0-30V Stabilized Power Supply

[liquibyte]

Don't knock yours, it's a beauty.  If you built two and put it in bigger box, you'd have what I'm doing.  I wish more people would post their pics.  I like to see how different people design and solve problems.

I don't even have my displays designed yet but they're coming.  I'm trying to decide on a temperature control system for the fans right now and am doing testing to find the "hot spots".  So far, the component that gets the hottest fully loaded is R7 so I'm thinking of using a fan for those as well and sensing the temp at that point to control the fans for the whole system.

I've decided to have boards made for all of the different sub-systems because it's really starting to come together and they are cheap enough to get made though the wait is fairly long after they ship.

 

[tcjeep]

Liquibyte : Quote.."Don't knock yours, it's a beauty.  If you built two and put it in bigger box, you'd have what I'm doing.  I wish more people would post their pics.  I like to see how different people design and solve problems...."

The idea of a split supply was considered and that's why I have a green connector between the + and - connectors on the front. Its supposed to be a virtual earth. I'll build a supply splitter with an op-amp and a couple of power transistors one day. Then I'll have a 0-15 v split supply for op-amp circuits. Its enough for the work I do, I suppose.

Thanks for the kind words. I only jumped into this forum to encourage others to not hesitate, but to go ahead and do it. I know there are at least two readers who are worried about the voltages on the op-amps and other issues. All the problems seem to have been resolved on the basic circuit. Now we only need to worry about our own design mods. I hope we see lots of design mods and addons. Maybe you can get a Flir Infrared camera to see those hot spots. I managed to borrow one and they work really well. I left mine with a short circuit and everything cranked to max for half an hour. Yep, R7 glows in the dark, the rectifier smokes a bit and the heatsinks smell of hot paint, but it keeps going. Probing the BY127 bridge with a temp probe proved that, although it was very hot it was still well within its max temp spec.

Looking forward to hearing how you approach your variable speed fans and temp sensors. Please continue to keep us in the picture.

 

[liquibyte]

I just got an IR thermometer to monitor temps and it's accurate enough to keep an eye on things so I'm good to go there.

R7 was the only part that I'm even a little concerned about because I tend to over engineer things.  My R7's are 16W aluminum clad versions.  My bridge rectifiers are hefty (Digikey) and mounted to the all aluminum case that acts as a big heat sink.  I jumped the voltages on the caps up to 100V except for the big filter caps at the input and output.  All of my pots are 10 turn including the trimmers.  Let's just say that I spent too much on this and am glad it works as well as it does.

I suppose I should share my parts list for all the world to see.  These are actually two separate orders.  The off board parts were on my first order that had the boards that didn't work and the on board parts were the second order in addition to the boards I had made.  This is a complete parts list for one power supply minus heat sinks, switches, case, plugs, wiring etc.  You get the idea.

Off board parts
1		237-­1281­-ND		XFRMR LAMINATED 130VA CHAS MOUNT
1		GBPC15005FS­-ND		RECT BRIDGE GPP 15A 50V GBPC
1		A102127­-ND		RES CHAS MNT 0.47 OHM 5% 16W
2		RSF2JTR330CT­-ND		RES MO 2W .33 OHM 5% AXIAL
1		399-­5658­-ND		CAP ALUM 15000UF 63V 20% SNAP
2		2N3055GOS-­ND		TRANS NPN 15A 60V TO3
1		BD13910S-­ND		TRANSISTOR NPN 80V 1.5A TO­126
2		987­-1523-­ND		POT 10K OHM 7/8" 10TURN WW
1		365-­1189-­ND		LED 631NM TAPERED RED 5MM

On board parts
1		RSMF2JT2K20CT­-ND	RES MO 2W 2.2K OHM 5% AXIAL
1		RSMF2JT82R0CT­-ND	RES MO 2W 82 OHM 5% AXIAL
1		RSF1JT1K50CT-­ND		RES MO 1W 1.5K OHM 5% AXIAL
1		220H­-ND			RES 220 OHM 1/2W 5% AXIAL
1		CF14JT56K0CT­-ND		RES 56K OHM 1/4W 5% CARBON FILM
1		CF14JT33K0CT­-ND		RES 33K OHM 1/4W 5% CARBON FILM
2		CF14JT27K0CT­-ND		RES 27K OHM 1/4W 5% CARBON FILM
4		CF14JT10K0CT-­ND		RES 10K OHM 1/4W 5% CARBON FILM
2		CF14JT2K20CT-­ND		RES 2.2K OHM 1/4W 5% CARBON FILM
3		CF14JT1K00CT­-ND		RES 1K OHM 1/4W 5% CARBON FILM
1		CF14JT68R0CT-ND		RES 68 OHM 1/4W 5% CARBON FILM
1		BZX85C10­-ND		DIODE ZENER 10V 1W DO41
1		BZX79C5V6­-ND		DIODE ZENER 5.6V 500MW DO­35
1		1N4001FSCT­-ND		DIODE GEN PURPOSE 50V 1A DO41
5		1N4148TACT­-ND		DIODE SMALL SIG 100V 0.2A DO35
2		399-6631-ND		CAP ALUM 47UF 100V 20% RADIAL (ESW476M100AH2AA)
1		399-5999-ND		CAP FILM 10UF 63VDC RADIAL
1		399-­4173­-ND		CAP CER 330PF 100V 5% RADIAL
2		399-­9707­-ND		CAP CER 100PF 100V 5% RADIAL
1		399-­6037­-ND		CAP FILM 0.22UF 100VDC RADIAL
2		399­-5861­-ND		CAP FILM 0.1UF 100VDC RADIAL
3		A98334­-ND		TERM BLOCK 3POS SIDE ENT 2.54MM
6		2A98333­-ND		TERM BLOCK 2POS SIDE ENT 2.54MM
1		3296W­-502LF-­ND		TRIMMER 5K OHM 0.5W PC PIN
1		3296W-­203LF-­ND		TRIMMER 20K OHM 0.5W PC PIN
1		3296W­-104LF-­ND		TRIMMER 100K OHM 0.5W PC PIN
1		BC557BGOS­-ND		TRANS AMP PNP 45V 100MA TO­92
3		296­-10456-­5-­ND		IC OPAMP GP 5.9MHZ 8DIP
3		3M5473­-ND		SOCKET IC OPEN FRAME 8POS .3"

 

[cscott719]

Hello everybody.  I used to muck around a bit with electronics back in high school, but have not done much in the past 30 years.  Recently, I regained my interest because my son has shown a big interest.  So I figured I should try to learn something so I don't look like a total idiot to him.  I have been lurking around this board and many other sites for a couple of months, trying to decide on which basic design I want to do.   This one won out because of the involvement of the members, the many questions asked and answered, and the various projects that users have discussed and illustrated.  

I decided to build this circuit with a few upgrades to it.  After trying to etch a double sided board for my first time and failing miserably because of the inability to plate vias in a garage, I got in touch with Liquibyte, who graciously sent me two of his beautiful pro fabricated boards.  So far, that side of things is going well.  Since the high temp parts are all off board, I decided to try to design a second board to hold those components (using TIP3055, btw) along with some sort of large heat sinks.  Because of the heat, I wanted a way to get it out of the case, so I added a fan speed control circuit to the board, along with a 7805 to power a pair of USB charging ports and a 7812 because I wanted to.  Next, I will be working out the display issues.  I could go with 16x2 LCD, but everybody does that, probably because it just works.  But I am leaning toward a 4.3" TFT driven by an Atmega MCU.  I already have the display working, just working on the code, and the circuit for the board.

I will post pictures of the fire when I am done.

 

[Canyoncruz]

Hello all,

I bumped into this power supply. I have been working on a power supply for my turntable. I originally was going to go with a shunt, out baord, but it posed problems especially if I unplugged it. My TT has a local regulator.

I need 26 to 27 VDC. The current draw is 250 mA transient and 45 steady state. the rest of the controls and lights etc. draw another 95 mA. So the total operating current is from ~ 150 to 350 mA.

Does anyone foresee a problem with using this supply as a steady power source?

Thanks in advance,

CC

 

[tcjeep]

Hello all,

I bumped into this power supply. I have been working on a power supply for my turntable. I originally was going to go with a shunt, out baord, but it posed problems especially if I unplugged it. My TT has a local regulator.

I need 26 to 27 VDC. The current draw is 250 mA transient and 45 steady state. the rest of the controls and lights etc. draw another 95 mA. So the total operating current is from ~ 150 to 350 mA.

Does anyone foresee a problem with using this supply as a steady power source?

Thanks in advance,

CC

Hmm.. Overkill? Why not just use a fixed regulator IC based PS? There are many available.

 

[Canyoncruz]

It is. But it turns out the better the supply, the better the low end and the pitch. locally there is a regulator. I am switching the LM317 with the newer and quieter TPS7A4701. I am also swapping out the components on the motor driver pcb. metal film for older carbon etc.

 

[tcjeep]

I just got an IR thermometer to monitor temps and it's accurate enough to keep an eye on things so I'm good to go there.

R7 was the only part that I'm even a little concerned about because I tend to over engineer things.  My R7's are 16W aluminum clad versions.  My bridge rectifiers are hefty (Digikey) and mounted to the all aluminum case that acts as a big heat sink.  I jumped the voltages on the caps up to 100V except for the big filter caps at the input and output.  All of my pots are 10 turn including the trimmers.  Let's just say that I spent too much on this and am glad it works as well as it does.

I suppose I should share my parts list for all the world to see.  These are actually two separate orders.  The off board parts were on my first order that had the boards that didn't work and the on board parts were the second order in addition to the boards I had made.  This is a complete parts list for one power supply minus heat sinks, switches, case, plugs, wiring etc.  You get the idea.

Off board parts
1		237-­1281­-ND		XFRMR LAMINATED 130VA CHAS MOUNT
1		GBPC15005FS­-ND		RECT BRIDGE GPP 15A 50V GBPC
1		A102127­-ND		RES CHAS MNT 0.47 OHM 5% 16W
2		RSF2JTR330CT­-ND		RES MO 2W .33 OHM 5% AXIAL
1		399-­5658­-ND		CAP ALUM 15000UF 63V 20% SNAP
2		2N3055GOS-­ND		TRANS NPN 15A 60V TO3
1		BD13910S-­ND		TRANSISTOR NPN 80V 1.5A TO­126
2		987­-1523-­ND		POT 10K OHM 7/8" 10TURN WW
1		365-­1189-­ND		LED 631NM TAPERED RED 5MM

