/diy/ - Do It Yourself

>>2049648
from each according to his ability, to each according to his sneed

>>2049654
You're quick. Best I could find was the JT9648-AS, an integrated thermometer and multiplexed LCD driver. A bit like how last thread there was that guy wanting a voltmeter with integrated LCD driver. But this thing isn't very versatile since the LCD needs the segments to make an F or C, and the degrees too.

>>2049648
how many labor vouchers did a KT315 transistor cost?

>>2049661
that's a nice one. oddly specific. I'm pretty sure nobody would have been interested in a 105Msps ADC anyway

>>2049670
>105Msps
holy shit it was
the AD9648 going on my "maybe put in an SDR" list

>>2049682
wouldn't a 105msps sample rate mean you can only sample signals with a frequency of 53mhz before you hit aliasing?

Does anyone have a good suggestion for a fairly simple but functional DIY lab supply project with voltage control and current limiting?

Bonus if it has one adjustable output and a couple of fixed 3.3v and 5v outputs.

>>2049692
Yeah, but that’s still pretty good for a 14-bit ADC. IIRC the best one I found was the WM8224.

>>2049698
Do you need the noise immunity that comes with a linear front-end? If not, power-brick + LM2596 module + replacement pots + panel meter is a very cheap way to get a decent power supply. Not sure what the minimum voltage is though, I think I ran into issues around 3V.

For a linear front-end however, I couldn’t tell you what the best option is, besides something with a bunch of op-amps for fully-linear. Switching with 1-2V of linear should be possible with an existing switching controller and some op-amps if you know what you’re doing. Might require a variable-reference switcher though, as opposed to a fixed-reference one where you vary the feedback divider. Autoranging multi-tap transformer with linear fine adjustment is ideal, and probably not too difficult for a hobbyist to make or design, but I can’t think how I’d design one besides a whole bunch of comparators, and the BOM-cost gets up there. Also the output ripple when current limiting kicks the voltage down to a different winding might be difficult to handle.

>>2049728
I'm mostly looking for something that'll be used with digital circuits including some PWM stuff and analog inputs, right now, so I think a little noise is okay?

I like the idea of being able to put a few modules like that together - would I be okay powering it and a couple of linear regulator based 5/3.3v lines off of the same power brick?

Pretty much an amateur with only one previous "big" project under my belt. I have been reading up on more bespoke power supplies, would like to try one later, but for now something simple would be better, I think.

>>2049735
>digital circuits
Purely switching cc+cv would arguably be better than using a 7805 or whatever for digital, since the current regulation would be pretty handy in the event that you shit up. But if you're using an ADC at all, or otherwise doing analog or noise-sensitive stuff with your logic rails, you'll benefit from a linear front-end. Do the calculations on power supply ripple rejection (at different frequencies) of your analog inputs compared to prospective power supplies to see what's worth doing.
It might be worth pursuing a simple op-amp+darlington high-side current limiting circuit you put before the linear regulator for this reason, wouldn't be that complicated at all. I think if you used an adjustable regulator you'd be able to put the current sense resistor after the regulator itself without getting any nasty dropout and removing the need for rrio op-amps. Throw some designs at a sim and see what sticks first though.
There's also always the option of going full-ham with LC pi filters, on both switchers and linear designs. They also rack up the BOM-costs, but if you've got a bunch of salvaged caps and/or inductors it isn't a bad idea.

The current limiting is also hardly necessary if you're confident not to accidentally short something, they're mainly useful when you've got a dumb load like an LED or whatever. In the case of such a dumb load, it's fine to just use the cheap LM2596 solution or equivalent.

Personally I'd make a dedicated digital power supply seperate to a dedicated analog power supply. For the analog supply I'd want split-rail, but that isn't very necessary for digital at all. If you don't plan on doing audio stuff that would warrant the analog supply, leave it out.

vector lainscope anon (>>2049010) again
i have realised my shortcomings
i failed level 2 computer science
i am a pajeet programmer
i will try to interpret existing svg files instead of that garbage
but the lainscope will live

>>2049682
beauty, isn't she. $75 tho. not the sort of thing you love to see burned