Sh5 Roland

[9 volts]

Hello, my roland sh5 seems to drift (frequency),when I use the hold or sustain function. If I keep a keyboard note held down, drift does not occur. I have looked at schem, but a little confused as the hold and sustain/decay knobs are on different pcbs. I know first off to check voltages from the power pcb. But can't get my head around where to check voltage for the after key hold hold function. Any help appreciated. Thanks

[JingleJoe]

I asume the sustain and decay are the envelope controls? They will not affect the pitch of the oscillator, I'm going to propose that you have a bad capacitor or some bit of circuitry is drawing too much current from the hold capacitor. Assuming it uses the standard sample and hold arrangement.

[9 volts]

Thanks, it's not the sample hold function, it's the hold function that keeps the note playing when you take your fingers off the keyboard. So I suppose it's may be the voltage to oscillators when the keyboard voltage is not being used.

[Peake]

Most keyboards use a sample-and-hold to maintain keyboard pitch once a key has been released. Check the schemos for main capacitor involved in this, maybe an FET. The portamento circuit as well.

[stewpye]

I'd try replacing C103 in the keyboard circuit unit. (page 13 of the service manual).

If this doesn't fix it check pin 1 of IC102 on the same board. If the voltage is drooping there I'd replace the FET pair and op amp with a low bias input FET op amp.

[9 volts]

Thanks a lot. Given my some direction, and a place to start. I'll study these two areas of the service manual a little more and clear a space to pull the synth apart in the next week. You guys are great, thanks again.

[radek tymecki]

try to calibrate synth first

[stewpye]

[9 volts]

I finally got some time to dismantle the sh5 today. The c103 appears to be ok- keyboard circuit as mentioned above(tested). So now I am looking at the sample hold circuit and 10v circuit. Thank again for the advice

[9 volts]

Checked the 10- and 10+ voltage which were ok.

[9 volts]

I have just tested the sample and voltage voltage adjustment (pg 35) and the output voltage definitely gradually drifts upwards. So I suppose now I test all the the capacitors in this area......I'm getting a little out of my depth so if anyone has any advice greatly appreciate it. Thanks

[ian-s]

A little drift is normal. It is supposed to just hold the pitch while the release times out. The minimoog drifted a lot but it was only a problem when external triggering was used. I used to wedge a key down with a patch cord to fix that problem.

[9 volts]

I replaced a faulty diode in the sample hold section(covered in glue). This got the samplehold adjustment voltage stable, however didn't alter the drift problem.

But I am going to recheck the portamento section again as rereading the manual I realize there is also a hold function around there also......so will recheck voltages around there tonight. (my misunderstanding)

Thanks for the comment. I have used a basic sh1000 which hardly drifted- the sh5 is rising a little faster than I would expect.

[9 volts]

Tonight I tested c103 (capacitor mentioned above) again. It tests ok with capacitance tester (dm meter). I tried to do a voltage test across it but the meter influenced the voltage- when meter was connected the voltage kept dropping-opposite of what normally happens- frequency lowered. I then tested pin 1 of ic102 as mentioned above- the voltage continues to rise after a key is pushed at about .01 volt per second. (causing the frequency to rise). I will buy a .33mf capacitor to try (c103) and maybe c104 also? (10uf). Definitely seems like the problem area.

[9 volts]

Well, tonight I tried a new c103 (hold capacitor) which didn't any difference. So I suppose I will have to look at it in a little more depth (learning curve)

[JingleJoe]

I've been followng this thread to see what you come up with but I feel I should pipe up now, if the voltage is increasing then the hold cap is getting voltage from somewhere. What op amp is it connected to? I assume it is connected to an op amp wired up as a voltage follower, or perhapsc some sort of dedicated analogue buffer?

[9 volts]

A ca1458 (1c101B) then a 2sk30a (tr101) then the hold capacitor. I have just done a few more tests. The voltage coming out of the ca1458/into the 2sk30a is constant, the voltage leaving the 2sk30a is not (everything beyond this point rises etc). I'm thinking it's probably worth trying a 2sk30a replacement.
(B+ and- and 10- and+ at the entry point on this board have been tested and are ok).
Thanks for getting back to me.
My basic understanding is that the 2sk30a should lock the voltage?

[JingleJoe]

The transistor in question (or perhaps FET) charges or discharges the capacitor to the voltage coming from the keyboard. However your layout description makes mea little dubious.
The op amp, which I suspected might be a CA series op amp, is a probably high impedance buffer AKA voltage follower which can output the same voltage as the capacitor, without discharging it. However from the datasheet it looks like the input impedance is a bit low for that sort of application. The CA3140 is a good high impedance input op amp.
I should have asked right away, but do you have a circuit diagram to refer to? It would clarify a lot.

[9 volts]

Yes this is the page I'm referring to

---

SH-5_SERVICE_NOTES.jpg
Description:
Filesize:	856.73 KB
Viewed:	231 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[9 volts]

I've got a few tl072 ics here, would it be worth trying one? I figure at least to see if it changes the synths behaviour. Thanks for the description, I understand now that if the k30a transistor was faulty probably more issues would arise than what I have. Thanks again

[stewpye]

There's a good chance the problem could be the buffer consisting of TR102 and IC102A. Since you probably can't a replacement for TR102 easily I'd suggest carefully removing it and IC102 and replacing IC102 with a low bias input, low offset FET dual op amp.
Remove R106 and R107.
Leave out TR102.
Join gate-source pads of both FET's. (ie:C103 will go to pin 3 of the op amp, join pin 1 and 2 of the op amp).

If you put a socket in for the op amp you could try the TL072. The TL072BC actually has better offset voltage specs than the CA1458. The OPA2134 would be better, and there would be even better choices than that. You want the lowest input bias possible so as not to drag down the S&H cap voltage. This was the reason for having the dual FET there.

Regards,
Stewart.

[JingleJoe]

[9 volts]

Thanks for the advice. I will try it out next weekend (buy an ic socket and also a burr brown).
I've just done a few more voltage tests. By constant voltage I meant stable (does not drift).
ic101
pin1 = 3.66
pin2 = 0
pin3 = 0
pin 5 = 1.29
pin6&7 = 2.65

fet (tr101)
gate (from diode 101) = -13.57
source (from ic 101 pin7) 2.65 (influenced by note on then returns to 2.65 after note released)

diode 101
-13.94 (when note is pushed becomes 13.94)
-13.57 (to gate)
Thanks again

[9 volts]

Just a quick note to say I tried the mentioned rewiring tonight (got rid of tr102 etc) and it has fixed the problem! Thanks to everyone for your advice! It now losses about .01v per 1.30 mins which seems alot more like the kind of behaviour i would expect (slow droop). I may experiment with cap size while I have the synth pulled apart. Thanks again there is no way I could have done this without your help!

[stewpye]

Good to hear. I'll remember that in case I ever have the same problem with my SH5.

Regards,
Stew.