Need eyes on a 3340 [SOLVED]

OK, I need some eyes (@fredrik? Anyone?) on a schematic and a breadboard because I obviously am doing something stupid and am utterly unable to find it.

I have this

on a breadboard:

The schematic’s pretty much the datasheet, with some unused functions unconnected: Only a single CV, no HF track, and no sync inputs. But ±12 V instead of ±15 V, and for R1 I have 680R instead of 620R and for R7 I have 330k instead of 360k because that’s what I have.

With the trim pot near center, when CV was 0 V I got 1.9 Hz on the outputs. And when CV was 12 V I get… 19.2 Hz. No, not kHz. Hz.

I cranked the trim pot up all the way and now I get:

CV (V) Freq (Hz)
0 2.23
7 2.7
12 263

All component values have been checked two or three or more times, and before turning up the trim pot I took out the 3340 and measured 29k from pin 1 to -12V, 6.3k from pin 2 to -12V. I also tried a different 3340, same result.

I can’t see any mistakes on the breadboard, and I can’t see any differences with the datasheet circuit other than those noted above.

So what’s wrong?

I’d start by hooking hard and soft sync up to something just to keep them at some known value, probably GND.
Also the HFT on pin 7 looks like it’s connected to some important internal circuitry, current mirrors and the like - try hooking that up too.


1 Like

I just did connect HFT to the CV node through a 1M, no change.

With hard or soft sync grounded I get no output at all. (And in the LMNC 1222 hard sync is unconnected, and soft sync floats if no sync input plugged in.)

I read the data sheet as having the HFT grounded through 20k - you probably don’t need the CV feedback path through the 1M.
Maybe the sync goes off when the driving oscillator goes through zero? I wonder if there are any example circuits that don’t use that bit?

OK, HFT grounded through 20k — still no change.

Yeah - looking closer at the datasheet both sync inputs have some pullup/down resistors so they sit where they should if they’re floating, so it shouldn’t be that.

Shouldn’t the circled connection be one hole to the right (to pin 15 of 3340) ?


That’d do it! Thanks!