CMOS LFOs

[Werebear]

Hi All

I'm having a problem. I wanted to build a drone box running off a single 9V battery. There are dozens to choose from but I thought no - make your own. So I went with LM13700 VCOs from the data sheet, the single amp one with the two diodes which I *think* gives a triangle rather than a square wave. I've seen this design used for LFOs. Anyway it works. One LM13700 two VCOs.
Now I thought I would use 40106 oscillators for LFOs, which I have used before with 40106 VCOs quite happilly.
I am using a passive CV mixer - 47K for the frequency and 33K for the modulation with mod being added via pots wired as dividers.
I'm using two 100K resisters to give 4.5V reference and a TL 081 to strap it down. A 100microF cap across the battery and a 47microF from 4.5 to ground.
The problem I am facing is that I can't quell the modulation. Even with the mod level pots "off" I'm still getting modulation through the VCO.

Questions:

Do I use reference voltage as ground fro everything - i.e. my divider pots go to battery ground or Vref, since the LFO output is (I assume) DC close to 9v?

Do I need to put a DC blocking cap on my LFO o/p?

The LFOs and VCOs are on seperate boards. Do I need to seperate them out with 22 Ohm on the live and ground to each?

Or is it that I can't use CMOS for that purpose? Something is going into the power supply causing my VCOs to wobble and I can't figure out what it is.

All this over a 9V battery. Just wait 'till I start building things that run off the mains!

Ta

The Bear

[Cynosure]

A 9V battery is fine for simple cmos ciruits, but even then you should be using a 7805 to regulate the power supply.

As the circuit gets more complex and you start adding opamps, it can cause problems.

I had a very complex cmos system with a few opamps running on a 12V DC supply that that was regulated to 9V with a 7809. I ended up changing it to a simple +5/0/-5 circuit. I am very happy that I did. A lot of the noise went away and stuff started working like it should.

More info on what I did here:

http://electro-music.com/forum/topic-50293.html

There was some confusion at first about the AC adapter, but I ended up using one with a 9V AC output.

[Werebear]

Thanks for that Cynosure. You are of course right, it's time to ditch the PP3!
What I am building is quite complex: two VCOs, Two CMOS LFOs with switchable square tri waves, one psycho lfo with five fixed oscillators and one variable oscillator and - I'm rather proud of this - an MS 20 type filter using two 2N3904s as resistive elements. This was "designed" on breadboard by building the basic topology and then adding and subtracting components until it worked!
The problem I face is that all these "modules" work beautifuly independantly but when I try to normalise them together with pots and switches it all turns to crap.
I have a 12V supply and 5V regulators so I'll build that today and start again.

One question: I assume that I power the CMOS elements from +5V and ground - er of course I do or else the chips will die from dual supply!!! #

Doh! You see what I'm up against...

Thanks

The Bear

[Cynosure]

You can power cmos from +5 and -5. The differential voltage is 10V, which they handle just fine.

Then you have 0V to use for ground on your opamps.

Make sure your 12V supply outputs DC voltage. Also, you need both a 7805 and 7905 for the schematic in the other thread.

Just a warning though - some of your hacked together stuff that worked without a proper power supply will probably behave differently once it is on a proper power supply. It is annoying at first because you have to go back and edit things you already made, but going forward everything should work how it is supposed to work - which makes it easy to follow schematics.

[Werebear]

Hi Cynosure

Well today I made up a + and - 5V power supply using two - count 'em - two PP3s and 5V regs. As you say the total is 10V, only 1V higher than my original 9V split supply but the difference is astonishing. My VCOs are crisp with a much larger frequency range. I breadboarded a new LFO and ran it off +5V and ground. Worked fine and was able to fade it in and out with no weird oscillations. Bit of a thump when I use it to modulate the filter cut off but I'm only using a passive resistive mixer for the CVs. LFO output a bit low but to be expected from 5V. Tomorrow I may try running it from + and - or maybe just use some amp stages with op amps to bring the level up.

I decided not to use the 12V supply - I'll save that for a more "worthy" project and I may just use two batteries and ditch the regulators. That would give me lots of head room but I'll have to watch the CV inputs on the filter.

The schem you showed me uses an AC output from the transformer. One half is used for + and - and the other for ground. You can use a DC but you use a voltage divider for virtual ground.

Thanks for helping out

Bear

[Cynosure]

I'm glad it is working better for you. Just beware that even when done right, a virtual ground is never as stable. It could cause some issues too.

I think the ac adapter was just $9, it was $1 for the 7905, I already had the 7805, and I salvaged the diodes from an old toaster. So it barely cost me anyhting and I got a proper supply that is stable and can be used with a ton of schematics without having to modify it.

Depending on how large the project is, it might be worth it for you to try the ac schematic. Even if you dont end up using it with this project, you will have a stable supply to start your next project.

[JovianPyx]

I bring you some "voice of experience"...

A ground in an electronic system is most often the most commonly used reference in the system. As such, the system works best and is most reliable when the ground is stable. Stable means that it's voltage with respect to other regulated sources never changes (or in reality changes very very little). A virtual ground doesn't fit the description of a "good ground" except in extremely small projects and under the right conditions. If the ground is jumping around then things will act strangely and usually not reliably so. An audio system with a virtual ground is likely to be noisey. "noise" being defined as "that which you do not want".

I'm not saying to never use a virtual ground, but understand that it has limitations. A virtual ground can be "bolstered" with an opamp, but while that extends it's usefulness, it is still limited. Small projects, especially those that do not try to push or pull very much current into or out of the ground can work. When the circuit tries to push or pull too much current from the virtual ground, it's voltage will change. That voltage change is copied everywhere in the circuit that uses ground as a supposed "zero volts" reference.

[Werebear]

Thanks Lads

I've learnt a lot from this project, more so than from other projects that have worked (or not!) and though I've always realised the importance of a stable power supply this is the first time I've had problems. Then again I,m building a box that has two VCOs, two LFOs, a random step generator and a VCF all off a 9V battery I should have realised I would have problems.

Thanks for all your help and I'll let you know how this turns out

Bear