Making an analog power supply digital...

Ductapemaster

I am starting to build this power supply (working on obtaining parts): http://www.uoguelph.ca/~antoon/circ/ps3010/ps3010a.html and I would really like to make it digitally controlled.

I don't like having pots a the user interface so I really want to convert this to a digital design. I'm planning on having an LCD display the voltage and current values, as well as the temperature and whether or not the fans are on. Also, I would like to use buttons for the interface, or maybe even a rotary encoder. I really like the design of this supply (better than others I have seen in the past) and it has very detailed construction information, which helps a lot.

So, my question is how would I go about controlling this digitally? I am thinking of using digital pots from Maxim for the voltage and current adjustments, but I'm not sure what kind of ratings they would need. Is it even possible to do it? Also, I need some help with voltage/current measurement. I am planning on using the MAX187 5V ADC to measure voltage and current. I know how to hook it up for measuring voltage, but I am clueless as to how to do it to measure current. Also, I would like to do away with the ranging function the designer implemented. I don't need power below 3 volts, and if I did, I would just add some diodes in line with the supply to drop the voltage to the desired level. That would clean things up a bit.

I can do most of the programming myself, but help with those little tidbits I asked for above would be greatly appreciated.

Thanks!

-Dan

Post Edited (Ductapemaster) : 7/21/2008 8:48:06 PM GMT

Guenther Daubach

Dan,

before looking for "external" components, like digital potentiometers, or ADCs, take a look "under the hood" of the SX controller - it has a lot of horsepower. "Building" ADCs and DACs in software requires only a couple of instructions and only some external components, like resistors and caps.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Greetings from Germany,

G

Ductapemaster

Hmmm...you are right. The reason I suggested the MAX187 is because I have a bunch of em laying around and they're calling to be used...

Also, the other reason I prefer to use external components is because they have really good resolution. For example, I can get 12 bit resolution with the 187 with a serial command. From what I have read, the SX has the capability for 8 bit resolution. In my application, that equates to about 0.1 volt resolution (measuring 30 volts with a divide-by-6 voltage divider) so more resolution is welcome. With 12 bit resolution I can get within 0.07 volts, if my math is correct.

However, I am more open to using the SX's built in capability for adjusting the voltage/current output of the LM723 used in the design, but I have no idea how I would go about doing so. Also, if the output is limited to 8 bits, having 12 bit measuring capabilities is completely useless...

Any suggestions on how to replace those two pots in the schematic with something hooked to the SX would be helpful! When it comes to this stuff, I'm more software than hardware...

-Dan

pwillard

The problem with Digital Pots in a Power supply is you end up with a RUBE GOLDBERG design trying to deal with their limitations. A Digital Pot can handle only tiny amounts of current and the average power supply analog adjustment pot may have to deal with 10 times more current than the digital version can handle.

Seems amazing that you might drive an antique LM723 with an SX, but hey... the 723 is a tried and true dependable analog IC. I love them and have many myself. I understand Gunthers statements, but I'd probably prefer to just have th SX feed a good quality DAC with bits rather than deal with discrete Resistors and the variance you might get with them. I have some nice 723 designs in my collection... I might dig them out.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


There's nothing like a new idea and a warm soldering iron.

Ductapemaster

Yeah that's what I gathered when I started looking up my options in that area...I guess I will look into other possible options for control. It's great to hear from someone experienced with the LM723 however! I am still clueless how to hook up a DAC to that chip...I plan on spending some time on that tomorrow though. Do you have any ideas?

Thanks,

-Dan

pwillard

Well, I haven't found my "stash" of reference designs yet... but, here is what I'm thinking. (I just pulled up a datasheet from National for reference)·Totally off-the-cuff and untested statements follow...

You use the SX controlled DAC to feed·IN+(pin 3)·input pin, take the V-(pin 5)·input pin below ground potential (like -2.5V) by taking VREF from pin·4 and using an OPAMP in an inverting mode, giving you a quick -2.5v from +2.5. (You feed the·-input·pin on the opamp with VREF output from pin 4 and take +pin on the opamp to GND with 1:1 feedback loop.)

The limitations are going to be around how much voltage you can feed the DAC (or rather, how much voltage the DAC can present to the regulator).

The rest is a normal design for the output stage and current sense. (you could use differential op amp accross the current sense resistor to provide AtoD feedback to the SX for current monitoring)

·Like I said, I'm a big fan of the 723 and you have me wanting to play with one again.



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


There's nothing like a new idea and a warm soldering iron.

Ductapemaster

I like the idea...However, I do want 12 volts (or more) to be available. Also, if I feed the 723 with DAC output, wouldn't the output current also be limited by the current capacity of the DAC? That would also pose a problem, as I am trying to pump 6 amps max out of this.

How would you suggest I control the current? I understand how you suggested to monitor it, but I'm not sure how I would set it without varying the current resistor...

