I was going to use a TL431-type regulator like this (view in Courier or fixed font)
7 volts | ,----+---------, | | pnpx2 | | \e e/ 1k |--+--| | /c | c\ | | | | '----+----' | | | | | _|_/ | / ^ +--Vout / \ | Vin---/ \ | ------- 100k | | | | '----+----' | ///or this
7 volts | +---------, | pnp | | e/ 1k ,--| | | c\ | | | +----' | | | | | _|_/ | / ^ +--Vout / \ | Vin---/ \ | ------- 100k | | | | '----+----' | ///I just wanted to figure out whether I should mirror the cathode current like in the first drawing, or could the circuit use some gain. So I started looking at datasheets for three-terminal shunt references, and the information in them is all about how they behave with the cathode tied to the reference pin, whether directly or by voltage divider. Nothing that would help me. So I thought I would simulate it. LTSpice has a model for the LT1009, which is a conventional 2.5 volt adjustable shunt regulator. Now, I know you can use a three-terminal shunt regulator like an op amp, because I have. But LTSpice's simulation was really screwy. In the sim, the cathode voltage doesn't float. I'm pretty sure that's wrong.
Version 4 SHEET 1 880 680 WIRE 64 32 -176 32 WIRE -16 112 -96 112 WIRE -176 144 -176 32 WIRE -96 144 -96 112 WIRE 64 160 64 112 WIRE 32 192 -16 192 WIRE -96 224 -176 224 WIRE -80 224 -96 224 WIRE 64 224 -80 224 WIRE -80 256 -80 224 FLAG -80 256 0 SYMBOL References\\LT1009 64 192 M0 SYMATTR InstName U1 SYMBOL res 48 16 R0 SYMATTR InstName R1 SYMATTR Value 1k SYMBOL voltage -176 128 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V1 SYMATTR Value 5 SYMBOL voltage -96 128 R0 WINDOW 3 42 121 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value PULSE(2 3 0 1) SYMBOL res -32 96 R0 SYMATTR InstName R2 SYMATTR Value 10k TEXT -114 280 Left 2 !.tran 1
Maybe the moral of the story is you can't trust simulations. But it would be nice to get some real info about how these chips behave in more than the limited way the datasheets address, or a faulty sim. Is there some literature out there, or maybe a functioning model?
Here is the datasheet for the LT1009: