LTSpice

Reverse breakdown rating of the gate-anode is also much higher in a PUT compared to an SCR gate-cathode (eg. +40V vs. -5V)

Best regards, Spehro Pefhany

I gathered that from the IKF/ISE numbers. ...Jim Thompson

In my case, the absence of UJT model gives me a decent excuse for delving deeper into LTSpice. My UJT forms a relaxation oscillator in hardware designed during the early 1970s. It's my understanding that the relaxation oscillator was a classic UJT application for that era, but others know best.

Thank you again Jim for sharing all of your Spice data both in this thread and other threads. :)

None of the above accurately models the DC component when a 2N4870 is used in the test circuit. The measured 2N4870 DC excursion is from about

-1V to about +0.25V (IIRC). The 2N2646 in THYRYSTR.LIB comes closest to the AC component, except the observed peaks are much flatter, which makes the waveform look like a pulse with about a 90% duty cycle.

Bordodynov keeps his 2N2646 in a file named Ujt&Put.lib. I may fiddle with this obsolete technology a bit more for the sake of delving deeper into SPICE modeling. Is the Common 2 Transistor Model worth describing?

The common two-transistor version models PUT's quite well,

I have to research the UJT curves and see if I can fit a parametric equation. ...Jim Thompson

But I did just check PS currents. Dandy match to load current _UNTIL_ railing, then bizarro... giga-amps >:-}

I think they don't check their modeling work there at Analog Devices... just sling s*it together. Studying their OpAmp models, it appears to be just a fill-in-the-blanks netlist :-( ...Jim Thompson