You need to get one of these:
View attachment 286146

That would neaten things up - I was trying to move the wires so the connections could all be seen.

Here's a breadboard with 3-4 different circuits.

I'd like just one circuit that works!!!

 

The pull down on the PNP gets hot, not the one on the NPN that you suggested.

Pull down on what? I suggested putting something smaller directly across the LED, not the LED and it's current limiting resistor. When the LED is on, a 1k resistor would be dissipating 6mW. When off, it would be around 1.5mW (at 1.2V).

 

That's not right. The LED should be ON when the Collector of Q3 is 0v and OFF when the collector is 23.9 volts.

I don't know what to say - I've just rechecked.

 

Are you sure your 2N3906 is the right way round...

Here's my version of your cct, and I get a few mV and <1uA when off and 24v and 43mA s/c when on, exactly as predicted.

 

I see you're using one of those inexpensive, to put it kindly, solderless breadboards.

You can use a 5 pin male 0.025" square header to align the clips with holes for easier lead insertion.

For holes indicated, slant the header pins to the left, then lean them to the right to force the clips to align with the holes. I use a male header with long leads and pointed ends.

This is the part I use:

 

Are you sure your 2N3906 is the right way round...

I was about to ask that question!

Here's my version of your cct, and I get a few mV and <1uA when off and 24v and 43mA s/c when on, exactly as predicted.

I will try to reverse engineer that picture - unfortunately I have to go out for a bit now.

 

I was about to ask that question!


I will try to reverse engineer that picture - unfortunately I have to go out for a bit now.

Would you mind posting a top down picture of your board so I can copy it in the morning? Please!
(are you using 5v & 24v - because my circuits worked at 5v only)

 

I will try to reverse engineer that picture - unfortunately I have to go out for a bit now.

Transistor orientation looks correct to me.

 

I was about to ask that question!


I will try to reverse engineer that picture - unfortunately I have to go out for a bit now.

The 2N3906 is the nearest transistor, emitter to +24 (EBC looking at the flat side). I used a 2N3904 for the NPN which is the same pinout so will appear reversed BUT the 2N2222 is CBE (I didn't have a 2N2222 but other than pinout they are similiar) so it will orient the same way - assuming you maintain good breadboard practice and keep supply and ground roughly on opposite sides!

Yes, 24v (red trace) and 4v drive (blue trace) on scope...

 

BUT the 2N2222 is CBE

That brings up an interesting point. Technically 2N2222A are in TO-18 cans. TO-92 should be called PN2222A.

onsemi PN2222A, 2N2222A, respectively:

 

Transistor orientation looks correct to me.

Agree and from the photo I drew this updated schematic. However the TS has stated twice that the LED was OFF when the collector of Q3 was 0v. This is the complete opposite of what should be happening. So something is wrong and this simple circuit is turning into some kind proverbial nightmare.

 

That brings up an interesting point. Technically 2N2222A are in TO-18 cans. TO-92 should be called PN2222A.

onsemi PN2222A, 2N2222A, respectively:
View attachment 286157 View attachment 286158

I think the mystery might be solved... the newer lead-free devices actually labelled P2N2222 are TO92 style 17, CBE as I asserted NOT EBC style 1 like the 2N3906

 

I tested the circuit using a 2N3904 reversed and the LED is constantly ON.
In fact just for fun I switched Q2 and Q3 around. No matter what incorrect combination I used and even installing the transistors backwards the LED was constantly ON. The SPST switch had no effect other then slightly dimming the LED in some configurations as mentioned.
So something is still wrong.

 

In fact just for fun I switched Q2 and Q3 around. No matter what incorrect combination I used and even installing the transistors backwards the LED was constantly ON.

That's because the base-emitter junction breaks down at around 5-6V. Make sure you mark those transistors or throw them away because you've destroyed current gain.

 

That's because the base-emitter junction breaks down at around 5-6V. Make sure you mark those transistors or throw them away because you've destroyed current gain.

They check OK I was testing at 5 volts.

 

They check OK I was testing at 5 volts.

I'd still mark or toss them. You had to break down the BE junction if the transistor was always on.

 

Photos of the wires all over the place on a breadboard is a nightmare. If the resistors were neat and tidy then we can see what they are connected to.
The resistor color codes are not correct. Blue, grey, black, black black for 6.8k?? Where is the orange band for the k?

 

The resistor color codes are not correct. Blue, grey, black, black black for 6.8k?? Where is the orange band for the k?

I see blue, gray, black, brown and brown. 5 band resistor code for a 6.8K at 1%.

 

Can I just say a HUGE thank you to everyone who helped me resolve this. I had one transistor image on my phone and one on my laptop - one was for a 2N2222A (the one I'm using) and the other a P2N2222. The collectors and emitters are opposite of each other - I was blaming my brain fog for not being able to remember the pinouts - thank you @Irving for making me re-check, double check and then check again.
I also took @dl324's advice and threw out all the transistors that I may or may not have wired incorrectly and lo and behold, it works as expected.
So thank you again for your patience - I have learned a lot: cutting down resistor leads makes the breadboard connections a whole lot more reliable, badly connected transistors won't work as expected but mostly I've learned that very similar component numbers can really bite you on the bum but AAC is still a really friendly place!

 

I had one transistor image on my phone and one on my laptop - one was for a 2N2222A (the one I'm using) and the other a P2N2222. The collectors and emitters are opposite of each other

There are 3 pinouts for TO-92: EBC, CBE, and BCE. 2N and BC are reversed from each other, 2SC put the collector in the middle. But watch out for reversed pinout devices. I have a bag of some BC transistor that was marked reverse pinout.

That 2N2222A in TO-92 always gives me heartburn. 2N2222A was always in a TO-18 package. What were they thinking?

From Texas Instruments 1973 Transistor and Diode Databook: