Jump to content
Electronics-Lab.com Community

0-30V Stabilized Power Supply


redwire

Recommended Posts

Guest vakeiros

Hi Vakeiros,
You can reduce the value of R7 to 0.33 ohms for a maximum output current of 4A but when the output voltage is low or is shorted then the driver and output transistors will get very hot.
For a 5A circuit we recommend using 3 output transistors (each with a series emitter resistor), then they share the heat.

With a 25V transformer you probably will not have an output as high as 30VDC when the current is 4A.


HI AudioGuru, yes I know about that, I'd read most of this thread already.

To start I'll make one 3A with a 25V 100VA Transformer i have here sitting around and see how this PSU behaves, if i like it (I'm sure I will) I'll make a few more. For now I don't need 30VDC, only will need about 24-25V so the 25VAC will give me that I think.

I Know for the 5A I have to do a few changes (3*3055 with .33 resistor, lower sense resistor .27r, etc..) like i said, for now the 3A to start, then we'll see XD

thanks for the reply  ;)
Link to comment
Share on other sites


Guest rvendrame

Hello gurus,

I will build one of these.  I have a toroid transformer of 28V  3A.  So I think I need to run it with maximum of 2A, in order to not overload the transformer, right?  And in this case the only modification would be increasing R7 to 0.67 ohms/10W?  And whould I keep the 2x 2n3055,  or just one with big heatsink will do the job?

Thanks!

Link to comment
Share on other sites


Hello gurus,

I will build one of these.  I have a toroid transformer of 28V  3A.  So I think I need to run it with maximum of 2A, in order to not overload the transformer, right?  And in this case the only modification would be increasing R7 to 0.67 ohms/10W?  And whould I keep the 2x 2n3055,  or just one with big heatsink will do the job?


The 28V transformer produces a peak voltage of 39.6V. If the maximum output current from your project is 2A then the transformer must produce 39.6V x 2A= 79.2VA.
But the transformer is rated for 28V x 3A= 84VA so it will be very warm but be fine.

0.67 ohms is not a standard value, use 0.68 ohms for R7. Its max dissipation is 2 squared x 0.68 ohms= 2.72W. Use a 5W resistor.

The unregulated loaded voltage will be 39.6V - 2V - 1V= 36.6V. If the output is shorted at 2A then the output transistor (s) must dissipate 36.6V x 2A= 73.2W which is a lot for a single 2N3055 transistor even if it has a big heatsink but it will be fine if a fan is added. Use two output transistors on a pretty big heatsink without a fan. 
Link to comment
Share on other sites

Guest thiagoas

OK, now that I tested a little bit this power supply, I have some points.

BD139 don't get even warm, it was those shit 2n3055 overcharging him.
When I set, for example 1.5A limit, after a while that current starts going down, until it stabilize in 1.39A. What could cause that?
There is a display picture in the files on the first post of this thread, what project is that from?

thank you all guys ;D

Link to comment
Share on other sites


OK, now that I tested a little bit this power supply, I have some points.

BD139 don't get even warm, it was those shit 2n3055 overcharging him.

The output transistors and the BD139 driver transistor get warm when they dissipate power. Power=  x voltagecurrent. So they heat when the current is high and the output voltage is low (because then these transistors have a high voltage across them).

When I set, for example 1.5A limit, after a while that current starts going down, until it stabilize in 1.39A. What could cause that?

If you set the voltage to 30V, set the current to 1.5A and have a 12 ohms load then the 30V across the 12 ohms would normally cause a current of 30V/12 ohms= 2.5A. but the current regulator causes the output voltage to drop to 1.5A x 12 ohms= 18V.

Since your current drops after a while then maybe the current sense resistor R7 increases its value when it heats (wirewound?). Or maybe your current-setting pot changes its value?

There is a display picture in the files on the first post of this thread, what project is that from?

The photo of the circuit board is this modified project and the pcb was designed and built by Redwire.
Link to comment
Share on other sites

My pot is a 10-turn, I think it is OK.

Yes, most 10-turns trimpots have excellent quality.

R7 is wirewound, it varies the value with the increase of the temperature? What would be the most recommended?

Yes, a wirewound resistor increases in value as it heats. Maybe yours is too small so it gets too hot? Maybe it is enclosed and gets hotter and hotter? Maybe it has poor quality?
Link to comment
Share on other sites

Guest thiagoas

I will make some temperature tests with it, and put something bigger if it doesn't work well.

Today I bought the case for the power supply, it's gonna look pretty.  ;D

Link to comment
Share on other sites

There are a few threads about this project. The latest BOM and schematic are many years old and have been posted hundreds of times in the threads. A few pcb designs have been shown.
Recently a member organized these files and requested that a Moderator post them at the beginning of one of the threads but I don't think it happened.

