Depletion FET regulator

Should be extremly clean, what were you fearing?

piglet

Startup, the fet acts lilke a resistor until it regulates and flattens out. Regulation will be mediocre.

Shutdown, as the input drops it may discharge the load through the fet substrate diode.

May as well buy an LM78xx or something.

Depletion fets are great in places, but that's not one. It might be OK for a high voltage reg where ICs aren't available.

It might fry if the load shorts.

I don't understand Fig 12. Why two resistors?

It's a solution looking for a problem.

I like the SOT23 Supertex part and an LED as a power supply cap discharger and warning. Nice linear discharge.

I had similar thoughts - only I thought jfet, 4391 or so. Until I saw the 600V D-S voltage.

The LND150 (or 250, looks the same) is rated 500 volts. Unlike jfets, Idss is very consistent between parts, about 1.6 mA for these.

Supertex has a bigger part, DN2530, in SOT-89. Nice.

Real low voltage zeners are awful too. Who wants a 5 volt regulator that might be 4, might be 6?

One could hang a depletion fet on top of a 3-terminal regulator to extend its input voltage range. That might need zero additional parts, ignoring the heat sink.

I'm designing a kind of evaluation board for a SPLC where it might be nice to accommodate a wide range of input voltages, since the part is mixed-signal and could be interfaced to higher voltage stuff, relays, etc..I may put some kind of open-collector output drivers on the board also so it might be nice if the whole thing could be powered by whatever the driver's output is tolerant to, say 50 V.

The SPLC handily supplies some spare internal comparators with a selectable Vref on their inverting "pin" so I was thinking just add a depletion FET and the part could become its own wide input range linear regulator.

But I have concerns about how a depletion FET would behave in that role, particularly on start-up, with the chip trying to boostrap itself to be part of its own regulator. Maybe that's too clever.

Hmm, actually if the pass device is depletion N fet I guess the error amp needs to be configured like a emitter/source follower, not a PNP ldo.

I don't know about SPLC and how cleanly its vref starts up but a plain zener and depl fet would be clean. Imprecision of zener could be solved by using TL431 instead.

piglet

That app note circuit is fine for a simple pre-regulator but not going to be useful for getting a well defined output voltage because in addition to the zener tolerance there is the poorly defined Vgs.

Much better use a TL431 or LM4040 with feedback from the output.

piglet

SPLC?

Ope, SPLD, rather

Small Programmable Logic Controller Simple Programmable Logic Device Simple Programmable Logic Controller

Etc...is the term "controller" reserved for integrated modules but an individual chip is a "device"?

Take a look at SRH05. It's a little potted 3-pin switcher that tolerates up to 72 volts in. It's very well behaved.

When under-volted the FET channel is conductive. also the internal diode will allow current to flow from the regulated side out through the supply terminals.

for repeatable results probably it wants TL431 or something like that instead of a zener.