Are UJT's used anymore?

[NMNeil]

I have a circuit diagram for a CDI trigger circuit circa the 1970's, that uses a 2N1671 UJT. Newark is the only place I can find them but at $84 each!!
Looking around for an alternative that's not solid gold and jewel encrusted it seems that UJT are a thing of the past. Mouser only lists 2 and Digikey list 34 but only 3 are kept as stock and the rest are shown as obsolete. Have UJT's been superseded by something else?

 

[Harald Kapp]

9.60 € on Ebay + shipping.

it seems that UJT are a thing of the past.

I agree.

Have UJT's been superseded by something else?

The venerable 555 for one plus everything (well, almost) today is controlled by a microcontroller, e.g. an ATTINY which comes in an SO-8 size package for around 0.60 €.

 

[NMNeil]

9.60 € on Ebay + shipping.

I agree.

The venerable 555 for one plus everything (well, almost) today is controlled by a microcontroller, e.g. an ATTINY which comes in an SO-8 size package for around 0.60 €.

I was considering using a microcontroller, but have read that the interference from generating a spark in a CDI unit does odd things to the MCU unless you design the voltage regulator properly. Uncharted territory for me as a newbie.

 

[Alec_t]

RS has some RS has some RS has some at a modest price, especiallly if you want 5000 of them .
If you post the schematic perhaps we can suggest a work-around.

 

[NMNeil]

RS has some RS has some RS has some at a modest price, especiallly if you want 5000 of them .
If you post the schematic perhaps we can suggest a work-around.

I'm not sure how to post photos or circuits, but I'll try and find out. I have already completed and simulated a couple of circuits as an alternative. One using a 2N3906 the other a CD40106.

 

[Alec_t]

I'm not sure how to post photos or circuits

Use the 'Upload a file' button and browse to the wanted file. Keep graphics files below 1Mbyte and preferably use .png format for circuits.
If you run a sim you can post the .asc file.

 

[Nanren888]

I have not tried to buy ujts since 2015, but they (2n2646) were $16 for a bag full on aliexpress at that time.
A much underrated device in my opinion.
And, yes, I too bought attinys more recently, but not for the same jobs.

A quick search suggested 2N1671 listed at about $7 for ten.

 

[Bluejets]

Futurlec have 2n6028
https://www.futurlec.com/Transistors/2N6028.shtml

Ebay.au have a china source of 2N1671A Motorola @ around $9.00 with post.
https://www.ebay.com.au/itm/1PCS-2N...456467?hash=item3fab7429d3:g:uRQAAOSwQwZacZMT

 

[hevans1944]

UJTs may not be exactly obsolete, but they have had their "day in the sun" as far as popularity goes. IIRC, their main function was to establish timing pulses for triggering SCRs in so-called phase-controlled dimmer circuits. They also were used to make simple oscillators covering the audio frequency range. I enjoyed "playing" with them on the bench in the 1970s, but as @Harald Kapp noted in his post #2, the ubiquitous 555 "timer" chip pretty much wiped out the market for UJTs... while opening the flood gates to millions of electronics enthusiasts in the process, who embraced the 555 wholeheartedly.

And, yes, pair one (or several) with microcontrollers allows just about anything possible as far as digital-controlled timing is concerned. Sure, the 555 is an analog timer and not particularly accurate, but let's get creative, people! @NMNeil, post your schematic for the capacitor discharge ignition circuit. These were popular projects "back in the day" before cars got them as "standard equipment" and controlled by microprocessors too. As @Alec_t says, given the right information it can be built. Shielding microprocessors from ignition interference is just part of normal electronic design for the automobile environment. You want to play in that field, you need to gather the tools and experience necessary to succeed.

 

[NMNeil]

O.K. My very first circuit design so be kind

Starting at the top left C2, R3 and U1A are set to oscillate at about 30Khz. I calculated the frequency to begin with then went to trial and error to get it right. R7 limits the current to the gate of Q3. It also acts as a current limiter later.T1 raises the voltage and it's rectified by D2 then stored in C3 which is a 400V mylar cap. D8 helps it to charge faster by bypassing the ignition coil primary.
As the rectified voltage increases the two zeners D3 and D4 begin conducting at about 300V and when the voltage between D4 and R8 gets to about 3V it turns on Q2 which grounds the gate of Q3 so the oscillations stop. R7 is used to limit the current through Q2.
This part of the circuit works perfectly even though I don't have a snubber in the circuit and Q3 doesn't even get warm.
S1 mimics a hall sensor which has an open collector. When the sensor is activated the S1 side of C1 is pulled to ground and this causes the input of U2B (This should be another one of the U1A inverters) to give a pulse of about 50uS which turns on U4 and discharges C3 via the coil primary giving the spark. R5 is to keep the gate of U4 to ground and C4 is to smooth any odd spikes that could cause it to turn on early. As C3 is discharging through U4 the diodes D6 and D7 give a voltage drop of about 1.5V which keeps Q1 turned on thereby grounding the gate of Q3, again with R7 limiting current. When C3 has fully discharged Q1 is turned off and the oscillator can start up again as U4 is now completely off.
I built the circuit on a bread board and it works just fine but any suggestions on improvements would be gratefully received.
Almost forgot, there is no timing control as the original bob weights and moveable baseplate have been retained. This just generates a spark.

 

[Harald Kapp]

ImHo your circuit diagram is incorrect. When S1 is open, there is no DC connection from batter y (V2) minus to ground. Without this connection the circuit cannot work.
Or am I missing something?

On this part:

when the voltage between D4 and R8 gets to about 3V it turns on Q2 which grounds the gate of Q3 so the oscillations stop.

To be precise: the oscillations don't stop. U1A will continue to oscillate. What Q2 does is short circuit the gate of Q3 to ground thus preventing the oscillations from reaching the gate of Q3. But that is only a minor defect in the description. The circuit should do what you need.

 

[NMNeil]

Sorry, I go the diagram wrong as I messed with it to post.