On board parts
1		RSMF2JT2K20CT­-ND	RES MO 2W 2.2K OHM 5% AXIAL
1		RSMF2JT82R0CT­-ND	RES MO 2W 82 OHM 5% AXIAL
1		RSF1JT1K50CT-­ND		RES MO 1W 1.5K OHM 5% AXIAL
1		220H­-ND			RES 220 OHM 1/2W 5% AXIAL
1		CF14JT56K0CT­-ND		RES 56K OHM 1/4W 5% CARBON FILM
1		CF14JT33K0CT­-ND		RES 33K OHM 1/4W 5% CARBON FILM
2		CF14JT27K0CT­-ND		RES 27K OHM 1/4W 5% CARBON FILM
4		CF14JT10K0CT-­ND		RES 10K OHM 1/4W 5% CARBON FILM
2		CF14JT2K20CT-­ND		RES 2.2K OHM 1/4W 5% CARBON FILM
3		CF14JT1K00CT­-ND		RES 1K OHM 1/4W 5% CARBON FILM
1		CF14JT68R0CT-ND		RES 68 OHM 1/4W 5% CARBON FILM
1		BZX85C10­-ND		DIODE ZENER 10V 1W DO41
1		BZX79C5V6­-ND		DIODE ZENER 5.6V 500MW DO­35
1		1N4001FSCT­-ND		DIODE GEN PURPOSE 50V 1A DO41
5		1N4148TACT­-ND		DIODE SMALL SIG 100V 0.2A DO35
2		399-6631-ND		CAP ALUM 47UF 100V 20% RADIAL (ESW476M100AH2AA)
1		399-5999-ND		CAP FILM 10UF 63VDC RADIAL
1		399-­4173­-ND		CAP CER 330PF 100V 5% RADIAL
2		399-­9707­-ND		CAP CER 100PF 100V 5% RADIAL
1		399-­6037­-ND		CAP FILM 0.22UF 100VDC RADIAL
2		399­-5861­-ND		CAP FILM 0.1UF 100VDC RADIAL
3		A98334­-ND		TERM BLOCK 3POS SIDE ENT 2.54MM
6		2A98333­-ND		TERM BLOCK 2POS SIDE ENT 2.54MM
1		3296W­-502LF-­ND		TRIMMER 5K OHM 0.5W PC PIN
1		3296W-­203LF-­ND		TRIMMER 20K OHM 0.5W PC PIN
1		3296W­-104LF-­ND		TRIMMER 100K OHM 0.5W PC PIN
1		BC557BGOS­-ND		TRANS AMP PNP 45V 100MA TO­92
3		296­-10456-­5-­ND		IC OPAMP GP 5.9MHZ 8DIP
3		3M5473­-ND		SOCKET IC OPEN FRAME 8POS .3"

Yeah, quite a shopping list, Liquibyte. Call me curious, but my feeling is that you must have a very sophisticated bench. If this PS is an example of the stuff you use everyday, then the likes of lil 'ol me can only dream.

About R7. Is my Ohm's law flawed? I calculate that that resistor should be 4.2W! Mine is only 5W but its stood way off the board with heat sleeves. If it goes open. it'll put 30v across my display. If my maths is correct, then yours is overrated by 380% Please tell me I'm no good at maths.

Thanks for your component list, very interesting. I don't recognise that op-amp. Is there a datasheet you could point me to?

 

[Canyoncruz]

TCjeep, it is the TLE2141, an equivalent to the original spec'ed opamp.

 

[liquibyte]

Yeah, quite a shopping list, Liquibyte. Call me curious, but my feeling is that you must have a very sophisticated bench. If this PS is an example of the stuff you use everyday, then the likes of lil 'ol me can only dream.

Heh, I wouldn't know what to do with a sophisticated bench.  My bench is usually my living room floor but I do have a table in another room that I do my soldering on.  I own 1 Craftsman multimeter and an old Conar 255 scope that I haven't even used yet.  That's it.

About R7. Is my Ohm's law flawed? I calculate that that resistor should be 4.2W! Mine is only 5W but its stood way off the board with heat sleeves. If it goes open. it'll put 30v across my display. If my maths is correct, then yours is overrated by 380% Please tell me I'm no good at maths.

You're ohms law is spot on (4.23W), the original schematic called for a 10W R7 but they didn't have them in the value I needed so I went up.  As I'm finding out that it's the component that gets the hottest, I'm kind of glad I did.

Thanks for your component list, very interesting. I don't recognise that op-amp. Is there a datasheet you could point me to?

TI Datasheet.  It's one of the recommendations for the latest iteration of the schematic.  Works nicely with an LM35 for a temperature controlled fan as well.

 

[liquibyte]

I've been working on a temperature controlled fan circuit to run the fans based on the temps of R7.  The circuit for the fan control consists of an op amp, an LM35, an NPN transistor, and a resistor, in other words, everything to the right of the last cap.  I want to add that I didn't come up with this, this guy did.  All I did was test it out, though I did change the op amp and transistor used.

 

[audioguru2]

Your schematic is missing a temperature sensor and parts for the opamp.

 

[liquibyte]

The temp sensor is off board, hence the VIN, OUT, and GND. The pinouts for the TO-92 and the TO-220 are different so doing a board for one precludes the other unless you just use wire pads, a terminal block or some sort of connector.  Pins 1, 5, and 8 aren't  used so I didn't invoke them.  I just cobbled this together quick just in case someone wanted to try it out.

