Building the Elektor 1982 power supply ,what pcb?

[BeetleJuice]

i have almost all the parts to build that power supply.
the plan is to make the elektor 1982 my second diy psu.
i got the original pdf from the magazine , edited the pictures for print, but i wanted something more pro (some trim, soldermask etc)
if i design everything in easyeda there is a big chance for errors since i am not a pro, mostly i do small projects in easyeda,but this is a complicated pcb.
so i thought that maybe a pro guy might have done the same already.
 
i searched everywhere i knew to find a "modern" pcb , so far i only found 2 guys that made this pcb from scratch.
one is in easyeda and one in github.
my plan is to trim a little (or not at all) the pcbs and order some ready pcb from jlcpcb

with respect to the designers i will upload the two links and the photos from the schematics and the pcb
i need someone to take a look at the schematics cause i am just a hobbyist not a professional electronics guy.

psu 1 link in github
https://github.com/stonetronics/powerSupply
seems ok but has some extra mods. cant find a forum or blog link to see the progress.

[BeetleJuice]

psu2 link in easyeda
https://easyeda.com/Gregi/Lab_PSU_0_30V_0_3A-242aadfa16604425b9cd9c255115aed4
gregy says "Still in progress.... need to doublecheck schematic", no further link for his design, blog etc

so what do you think i should do?
can i trust some of these boards and print them or stick with the original one?

thank you all

[CatalinaWOW]

First, the caveats.  There are those here far more knowledgeable than me and I have only taken a quick glance through these.  Overall I don't see any glaring mistakes.

1.  Your design appears to have terminal blocks adjacent to each other in opposite orientation on the upper left side of the board.  If this is intended as a simple form of keying to prevent incorrect connection, experience has shown that it is totally inadequate.  If it was simply to simplify layout you will probably end up regretting the access issues and confusion that result.

2.  The boards you found on the net were designed to be cut on a PCB mill.  The copper pours are result of the process, and AFAIK are not needed for the supply.

3.  The partitioning of the circuit onto separate boards as done on the web design is worth considering.  It may make packaging in a small enclosure, mechanical support of heavy components, and heat sinking of the transistors enough easier to compensate for the extra cabling between boards.  It will depend on your plans for putting this in an enclosure.

[BeetleJuice]

after some hours i found the differences in the first pcb

1)it uses a fan control that can be removed without harming the rest circuit ,am i right?
2)it uses for some reason one 3055 and 4 BD243 , cant understand why
3)it has some extra led and circuit for that led
4)one resistor is missing

i marked all the changes and i upload the original schematic
can you please check if the changes are ok ?

[BeetleJuice]

any chance to help me with the new schematic above?
thank you!

First, the caveats.  There are those here far more knowledgeable than me and I have only taken a quick glance through these.  Overall I don't see any glaring mistakes.

1.  Your design appears to have terminal blocks adjacent to each other in opposite orientation on the upper left side of the board.  If this is intended as a simple form of keying to prevent incorrect connection, experience has shown that it is totally inadequate.  If it was simply to simplify layout you will probably end up regretting the access issues and confusion that result.

2.  The boards you found on the net were designed to be cut on a PCB mill.  The copper pours are result of the process, and AFAIK are not needed for the supply.

3.  The partitioning of the circuit onto separate boards as done on the web design is worth considering.  It may make packaging in a small enclosure, mechanical support of heavy components, and heat sinking of the transistors enough easier to compensate for the extra cabling between boards.  It will depend on your plans for putting this in an enclosure.

[mvs]

the plan is to make the elektor 1982 my second diy psu.

Is there any reason to build such old design?

1)it uses a fan control that can be removed without harming the rest circuit ,am i right?

IC3 is used not only for a fan control, but also for the thermal protection (IC3B) and to adjust 0 for current limiter (IC3A).

2)it uses for some reason one 3055 and 4 BD243 , cant understand why

Schematic/PCB was redesigned to allow different power stages. A single transistor will be sufficient for a 1A supply.
However 2N3055 has low current gain (hFE) and will not work as intended. It should be darlington pair like TIP142.

3)it has some extra led and circuit for that led

This led indicates CV mode. If you do not need this function, you can leave R24, R33, R34, D11, Q2 and LED2 unpopulated.
 

4)one resistor is missing

R5 is placed in parallel with R4. You need to populate only one.

