[sdiy] saw core VCO tuning/scaling (Roland SH-2)

[Adam Inglis]

Thanks Magnus and Juergen for the tips:

Thermally challenging a MF resistor in the range CV input reproduced  
the problem, however 2 hours after replacing it, the problem was back!
I then got out the hair dryer. Heating the circuit brought the  
problem on again from a cold start, but it was hard to localise - it  
seemed to be around the integrator though. Refreshing all the joints  
made no difference, so I finally replaced the integrator/comparator  
op amp (IC 18), and now after 4 hours she's rock-steady, and sounding  
sweet.

I'd still be looking for the problem come next year, without this help!
And thanks for reminding this dummy about Ohm's Law...  ;-)

cheers
Adam


http://www.adambaby.com/images/tech%20pics/SH_2/sh2vco.html


On 20/08/2011, at 7:52 PM, Magnus Danielson wrote:

> Adam,
>
> On 20/08/11 10:07, Adam Inglis wrote:
>
>> Thanks Magnus, Harry and list
>>
>> I've spent a few more hours looking at this..
>> After a full calibration after 30 min warm up, VCO 1 tracks and tunes
>> nicely with VCO 2. After 2 hours, the highest notes are still in  
>> tune,
>> but the lowest are sharp by a full tone (e.g. from 175 hz to 198 hz).
>> After 4 hours the highest notes are just slightly sharp, but the  
>> lowest
>> are about the same, a full tone sharp.
>>
>
> Sounds pretty stable to me.
>
>
>> DMM measurements at the output of the CV summing opamp show barely  
>> 20mV
>> change over this time, at various octaves, and the voltage  
>> differences
>> remain linear. The off-board trimmer in the feedback loop ("VCO  
>> width")
>> and it's contacts are sound. The "Range" voltage divider and it's  
>> opamp
>> scaler that connect to the summer show no variation in voltages  
>> hot or
>> cold.
>>
>
> What about mechanical stability? Push and shake it around a bit.  
> Both connectors and pot.
>
> Sometimes you need to resolder things, since there can be a hidden  
> creep so that a pin may be in a tube of solder and depending on  
> temperature the contact varies. This is typical of physical stress  
> and temperature differences. Not too obvious at times, but can show  
> up for older equipment.
>
>
>> The "temp adjust" pin, no.6, of the uA726 shows -6.75 V for VCO1, and
>> -6.54 V for VCO 2. I haven't yet measured the current here - it's  
>> a bit
>> of a fiddly old board, looms just soldered straight on, making  
>> access a
>> bit tricky.
>>
>
> You have a resistor, measure the voltage over the resistor and the  
> current comes out by dividing the voltage by the resistor value...  
> no need to "cut in".
>
>
>> Questions, given the above
>> 1) Can I assume that the problem is NOT in or around the CV summing
>> circuit, given the stability of it's voltages over time/heat?
>>
>
> Until you found your problem, you should not use strong words like  
> that, you should just say that it is less likely. I've found both  
> myself and fellow friends and colleagues to lock their mind up to  
> early and not seeing the actual fault mechanisms. Instead they go  
> into more and more obscure things rather than checking easy stuff  
> like contact problems and power supply voltages.
>
>
>> 2) Does the low frequency drift point to a typical problem?
>>
>
> The scale of the problem is fairly low, Could be many different  
> sources.
> Solder joints, connectors, bad cap, bad op-amp, bad pot.
>
> Cheers,
> Magnus
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>
>
>