Edit:  To be honest, I wanted to see how small I could get the board.  I know it's ugly and not finished yet, but this is what I have so far.

 

[tcjeep]

The temp sensor is off board, hence the VIN, OUT, and GND. The pinouts for the TO-92 and the TO-220 are different so doing a board for one precludes the other unless you just use wire pads, a terminal block or some sort of connector.  Pins 1, 5, and 8 aren't  used so I didn't invoke them.  I just cobbled this together quick just in case someone wanted to try it out.

Edit:  To be honest, I wanted to see how small I could get the board.  I know it's ugly and not finished yet, but this is what I have so far.

Liquibyte, your temp sensor board looks really good, but sort of complicated. I've seen lots of sensor circuits, most are really simple, like this (attached) Sorry about the rough drawing. I'm using something like this and it works fine, except I've adjusted the pot to only come on if my heat sinks get really hot, because I hate noise when I'm trying to work. I 'm not criticising your circuit, I'm just interested in what makes it better.

 

[liquibyte]

I'm just interested in what makes it better.

The LM35 and its accuracy.  I searched this one out because I wanted one sensor to control 3 fans.  The resistor can be changed according to the total current draw.  If you look at the schematic, you'll see that I've included a power supply in addition to the temp circuit.  Here's just the temp circuit.  VIN, OUT, and GND are the LM35 connections basically using any sort of 3 wire to board connector you have on hand.  I used a TLE2141 because I had some handy but any bog standard op amp will do and any bog standard NPN transistor will do as well.

 

[tcjeep]

The LM35 and its accuracy.  I searched this one out because I wanted one sensor to control 3 fans.  The resistor can be changed according to the total current draw.  If you look at the schematic, you'll see that I've included a power supply in addition to the temp circuit.  Here's just the temp circuit.  VIN, OUT, and GND are the LM35 connections basically using any sort of 3 wire to board connector you have on hand.  I used a TLE2141 because I had some handy but any bog standard op amp will do and any bog standard NPN transistor will do as well.

Yes, I know the LM35, and I see your circuit. I am confused with the logic involved. You said you need to control three fans. but from how many points of reference? You will have to go easy with me. I'm a thrift designer/developer. the guy who strips good designs down to economically viable production prototypes. I am struggling to understand your need to over engineer this good, but basic concept. Its a good design, made better because of the open initiative of many hobbyists assisted by experts within this forum. We need, however, to be pragmatic and explore the fact that
this design is probably not the best choice to bet all your money on. I have an old power supply based on the old LM723 and 2N3055 parallel solution that out performs this in every way except that it won't go to zero volts. I could add, that this is something that could be fixed without a major outlay. Your project is impressive. but what will you do with it? What is your end game? You have a dual supply with three fans. Impressive in every detail, but compared with what? Liquibyte, my questions are not intended to annoy you. I am simply interested in understanding your motives. why are you investing so much into something that has little or no return? What will you do with your masterpiece?

 

[liquibyte]

Honestly?  I was bored and wanted to get back into electronics.

I've spent more in money and time than any reasonable person should have doing this but I'm having fun.  Isn't that why we do these things?

The reason I'm using three fans is because I had them (2 80mm and 1 40mm).  I was going to just do passive cooling but figured active would keep things cool and working for longer, so...  The decision to cool the 2N3055's was a given due to where they were mounted.  I decided to add a third fan to keep the R7's cool because during testing they were getting the hottest and, as I'm sure you know, resistance changes with heat.  The LM35's were samples because I wanted to see how good they worked and if I could use them to do what I needed.  I plan on only using one (TO-220), mounted between both R7's to act as a fan speed controller from there.

I think the project started out with the intention to make one or two of these that were going to be good enough.  I had the case and things just sort of ballooned from there.  Oddly enough, I was looking into LM723 designs when I found this project and the accompanying thread(s) and decided to give it a try.  Want to share the schematic for the 723?  Before anyone thinks I'm some kind of guru, I'm not.  99% of people on here will know more than me by a long shot.  The one thing I do have experience in is hardware design.  I've been cobbling things like this together since I was a kid.

