Jump to content
Electronics-Lab.com Community

New Power Supply


Guest liquibyte

Recommended Posts

Guest liquibyte

So redwire keeps talking to me about this new power supply over at youtube and I thought I'd draw a schematic and board out because the guy hasn't seen fit to publish anything downloadable or static, just videos.

I don't like video tutorials because they aren't searchable so I don't know if the transformer represented is the right rating.  I just did a freeze frame of the video in HD and a screenshot of the schematic and went from there.

I do, however, get a kick out of working with schematics and turning them into boards and this one was fun.  I've kept the big parts and the parts that get hot off the board so what we have is the control circuitry.  I managed to get this board down to a 5cm square which was really challenging and took me two tries.  This could be made really tiny if surface mount parts were used.

I have not built this yet and can't tell you if it works right or how it performs and I'm just putting it here in the hopes that someone might find it useful since schematics or boards don't seem to have been published for it.  I eventually plan on doing some testing with this to see how it works but right now I've been busy with the other one in the big thread(s).

Anyway, here are pics of the board and schematic and the .brd and .sch Eagle files zipped up for your consideration.

Youtube links to each section:
Part 1
Part 2
Part 3
Part 4
Part 5
Part 6

post-107142-14279144760387_thumb.png

post-107142-14279144761236_thumb.png

ps.zip

Link to comment
Share on other sites


I also sketched this a few days ago up but your Eagle skills are much better.  Please check on the following items:

1. The LM358N has a range of 3V-32V which is out side of the needed operating range of the voltage sense amplifier.  Suggest using a TLE2141.
2.  The current sense amplifier has no connected rails.  In this mode the high would be 0 to 1 V, the LED would always be on,  and Q3 might never conduct.  I suggest a TLE2141 with a cap between the output (pin 6) and Pin 2.
3.  I think R9 should be 2K2 instead of 2K7.
4.  Is R13 47 ohms?
5.  RSense is  0.22ohms

Suggestions for discussion:  Changes I intend to implement-
1.  Use a 15V Zener to feed a 7812 linear voltage regulator instead of the fet and zener combo used to set up a reference voltage.  Using a 15V zener would reduce the rail to rail voltage on the current opamp  as well as the voltage to the linear voltage regulator.
2.  C3 is correctly listed as 100uf but this seems high.
3.  The unnamed main filter cap (4,700uf) seems small for 3A.
4.  A small cap between the output of the voltage control opamp and inverting pin may be helpful.

Link to comment
Share on other sites

Guest liquibyte

To be honest, I just did this up for the practice.  I didn't really think about calculating it out and just used a pic of the schematic from the video to sketch things up.  I hate it when people do videos but don't release at least an image of the circuit.


I also sketched this a few days ago up but your Eagle skills are much better.  Please check on the following items:

1. The LM358N has a range of 3V-32V which is out side of the needed operating range of the voltage sense amplifier.  Suggest using a TLE2141.

I knew that from the datasheet and to be honest, knew that the transformer voltage was probably too high for the op amp used.  I might redo it with 2141's in place.  I had originally thought of doing that but I just wanted to replicate the circuit first. I should probably just not be lazy about it and calculate out the right transformer though.


2.  The current sense amplifier has no connected rails.  In this mode the high would be 0 to 1 V, the LED would always be on,  and Q3 might never conduct.  I suggest a TLE2141 with a cap between the output (pin 6) and Pin 2.

The 358 is a dual package so the rails connected on the voltage side will take care of that.


3.  I think R9 should be 2K2 instead of 2K7.
4.  Is R13 47 ohms?

That's the problem with doing a screenshot of a video, hard to read some of the values.  It says 2K7 in my screenshot.  I didn't think twice about the value other than replicating it.  I thought that R13 was 47K but if you zoom in on it, it seems to say 47R.  I'll attach my screenshot so you can have a look


5.  RSense is  0.22ohms

You're right, I forgot to put that value in there.  My bad.


Suggestions for discussion:  Changes I intend to implement-
1.  Use a 15V Zener to feed a 7812 linear voltage regulator instead of the fet and zener combo used to set up a reference voltage.  Using a 15V zener would reduce the rail to rail voltage on the current opamp  as well as the voltage to the linear voltage regulator.

