EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Bortmoun on May 23, 2018, 05:58:32 pm
-
Hi there :) . I need a good bench power supply, which i want to build myself, for my hobby lab. The thing is, i have been thinking about LM-317 as my linear voltage regulator (and because here where i live i don't have no access to better ones). The design must have these features:
a)Current control from few mA to 5A
b)Voltage control from 0V to 30V.
For this purpose I've built this schematic, which i attached here. Please, tell me what you think about this project or if does work or does not, for i am a beginner in these things. R8 is just a dummy load.
The things i need help: :scared:
1) How can i add a minimum load (10 mA as by LM-317 datasheet) to keep the power supply operating in different potentials?
2) I need a replacement for 2n3822, i can't find it in my country too.
3) Can I run it from a fixed switched power supply (way more cheaper than buying a transformer here)?
4) In the future, I'm thinking about adding digital control with PWM to this circuit with Arduino or something like that. Do you think this is feasible?
Thanks in advance and sorry for my English.
-
2N3822 is current source
Bridge rectifier and capacitor have bad PF , more loses in rectifier and transformer.
Transformer is more durable than cheap SMPS
-
The current through the pass transistors is not going through the current shunt. Regulating the drive current to the transistors is called beta current limiting and it has poor accuracy which is sometimes suitable for overload protection but not for current regulation.
-
A beginner? :wtf:
I've been messing about tinkering for over a year now and wouldn't tackle building a power supply myself! My head doesn't work properly now though (official!! :scared: ) ;D
I hope you get yours built. I'm sure the guys here will keep you right where you need help.
:-+
-
Hi, I started a topic a while back about a µC based PSU, haven't got to test it yet but simulation runs fine. I'm thinking in using an STM32 and it's own ADCs and PWM to make the interface, not much to it, easy to get parts.
The circuit is lacking a few things as I mention in the post but I wanted to be sure about stability and regulation of the basic schematic before getting into adding more stuff to it. I'd like to add OCP as it's easy enough using an extra PWM signal and just turning the voltage low to the output device. I don't know about OVP as it should have an extra power transistor to clamp it down, but I might get there. The idea would be a modular design, this would be the analog control board, having a separate power drive board and µC board would make it easy to change power capabilities with some changes in the power module, speed and precision changing the µC board. This board should manage output sensing and feedback loop.
https://www.eevblog.com/forum/projects/lab-psu-(analog-circuit-for-c-driving)/msg1542611/#msg1542611 (https://www.eevblog.com/forum/projects/lab-psu-(analog-circuit-for-c-driving)/msg1542611/#msg1542611)
I don't know when it will be ready, I'm waiting for my new scope to arrive to be able to test some things (should be less than a month for that) and to have some spare time, while I do need something like this for the lab and I can't find anything around, while designing it I have in mind to be easy-ish to build (it's a lab PSU, nothing really easy there, just not using BGA mosfets or funny ADCs) and components easy to get. I'm not aiming for high resolution, high precision or high accuracy. A functional PSU so anybody can build it.
Note that the ADC board could be swapped for a few pots and panel meters if desired, as the µC is just feeding reference and measuring output. I do have a firmware already tested for the basic stuff, so when I build the board I have right away something to play with. I should just build the power board as is the one that can't be bodged in a protoboard.
JS
-
You don't need rare JFETs for the classic LM317 'lab' supply. See https://www.eevblog.com/forum/projects/lm317-_labratory_-power-supply/ (https://www.eevblog.com/forum/projects/lm317-_labratory_-power-supply/)
However if you can get a LM723 chip I strongly recommend using that, as its far far easier to add external pass transistors reliably than with a LM317. There are some old Philips LM723 based PSU schematics that have been previously posted here. See https://www.eevblog.com/forum/projects/how-do-make-a-current-limiting-knob-using-lm723-for-linear-reg-power-supplies/ (https://www.eevblog.com/forum/projects/how-do-make-a-current-limiting-knob-using-lm723-for-linear-reg-power-supplies/) - Oldway's reply #7 for the schematics and my reply #29 for a LTspice sim of one of them.
-
I would say ...take any hp power supply manual that has schematic in it and start fro there as I did
-
You don't need rare JFETs for the classic LM317 'lab' supply. See https://www.eevblog.com/forum/projects/lm317-_labratory_-power-supply/ (https://www.eevblog.com/forum/projects/lm317-_labratory_-power-supply/)
However if you can get a LM723 chip I strongly recommend using that, as its far far easier to add external pass transistors reliably than with a LM317. There are some old Philips LM723 based PSU schematics that have been previously posted here. See https://www.eevblog.com/forum/projects/how-do-make-a-current-limiting-knob-using-lm723-for-linear-reg-power-supplies/ (https://www.eevblog.com/forum/projects/how-do-make-a-current-limiting-knob-using-lm723-for-linear-reg-power-supplies/) - Oldway's reply #7 for the schematics and my reply #29 for a LTspice sim of one of them.
Actually, my first option was the LM-723, but I was told it was old and obsolete. Oh well, I think i'm gonna rethink the hole project again haha
Thanks for the replies, people!
-
Its old, its internal pass transistor cant handle much current or dissipation and its current limiting function isn't very sharp, so when its in current limiting mode, its output current is still somewhat dependent on the load voltage, but its still one of the easiest ways of implementing a wide voltage range linear PSU with external pass transistors.
One thing to be aware of is it doesn't offer any sort of thermal overload protection, so a robust design would need a temperature sensor on the heatsink and a latching comparator circuit to force the LM723 off by pulling its Comp pin low (or tricking the internal current limit to do so) if the heatsink temperature threatens to exceed the maximum temperature at which the pass transistor(s) can handle the min. voltage, max current dissipation. This could easily be done with a NE555 with its Discharge pin connected to LM723 Comp and a thermistor in a small hole in the heatsink + a resistor in a potential divider feeding its Trig pin, and a pullup resistor and cap on Reset so it powers up with its output high.
-
There is nothing wrong with the 723 but to add precision constant current operation, a second error amplifier is needed which might be another 723 or perhaps an external operational amplifier.
3 terminal integrated regulators like the 317 are not really suitable for a constant voltage and constant current power supply either without external help. They are however useful as power pass elements with built in protection or for driving discrete power transistors and extending protection to them. National made a variation of the LM317 in the form of the LM395 "ultra reliable power transistor" intended for exactly this but they are expensive compared to LM317s. The first schematic below shows an example of a power supply designed for these.
The second schematic shows how to extend the protection of an LM317 to discrete power transistors. It relies on beta current limiting to extend the current limit of the LM317 to the discrete transistors and properly mounting the LM317 and discrete power transistors in the same thermal environment to extend the LM317's thermal protection to the transistors.
The last schematic shows a simplified version of what you were attempting and I assume it was the inspiration. There is nothing wrong with this design but it is difficult to extend the maximum output current and it has a high dropout voltage. The minimum required 10 milliamp load current is conveniently supplied by R3 which is part of the output feedback divider so no separate circuit is required for this.
-
Actually, my first option was the LM-723, but I was told it was old and obsolete.
The 723 is part of the walking dead. It will, just like the 555, 2N3055 and 1N4148 never die.