[BeetleJuice]

thank you for the reply
do you think that this circuit is ok ? and the pcb to send it for printing as is?
i have a link with the eagle files

1) you are right, i will keep the ic3 and the whole fan circuit but i guess we can leave unpopulated the parts for fan and thermal control while keeping the ic3 in place.
2)how many and what kind of transistors do i need for 30v and 3amps max?
3) we need this function, i don't know if this extra circuit harms the rest of the original circuit.

i am open to new designs if we know for sure that the psu will work correctly. i just need to reuse my case, my analog meters,my transformer etc, cause right now all these need only one good pcb to make a whole psu unit. i could also buy a good pcb,a kit etc the point is not to trash everything cause my old psu pcb was not working correctly..

thank you so much!

the plan is to make the elektor 1982 my second diy psu.

Is there any reason to build such old design?

1)it uses a fan control that can be removed without harming the rest circuit ,am i right?

IC3 is used not only for a fan control, but also for the thermal protection (IC3B) and to adjust 0 for current limiter (IC3A).

2)it uses for some reason one 3055 and 4 BD243 , cant understand why

Schematic/PCB was redesigned to allow different power stages. A single transistor will be sufficient for a 1A supply.
However 2N3055 has low current gain (hFE) and will not work as intended. It should be darlington pair like TIP142.

3)it has some extra led and circuit for that led

This led indicates CV mode. If you do not need this function, you can leave R24, R33, R34, D11, Q2 and LED2 unpopulated.
 

4)one resistor is missing

R5 is placed in parallel with R4. You need to populate only one.

[BeetleJuice]

what do you think of this psu and pcb?
bad news is it has smd parts and no eagle files to customize it.

http://diyfan.blogspot.com/2013/03/adjustable-lab-power-supply-take-two.html

[mvs]

what do you think of this psu and pcb?
bad news is it has smd parts and no eagle files to customize it.

http://diyfan.blogspot.com/2013/03/adjustable-lab-power-supply-take-two.html

It is quite similar with a chinease lab power supply kit, that you can get for around $5. It does not require separate supply for control circuit as elektor design, but max. output voltage is limited by supply voltage limit of opamps. You can replace TL082 with some 44V opamps, like in this blog
http://www.paulvdiyblogs.net/2015/05/tuning-030v-dc-with-03a-psu-diy-kit.html

[mvs]

do you think that this circuit is ok ? and the pcb to send it for printing as is?
i have a link with the eagle files

I do not use eagle, so i can not check pcb for mistakes. Schematic looks ok.
 

2)how many and what kind of transistors do i need for 30v and 3amps max?

The worst case is short circuit with 3A limit. Output stage will dissipate around 100W of heat.
If you take one transistor in TO-247 case, like TIP142 with Rjc = 1 °C/W, Tjmax=150 °C, then max allowed case temperature Tmax = 150 °C - 100W * 1 °C/W = 50 °C. With good heatsink and thermal shutdown at say 45 °C it should be safe.

If you take three transistors in TO220 case, like BD243 with Rjc = 1.9 °C/W, Tjmax=150 °C, then you get  Tmax = 150 °C - 35W * 1.9 °C/W  = 83,5 °C.
You can set thermal shutdown much higher.

[mvs]

In China one can find just anything...
https://aliexpress.com/item/32946735505.html
Basic topology looks similar to Elektor design, but it has lm358 opamps, voltage range switching and fan control. TL431 serves as a reference.

[xavier60]

This is the latest version of a design that's working well in one of my bench supplies.
I suggest that it should be bread-boarded first to prove to yourself that it works properly before committing to having a PCB made.

https://www.eevblog.com/forum/beginners/lm324-power-supply-with-variable-voltage-and-current/msg3345000/#msg3345000
https://www.eevblog.com/forum/projects/linear-lab-power-supply/msg2388873/#msg2388873

[mvs]

This is the latest version of a design that's working well in one of my bench supplies.

Is it stable with generic parts tolerances, or you need to tweak control loops for every unit?

[xavier60]

This is the latest version of a design that's working well in one of my bench supplies.

Is it stable with generic parts tolerances, or you need to tweak control loops for every unit?

Yes, it works fine with an LM358. There is some useful detail in the last few pages of the thread.
With the TLC072, I have set the CV response to about 10uS, it can be set to about 3uS but no need to.
The design also has fast current limiting. D5 speeds up the response of the CC op-amp when the current spikes above some level.

[xavier60]

It is possible that the CV compensation has to be readjusted if the output transistors used have a very high hFE because of the increased gm of the output stage.

[blackdog]

Hi,

Yes the power supply from Elektuur 1982 is an "old design"
But the design is based on the developments of Harrison / HP from the end of 1950.
As far as I am concerned this is still one of the best designs for linear power supplies.