Still, you can see why people would want to do a discrete version though.  I tend to think of regulators as black boxes and op amps as more of a discrete component even though it's not.


2.  C3 is correctly listed as 100uf but this seems high.

I thought that too but, like I said, I just wanted to replicate the circuit in the video.


3.  The unnamed main filter cap (4,700uf) seems small for 3A.

In the original, it was two 2200uF caps but I don't see the need for separate caps and opted for something a little larger but of a common value.


4.  A small cap between the output of the voltage control opamp and inverting pin may be helpful.

That's something I noticed that was missing too.  Doesn't it prevent oscillations or something like that?

post-107142-14279144761614_thumb.png

Link to comment
Share on other sites

Guest liquibyte

I have a Radio Shack 25.2V 2A transformer that I think would work with the LM358 but I think that's about the biggest transformer I would use as far as voltage goes.  I got the part number from the first video when he showed the board that he was reversing.  I distinctly saw a 358 but I didn't watch enough of it to say if there was a suggestion for another op amp.  I have all the parts for this except the current sense resistor and I think I can parallel enough resistors to get that value so I'll breadboard it up in a week or two and see how it goes.  I had also ordered several different types of op amps on my last order so I may give a few different ones a try as well.  I still have some 2141's too so I can try those out as well.

Link to comment
Share on other sites


I think a TL2141 opamp driving the amplified output section will oscillate because the slow 2N3055 output transistor delays the negative feedback.
The single output transistor will get fried if the output voltage is low or shorted and the current is high.


Do you have a suggestion on a different output transistor or opamp?

EDIT:  This combination exists on the other PS project but I don't recall anyone mentioning a problem with the tLE2141 getting fried.
Link to comment
Share on other sites

  • 5 weeks later...

Good News.  I built the subject PS with the following changes:
1.  Used 10,000uf  main cap instead of 4,700 uf
2. Used TLE 2141 op amps with caps similar to Project Power supply
3.  Deleted funky zener/Fet/Resistors reference voltage an used a 12V linear regulator as reference voltage for better thermal coefficient
4. Added a few other caps similar to Project Power supply.
5.  Changed the value of a few resistor for a greater current range and to limit output until a trimmer can be added to negative feed back loop.
6.  Used MJ11016 single output transistor.

Outcome:  Using 12 v setting being controlled by current  with 0.5 amp bulb I didn't see any over shoot at start up and the design has the ability to turn voltage down to less than 5mv without a load. 


EDIT 11/22/14    Note that voltage dropped with load on this design.  See later design

Changes still needed:  the 100uF cap at the output provides sluggish voltage drop with no load.  I added a 10k resistor to output which helps significantly but I intend to change the 100uF to 10uF.

Overall ,  very impressed with function.

NewFile12:  is during shutdown
NewFile13:  when plugged in with load

post-34537-14279144809535_thumb.png

post-34537-14279144809666_thumb.png

post-34537-14279144809807_thumb.png

Link to comment
Share on other sites

I used thru hole parts for this trial board.  I don't see any reason why smd parts wouldn't work just fine.  The only parts that get warm are the bridge diode, resistor for power on led and the .22 ohm 5W non inductive resistor when pushing a heavy load.  I built the board to also fit a 10W 0.27 ohm resistor, but unless you are trying to get more than 3A the 5W resistor should be fine.  Using the darlington power transistor puts very little load on the BD140.  Just ordered 3 PCB's for the design shown.

Link to comment
Share on other sites

Guest liquibyte

I haven't forgotten about this, it's just the work is coming along slowly.  I'm trying to see if I can get all the components on the top layer and use the bottom for one massive ground plane.  I'll eventually get it but since I haven't worked with surface mount before, I'm having to learn a few new techniques.

post-107142-14279144824458_thumb.png

Link to comment
Share on other sites

I wanted to note that R11 (the 18K resistor) that provides negative feed back to the voltage regulator and  is tied to the output side of the sense resistor.  The original project shows it connected the the transformer side of the resistor.  The youtube video for the original design showed a voltage drop at the output when a load was applied.  I believe by moving the ground connection to the output side of the sense resistor will fix this problem.

