I sourced 100 LT3080 T0-220 Parts, will list on ebay.

[nctnico]

Nonsense. 99.99999% of the circuits don't really care about their input voltage. Within 0.1V is no problem. Using a LM317-like to built a simple lab supply is perfectly OK. You've got the whole regulation, short circuit current limiting, current fold back (thermal shutdown!) and thermal protection for very little money and effort. When I was a kid I built a triple output power supply with 3 LM317s and a couple of potmeters. I've used it for years.

[Short Circuit]

...
The goal of hobby electronics should be to learn something nad to make things better. Not to blindly solder some parts on a board to make an LED blink. That is not hobby electronics. That's just playing and trampling over a beaten path.

I think you make the mistake of projecting YOUR view of hobby onto others too much.
The goal of a hobby is just doing something you enjoy, or which results you enjoy.

[free_electron]

Using a LM317-like to built a simple lab supply is perfectly OK. You've got the whole regulation, short circuit current limiting, current fold back (thermal shutdown!) and thermal protection for very little money and effort.

Make that 'an adjustable supply' and i agree. Just becasue it is adjustable or thermal protection does not make it a lab supply.

A lab-grade supply needs adjustable voltage and current limit, and bneeds to be able to work in both CV and CC mode. It needs to be robust , protected against all kinds of abuse. It also needs good load and line regulation and remain stable even under strange loads.
Preferrably there is a kind of meter showing set and real values that is trustworthy.

Here's food for thought : even the cheapest chinese 'bench-supply' ( i'm not even going to call it a lab supply ) doesn't use an lm317 ...
Making a 'lab' supply with an lm317 is like bolting one of those fake ferarri or lamborghini body kits on a volkswagen beetle chassis and pretending its the real deal. It just isn't.

[Tepe]

If you can spend the money, get a lightweight PSU like these:
http://www.delta-elektronika.nl/en/products/es150-series.html

They appear to use the "special price for you, sir" principle. No thanks.

[Short Circuit]

If you can spend the money, get a lightweight PSU like these:
http://www.delta-elektronika.nl/en/products/es150-series.html

They appear to use the "special price for you, sir" principle. No thanks.

Know what you mean, but these are not that excessive actually;
from the 2011 pricelist: 420 euro ex VAT for the 30V version, 430,- for 15V and 60V versions and 460,- for the 300V.

[T4P]

I fainted when i saw 420 Euro ...
Okay ... Already so close to TTI

[minime72706]

free_electron, my supply design will have settable voltage AND current and it will use a high-side monitor to monitor the real current/voltage (I won't use a feedback control loop to actually try to correct the output, I can deal with the offset and component variability errors).

As for robustness, the LT3080 is going to be hard to kill, especially if I add a reverse protection diode external to it. It's also supposed to be pretty damn stable, too.
For safety, I'll probably have the microcontroller shut down the PSU if the output voltage exceeds like 10% of the set-point, but I won't be using any sort of control loop otherwise. Unless the DAC flips out, it should output what I tell it to output!

EDIT: Also, yeah, I don't trust the MC34063 at all anymore. I think I saw a similar situation to what you describe, free_electron, where the switch turns on and stays on.
My initial prototype that I made STILL WORKS, but pretty much every other one I built failed instantly, and spectacularly since I had a 24VAC transformer attached to it.

I'm taking what you're saying into consideration. I would like to see just how poorly behaved the pre-current-limiter design is before I decide against it completely, though.

Though I do like to joke about how older (40+) engineers all seem to rage about the overuse of micro-controllers, I completely agree with you. To this day I am unsure if the existence of Arduino is a good thing. I'm leading toward NO because I did my senior project a few years back using AVR GCC just fine and I am NOT a programmer. I would only use the Arduino IDE if I was doing something REALLY SIMPLE and/or was in a hurry for a good reason.

I didn't initially appreciate your ranting and raving, but I'm seeing where you're coming from. Thanks.

[T4P]

It's still wrong to shoehorn a LT3080 into a box with knobs and a high side monitor and pretend it's a lab PSU
It's ... JUST WRONG!

[minime72706]

I've repeatedly said that I don't expect the performance of a lab power supply out of this thing... In the very least I'm going to consider using a feedback loop. I'm not sure how to digitally control such a thing unless I use a digital potentiometer, which I get the impression are not terribly robust.

EDIT: The instance I mentioned where I might use a similar design at work in a general IC test board speaks against this, but we're talking about low voltages and the only thing I need it to do is current limit. If you're drawing a lot of current, the chip is probably FUBAR anyways (silicon bug shorting it out, bond wires shorting out, bond wires incorrectly placed, etc.).

[free_electron]

One of the things that get my stomach in a knot is the classical design where they use an lm317 with a 3 ohm power resistor between its output pin and the input of another lm317.
they put a potentiometer across this one ohm and connect the wiper to the adj pin of the first regulator.

The current drawn creates a drop across the power resistor and they use that to force the first lm317 to scale back its output voltage.
so the current limit works on the principle of denying the second regulator 'breathing room'. This creates all kinds of strange effects. the output can go into oscillaton , the second lm317 doesn't have enough forward drop to regulate properly and it all goes to hell.