As it is shown in the "real" schematic with relatively old components also has a big advantage.
The low bandwidth of the 2N3055 and the uA741 opamps makes it very stable for rebuilding.

This same design can also become an even better power supply with faster components.
Without knowledgehow how to do this, I would not start modifying the design/schematic!

Here some advice.
1e
So don't go modifying this design if you have no knowledge of phase margin and loop control!

2e
Stick to the original schematic, good versions of the 2N3055 transistors and the uA 741 ic are still available.

3e
Don't use China junk bought through ebay/Ali.

4e
Gives a lot of attention to your wiring technique, wires are also inductances and resistors.

5e
Places the electronics far enough away from the transformer and the wiring of the bridge rectifier and buffer electrolytics.

6e
It is possible to get very low hum and noise values with this power supply design.
But if the above regarding the transformer and wiring is ignored the good features are thrown out the window.

7e
A toroidal transformer gives and much smaller interference field and that is my preference to use.
Twist the transformer wires and also the wires to the buffer elco's.

8e
Do not place the power supply in a cabinet that is too small, the power supply will become much too hot under maximum load.

9e
I regularly build power supplies which I then equip with a CPU cooler with good fan control.
Most of the time the fan is inaudible but still cools well because it is set slightly above the minimum speed for the fan used.
This way you always have airflow and it keeps the power supply cool and the fan is only audible when needed.

10e
Most who build a power supply's choose the heatsink too small, give this attention!
If possible, mount the 2N3055 directly on the heatsink and then isolate the heatsink, for optimal cooling

If this is not possible, use a "thin" mica plate, with a thin layer of a good heatsink paste on both sides, that is also used for CPUs, often these is heatsinks paste with silver in them.

11e
Use a snubber over the transformer windings. and your Power Switch.

12e
You want a 5A continu power supply? then you need a 5A/0.63 transformer, that wil be a transformer that can deliver 8-Ampere!

13e
Mmmm
I see that schematics are " redrawn " these are to me horribly unreadable. 
Look at the original diagram, which is not optimal but can be understood reasonably well at first glance if you are familiar with the components used.
Dividing the schematic into blocks is the horror taken from the draftsmen of digital electronics, bad practice.

14e
Sorry for the sometimes crooked sentences, I am dislectys and use a translation machine, so that's is not win-win but lose-lose. 

I hope this helps

Kind regards
Bram

[BeetleJuice]

do you recommend this psu? can we mod it to have sense output like elektors psu?
i really try to understand how the dual ac works. what happens to the high voltage the big cap has when you switch to lower voltages (so the tap changes)?
how can i find more info of this feature?

In China one can find just anything...
https://aliexpress.com/item/32946735505.html
Basic topology looks similar to Elektor design, but it has lm358 opamps, voltage range switching and fan control. TL431 serves as a reference.

[mvs]

do you recommend this psu?

It is not perfect supply, there are some cut corners and things to improve.
For instance i see no hysteresis on tap switching opamps. If output voltage is set near switching threshold, they may switch relays back and forth nonstop.
100uF output capacitor is also quite large, i would prefer a smaller one.
Preload with a resistor is easy solution, but is not so efficient as constant current sink.

can we mod it to have sense output like elektors psu?

Sure, but you will probably need to cut some traces on PCB.

i really try to understand how the dual ac works. what happens to the high voltage the big cap has when you switch to lower voltages (so the tap changes)?

Diode bridge will be closed and bulk caps will be discharged throught load and discharge resistor R17.
This supply can work also without tap switching. It would just dissipate more heat at low output voltages.

[xavier60]

This is the latest version of a design that's working well in one of my bench supplies.

Is it stable with generic parts tolerances, or you need to tweak control loops for every unit?

Replacing the output stage with a PNP Darlington might reduce the affect that hFE has on gm.
The use of Darlington pairs in the Harrison topology doesn't seem to have issues caused by hFE parameter spread,

[Gregi]

I never got the time to recheck but i think i made a mistake in drawing a schematic. I am building my own house and a new workshop for the past 2 years, and with my regular daytime job I just dont have time to do it. I will try to do it by the end of the year.

PS. for the copper area, someone replied to you that it was made for milling on cnc and that is correct (i make my own protoboards using photo etching). I will modify the PCB for normal manufacture.

btw its today when i saw this post. If you need any more info, just ask. And yeah... You can make two of these and use them in series or in paralel, depending on your needs. It can give more current with small mods (more output transistors).

this is the version with some smd components made in kicad while i was learning how to use it.

[simba15]

Can you post the schematic or files for the smd version?

Looks quote nice.

Thanks.