EDIT:  C5 should be 200nF

This topic is corrupted. Please continue discussion on this topic.

post-34537-14279144894348_thumb.png

Link to comment
Share on other sites

  • 2 weeks later...

After running some tests, here is the latest version.  Seems to be working pretty well.  I was able to set the voltage to about 3.5 V -unloaded  and then put a load about 1.5A with no drop in voltage.  I will continue to do more tests.  I found that the leakage voltage to be about 3-4mV which is very good and the diode did nothing to change that.

post-34537-14279144855731_thumb.png

Link to comment
Share on other sites

Guest liquibyte

After running some tests, here is the latest version.  Seems to be working pretty well.  I was able to set the voltage to about 3.5 V -unloaded  and then put a load about 1.5A with no drop in voltage.  I will continue to do more tests.  I found that the leakage voltage to be about 3-4mV which is very good and the diode did nothing to change that.

Your R9 82R resistor is hooked up the the wrong rail in that schematic I think.  I'm also getting a lot of oscillations a low voltages when I simulate this and can't get the output to go above 12V with a load.  You say you've got this built?  What's the output look like on the scope at lower voltages?

post-107142-14279144848687_thumb.png

Link to comment
Share on other sites



Your R9 82R resistor is hooked up the the wrong rail in that schematic I think.  I'm also getting a lot of oscillations a low voltages when I simulate this and can't get the output to go above 12V with a load.  You say you've got this built?  What's the output look like on the scope at lower voltages?


If you connect the 82R resistor to the other rail then you will never measure the voltage drop across the current op amp and the PS will never current regulate.

I had 3 boards made based on the first schematic that I posted.  I  did notice a few problems 1) voltage drop when I added a load  2)I changed to a 10uF cap for faster response but discovered significant noise.  I changed back to a 100uF and it is smooth.  3) It had some noise on pin 3 of the voltage op amp so I put a cap on it and it was quiet.    4) I initially had a cap betwen the ground and ground pin on the 10k potentimeter and that didn't work.  5) There was an grounding error on the old .brd file.    As far as I know the sketch shown is correct. 

I get 37V output with a 28AC transformer.  If you only get 12 v then you likely have a short (or equal value) on the feed back resistors on the voltage op amp. Considering you are feeding 12V and only getting 12V out.
Link to comment
Share on other sites

Guest liquibyte

I can only get the sim up to 1.1A and it drops the voltage below 12V when it gets there.  I zipped it up for you to try because I may have gotten something wrong.  I've been working on the other one and it's gotten my eyes tired today.

Edit: forgot to add the 2141 models to the zip.

new-ps.zip

Link to comment
Share on other sites

Liquibyte  here are some results:  The blue line is the reference voltage to the voltage control op amp and Yellow is output voltage.

NewFile 3 is startup 12v, under load
NewFile 4 is startup 12v, no load
NewFile 5 is shutdown with load  (no load takes a long time to shutdown for scope)
NewFile 6 is startup 3-4v, under load

post-34537-14279144852505_thumb.png

post-34537-14279144852634_thumb.png

post-34537-14279144852755_thumb.png

post-34537-14279144852875_thumb.png

Link to comment
Share on other sites

Guest liquibyte

I wonder if this is being caused by the zener.  I encountered something like it in the sim I was doing for the other supply.  The spike was cascading through the downstream components and took me forever to figure out.

post-107142-14279144852998_thumb.png

Link to comment
Share on other sites

  • 3 weeks later...

Looking for suggestions.    I built the board as shown in the above post with quite a few wires and jumpers and things seemed to work very well.  I then had 3 boards made.  I populated 2 of them.  Strange things happend on the new boards.  With no load the output would spike about once every second.  I changed C5 to 200nf instead of 100pf and the PS worked smoothly with no load.    If I set the voltage to about 4V and connect a load,  the output voltage instantly surges and oscillates between max and 17v at a frequency of 800hz.  Neither the voltage or current control seem to work with a load.  If I set the current to minimum before attaching a load then the PS will not ramp to max and oscillate. 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
  • Create New...