Lab Power Supply Design / Build

[MJL21193]

Originally Posted by AndrewT

The coarse could be switched rather than continuous.

Ah, a DPDT switch to put a 10K resistor in series with the pot, either before for 'high' adjust and after for 'low' adjust.
I think I like the fine adjust pot idea better. A 1K pot will vary the voltage by about 2V over its range which would be perfect. I can get a smaller pot in the cow catcher no problem. The only obstacle is that I don't have any more 1K pots. On my next Digikey order, I get a pair and at that time I'll wire them in.
Snow is a-coming down here and it's a grand day to put some off the finishing touches on this project. Time to get it off the bench and into active duty.

[AndrewT]

I was thinking more like switch positions for 0-5Vdc, 5-10V, 10-15, 15-20, 20-25, 25-30. Would suit a 6way rotary. Then continuous fine to cover the spread of 5V at each position.

[MJL21193]

Originally Posted by AndrewT

I was thinking more like switch positions for 0-5Vdc, 5-10V, 10-15, 15-20, 20-25, 25-30. Would suit a 6way rotary. Then continuous fine to cover the spread of 5V at each position.

That is probably more practical for a bigger unit. The 2 pot solution will give the resolution I'd like but still make for quick adjustment. To see the accuracy of the fine adjust pot, I'd have to connect a multimeter - my analog meters will not show it precisely enough. I may, at some time in the future, change the scale on the voltage meters from 75VDC to ~37VDC for better precision.
I like multi position switches but only if they are board mounted, otherwise there are too many wires running willy nilly. There is enough of that going on now.

[MJL21193]

Originally Posted by MJL21193

Snow is a-coming down here

NPX_01098.JPG

Some additions that should have been on the PCB - resistors in series with the voltage control pots and current limit pots. I put these in place on the pots:

NPX_01100.JPG

The 'light tube' that channels the light from the bulb limiter out to the front panel:

NPX_01101.JPG

just a piece of 1/4" plastic tubing.

[MJL21193]

A look under the hood:

NPX_01108.JPG

Everything in place, neat and tidy.

In place in the gear and parts hutch:

NPX_01102.JPG

Ready for action.

A close-up of the bezel I added to the front panel:

NPX_01105.JPG

I just roughly laid out and hand printed (in pencil) the scale for the current limit.

It is done(!!) for now.
Final specs are ~0 to 35VDC at a maximum 2.2A.
A lot of fun to build, a great learning process and it looks to be a very useful tool indeed. I'm sure I'll put it to good use in many future projects.

[MJL21193]

Final schematic:

LAB SUPPLY SCH.png

[akis]

Originally Posted by MJL21193

I'm running it on the trial version of 11 and it isn't stable. It is running very slowly as well (a bad sign in my experience) with it taking taking 20-30 seconds to step thru .5 seconds of operation.
After I loaded it, I had to change the 2N5459 to 2SK170 (no MS model for that) and I ran it. It started to oscillate at .443 seconds and then halted at a convergence error. Maybe something was corrupted while zipping it.
This is not an overly complex circuit and that is not typical performance for MS.

Tit for tat, I'll attach mine and you can check out its performance. Give it about 20 seconds (MS timestep clock) for it to settle in. Adjust R18 for 10VDC after the LM317 and R19 for ~1.3VDC at the anode of D9.

Attachment 200413

Sorry I was out of London for a week. I am back and will try it out tomorrow-ish.

[MJL21193]

Originally Posted by akis

Sorry I was out of London for a week. I am back and will try it out tomorrow-ish.

Hey akis,
Did you get any further with your design?

[jaycee]

This thread caught my attention - been wanting to do a bench supply for a while! I had more or less set myself on a discrete copy of the MC1466L with some adjustments until I looked at this.

I simulated it in LTSpice and spotted a problem. I set the output voltage for 12V (or near as possible) off load. Then I added a resistor that drew 1A. The output voltage *rose* to 12.2V. Seems a bit odd.

This circuit does have one neat advantage over the 1466L, no need for a separate floating supply. OK small transformers are cheap, but it's still a pain. However the 1466L does use high side sensing, which means it'll pick up any short, not just a short to ground.

[MJL21193]

Originally Posted by jaycee

I simulated it in LTSpice and spotted a problem. I set the output voltage for 12V (or near as possible) off load. Then I added a resistor that drew 1A. The output voltage *rose* to 12.2V. Seems a bit odd.

This circuit does have one neat advantage over the 1466L, no need for a separate floating supply. OK small transformers are cheap, but it's still a pain. However the 1466L does use high side sensing, which means it'll pick up any short, not just a short to ground.

Hi jaycee,
What can I say about simulation? I use it extensively myself.
I can say that the circuit does not behave that way in Multisim and, more importantly, it does not do that in reality. The addition of a 1A load causes less than 10mV deviation in the supply voltage at 12VDC.
You might want to check the ground references for your measurements to ensure they are on the regulated side of the current sense resistor (0 volt side). This tripped me up a few times in the beginning.

I like to be efficient and run everything from one transformer. Less mains voltage wires to run, less space taken up inside the case. The implementation in this supply is in no way inferior to a separate supply and the results speak for themselves.

Not sure what you mean by 'any short, not just a short to ground'. The supply limits current on a dead short between the positive volts terminal to the 0 volts terminal to less than 2.2A. Here that is in action:

NPX_01113.JPG

That is a 1/4 watt, 0 ohm resistor across the terminals. It looks odd but demonstrates the effectiveness of the current limit and the foldback feature.
This was actually me testing some of the new cheap resistors I bought (5000 - E12 series from 0 to 10M ohms for $20) to see how much actual current they could withstand before failing. The 0 ohm resistor didn't fail, since its wattage wasn't exceeded, despite the ~2.1A current draw (near 0 volts).