Roland RS-70 (same keys across octaves don't work)

[elias]

Hi guys,

I have joined today, and first of all want to thank in advance any help you can give me.
My case is as follows:

A week ago my keyboard (Roland RS-70) was working completely fine.
Last sunday when I got it out from it's bag for the Sunday Mass, It came with lots of electrostatic,
and when I touched a spark came from the bottom of the keyboard to my hand (ouch !)

Then I turned it on.. and by suprise, G key of every other octave is not working, also B
and D...
This is an old keyboard used only to praise God.
I want to repair it by myself, I'm an IT guy with good electronic repair background.
But I don't know where to start ?
or if the part can be replaced ?

is it the keyboard itself (I can see some diodes and the kind of failure seems to be a serial... maybe one or two damaged diodes could produce this effect)
is it the CPU or any serious component ?

Please advise...
Any help will be really appreciated.

God bless you all.
Elias.

[jrfrond]

Unfortunately, God doesn't watch over the instruments used in religious settings, it would seem, or the Worship repair industry wouldn't be so big!

Before I opened ANYTHING, I'd perform a Factory Reset as follows:

[PATCH]-[UTILITY]-PAGE/CURSOR [<] or [>] to #5 "Factory Reset".

[elias]

I already did it

Dear JrFrond:

I tried that and got no luck... only my user presets erased (forgot to back them up...)
anyway...
any other ideas ?

warmest regards
Elias.

[Enzo]

God is concerned with souls, not Rolands...

If you are an IT guy, then you know about matrixed inoput switching arrays - for example the keyboard for your computer doesn't have a wire for each key running back to the CPU. Musical keyboards are matrixed. Keys are in groups. Some may be groups of six, and then every sixth note might be out. Others are grouped as octaves - sounds like yours is. So we call the columns and rows of this matrix something like strobes and returns. The CPU sends a strobe pulse down the strobe lines into the matrix. Then it watches the return lines to see if it sees any strobe pulses coming back. Any return pulses it sees must be the result of a key closure. Key switches connect a strobe to a return. Each key will have a diode in series with its switch. I won't explain in depth, but it prevents confusion when more than one key is pressed.

So if all the D or G or XYZ keys go dead, then I am led to think the return line for those keys has died. Or possibly the strobe line. All depends upon the design arrangement. The diodes tend to afect individual notes, not groups, though anything is possible.

I'd scope the strobe lines coming into the keybed. See if any are dead. Then scope the return lines, run your finger up the scale to hit all notes in an octave while watching each return. See if you see activity on each line. If activity is missing, then find out what is wrong on the keybed contact strips. If all strobes are present and all returns show return pulses, then I'd guess the return circuits have failed for those notes. I don;t know the keyboard offhand, so it could be discrete circuits in older boards, or some key assigner IC on more modern boards. Get the schematic.

[doctor]

God is concerned with souls, not Rolands...

If you are an IT guy, then you know about matrixed inoput switching arrays - for example the keyboard for your computer doesn't have a wire for each key running back to the CPU. Musical keyboards are matrixed. Keys are in groups. Some may be groups of six, and then every sixth note might be out. Others are grouped as octaves - sounds like yours is. So we call the columns and rows of this matrix something like strobes and returns. The CPU sends a strobe pulse down the strobe lines into the matrix. Then it watches the return lines to see if it sees any strobe pulses coming back. Any return pulses it sees must be the result of a key closure. Key switches connect a strobe to a return. Each key will have a diode in series with its switch. I won't explain in depth, but it prevents confusion when more than one key is pressed.

So if all the D or G or XYZ keys go dead, then I am led to think the return line for those keys has died. Or possibly the strobe line. All depends upon the design arrangement. The diodes tend to afect individual notes, not groups, though anything is possible.

I'd scope the strobe lines coming into the keybed. See if any are dead. Then scope the return lines, run your finger up the scale to hit all notes in an octave while watching each return. See if you see activity on each line. If activity is missing, then find out what is wrong on the keybed contact strips. If all strobes are present and all returns show return pulses, then I'd guess the return circuits have failed for those notes. I don;t know the keyboard offhand, so it could be discrete circuits in older boards, or some key assigner IC on more modern boards. Get the schematic.

Hi Elias,

Enzo is right.

Due to 8-bit matrix structure of the key assigning circuitry one in eight key
(but not each octave, i.e. 12th) may be out of order. Please examine attentively.

In your case, if keyboard’s print has not mechanical damages, the most probable
cause is malfunction of the key assigner chip IC33 (SSC1080F0B, Roland p/n is 00129278).
It is the custom device and it is possible to order only from Roland authorized service.

If you want to repair this yourself, you need to know how to replace 52-pin SMD chip.

Good luck.

[jrfrond]

One of the most common causes of issues like this is a poor header connection somewhere on the keybed ribbon cables. Now, the OP stated that it was OK BEFORE the ESD incident occured, but it COULD just be an unhappy coincidence. I'd buzz-out the keybed ribbon cables first, to see if there's an open line, or see if it got crimped in the chassis, etc. I've seen cases where the keybed ribbon was entrapped during factory assembly and eventually wore through years later and caused a short to the chassis. You just never know.