The more I look at this the more I think I want to use a power supply based on discrete components! Then I could just feed the power transistor with the DAC output and problem solved! But, I am not giving up yet.

Thanks,

-Dan

EDIT: Wait, I think I understand how to control the current if I can control the voltage. I would just sense the current and if it goes above a preset level then I could just lower the drive voltage (DAC output) or shut it down completely. That would be much easier.

Also, I found this article: http://tuxgraphics.org/electronics/200707/bench-power-supply-unit.shtml that talks about making a discrete power supply thats controlled digitally. I may end up going with this design, unless we come up with a better voltage control option...

EDIT 2: Added a proposed circuit diagram...let me know what you think. The design is a combination of the 723 output stage and the voltage amplifier from the link above. R1, 2, and 3 are calculated according to the formula Voltage amplification= (R1 + R2)/R3 (from link above).

Post Edited (Ductapemaster) : 7/22/2008 9:30:04 PM GMT

John Couture

Interesting topic.

Couldn't you use a LM117/LM317 ·(Adjustable Regulator, 1.5A, 1.2V to 37V range) and a digital pot (DS1803)

With R1 at 1000 and R2 a 10K variable pot you could go from about 2.0 to 15V

LM317 calculator
http://spazioinwind.libero.it/andreabinello/lm317-lm337/index.html

Higher amps available as LM150, 3A and LM138, 5A

A couple of reference articles:

High Side Current-Sense Measurement: Circuits and Principles (MAXIM)
http://www.maxim-ic.com/appnotes.cfm/appnote_number/746/

Making a shunt from 12ga wire
http://www.uoguelph.ca/~antoon/gadgets/shunts/shunts.html

LM117 Datasheet
http://cache.national.com/ds/LM/LM117.pdf

DS1803 Datasheet
http://datasheets.maxim-ic.com/en/ds/DS1803.pdf

... Or did I misunderstand the whole concept?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John J. Couture

San Diego Miramar College

pwillard

All those are possibilities John, no doubt. I STILL use an LM317 ( with a precision 10 turn full size POT for ADJ) in my lab for breadboarding that I made in the early 80's.

The LM723 is a GEM though. You can do some current monitoring and control in addition to voltage adjustments.

Oh·yeah, in my head...·the DAC is merely feeding the internal "op amp" in the 723, don't take the 723 out of·the design...·it can server a good purpose.· The DAC·should not have to carry the current that the Pass transistor is dealing with. (I'm now getting frustrated looking for my folder with 723 designs... lol)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


There's nothing like a new idea and a warm soldering iron.

Ductapemaster

I'm fine with sticking with the 723, however, I still don't see how I could control the current limiting on the chip via a DAC. I don't completely understand how the 723 limits current anyways, but I do know that the current limit pin takes voltage ranging from 0V to Vout. Would I just use another transistor on the output of the DAC to control pin 2? It woudl effectively create a darlington pair with the internal transistor of the 723...

If you could explain that I would be very grateful...

Also, did I set up the current monitoring OP amp correctly in my schematic? I'm not sure what value to use for the shunt resistor, as I don't know the upper current limit of my design. It's made for 10 amps , but I can't find a 25V 10A transformer anywhere...All I've found is a 6.25 amp one. I would really like the full 10 amp output, only because it makes the voltage drop easy to calculate...at full current flow it would be 5 volts, aka the max input of the ADC. That eliminates another OP amp I would have to add to scale the voltage up from whatever to 0V-5V. Any ideas here would be appreciated too...

Thanks!

-Dan

pwillard

Voltage Control.·· My thoughts.



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


There's nothing like a new idea and a warm soldering iron.

Ductapemaster

Actually, I have been thinking...I kinda want a power supply with more than one output, so what if I used a discrete transistor design for a variable output, and I use 723 to generate +12 volts as a fixed output? That way I can just monitor the current and shut it down if a settable current limit is reached. I'll just put a relay on the main input for shutdown, or maybe hook it up in such a way it makes the current limit set to like 10mA. What do you think?

Post Edited (Ductapemaster) : 7/24/2008 7:59:43 PM GMT

Ductapemaster

OK, here is my proposed schematic(s). Let me know what you think. It's broken into three parts so it's not a mess of wire (although it still sort of is...).

I found that I do not have to use a differential amplifier to find the voltage across the current shunt. I just find a value that drops 5 volts across it and then measure the voltage with the ADC. However, this only works with the adjustable supply, as I can lose 5V there without a problem, as it drops the max output to around 16.5V (15VAC * 1.44 - 5V = 16.6). I don't need the full range anyways. However, this only works for the adjustable supply. If I lost 5V on the 12V supply, I would have to be outputting 17V to start with. I am probably going to use a differential amplifier there, use a low value resistor, and then just adjust the gain to get 0-5V range. I still need to do the calculations there.

Other than that its all complete!

-Dan