A person can get too bogged down into ºC/W and a way to avoid that is to over engineer even if it costs a little more.  My original order including the expensive stuff like the big filter caps, transformers, transistors etc. came to $185.18.  The majority of that was for the transformers ($48.92), the big filter caps ($32.08), and the 10 turn pots ($47.96).  So, the initial outlay for someone that already had these things would have been $56.92.  My first go around didn't go well so I decided to have boards made and use better components.  My second order which just included the parts to redo the boards came to $66.43.  The boards themselves came to $28.80 for ten (four of which I've given away to two other people from here).  I've got about $30 in knobs because one of the things I heard complained about was small knobs being inadequate, so I bought knobs that are over an inch in diameter, solid aluminum and heavy.  I've also got a few bucks in the four TO-3 mounting kits.  The heat sink for the TO-3's cost me less than $16 on Ebay.  As I said, the LM35's were a TI sample order.  The case was a gimme from another smaller case sample order (Thank you OKW).  I've got two Atmel 328P-PU's that were samples.  I've got two AS1100's (pinout compatible MAX 7219's) that were samples.  I've got two MAX7219's that were samples.  I've got two LDT-N2807RI-USB and two TDCR1060M LED displays from Lumex and Vishay respectively that were samples.  Starting to get the idea?

I'll make you an offer.  I have four boards left over from my order and I only really need to keep two more.  Would you like to have two?  PM me your address and I'll send you out a couple to build just for the fun of it.  You could probably use stuff you have laying around to populate it with but the C7 footprint is for a large film cap so you'll have to rig something there if you don't have one (Digikey #399-5999-ND $2.70 each).  Just remember to flip the BD557 around backwards because the footprint is wrong on the silkscreen.  That's my fault for not checking before I sent them out to be made.

 

[tcjeep]

Honestly?  I was bored and wanted to get back into electronics.

I've spent more in money and time than any reasonable person should have doing this but I'm having fun.  Isn't that why we do these things?

The reason I'm using three fans is because I had them (2 80mm and 1 40mm).  I was going to just do passive cooling but figured active would keep things cool and working for longer, so...  The decision to cool the 2N3055's was a given due to where they were mounted.  I decided to add a third fan to keep the R7's cool because during testing they were getting the hottest and, as I'm sure you know, resistance changes with heat.  The LM35's were samples because I wanted to see how good they worked and if I could use them to do what I needed.  I plan on only using one (TO-220), mounted between both R7's to act as a fan speed controller from there.

I think the project started out with the intention to make one or two of these that were going to be good enough.  I had the case and things just sort of ballooned from there.  Oddly enough, I was looking into LM723 designs when I found this project and the accompanying thread(s) and decided to give it a try.  Want to share the schematic for the 723?  Before anyone thinks I'm some kind of guru, I'm not.  99% of people on here will know more than me by a long shot.  The one thing I do have experience in is hardware design.  I've been cobbling things like this together since I was a kid.

A person can get too bogged down into ºC/W and a way to avoid that is to over engineer even if it costs a little more.  My original order including the expensive stuff like the big filter caps, transformers, transistors etc. came to $185.18.  The majority of that was for the transformers ($48.92), the big filter caps ($32.08), and the 10 turn pots ($47.96).  So, the initial outlay for someone that already had these things would have been $56.92.  My first go around didn't go well so I decided to have boards made and use better components.  My second order which just included the parts to redo the boards came to $66.43.  The boards themselves came to $28.80 for ten (four of which I've given away to two other people from here).  I've got about $30 in knobs because one of the things I heard complained about was small knobs being inadequate, so I bought knobs that are over an inch in diameter, solid aluminum and heavy.  I've also got a few bucks in the four TO-3 mounting kits.  The heat sink for the TO-3's cost me less than $16 on Ebay.  As I said, the LM35's were a TI sample order.  The case was a gimme from another smaller case sample order (Thank you OKW).  I've got two Atmel 328P-PU's that were samples.  I've got two AS1100's (pinout compatible MAX 7219's) that were samples.  I've got two MAX7219's that were samples.  I've got two LDT-N2807RI-USB and two TDCR1060M LED displays from Lumex and Vishay respectively that were samples.  Starting to get the idea?

