I hesitate at the word "always". I think there was significant variation between samples. What most sticks in my mind was asymmetry in the waveform, and I think I sometimes saw spikes at the peaks and valleys. As I remember, the circuit operation seems to start with the timing capacitor which is connected differentially (not one end connected to ground.) There is a significant "switch" which happens when the differential voltage across the timing capacitor stops slewing in one direction and starts slewing in the other direction. Mismatches in the circuitry seem to cause waveform asymmetry and probably also spikes. But I also think that if you have a better sample, the output waveform looks pretty good.

Somewhere I might have a hard copy of the datasheet. This datasheet would have predated my awareness of .pdf files by 25 years and (for that matter) also would have predated the IBM PC by at least a few years.

Many years ago, when this chip was new, I bought a kit (I believe from EXAR) that used two chips in a signal generator. One chip modulated the other to provide variable frequency or voltage. It used off-board rotary switches to select the operating mode. I still have the board with the two chips, but I have lost the schematic for the kit. Do you remember the kit, and would you know of the schematic?

Moderator edit: New threat created from this.

 

Post photographs of both sides of the board and maybe we can make you a schematic.

 

Have you checked the Exar's datasheet? I vaguely recall it showing more than one circuit.

 

XR2206? The data sheet has suggested circuits, but googling should find schematics of commercial function generators based on it.

The other popular chip was Intersil ICL8038; same applies to that.

 

I have the EXAR data book back in the archives. Give me a day or two to locate it and I may have the circuit on hand.

 

Hello,

You can find a lot of databooks and sheets over here:
https://bitsavers.org/components/exar/

Bertus