Dual VCA thumping with LFO when both sides are powered (CV feedthrough?)

[Joncaster]

So, I have a VCA tremolo giving me problems at the moment.

Based on the Tremolomatic-X:
http://moosapotamus.net/files/stompboxology-mo-tremlo.pdf

But I've substituted the NE570 for a Mitsubishi VCA i found:
https://www.mantech.co.za/datasheets/products/M5283P_M.pdf

When I connect only one side of the VCA to power, that side works perfectly. No thump, great sounding tremolo, depth pot has perfect range.

When I connect both sides to power:
I get a thumping in time with the LFO, and loss of volume, and the LFO depth control gets a bit weird.

Different from the Tremolomatic-X:
I've changed the LFO to an opamp Phase Shift Oscillator
I've also had to incorporate a DC offset on the LFO drivers to give the CV input the right static level.
This is using the same DC Offset method that the Tremolo-matic (standard vca trem) uses, with a feed from 0V-9V, split via a 100k resistor, into the inverting input of the LFO driver.
I've also had to AC couple the inverted LFO driver because of the DC offset needed.

None of that seems to make much difference to the issue when I disconnect things, etc.

Is this a CV feedthrough problem?
If so, how would i correct for that if the VCA doesn't have differential inputs?

Could it be a damaged IC? I did bend the pins when I moved it once...but it works fine when a side is isolated...

Could it be a funny breadboard issue?

The VCA V+ and GND ports aren't connected internally, but the V- is...i've tried messing around there, but no dice

Any ideas?
Thanks

[ElectricDruid]

I feel for you. I've got an AS3360 Dual VCA project on stripboard at the moment with similar problems, can't get the thumping out of it when all I want is a smooth silent crossfade from one to the other.

Remember the CV for the second VCA is getting flipped, so any DC offset could get flipped too. That might be why it works for one half, but not with both.

T.

[PRR]

M5283P is *designed* to use LOTS of current.

Is your power voltage steady?

Is the chip getting HOT?

[Joncaster]

M5283P is *designed* to use LOTS of current.

Is your power voltage steady?

Is the chip getting HOT?

Chip is running cool, but I wonder if it prefers to run at higher Voltages. I might try 18v.

Ive decoupled everything I can think of and run sewer grounds, etc...no dice.

I need to try a different lfo maybe, go back to the triangle one and see what happens.

The inverters don't make much difference, even if they're unpowered, I still get thumping.

We'll see...

[duck_arse]

M5283P is *designed* to use LOTS of current.

Ive decoupled everything I can think of and run sewer grounds, etc...no dice.

We'll see...

this is something I was thinking about in the dipthonger debugging - if you have [too  much] supply R//C decouple on the VCA's, when it wants whacks of current it will also sag the voltage.

[Joncaster]

this is something I was thinking about in the dipthonger debugging - if you have [too  much] supply R//C decouple on the VCA's, when it wants whacks of current it will also sag the voltage.

I've added and taken them out in various places, fed the LFO a separate V+, all kinds of combo's...not much changes (maybe a bit of frequency filtering).

I haven't noticed any jumps in the power supply, but my multimeter isn't the fanciest thing.

For the record:
using TL072's all round (4 dual's...is that too much current on top of an already hungry VCA?
I'm using a boss 9V 100ma supply
The virtual ground is from one half a TL072 (like in the tremolo-matic-x)

[anotherjim]

TL072 are rather greedy. I suggest MC1458 for the LFO oscillator, RC4558 for everything else. If you must, a TL072 for the input.
The original idea used TL064 which are very low current but not particularly good for audio. Great for LFO though if you have any TL062 duals.

The VCA's probably needs a very firm power supply, so the "R" in RC power filter can't be large (<100R) but the "C" will be big (>100uf) and the rest of the circuit should have it's own RC filtering (larger "R") and the LFO's are best having their own too. Star distribute the power feeds and grounds and don't run any indicator LED's from the RC filters.

[ElectricDruid]

You could stick TL062s in it just to test the theory. Their noise performance isn't as good, but they are low current. If it's the current draw that's causing the issue, it should improve. If it doesn't, look elsewhere.

T.

[Joncaster]

Cool thanks guys, will work on power some more, have some 062's on the way, and a bigger damn breadboard!

[Joncaster]

Success!
It needs way more power.
I plugged in a 15V desktop supply and rebiased the CV offsets...works perfectly.

I saw on the datasheet that it's recommended +-7 to 18, so that's obviously so the headroom can get above the CV input (GND to +5V) and keep the internal circuit happy.

There was a current draw happening at the CV input at 9v single supply, when both reference grounds were connected. It was dropping 0.5V across a 47k (0.01ma?)
So the whole thing was just squishing against itself.

Now there's no change in CV static voltage, and no thumping at all, up to max depth and speed.

I suppose I'll have to build it as an 18V pedal, or just have a dedicated supply for it.