I'm planning out my first bench power supply, just ordered

Oh, and no short circuit protection. The circuitry on that is a hassle too. The LM317 and LM337 can handle have the output shorted for a little while as long as the Vin - Vout is less than 40V. So I guess I would only need some sort of protection between -V and +V

And I will be protyping on veroboard

File: 0-60-volt-dc-variable-power-supply-using-lm317-lm337.jpg (320KB, 1224x770px) Image search: [Google]

320KB, 1224x770px

you can use the 2-diode trick to make sure the output goes to 0V instead of 1.2V minimum. also, ten-turn pots are a pain in the ass. i'd go for the more traditional 2-series pots, for coarse and fine adjustment.

>>1073907
yup, definitely gonna go with this schematic. nice

I think the pots are just a personal taste, I much prefer the ten turn pot

>>1073907
why are those two diodes in series there? in case one fails, or...?

>>1074227
I think it's to increase the voltage drop. Diodes dissipate very little power but they drop the voltage from your source by about 0.4-0.8 VDC

>>1073902
You need short circuit protection. The lm317 does not like shorts. I have had to replace mine 3 times in my power supply and ended up adding short protection.

File: 5vdc-power-supply-circuit-diagram-featured-short-circuit-protection.jpg (78KB, 567x560px) Image search: [Google]

78KB, 567x560px

>>1074355
something like this look good?

>>1074359
And I guess I'd have to do three of these, two for -V and +V to GND and one +V to -V

I think the BC547 can be replaced with a general purpose NPN, but I don't have anything like the SK100

http://electronics.stackexchange.com/questions/85187/can-i-use-tip42-as-q1

Know any circuits with more common components? LED indicator would be nice too

>>1074361
TIP42 should be fine. You may want to do the Darlington thing like in the SE answer, just use a bog standard PNP like 2N3906.

Or for extra points, go for a P type MOSFET pass element.

File: lm317 current regulator.png (10KB, 550x300px) Image search: [Google]

10KB, 550x300px

>>1074227
>why are those two diodes in series there?

each diode has a 0.65V drop so two in series creates a -1.3V reference. this is used to cancel out the +1.25V min voltage between the ADJ and OUT pins on the LM317 so the output can go to down to zero volts.

>>1074359
a current limit is suppose to kick in automatically and instantly to save your components from getting fried. just flashing a light doesnt cut it. also, it's supposed to be adjustable, so you set the limit it to, say 100mA, when dealing with a sensitive circuit, and 3A when you're charging a car battery.

one easy way to accomplish it is to stick another LM317 wired as a current regulator before the voltage regulator. the problem with that is you need to use something like a 10-ohm high power pot to adjust the current. or you could use a rotary switch with multiple power resistors.

>>1074413
the circuit I posted should stop current from the 317 and it just indicates the short with an led

on my next psu I'll have a current limit. I think there's a way you can get a power resistor and a 1k pot to adjust

File: minivoeding.jpg (32KB, 579x347px) Image search: [Google]

32KB, 579x347px

Not gonna help OP but maybe inspire others that don't feel like making a fancy bench sup but do want something simple.

I made a small power supply for a couple of bucks from aliexpress parts. Everytime I needed some power in some device or in some project I never had the right power supply or plugs. So I made this thing that takes anything as input (micro usb/bananaconnector/3 types of dc connectors) so I always have something to power it. Input 0-50v. Output can be 0-50v independent of the input.
Output are banana connectors. Made a cable that takes several types of output dc connectors (like 30 of em) and one with crocodile clamps.

Sure it doesn't handle a large amount of power and theres probably some shitty ripple but nothing bad. I'll always have something lying around for input. Great for little devices whose power supply is always lost or some arduino project and so on.

File: my variable power supply.jpg (798KB, 1936x1288px) Image search: [Google]

798KB, 1936x1288px

I always had a preference for the L200 when making basic CC/CV power supplies. It requires fewer parts than the LM317, and usually gives you a bit more current. I don't think there's a negative voltage variant, though.

Picture related, my first DIY L200 based bench supply, from many years ago.

>>1074604
>the circuit I posted should stop current from the 317

it clearly doesnt. there's no signal sent back to the regulator to tell it to shut down. the current limiting circuit is 100% passive. it only lights the orange LED when output 2 is shorted (but not so badly shorted that it brings down the regulator as well). in other words, it's a silly circuit.