I'll make you an offer.  I have four boards left over from my order and I only really need to keep two more.  Would you like to have two?  PM me your address and I'll send you out a couple to build just for the fun of it.  You could probably use stuff you have laying around to populate it with but the C7 footprint is for a large film cap so you'll have to rig something there if you don't have one (Digikey #399-5999-ND $2.70 each).  Just remember to flip the BD557 around backwards because the footprint is wrong on the silkscreen.  That's my fault for not checking before I sent them out to be made.

"having fun, isn't that why we do things"..Yes. Right. That's the answer. I suppose if we really analysed this project, we would have to agree that it has to be a fun project. You have a 'scope, have a good look at the ripple and noise at max load. Jot down your findings and then compare them with a basic commercial unit with similar specs. I not saying its bad. In fact your project will be better than mine, for lots of reasons. R7 is just one reason. The bigger, the cooler and yes, more stability and less of the infernal drift. A precision low temp coefficient, here, would work better. Also the balance resistors in the emitters, low ESR caps..ceramic, polycarb..the list is endless. The LM723 is old, I think Fairchild first made it in the 80s It has a few revisions, but no real alternatives. I've built a few power supplies in my time but when you need variable voltage and current sense, I usually ended up with a 723. This is why I wanted to give this one a try. I really wanted to know why the original designer didn't take this easy route. I remember seeing an old version of this circuit, using only two op-amps. I have a feeling that this circuit simply developed over, many years, as circuits often do.

Your offer to send me your spare boards is a kind gesture. I am tempted to accept, but firstly, I live at the end of the world. New Zealand. Secondly, there are others, on this board, that are far more deserving than me. Finally, I would never build another power supply using this type of circuit. I would use a fat MOSFET driven by a 723 and for R7? Well I guess we are stuck with this and the transformer. I hate tranformers! They are so big and heavy and very expensive. If you come across a circuit idea that uses a variable switch mode concept I would be very interested. What do you think of this? I have not seen a full spec variable PS with current sense based on the switch mode, buck regulator. Liquibyte, I want one of those. 
A 723 and a MOSFET would do the trick. But we still need a current sense..R7   

 

[liquibyte]

About the only thing I can point you at that will get you on your way is this book.  There are many designs and ideas out there that can be had but, as you know, the level of complexity increases with switchmode stuff.  I was just looking at this app note after having read your post.  I'll keep my eyes peeled but I'm sure that others here could offer more suggestions than I can, I'm just a hobbyist and some here do this for a living.  I wanted a linear supply due to the simplicity and, honestly, this one worked out for me for the most part.

I will probably eventually hook up to a scope to see what's going on but to be honest, I don't have a very good scope and have never even used it before so I'll have to study up on it first.  I hope anyone that has done this here can chime in with some results.  As for the postage from me to you, it would cost around $5 and I'd be happy to do it just to have another person out there to test things out although it may take a few weeks to get there.

Quick edit:  I did find this 0-30V 2A switchmode supply

 

[redwire]

I just received my Rigol 1074z scope.  I'm probing around the circuit and learning how to use this thing at the same time. Here is a screen shot of the output with the voltage set at zero- there is about 35mV output.  I am using a 10uF electrolytic cap, when I get a film cap i'll post the results.    The only unusual thing I noticed was that Pin 3 of U3 had substantial noise, this could be my set up and perhaps having 12" of wire from the pot to the board had an effect.  I had a 1uF cap handy so I hooked it between ground and pin3 and it tamed things down substantially.  I didn't notice any substantial change with the output whether or not this extra cap is utilized.

View attachment 42088