Yusynth VCO not tracking 1V/Oct - With built pics

[Volter]

Hello everyone.

UPDATE: I may have theory to the problem, I've added a copper fill to both the top and bottom of the board without thinking. I've roughly estimated that it will add 0.88pF to the trace connecting the JFET, 220pF cap and op-amp buffer. Therefore at higher frequencies this will cause notes to be flat.

I may test this by removing said trace and running a wire to help reduce the added capacitance. But if this is the problem I'll have to remake the boards without a copper fill...

UPDATE END

I've been having an Issue with a set of 5 Yusynth VCOs I built. The problem is I can't get more then one or two octaves in tune at any given time. I am using Ableton live's inbuilt tuner and a Roland MPU-101 Midi to CV converter. Before I attempted to tune it, I left the VCO on for 10-20mins to warm up fully.

I've also previously tried to tune these using a set variable voltage input for the 1V/Oct, but still no luck.

Here is a link to the VCO schematic by Yusynth: https://www.yusynth.net/Modular/Commun/VCO/new/VCO-V2.gif

For the PCB layout I created my own trying to match the given layout as closely as possible:



Here are some pictures of the built VCO:




I've purposely have left D5 and D5 unpopulated until the tuning is fixed. I have used the BCM847 matched NPN pair.

It is being powered with +-15V. Here are some measurements from the board:

+Vcc is 13.57V
-Vcc is -14.33V
For the "+5V +/-2%" I get 4.52V

I've checked that the tempco is indeed 1k. And it is placed directly on top of the BCM847 with some thermal paste.

Following the tuning guide on Yusynth, as I start tuning the A octaves. The A octaves between the two I have tuned are too Sharp. i.e:

0V = A 55
1V = Bflat 114.4
2V = B/Bflat 227.1
3V = A 438.6


Thank you very much for reading.
-Volter

[Grumble]

This will not help you to get the tracking done, but it seems you are soldering from the component side. WHY?
NOT a good idea...

[Volter]

Hello,

Would you have any recommendations on what better details I need to include?

I haven't soldered from the component side. The PCB manufacturer coats the inner wall of the component holes connecting the top and bottom pads. When soldering this causes gravity to pull some of the solder down through the hole to the other side.

Thanks
-Volter

[gabbagabi]

Hello Volter,

Sir Yves states about the 78/7915 =" providing a reference voltage of +/-14.2V. This helps very much in increasing the stability of the VCO. "

at least your V++ looks like not meeting this values.

this is also affecting the 5V (2%), which should be 5,1V to 4,9V
which of course left us with the question how this could be archieved with the 100K/50K combo cos 14.2 /3 is 4.73V.


in the end i would start try to get "better" even values from the 78/7915
an other first try would be to test what happens when you bridge the xx15