>>1074857
It's not the greatest circuit but the pass element T1 will turn off in the presence of a short. Of course that's outside the control loop, so regulation is gonna suck.

File: 20161024_092357.jpg (4MB, 5312x2988px) Image search: [Google]

4MB, 5312x2988px

My home made one is a two parter, 5 amp transformer under desk.

File: 20161024_092423.jpg (3MB, 5312x2988px) Image search: [Google]

3MB, 5312x2988px

>>1074946
Has a fixed 12v output and and adjustable output.

>>1074857
I haven't taken a class that discusses transistor analysis yet, but my prediction was that if the output was grounded it would stop the gates of transistors from allowing drain to well.. drain. Transistors are non-mechanical switches

I have to, of course, test this out, or simulate in spice

(if anybody is good with spice I would appreciate someone testing this out)

>>1074946
huh I didn't know you could use a volt meter as ammeter. did you just put it in parallel with a shunt like usual?

>>1074995

a shunt is the easiest way. (well, besides putting a light bulb in series with a circuit to see how bright it glows)

>>1074995
Volt meter is across a shunt yes. I used a couple low ohm resistors in parallel, mounted on that heat sink and fine tuned with a internal trim pot.

>>1073888
Linears are not very efficient at low voltages as current flowing thru them times voltage drop can be 20+ watts in dissipated heat in your case. Also to220 package ones are limited to 1A - too low for PS imho. Id stick with LM2576 which is little bit more expensive but switched (may easily operate w/o heatsink) and serves more than 3A (up to 5A in my experience, depends on LC setup, also has internal current limiter). I am also fan of multiturn pots, just removed fine/coarse setup today :)

>>1075107
can I use negative voltage with a buck or boost converter? Now sure how I could have a dual supply. I wouldn't worry about except I'm experimenting with op amp as the moment

Yeah and I can see why people like course/fine if you're going back and forth to low and high voltages, I think a 5 turn pot would be just right for me, couldn't find any tho. Plus with multiturns you can get those cool dial knobs that count the # of turns :)

>>1075107
Apart from reducing heat dissipation, a switching regulator lets you source more current: e.g. given a 12V 1A input, you can source 2A at 6V or 4A at 3V, etc. A switching regulator conserves power, a linear regulator conserves current (so lost voltage = lost power).

>>1075113
> can I use negative voltage with a buck or boost converter?
A buck-boost converter generates a negative output from a positive input.

You can make a buck converter which generates negative from negative, but it's more complex (you typically want the logic ground to be "ground" rather than the negative rail, particularly if you're generating both positive and negative voltages).

>>1075143
ok so how to make dual supply from a buck/boost converter?

>>1075107

Switched supplies are generally avoided in lab/benchtop supplies since they introduce more noise into the output. The wasted power often isn't worth worrying about, since it's not expected that you're going to use it to power something long term.

>>1075191
Use a buck converter for the positive rail, a buck-boost converter for the negative rail. The power input only needs a positive rail, but you need a split supply for the sensing (i.e. you need an op-amp to convert a negative sense voltage to a positive voltage for the control logic).

>>1075198
It depends upon the type of the supply.

For sensitive analogue stuff, where even <1mV of noise would be an issue, then linear is preferable.

But bench supplies tend to have lower switching frequencies and somewhat better filtering than consumer-grade wall warts.

For high-power supplies, you don't want to be dissipating hundreds of watts as heat, and the size and weight of a conventional PSU can be an issue (although in colleges with open labs, these can serve as an anti-theft measure).

>>1074359
Wouldn't it make more sense to current sense on the input side so it can regulate right to cut off current?

>>1075231
like a fuse?

>>1075242
No. The current shunt resistor.

>>1075380
There isn't one in this circuit, just a simple short circuit protector

A current shunt wouldn't stop anything anyway

I have not read all comments

But current limitation is a must for a lab supply. It can save your ass many times when you are testing and troubleshooting!

Short circuit protection should be a part of the current limit, the supply should be capable of sustaining full current in to a short!

