philba said:
What is the concern about oscillation of a linear vreg? low load?
Eh, there was an
element of that, but I figured I could solve that with just an extra resistor in parallel
The main issues were (
deep breath):
-For the LM317, the dropout is generally listed as around 2V or so, somewhat less if it's hot or is using little current (factors which sadly don't tend to go together!!); The big PSU brick that I refer to as "about 15V" is really rated at 13V@2A, but merely happens to give more when it doesn't have as much load. I can't really trust it to be above 15V in use, especially when I don't know how much the programmer will draw and my voltmeter is somewhat skewed. So I'd expect a good chance of dropout there- hardly a certainty, but not something I could reasonably rule out. I'm not certain if dropout leads to oscillation, but regardless it seems something to avoid. I should be able to test this later though.
-I'm not buying
any more components if I can possibly possibly avoid it-I've kinda drawn a line in the sand by now. I'm probably more of a Scrooge than Audioguru there
My inner miser will not allow me to spend £5 or so for delivery of one 10p component, and I know that if I go to buy one thing, I'll want to buy a ton of other things that I feel I need too, etc- I know this because the list is already huge
Can't afford to right now, so by refusing to buy anything more, I've been exercising my ingenuity, knowledge, etc rather than my wallet
-The obvious answer to the potentially low headroom would be an LDO or quasi-LDO regulator, but I haven't got one already and as you see, will not buy one if I can avoid it. So the plan WAS to try to build one of my own, with a 358 op-amp, a 3v6 zener, and a few resistors and transistors
I've worked out a few different designs for these, but have little confidence in their stability, and even less in my capacity to work out the bode plot for them!!
My best guess is that I could try sticking in a pole at about 0.1Hz and then maybe a zero at around 10Hz, and forget about having any sort of bandwidth, relying instead on the capacitor for filling in transients. I still feel like I'm in way over my head with this part though, I don't even know how I'd
put poles and zeroes into the circuit! As I said, I'm not too good with the AC side of things. It'd be interesting to try once I can
check if it works as intended though
I've been wrestling with this one (on and off, I do other things too!) for some months now, and as you can imagine I'd now much rather put it aside for the time being and simply start off with something quick-and-dirty that I can make work and then just dismantle again once it's served its purpose.
If I had a stopgap power+Vpp board, I could program up some chips to help check all sorts of things: the total programmer load, various voltages, and testing for signs of bad oscillation (like big spikes and dips) in any prototype LDO. I could probably do some of those things without any PICs too, but I'd much rather do them with
Then I can make a good, informed decision about a replacement power board.