Roland CUBE-01 schematic? No output on speaker; has output on headphone jack; speaker works

[dchang0]

Hi, everyone--

I've got a Roland CUBE-01 that I'd like to repair and give to my niece, who is learning electric guitar after having started on acoustic.

It's got output on the headphone jack and the speaker is good, but zero output (not even hiss) comes out of the speaker. All the controls and the reverb effect work fine.

After unsoldering and carefully checking the Jalco 1/8in stereo switched headphone jack, which is supposed to cut off output to the speaker when the headphone is plugged in, I will need a schematic to do any more troubleshooting.

Does anyone have a schematic or the Service Manual? Roland of course refuses to give me either. Google searches turn up nothing for this particular model of Cube.

If another model of Cube is close enough, perhaps I can use that one to deduce what's going on.

UPDATE: I just noticed that the mainboard has silkscreening that is blacked out with a Sharpie. The mainboard apparently is the same as that used in the CUBE-10GX, possibly minus some components. So, a CUBE-10GX schematic or Service Manual will likely work too.

Thanks very much in advance!

[dchang0]

While troubleshooting, I noticed some unusual behaviors (unusual to me--maybe they are normal). Can anyone confirm that these are supposed to happen or not?

a) When the power supply is plugged into the back of the amp, the green (Clean) amp type selector light comes on immediately, even if the rocker power switch is turned off. No guitar sound comes through to the headphone jack if the power switch is turned off, but it does come through if the power switch is turned on. Seems weird that any part of the amp turns on when the power switch is turned off.

b) If I am playing an MP3 into the AUX IN, the audio comes through the headphone out even if the power switch is turned off, but it is quieter than if the power switch is turned on. Again, it seems weird that any sound would be passed through at all with the power switch turned off.

[g1]

This is from a 20X, is it anything like it?

[sugeda]

...

[dchang0]

This is from a 20X, is it anything like it?

Hi, g1! Good to see you again--I was looking at the 20GX and no, it seems too complex compared to the 10GX.

But lucky me, sugeda just posted the exact stuff I needed!

[dchang0]

...

Thank you very much, sugeda!!!!


I will update this thread as I proceed with testing.

[dchang0]

Looking at just the block diagram alone:

I suppose the JFET is bad based on the fact that the LEDs turn on right away. Essentially, the DSP/CPU must be getting enough power through the JFET to boot up and light up the LED at all times, regardless of the state of the rocker switch.

Also, the CPU MUTE or MUTE could be cutting off the MAX9710ETP power amp block that drives the speakers, since the headphone out is before the CPU MUTE and MUTE.

UPDATE:

I followed the instructions in the CUBE-01 Service Manual to enter test mode, and they didn't work. I suspect I have to fix the power problem(s) first before the DSP/CPU will behave properly. Up next will be measurements at various test points in the power supply block.

[g1]

Looking at just the block diagram alone:

I suppose the JFET is bad based on the fact that the LEDs turn on right away.

Not necessarily as the block diagram is quite simplified. The Q17 fet could be getting turn on commands via Q4/Q6 circuitry.

[dchang0]

Thanks again to sugeda!

I printed out pages 22 and 23 on my laser printer, but the text ended up being too tiny to read, so here is a zoomed-in PNG image of the power supply portion of the mainboard. It is very wide--too wide to print on one piece of paper, but it'll be awesome for viewing on a monitor where you can scroll around.

UPDATE: added the P-ON TRIGGER portion of the schematic. This might be how power supply problems end up muting the final power amp chip.

[dchang0]

Not necessarily as the block diagram is quite simplified. The Q17 fet could be getting turn on commands via Q4/Q6 circuitry.

Good point. I will probably replace all three, as the total cost is under $1 for all three of them.

After that, I just need to make sure P+5V, A+5V, A+3.3V, D+3.3V, and D+1.5V all measure correctly. It might be a while before I reply--the parts won't be ordered until next month due to a backorder for something else I want in that shipment, to arrive near the end of December...

Guess I won't be able to present the amp to my niece for XMas, but then again, I have other broken amps that could be fixed before then...

Thanks again for the help!

UPDATE:

Q4 = 2SC4116SU-GR.LF(D
Q6 = 2SA1586SU-GR.LF(D
Q17 = SSM3J328R.LF(T

Diodes D8 = RB051L-40 TE25 (PB FREE) and D18 = UDZSTE-175.6B could probably be replaced too--might as well since they're merely 17 cents each.
Also, I will probably replace C136 = 4700uF 6.3V.