[elias]

Thank you All

I just want to thank you All for your prompt responses...
I will check what you mentioned and then see what is the problem... maybe it's just a bad coincidence as doctor said.
I personally don't think coincidences exist but that is another story for another forum.

Thank you all and my warmest regards to everyone,
God bless you and your families.

Thanks.
Elias.

[Enzo]

OH my friend, coincidences most certainly exist. An important axiom of science is that correllation is not the same as causality. In troubleshooting, I have seen many people get confused by assuming so and so MUST be involved with their problem, because it happened at the same time. Ther is a logical fallacy called "post hoc, ergo propter hoc." That more or less translates to " since it was after this, it must be because of this." It may be the case, but you can;t count on it as proof.

[elias]

New Developments

Ok, minutes ago, I disassembled the keyboard... Cleaned the connectors... and got no luck...
same problem, but this time I noticed something that last time I didn't

every 8 notes... one note doesn't work.
that means...

C-C#-D-D#-E-F-F#-G-G#-A-A#-B <--- from the first octave just G (8th note)
from that note.. counting eight... the next note not working is... D# then... B ... and so on and so forth...

every eight notes one note doesn't work...
So I think its the problem that you guys mentioned earlier (matrix scan chip)
and No, I don't know how to replace that kind of chips...
and can't get Roland to sell it to me...
So I have to take my keyboard to service... any ideas of how much am I going to pay to fix this issue ?

Warmest Regards
Elias.

[bkahuna]

every 8 notes... one note doesn't work.

That problem is way more common (in fact I don't think I've seen a keyboard that skipped the same note every octave), and could very easily be an open scan line. You need to check the connections to the key contact pcb very closely. If no problem is found, you'll need to pull the key contact pcb and inspect it carefully. Sometimes a liquid spill gets in there and corrodes one of the scan lines, causing this problem. It could also be a broken solder connection.

This isn't one of those Rolands with the "plastic membrane" key contact circuit board, is it? If it is, I can say with almost 100% certainty where the problem is...

[jrfrond]

Yes, the unhappy coincidence..... it happens at least once a week here.

If this is, in fact, and open scan line, it could also be a cold solder on the flatpack IC. But really, I am suspecting something of a more mechanical nature, and if this was in my shop and on my bench, I'd be recrimping the keybed ribbon cable headers right now, then buzz out the cables from the pins on both sides. I'd also remove the key contact strips to check for contamination/corrosion. I'd suspect the IC LAST.

[doctor]

Yes, the unhappy coincidence..... it happens at least once a week here.

If this is, in fact, and open scan line, it could also be a cold solder on the flatpack IC. But really, I am suspecting something of a more mechanical nature, and if this was in my shop and on my bench, I'd be recrimping the keybed ribbon cable headers right now, then buzz out the cables from the pins on both sides. I'd also remove the key contact strips to check for contamination/corrosion. I'd suspect the IC LAST.

Hi elias

If only one of each 8 notes is out of order, there is high probability
that one trace on the keyboard print is open (as a rule it is corroded
as bkahuna writes). Find common trace for all defected notes and
check its continuity up to connector by means of ohmmeter.

If this trace is intact, perform the following checks:

a) If this trace connected to the one of the following contacts
on the main board (connector CN7 - pins 1, 2, 3, 4, 5, 6, 7 and 8;
or CN9 - pins 9, 10, 11, 12, 13, 14, 15 and 16), check corresponding
resistor of the RA38 array connected to this contact.
Resistor’s value should be equal to 22 Ohm.

b) If this trace is connected to the other contact of these connectors

1. Check corresponding resistor of the RA32 or RA34 array (it should
be equal to 22 Ohm as well), or resistors R144 and R145 (10 Ohm each).

2. Check pull-up resistor connected to this pin in the one of the RA30,
RA33 or RA35 array. Resistance between this pin and +5V should be 10K.
+5V is connected to the pins 7 and 33 of IC33.

c)If these tests are OK, check how corresponding IC33 pin is soldered.

If all tests are OK, IC33 is defected.

Good luck,

[elias]

Another thing I noticed

Later on today I'm going to perform the checks that you mentioned (thank you very much for the specifics of doing that)
I noticed that there are two connectors from the keyboard itself to the Main board... one correspond to the upper part and one to the lower part... (I disconnected one, and just half keyboard worked... but with same defects) did the same with the lower part and the same happened.

Damage in the same line of the two sections of the keyboard at the same time ?
or
damage in the IC33 ?

warmest regards
Elias.

[Enzo]

Unlikely both keybed boards have identical defects. However, neither of those connects DIRECTLY to the IC. They connect to the circuit board, where ther are traces and solder connections etc etc. So we really need to pinpoint the failure.

[elias]

What I checked...

HI, I checked the cables for continuity and both are good.
inpsected the key beds, and the surface in which they are mounted, also I performed a basic maintenance (cleaning with qtips and isopropyl alcohol)but the failure remains, I also checked both connectors (in the motherboard and in the keys) but everything seems fine.
I contacted roland and the spare IC33 is 15 bucks only...
what probability do you think I have to solve the problem replacing the IC ?

warmest regards
Elias.