Next important thing is reverse input protection diodes at the output! These should be high current schottky diodes or TVS.
In case you want to charge a battery and mess up polarity, the the diode will die, and not all of the circuit. It will also allow for series connection of several supplies with different current ratings..

Decouple your supply sufficiently, both with electrolytics, film and ceramic caps.. A noisy supply is a pain in the neck.

Maybe add possibility of remote sensing. To eliminate R*I^2 voltage drop of long wires..

Voltage tracking of positive/negative rails.

I think the 555 as a signal reference is a great idea. Perhaps expand it s bit so it would do 1 Hz - 100 kHz at variable duty cycle and amplitude.

I would also recommend to do it on proper Pcb, it is cheap!
Try Pcbway.com

>>1075458
> the supply should be capable of sustaining full current in to a short!
That precludes the use of most off-the-shelf linear regulators, which use foldback current limiting.

OTOH, buck regulators are ideal for constant-current supplies (just remove the filter capacitor, but don't forget to add some over-voltage protection in case the load is removed).

>>1075468
Just a single voltage regulator will not be capable of sustaining that no. But if you supply a constant current source before the voltage regulator then you should be home safe. This is of course a crude simplification.

I'm also curious to why no one have mentioned the combination of a LM78xx and a power npn like 2n3055

Switched topologies are a possibility, but I would add filters and ideally a post linear regulator, this would help reduce the switching noise.

File: 2016-10-25-12-25-05-807791027.jpg (7KB, 300x168px) Image search: [Google]

7KB, 300x168px

I've got a couple of these on my bench and really really like them.
They can be had for around $60 for you non-poor fags.

>>1075522
What is the P, I and V ratings?

>>1075499
> But if you supply a constant current source before the voltage regulator
What you put before the regulator won't help. Most linear regulators have foldback current limiting, which means that the current limit drops as the output voltage drops (the result is to limit the power dissipated within the regulator). So they'll supply significantly less current into a short than into a normal load.

> I would add filters and ideally a post linear regulator, this would help reduce the switching noise.
That depends upon the nature of the noise. A linear regulator will reduce ripple caused by a low switching frequency and too small a filter capacitor. It won't help much for high-frequency noise which tends to be capacitatively and/or inductively coupled.

Bench supplies use lower switching frequencies than PSUs for consumer electronics. The latter accept switching noise as the trade-off for minimising cost and size, the former don't.

>>1075547
As you mention, most regulators have foldback, not all. As it is a DIY project, it is also possible to make a linear regulator with some opamp and discrete components.

Im well aware that a linear regulator can't suppress 300 kHz switching ripple, at 20 khz it might have 40 dB dampening, but at 100 kHz it might have none.

But consider a dynamic load, which loop is faster?
The switch mode which is "discreet" due to PWM nature, or the linear regulator?
It is not a coincidence that post regulators are used where very low noise is required..

>>1075541
0-30v
0-5
It's a china uni-t

>>1075522
>post crappy chinese power supply in DIY thread
>implying DIYers are poor fags
whew lad

>>1074359
Haven't read today's comments yet but I found someone posted this circuit on the ee stack exchange, and they're adapting it for the LM317. Might have some issues I'll have to check out for the negative supply

http://electronics.stackexchange.com/questions/16608/power-supply-short-circuit-protection

File: 1FCOv.jpg (50KB, 906x356px) Image search: [Google]

50KB, 906x356px

>>1074604
Here's the current limiter I'm gonna use. I think the pot can be anything cuz it's just being use as a voltage divider. And I'm ordering a 1ohm,100W power resistor to set the max current

File: unnamed.jpg (144KB, 1280x720px) Image search: [Google]

144KB, 1280x720px

Here's my rats nest construction of the short circuit protection. Really surprised it worked with just a 2N3904 and 2N3906, tho I'm sure I'll have load limitations and other problems. Oh well

Really tricky to put together on a breadboard layout, hopefully I can get a system to lay it out nicely on veroboard, or point-to-point if I'm desperate. My parts are coming in on the slow boat from China, so I got plenty of time to plan it out ;)

>>1076044
oh and I tested it with my multimeter (alligator clips)

I measured the current from the output of an LM7805 to the input of the transistor, which went from 25mA to about 7mA, with no load, when the red short circuit led indicator turned on

the middle green wire connected to ground is the short