Another Bench supply question.. :-)

[N TYPE]

Guys what do you think about this Dual Tracking Linear Power supply kit http://www.altronics.com.au/index.asp?area=item&id=K3218 which was first featured in silicon chip magazine http://archive.siliconchip.com.au/cms/A_112014/article.html I'm sure I can make my own for cheaper than the price listed here as I have the schematics, parts list and board templates already, The thing which I like is that its dual tracking, and as it's a kit I can add extra outputs or modify the existing circuit to suit. Sorry I cant post the entire article as I dont have permission but the specs are as follows:
Output voltage : +- 0-19V
Output current : up to 1.6A (With current limiting feature)
Noise (0-1A) : <5.25uV peak to peak
Ripple (0-1A) : <1mV RMS <1.7mV peak to peak

[zehome]

Always good to build something. Go for it, even if it's crappy, you'll just learn a lot while having fun

[Rerouter]

that 1.6A rating leads me to believe this is a LM317/lm337 based power supply, using the other rail to allow for 0V output, nothing wrong with it, just as a good starting point for investigation into modifying it,

[N TYPE]

Rerouter: You're right its based on LM317/337 and it also has a 5V 7805 output too which is nice.

Wilfred: To me the Linear transformer version is preferable as you dont have to worry so much about noise coming through to the output as you would with a switcher.. but then again Im not sure its really worth worrying about as most laptop power packs are generally built high quality and have pretty good supression anyway, so whatever i suppose

[Strada916]

Its a great start. I am actually in the process of rebuilding My first ever power supply. It was sold at DSE about 25years ago. (Model ETI-162)

[blackdog]

Hi,

Take a look @ the schematic off a powersupply i designed.
Specifications are very good...
DC stability extreme good.
RI DC <0.001 Ohm.
Noise < 5uV RMS.
Hum does not exist.
Fast, small error puls @ dynamic load.
Stable, also in current mode.
Clean switch on/off

But... its all depend on how you build the power supply 
This applies to any power supply that you build.
I do a lot of explaning on this forum, sorry its in Dutch.
http://www.circuitsonline.net/forum/view/110029/5

First the schematic.
www.bramcam.nl/NA/NA-PSU-20.jpg

Testsetup and the fast Jim Williams Dynamic Load connected to the output.


This is the error puls on the output connector.
3mV error and perfectly stable within 20 uSec.
Look at the piece of line between the pulses, no DC shift visible...


This is the puls, witch is hitting on the output of the power supply, from 500mA to 2A.


The same puls, but with 2x 0.5M cabel to the load.
The ringing is the induction of the cables, there is no manufacturer of power supplies that shows this effect.


If you have any questions I will try to answer them.


Kind regarts,
Blackdog

[nctnico]

Its easy to make a clean power supply with a large capacitor on the output. The downside is that the capacitor will be discharged into a load BEFORE the short circuit protection kicks in. That may cause problems. I like how the two quadrant HP power supplies are build. They are basically a power amplifier. A common-mode choke also kills a lot of noise but may affect load regulation.

[blackdog]

Hi nctnico,

I can tell you that it is not easy to build a good power supply...
I dit a lot off work on the reference circuit to keep it clean.
And a lot of testing with wiring to keep in order to keep the magnetic field is small so
that the charging current does not disturb the error amplifier, enz.

The cap on the output off my design is not big, 220uF, you can bring it down to 150uF.
But, this wil give a higher error puls on the output with a dynamic load.
And, the current controll in my schematic is fast and stable.
If you do some testing on commercially available power supplies you laugh yourself silly 
They are in terms of current control very poor, especially slow and instable.

I have here a HP two quadrant power supply and the dynamic load stability is not so good, but i like is, its a good tool te test power supply's,
I made a modulation input on the front for easy testing, say rippel rejection of a uA7812 :-)


Kind regarts,
Blackdog

[Strada916]

Its a great start. I am actually in the process of rebuilding My first ever power supply. It was sold at DSE about 25years ago. (Model ETI-162)

I've got one of those. Geez. I must be getting old. Original design published Dec '82 and it was updated June '84 with a small circuit that fixed a problem with a destructive switch off transient that would fry Q1. That isn't the reason you're rebuilding it, is it?

Hi Wilfred

No. Rebuild is due to mods. 5amp instead of 1 and dual LED readout. And multiturn pot instead of single turn. I was going to do the fine coarse setup.  But found some cheap multi turn pots on ebay. Yes I know they are cheap. New board with the goof proof on board.

[BravoV]

Hi,

Take a look @ the schematic off a powersupply i designed.

...<snip>...

I do a lot of explaning on this forum, sorry its in Dutch.

...<snip>...

If you have any questions I will try to answer them.

Hey Blackdog, thanks for sharing ! 

Just a suggestion, make a new dedicated thread specially for this design and re-post with all the important remarks or comments from your Dutch posting here in this forum in English of course  , rather than buried in this thread. Please !

It will be a nice reference thread for PSU schematic collections in this forum, and also I believe will be useful for others (me included) to have inter-activities on your design here, hope you wouldn't mind.

[Rerouter]

i'm starting to think based on all these various approaches that the only way to pull off constant current without a large fault current is to have your bulk output capacitance switched in and out with the current limiting feedback relying on some half magnitude smaller capacitance for the transition...

[blackdog]

Hi BravoV

I will think about your comments about a new topic for my power supply.
Writing in English is very difficult for me because I'm a dyslect.
It takes a lot of time to translate my brainwave's into English 

Later this week I will start a new topic, it will not be as extensive as on the Dutch forum...

Kind regarts,
Blackdog

[jaycee]

Hi blackdog, just came across this post and your schematic. It looks pretty much like the circuit used by the Mastech supplies, with some refinements I have some questions

1) The voltage control part has me unsure, i know how a regular voltage regulator with opamp works but i guess this is a bit different due to the opamps supplies being floated around the output. The bit that I am unsure about is how the +ve input is defined between Vref and the -ve output of the supply. I am guessing the opamp is comparing this reference with 0V (as far as it's concerned) and adjusting its output until the +ve input matches 0V too?

2) What is the LM317 that's wired as a current sink (sinking 57mA according to your diagram) doing?

I hope these questions arent too daft. Your diagram has actually helped me understand it a bit better, particularly for layout out a PCB - I now know how to layout the sensing points to the opamps better.