Precision +/-15V @ 1A PSU

Uncle Krunkus · Posted: Wed Oct 01, 2008 6:29 pm Post subject:

Here you go,
Ask if you're unsure of anything.
This is only the 1Amp version. Exactly as shown in the schem on Laurie's (Elby) site.
Note the orientation of the ICs, and the links underneath them. I can mod it for rectifiers if you want.

PSU-1A-Strip.zip

Description:

Download (listen)

Filename:

PSU-1A-Strip.zip

Filesize:

161.29 KB

Downloaded:

235 Time(s)

Luka · Joined: Jun 29, 2007 Posts: 1003 Location: Melb.

thanks

Uncle Krunkus · Posted: Thu Oct 02, 2008 4:05 pm Post subject:

I decided to put the indicator LEDs inside the Ammeters, to light them up. Red for the +ve rail, Yellow for the -ve.
If you use the same type of meters, be careful about the placement, so you don't foul the front cover. You'll have to pop the front cover off while you drill. The needle assembly is very fragile, so be extra careful, and get the cover back on ASAP. Wink

The material is a bakelite type of plastic which cracks easily, so drill in stages, ie 1.5,3,4,5mm bits.
I found that 2 LEDs in series work well if you change the current limiting resistors (on the PCB) to 820R. I still haven't tested them in the meters though, so this might change.
Here is a photo of the completed meters.

10-Meters.jpg

Description:

Filesize:

199.82 KB

Viewed:

134 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Uncle Krunkus · Posted: Sat Oct 04, 2008 5:07 pm Post subject:

Just a quick post to let everyone know that I powered it up this morning for the first time, and it seems to be working perfectly straight off the bat! Very Happy

Using 25 turn trimmers, there's about 1/8th of a turn per .01V, so it's dead easy to dial in exactly 15V per rail. I could just as easily dial up 14.99 or 15.01 if I wanted to! Very Happy

The LEDs in the meters are not bright enough for me (though I haven't seen them at night) so I'd probably suggest for two standard 5mm LEDs per side to go back to 680R as on the original schem. I also need to update the stripboard layout, as I added a screw connector for the chassis earth to 0V line.
I'm going to do some current tests now to see how accurate these meters are, and how warm my heatsinks get. I also want to check how stable the voltage is under load.
There will be some more construction pics too, in the next couple of days.

Uncle Krunkus · Posted: Sat Oct 04, 2008 5:17 pm Post subject:

Here's a couple of pics of the front panel.
It was actually heaps harder than the back one, as it's about 3mm thick, and my nibbler can't get it's jaws around more than about 2mm. A metal fretsaw, step drill, files etc. come in handy for this stuff.
The 4way mic-type connector on the front is an auxillary out for testing, prototyping, and possibly a handheld softpot ribbon controller. Wink

11-Front-Cut.jpg

Description:

Filesize:

109.03 KB

Viewed:

129 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

12-Front-Panel.jpg

Description:

Filesize:

78.79 KB

Viewed:

128 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Uncle Krunkus · Posted: Sat Oct 04, 2008 8:30 pm Post subject:

Okay,
Now that I've done some current tests, I must say I'm impressed in some ways, others not.
First, the meters are quite accurate, for what they are. Give or take 30mA, they reflected the DMM reading quite well. I didn't want them to be spot on, I wanted a quick guide as to how much space I had left to play with, and to look cool. They deliver that just fine.
But,.....
I thought the LM723s would keep the voltage spot on regardless of load. As you can see from the attached graph, they don't. Maybe I was being naive about it's ability to track the load and adjust the voltage dynamically. Still .5V drop for the first .4A isn't too bad. I'm sure the LM723 is capable of better though, and I'm going to share these results with Elby and see if I can't improve the design a bit.
On a very positive note, MJE3055s @ 1Amp are still virtually cold!

PSU graph.gif

Description:

Filesize:

3.97 KB

Viewed:

127 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Sine · Joined: Sep 10, 2007 Posts: 111 Location: Netherlands

I'll put mine on the bench later today and do some measurements

Where did you insert the ammeters ? if you put them after the psu they will introduce a voltage drop, a good place for them would be behind the current sense resistors and pin 3 of the LM and before R106, that way the voltage sense is behind the ammeters and the drop doesn't affect the output

Posted Image, might have been reduced in size. Click Image to view fullscreen.

Posted Image, might have been reduced in size. Click Image to view fullscreen.

Uncle Krunkus · Posted: Sun Oct 05, 2008 2:36 am Post subject:

Hmm,...
So I could test that by simply shorting out the ammeter hey?

Sine · Joined: Sep 10, 2007 Posts: 111 Location: Netherlands

Sine · Joined: Sep 10, 2007 Posts: 111 Location: Netherlands

If it's not the ammeters then something else is fishy ..

I plotted the load graph of my psu :
The transformer is a 18 volts brick, the filter cap is 4700uF, the current limit resistor is .22 Ohm and the transistors are BDV65B's

The rest is exactly the same as the elby PSU

Posted Image, might have been reduced in size. Click Image to view fullscreen.

The 723 is rock stable !

@2750mA the voltage drops to 14.843 as the current limits starts to affect the output

Uncle Krunkus · Posted: Sun Oct 05, 2008 3:46 pm Post subject:

Thanks Sine!
Now I know what I'm looking for. Very Happy

I'll short the meters and do another test soon.

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

Ooo plots! Smile

Uncle Krunkus · PSU Plot 7-10-08.gif

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

Why not measure the voltage across the 1W resistors?

Uncle Krunkus · Posted: Tue Oct 07, 2008 1:57 pm Post subject:

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

Yeah but...

1. Measure resistance across parallel 1W resistors with power off.

2. Measure voltage across parallel resistors.

3. Divide voltage by resistance and get current, ammeter free.

Those current sense resistors are there for a reason, use them. That's what the LM723 is doing with that sense input.

Smile

Sine · Joined: Sep 10, 2007 Posts: 111 Location: Netherlands

That would actually work .... but you can't drive those ammeters with that voltage, something like 100uA meters with a resistor in series would be possible

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

You don't need an ammeter or any additional resistors. Just an ohmmeter to measure the resistor once, and an voltmeter to carry out the tests.

Uncle Krunkus · Posted: Tue Oct 07, 2008 10:51 pm Post subject:

Yeah,.... but the ammeters,....... the ones which I've got mounted on the front of the case,....... are ammeters.
I'm trying to decide where to insert them into the design,......... permanently. Rolling Eyes

Sine · Joined: Sep 10, 2007 Posts: 111 Location: Netherlands

Wherever you put them, they must sit BEFORE the the feedback resistor

Uncle Krunkus · Posted: Wed Oct 08, 2008 3:57 am Post subject:

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

Where is the facepalm emoticon? I was so concentrated on your plots I thought you were measuring performance.

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

Sorry.

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

What is the resistance of the ammeters?

And does putting them between between the 3055's collector and the bridge affect line regulation? I don't think the reading should be thrown off too much.

urbanscallywag · Joined: Nov 30, 2007 Posts: 317 Location: sometimes

dupe, hit quote not edit.