'm impatient, so I'd remove the IC, clean the board up and go on to the headphone.
Of course, removing & replacing the amplifier IC in one go would save some effort.
This would also give you a guitar amplifier that you can feed into *something* while waiting for your part.

The headphone opamp only costs a dollar, so I think that I'd just order one.

Did you use stereo headphones and did both channels sound the same?

[paul_g_787]

It's mostly a question of how quickly you can get a replacement IC and your choice.
I'm impatient, so I'd remove the IC, clean the board up and go on to the headphone.
Of course, removing & replacing the amplifier IC in one go would save some effort.
This would also give you a guitar amplifier that you can feed into *something* while waiting for your part.

The headphone opamp only costs a dollar, so I think that I'd just order one.

Did you use stereo headphones and did both channels sound the same?

It will take approx 5 days to get one. The cheapest I can find is £5.50 GBP without waiting months for one from china.
I don't mind removing the old IC first. Just wanted to check I wouldn't damage anything by not having the IC in circuit..

at you mean about connecting headphone and speaker jack. Could easily have fried both! 

Yes, both channels on headphones were distorted. I see that the headphone op amp is a dual package with both channels in parallel. A NJM4556AM is £4.75 in the UK! But only £1.50 for 5pcs from china. And people wonder why everyone doesn't buy British.

I'll remove the power amp IC6 now and then with it out, test the IC17 voltages.

If both are faulty I might as well get both from china and wai patientlyt. I do have another amp I can use for now.

[Renate]

Well, whatever you do, be sure to write us an epilogue some time.
We get hooked into these things and like to know how it ends.

This thread caught my attention because I have a Fender Acoustasonic 40.
It worked mostly fine, but the turn-off popped.
They had put in a circuit for that, but it didn't work well at all.
I ripped it out and put in my own uP monitoring the +28V supply.
It uses both mute and standby in timed sequence.
I replaced the power LED with a red/green that shows what the uP is thinking.

[Hmm, I tried to post a photo, but it didn't work.]

[paul_g_787]

I will keep you posted  Thank you again for all the help. I have learnt a lot and would not have managed to get this far without your help.

Just managed to remove the IC. Bloody fiddly thing. Half the legs bent one way and half the other so I had to snip one half then bend it upright.

Just checked pins 3 and 4 and yes dead short between them.

This video clip illustrates my last 10 minutes perfectly!  https://youtu.be/c1QcjsjjtRc

Just having a cup of tea then will check the headphone op amp voltages.

RE: replacing the IC. Can I use some of my Arctic Silver CPU heatsink paste I used on my gaming computer?

[paul_g_787]

OK, I just checked the headphone circuit and there is +11.6VDC coming out of both tip and ring on the jack!!!!!! 

You can clearly see this on the op amp output pins 1 and 7.

Voltages of IC17 pins:
1. 11.16V
2. 9.81V
3. 9.72V
4. -11.84V
5. 9.72V
6. 9.80V
7. 11.16V
8. 11.79V

Also there are no direct shorts from pin 1 or pin 7 to the + or - 12V rails. And i cannot find any direct shorts between any pins on IC17.

[paul_g_787]

Have taken voltage readings from the headphone circuit. Had a think about it but still cannot work it out.

Both the left and right channels share a common input and both have the same problem with the voltage on the output and the levels are exactly the same on both channels. 

C175 and C180 are blocking the DC which is why the line out is working great. So the problem must surely be after those two capacitors.

It also looks like the voltages are being divided by the feedback resistors and also the biasing resistors, so I would hazard a guess that the unwanted voltage is coming from the highest point which is from pin 1 and 7 of the IC. 

So unless I have missed something; I can only conclude it must be something internal in the IC. What do you think? 

[SilverSolder]

I think your diagnosis is right - that chip is probably fried.

The voltage at the inputs 3 and 5 should be zero volts...  those pins are inputs connected to ground via R157 / R163, so no significant voltage should be present there.

[paul_g_787]

Well I finally manged to remove the damn headphone IC!

First time I have de-solder a SOP8 IC and I am in no hurry to do it again! Very fiddly. Much harder than installing and SOP8 IC.

I have tested the voltages now it is removed and there is 0v on the inputs and outputs now! So it was indeed the IC.

Will order replacements fro both ICs and post back once they arrive. 

[paul_g_787]

I have some questions:

1. Would a TL072CDR be a suitable replacement for the faulty NJM4556AM? I ask because I already have one.

2a. What is the white greasy compound used on the LM3886TF? It is still wet so is non-setting.

2b. Is this the same stuff used on computer CPUs? I already have some arctic silver. Can I use that?

[Renate]

You need a NJM4556AM because it can drive a low impedance load.

My philosophy is that any heat sink compound will be fine.
Most older things used standard zinc oxide, that white, greasy goop.
You can get fancier, silver containing or those rubberized thingies.
I'm sure somebody has a strong opinion on this, but I don't.
I think that somebody did a test using spit!
It worked fine, but tended to dry out.

[paul_g_787]

You need a NJM4556AM because it can drive a low impedance load.

My philosophy is that any heat sink compound will be fine.
Most older things used standard zinc oxide, that white, greasy goop.
You can get fancier, silver containing or those rubberized thingies.
I'm sure somebody has a strong opinion on this, but I don't.
I think that somebody did a test using spit!
It worked fine, but tended to dry out.

OK thanks. Will have to have a shop around then. Otherwise a chinese order and a month waiting it is.