This works fine if you connect a quasi dc load but with modern electronics the crest factors can be quite high. That regulation system is not stable at all. Attempt to power something that does pwm speed control of a motor and that supply will go totally wonky.

you can get away with this lt3080 if you are going to build a little thing that does 0.. 15 volts 0.500mA or so.
For me a lab supply is something that can do 0..20 volts 0..3 ampere, has sense terminals , correct setting and readout of what it does , and can be put in series or parallel.

there si a problem with putting two or three of these supplies in series. The output of these things cannot hold much in reverse . If a load dump occurs you will fry the regulator chip.
if you use a big fat power transsitor , those holds 120 volts or more in reverse. The little lm317 or lt3080 can't hold that. 40 volts and it's fried...

Making a decent power supply is not complicated. two opamps , one for voltage, one for current , a reference voltage ( which can be the lm317 ) to derive the setting. and a transistor to let the opamp pull on the base of the power transistor.

Take a look at the schematics of an E3410 for example. the simplest design possible , and there's even two lm317 ' in it. they use one to create a constanct current source to drive the power transistor. all the regulator does is deviate current from the base of the power transistor.

Anyway, to each his own. It just frustrates me that the internet is full of 'dumb' schematics being cloned over and over again without any progress. nobody goes through the effort of really working something out anymore. we are breeding a massive amount of 'recyclers' and 'duct tape ' engineers. ( i use the term 'engineer' to indicate someone who creates a new design . not in it's academical denotation). It's all so watered down these days. Very sad. Copy-paste without understanding what is going on. And with understanding i mean really understanding instinctively how something works. Not just being able to write it down in equasions. A trained monkey can write down equasions.

It's like music almost. anyone can copy the sheet music , some can jot it down on a tone ladder , but only a real musician can 'hear in his mind' how it sounds by reading it. sheesh... i'm getting lyrical now.... time for my dosage of prozac or zoloft i guess. i must be 'down....

I'm kicking and ranting and raving because i care...

[minime72706]

I understand; there are a lot of pretty incompetent people out there that call themselves engineers. I guess I didn't think much about what I was doing, but I'm not trying to design anything amazing anyways.

I don't expect to be driving a highly inductive load with this supply and why not protect it with a zener connected from output back to the input? (to prevent the output switch transistor from breaking down)

I intend to output 20-30V @ 2A for the adjustable side and probably a fixed 5V supply @ 1A. I still haven't found a transformer that makes me happy, but I'll settle for a 9V/18V secondary - it's a lot better than a single 24V secondary, which is a common voltage.

[minime72706]

As a random aside, I just sampled a bunch of parts to play with from Maxim-IC (DAC,ADC,current-shunt monitor, opamps, some other crap) and then it dawned on me when  I was looking through my random development boards. The ATXMEGA chips pretty much have everything I want integrated into one chip (assuming I can tolerate however crappy the internal DAC/ADC might be). I've been looking for an excuse to play with those chips for a long time.
It even has a battery backup system, which I thought might be useful because if there's a fault in the power supply, such as the output being shorted, I don't want the voltage droop (from the short) to shut down all the regulation and control circuits.
Using a MCU like the xmega with a couple op-amps and a current-shunt monitor seems like a more cost effective and practical way to do this, but meh, if I have some high performance parts from  Maxim, I might as well use those too.
I might even be able to find a use for the xmega's internal comparators.

[minime72706]

I have a feeling that I'll be dead and buried by the time I receive those parts...

[bingo600]

I have a feeling that I'll be dead and buried by the time I receive those parts...

From Maxim no ...
From Atmel yes ..

/Bingo

[minime72706]

Yeah I knew as much about Atmel, but I'm talking about the LT3080s from Linear Technology - I think they might not have them on hand right now.

[minime72706]

Oh neat; looks like they're being delivered today! I can finally build something... maybe I should make a different thread since free_electron and I derailed my thread... thanks for the advice, though!

EDIT: They faked me out yesterday, but I just signed for them at 9:15am EST (my time).

[minime72706]

THREAD SERIOUSLY DERAILED (with useful banter, actually), BUT THEY'RE LISTED ON EBAY NOW.
I'LL CONSIDER BEST OFFERS SINCE $25 ISN'T A HUGE SAVINGS, BUT IT DEPENDS ON HOW MUCH SHIPPING MIGHT BE.

I'm selling them in 5-unit quantities.

I hope I can get rid of at least some of these... if not they're good regulators to have in my gigantic hobby kit I guess!

[minime72706]

Yeah, I guess I'll have to generate a product and sell it to make it worth it.

[minime72706]

I'm going to use an external pass transistor with the LT3080 just to stick it to you guys. It'll basically be an op-amp at that point... if you ignore the weak current source.

[minime72706]

This thread was severely derailed long ago, but I just wanted to bring it back to the top in case there's any interest.

http://www.ebay.com/itm/Lot-of-FIVE-LT3080-LDO-Adjustable-Linear-Voltage-Regulator-TO-220-1-1A-LM317-/200792817644?pt=LH_DefaultDomain_0&hash=item2ec02f3fec

I'm not sure why such a long link is necessary, but there it is. I'm selling them in quantities of 5. MAKE AN OFFER IF YOU DON'T LIKE THE PRICE AS IT IS.

Thanks.