[g1]

I think you should be able to jumper across Q17 to continue troubleshooting while you are waiting for those parts?
There is one other line that seems triggered by the power switch but I didn't find where it went to (bottom of pg.22, 'P-ON_TRIGGER').

[dchang0]

I think you should be able to jumper across Q17 to continue troubleshooting while you are waiting for those parts?
There is one other line that seems triggered by the power switch but I didn't find where it went to (bottom of pg.22, 'P-ON_TRIGGER').

Good idea, I will jumper Q17, pin 2 (source) to pin 3 (drain).

Today I went through and checked many of the voltage test points marked on the schematic. The 5V ones all measured about 5.5V (just under the 5.6V of the power adapter). The 3.3V and 1.5V points came in just a tad under.

Here's the P-ON_TRIGGER portion of the circuit.

I checked the various Test Points and the important one, TP2, measured near zero volts. I don't know whether zero = MUTE ON or zero = MUTE OFF by design. I'll look up the data sheet for the power amp chip to see which state engages the MUTE.

UPDATE: Just checked the MAX9710 datasheet and it says "Drive MUTE high to mute the device. Drive MUTE low for normal operation." So, it would seem that the power amp chip is NOT being muted, which may point back to the headphone out jack switching the signal off. I have not desoldered it yet. It's a Jalco 1/8in jack with five total pins. One is obviously the sleeve. I can't tell what the other four pins do, but I can see that there is indeed some kind of plastic plunger on the side inside that must be related to the switching mechanism. I see the part number YKB21-5401N in the schematic--now I gotta find the datasheet to see the physical pin layout.

UPDATE 2: I found this datasheet for a Jalco YKB21-5406N. As far as I can tell, there is no available datasheet for the YKB21-5401N still on the web. From the photos, the YKB21-5406N is very close, if not the same. Assuming the physical pin arrangement is the same for both, I can figure out which pins to test for continuity that would cause the speaker to always be cut off.

Based on the datasheet and the schematic, pins 8 and 9, when shorted, signal to cut off the speaker output. Pins 8 and 9 are on the side of the jack, whereas pins 2 (tip) and 3 (ring) are near the center of the jack. I should be able to measure for continuity between 8 and 9 now without desoldering.

https://t-tatsumi.com/wp-content/upl...KB21-5406N.pdf

[dchang0]

Well, kinda bad news. The headphone out jack is NOT cutting off the speaker. I figured out the physical positions of Pins 8 and 9, which will short when a headphone plug is inserted into the jack and will be open if no headphone plug is inserted.

They tested exactly as they should, so the jack is not stuck and not broken.

See photo for the pin positions that I deduced from the official datasheet and the the Roland schematic.



It's bad news for me, because if the MUTE is off on the MAX9710 power amp chip AND the headphone jack is not stuck in cut-off-the-speaker mode, then that means I gotta check:

CPU MUTE (different than the P-ON TRIGGER to MUTE_LINE muting)
X-PHONES DETECT (not including the switch in the headphone out jack)
input pins 1 and 5 on the MAX7610 power amp chip (for audio)


WHOA, I JUST REALIZED SOMETHING LOOKING AT THE SCHEMATIC!

The PAMP_MUTE line doesn't connect to the MUTE Pin 4 on the MAX9710 as I had thought.
It connects to the _SHDN Pin 14 instead, which according to the datasheet behaves like this:

"Pulling SHDN low disables the device’s bias circuitry, the amplifier outputs are actively pulled low, and BIAS is driven to GND. Connect SHDN to VDD for normal operation."

So the fact that the P-ON_TRIGGER to MUTE_ON Test Point TP2 is zero volts is BAD--it shuts down the MAX9710 power amp chip!

That said, it looks like CPU_MUTE (Pin 44 on the DSP/CPU) can override MUTE_ON.

I need to measure zero volts at the top of C46 that goes to _SHDN Pin 14 (aka PAMP_MUTE) to be sure that the MAX9710 is being told to shut down. Pulling PAMP_MUTE up to VDD = 5.0V will tell the MAX9710 to turn on.

Another interesting tidbit: the max voltage of the MAX9710 is 5.5V, which so far is about where I've been measuring the various supposed-to-be-5.0V test points.... Not good. I will have to double-check VDD on Pin 12 of the MAX9710.

