Electronics > Projects, Designs, and Technical Stuff

My power supply based on Dave's design

(1/2) > >>

m12lrpv:
I would like to put up my design for a power supply based on Dave’s one. Now I’m just a new hobbyist so go easy :-)

This is my first mini-lab supply for use on my desk (or dining room table) so the field requirement of batteries is not a consideration. That means that reduced power consumption from the control circuit is not a big consideration but idiot proofing is.

The main design requirement was that it uses up stuff from my parts bin so as to relieve pressure from the wife over this new hobby.

Because I had a LCD display that is 5v (from another project that ended up being 3.3v) and because I want to run the atmega328 at 20MHz it’s a 5v controlled project instead of Dave’s 3.3v. I feel that it gives a better resolution going into the LM324 op amps.

For output I only want up to 5v and I’m happy with 1.25v minimum so I’m using the LM317.

I’ve used Dave’s designs and tweaked it to what I think I can get to work to suit my needs.

From Dave’s revision C, because of the rail to rail issue of the U12A buffer op amp, I merged his U12A and U12B into a dif amp across the shunt. I added a diode to the low side of the shunt before the voltage divider in order to lift the op amp inputs out of the noise at low current. This combination gives me an active current adjustment input range of approx 1.2v to 4.5v which seems a reasonable resolution for the current limit. I’m hoping that my theory is sound there. It seems to work but I’m not too familiar with the dynamics of diodes and dif amps so I might come unstuck in but it seems to give me some room at the bottom and top of the current limit adjustment with plenty of range in between.

I’m not doing the current sense because I cannot justify it at this stage although I should be able to add it later if I need to.

It simulated ok and seemed to breadboard OK. I don’t have a lab full of test equipment to accurately gauge performance however the basic concept seemed to work. I did struggle to get to minimum voltage which I suspect is because of noise on the voltage control op amp. No matter what input I set there was 0.7v on the adjust pin of the LM317. That is something I’m hoping to eliminate once I solder it to a board although I would love to be able to do something like I did with the diode in the current set in order to offset the minimum adjustment up out of the noise.
On the breadboard there was still some noise on the current limit input preventing limiting to zero current but hopefully soldering it up will get rid of some of that too.

Not visible on the schematic (but on the layout) is a bridge rectifier which should enable me to plug in various polarity plug packs without blowing it up. That is accompanied by a 1.1A poly fuse just in case I screw up the current limiting.

I’m doing a mechanical power switch because for a one off it suits my needs.

For adjustment control I looked at rotary encoders and push buttons but have decided to go with regular pots because they’re easy to panel mount. I will feed them into the ADC on the atmega328 and depending on the value read in I can use them for variable adjustment rate. I can do both fast and slow adjustment of voltage and current.

I’m thinking about putting in a mosfet to cut the supply to the regulator until the micro has started up so that I can ensure that it doesn’t start up in a state that will damage something that I have connected to the outputs.

Anyway I’m putting out there. If anyone can see anything completely retarded in what I’m doing then speak up but at this stage I think I’ve got a design that will meet my needs.

amspire:

--- Quote from: m12lrpv on April 17, 2012, 12:52:08 am ---For adjustment control I looked at rotary encoders and push buttons but have decided to go with regular pots because they’re easy to panel mount. I will feed them into the ADC on the atmega328 and depending on the value read in I can use them for variable adjustment rate. I can do both fast and slow adjustment of voltage and current.

--- End quote ---

I have no idea how you are going to make this work. If you start getting a pot to do variable adjustment rate rather then absolute adjustment, what are you going to do when the pot is at the maximum stop and the output voltage is still only 1V? How will you get to 2V output or more? The point about the rotary switches is that they have no stops so you can use them for variable rate. Same with using pushbutton switches instead.

Surely if you are going to use a potentiometer, use it to directly control the voltage to the LT3080, and then the micro just reads the resultant voltage. Eliminates the DAC, and if you use a 10 turn POT, you will get better resolution then the DAC could do anyway.

Richard.

m12lrpv:
The pot is used for specifying the relative increments for the micro to apply and not absolute output.

The pot can be connected to the ADC so the micro will read them as a value from 0 to 1024
If it's reading say 462 to 562 ( the pot is centred) then that tells the micro to make no change.
For 231 to 462 I decrement by a small amount. For 0 to 231 I decrement by a larger amount.
The same goes for the high side. When you've reached the value you want turn the pot to the centre and leave it.

So rather than sit there and spin a rotary encoder for a month to get from one end of the adjustment range to the other the pot can be used to input a variable relative adjustment rate that can get you from one end to the other easier and faster. Coarse, medium and fine adjustment is all possible depending on how much you turn the pot.

It gives better control than a rotary encoder ever could because it's entirely in software. It beats the pants of up/down adjustment buttons too. The only limitation is the DAC output and at the moment what the inbuilt PWM delivers is more than adequate for my needs and I know how to tweak that if I want.

amspire:
That makes sense.

IanB:

--- Quote from: m12lrpv on April 17, 2012, 03:07:21 am ---So rather than sit there and spin a rotary encoder for a month to get from one end of the adjustment range to the other the pot can be used to input a variable relative adjustment rate that can get you from one end to the other easier and faster. Coarse, medium and fine adjustment is all possible depending on how much you turn the pot.
--- End quote ---

I see how this can work, but as a human I tend to find this kind of "up a bit, down a bit" control a bit awkward to use. If you have not already built this design, you may find likewise when you actually come to use it.

Navigation

[0] Message Index

[#] Next page