Drive two 13700s with identical currents

dslamnig

I'm trying to control two 13700 OTAs with the same current. One OTA is part of a triangle oscillator, the other charges the cap of a slaved saw oscillator. The idea is to make a dual-core VCO instead of doing triangle-to-saw conversion (or vice versa).

The circuit is on the breadboard and works, however the saw waveform shows artefacts which are (very probably) caused by crosstalk between two OTAS.

The control current is coming from Thomas Henry's 555 VCO exponential converter:
https://electro-music.com/forum/topic-54623.html

... and goes directly to both 13700 amplitude bias pins - so they're connected. And I see a +/-10mv voltage fluctuation at this point, synced to the VCO waveforms, which is caused by Iabc input impendance fluctuations of _both_ OTAs.

So the question is - how to drive two LM13700 amplitude bias inputs with identical currents, from two independent current sources?

Ideally this would be some kind of "current-splitter" that I could plug in between the existing current source and OTA Iabc pins. Tried a pair of resistors, no luck...

blue hell · Posted: Fri Dec 18, 2020 4:24 pm Post subject:

Ok .. some thoughts that bubble up here ... and also welcome

to the forum.

Do you want the two oscillators to be in sync or should they be allowed to relatively drift away from each other?

Asking, as you'll never get the two currents perfectly equal.

So .. when they should drift, probably the best way would be to have two expo converters.

Or maybe stack a pair of parallel current mirrors on top of each other (one for the doubling and the other to let the current flow in the right direction), that would isolate the currents from each other .. but is seems a bit over complex, using 5 or 8 transistors (and some trimming resistors / pots).

Also ... the transconductivity for the two OTA's will differ from each other when running 'm on the same Iabc ... so you will need some trimming to get 'm begin and and at the same frequencies (on a sweep over their range).

And when you want 'm to be in sync then a tri to saw or a saw to tri converter would be the only thing properly working I think.

dslamnig · Posted: Fri Dec 18, 2020 4:54 pm Post subject:

blue hell · Posted: Fri Dec 18, 2020 6:09 pm Post subject:

elektrouwe · Joined: May 27, 2012 Posts: 146 Location: Germany

dslamnig · Posted: Sat Dec 19, 2020 3:52 am Post subject:

dslamnig · Posted: Tue Dec 22, 2020 10:46 am Post subject:

The previous circuit provides more current for each OTA, which is not bad, but doesn't do real isolation. The OTAs still talk to each other through emitters (because emitter voltage is floating). Here's a circuit which really provides two independent and hopefully identical currents.

I've tried both on the breadboard, they work as expected. The second one is better and tracking between currents is good enough for my purposes. But after all the experimenting I found that 2 LM13700s will consistently draw equal currents even if connected to just one current source. My saw waveform issues remain, and now I know that they're caused by something else. So I'll probably revert to one expo version. But knowledge is a reward in itself, at least in my book Wink

tri_saw 003.jpg

Description:

Filesize:

147.5 KB

Viewed:

156 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.