[dchang0]

Follow up after I finally got some parts in a much larger order of parts that was held up by backorders:

The JFET Q17 was indeed broken, as the red/green/orange mode LED was turning on regardless of the state of the power switch before Q17 was replaced. After Q17 was replaced, the behavior was as expected: the mode LED turns off when the power switch is turned off and turns on when the power switch is turned on.

However, the speaker output has not been fixed yet.

I also replaced D8, D18, and C136 so far with no change after replacing each part (they were good, in other words).

I have yet to replace Q4 or Q6.

At this time, I'm starting to suspect the power amp chip itself is fried, but I'll keep replacing the ultra-cheap components related to the MUTE signal before finally ordering the more-expensive power amp chip. It'll be a bit challenging to swap the power amp chip due to having to desolder four ends of two heat pipes that hold down an aluminum heat sink. I am not sure my soldering iron is up to the task.

UPDATE:

After replacing Q17, it is now possible to enter the Test Mode as specified in the service manual. Apparently I have version 1.01 and it passes all the tests available in Test Mode, except of course that no sound is coming out during the volume test (the pots do pass the test). Passing the Test Mode likely means the DSP chip itself is fine. I also confirmed for sure that unplugging the headphone out does switch on the speaker (the Test Mode detects this).

I'm still getting 5.56VDC on pin 3 of Q17. The schematic says P+5V--not sure if 5.56V is too high. I'll try replacing Q6 next, then Q4 after that.

UPDATE 2:

Replacing Q6 and Q4 did not bring back the output. Q1 is next in the power section, and then I'll go into the mute circuitry.

UPDATE 3:

Replaced Q1, no improvement. The power section is probably fine, though it still bugs me that the P+5V is too high. Moved on to MUTE switch circuit and replaced Q2 and Q3 to no avail. Next are Q10 and Q18 in the MUTE circuit, then after that, I will replace the power amp chip.

UPDATE 4:

No luck replacing Q10 and Q18. That's all the transistors in the MUTE switch section. Unless D4 is bad (it doesn't measure bad) or some of the capacitors in the power section or MUTE section surrounding the transistors are bad, I have to assume now that it's the power amp chip that's bad. I'll order one in my next big parts order.

UPDATE 5:

Now that I had reason to suspect the power amp chip, I looked more carefully at the schematic and realized that I could have tested the entire MUTE circuitry long ago by measuring the voltage across C46. If it is VDD, then the SHDN signal is not being sent to the power amp chip, and it should be working properly (not shut down). I checked, and the voltage across C46 is +5.56VDC, so all the prior MUTE circuitry is definitely working properly, including the DSP.

This means that the sole problem MUST be the MAX9710ETP+ chip.

It's a TQFN package measuring 5mm x 5mm. I am not sure I have the skills to desolder/solder this tiny a chip. The first challenge is to non-destructively remove the heat sink and heat pipes.

UPDATE 6:

Removing the heat sink non-destructively wasn't so difficult with the use of about 1cm of a stick of Chip Quik removal alloy and a Hakko 808 power desoldering gun and a flathead screwdriver to pry the heat pipes out of the PCB. There are no components surrounding the heat sink on the side of the board that has the heat sink, so I was able to insert the flathead screwdriver without fear of damaging nearby components. Looking at the TQFN chip, it looks like it will be straightforward to remove if I use the Chip Quik removal alloy again.

[dchang0]

I was able to remove the suspected-bad MAX9710ETP power amp chip, BUT it required the use of a Steinel professional-series heat gun with neck-down nozzle or hot air station because the chip is soldered to the PCB on its belly. I did some minor damage to the PCB trying to remove it by adding Chip Quik low-temp removal alloy to the pins on the sides, but since I couldn't get at the solder underneath the chip, all I did was heat up and damage the PCB.

That said, the pins on the PCB were not lifted up and seem to still be connected to the proper points on the rest of the PCB, so the repair is not doomed.

I will order a MAX9710ETP on my next Mouser or Digikey order. They are surprisingly cheap, at a mere ~$2.50 each. I wish I had known that--I thought they'd be $15. At this price I would have ordered one just in case in my last big parts order. Luckily, I am working on fixing someone's failed build of a Spaceman Sputnik II clone and that likely will need some small parts (probably the commonplace transistors and not the rarer Ge diodes).

I'll report back when I have finally repaired this amp. Gave my niece a different amp instead since this one was taking so long to fix. Good to get her hooked on electric guitar; she loves it over her acoustic.

Just wanted to warn anyone trying this repair that you will need a hot air station or a decent-quality heat gun with the proper neck-down nozzle and precise temperature and air speed controls.