I fixed the project but never built it. If I built it then I would have posted it as a new project.

I found it. Look at the reply from Peter K on August 15, 2014 in this thread.

Link to comment
Share on other sites

Guest liquibyte

There are a few threads about this project. The latest BOM and schematic are many years old and have been posted hundreds of times in the threads. A few pcb designs have been shown.
Recently a member organized these files and requested that a Moderator post them at the beginning of one of the threads but I don't think it happened.

I fixed the project but never built it. If I built it then I would have posted it as a new project.

I found it. Look at the reply from Peter K on August 15, 2014 in this thread.


Thanks for the quick answer (and for your contribution of course)

I'll have a look at the beginning of the threads

Paul

Done.  See this Sticky.
Link to comment
Share on other sites

Guest liquibyte

Thanks, Liquibyte.
I hope you or Peter K will select and add one of the excellent pcb designs and place it in the sticky.

I thought about that but how do I pick?  I didn't want to add mine because I thought it would seem pretentious to just put that one up.  I guess I should go through the thread(s) again and link them to the relevant thread plus include the files.  That may take awhile but I'll work on it.
Link to comment
Share on other sites

Guest paul18fr

Very interesting project ; nevertheless I'm trying to find the transformer (primary = 230V for EU & secondary = 28V - 120 W at least) and the only ones I've found are too expensive (close to the price of a 30V - 5A power supply of the market), even if I degrade the output voltage to 24V, and thus the features of the power supply

Last chance : does somebody have ever find it  (even in CN, HK, etc. ..)

Thanks

Paul

Link to comment
Share on other sites

Guest liquibyte

Very interesting project ; nevertheless I'm trying to find the transformer (primary = 230V for EU & secondary = 28V - 120 W at least) and the only ones I've found are too expensive (close to the price of a 30V - 5A power supply of the market), even if I degrade the output voltage to 24V, and thus the features of the power supply

Last chance : does somebody have ever find it  (even in CN, HK, etc. ..)

Thanks

Paul

I used this one.  I don't know if you can get it somewhere other than Digikey but it's a Triad Magnetics model VPS28-4600.  Try a search with that and see if something local comes up.
Link to comment
Share on other sites

Guest liquibyte

I tried to do my best including all of the projects from both threads that had a semblance of completeness to them in the data sticky.

Some are woefully incomplete as they are just PDF's and would require work to be printed out and used to make boards.  PicMaster's version is there several times, sometimes as just the schematic and finally in his full version.  He only used PDF format so there is that.  Some are just image files but their authors reported working versions based on those.  I did include my version as well and can report that it's the only one that has the gerbers that can be sent out and made with no extra work (they are Itead gerbers).  I included a through hole version and a surface mount version that redwire did.  These are complete enough to generate gerbers from and sent out.

I unlocked the topic so that those wishing to add their completed projects can do so.  This is NOT a discussion thread, it's a completed thread.  I don't think it needs to be a requirement to have generated gerbers but I do think that at the very minimum a pic of the schematic used and parts list should be included as well as completed project pics.  Not everyone uses Eagle but I think that adding those files would be nice.  The same goes for Kicad or Diptrace.  If you use Altium or some unrealistically expensive commercial product, at least have the courtesy to attach screenshots/exports in addition to the programs default generated files.  Remember also that not everyone uses windows and not doing so excludes Linux and BSD users from even wanting to try your version.

If I missed something, let me know.  I can assure you it wasn't on purpose, those are both really, really long threads.

Link to comment
Share on other sites

Guest liquibyte

Why was Q1 in the original schematic removed?  After researching the voltage spike issue a bit after participating in a thread over at eevblog and one of the members pointing me to a solution that was talked about recently and directing me to the Tektronics PS-503A manual that suggested this situation happening and showed this exact workaround, it seems that this solution was already designed into the original project but then removed with the update.

Link to comment
Share on other sites


Why was Q1 in the original schematic removed?  After researching the voltage spike issue a bit after participating in a thread over at eevblog and one of the members pointing me to a solution that was talked about recently and directing me to the Tektronics PS-503A manual that suggested this situation happening and showed this exact workaround, it seems that this solution was already designed into the original project but then removed with the update.

On the Tektronics circuit it looks like transistor Q15 shorts the output of the voltage error amplifier opamp U45 to ground when the main power is turned off. I did not see their discussion of it.
I do not know which opamp they use since they do not use standard part numbers.

The original circuit of this 0V-30V power supply here used the TL081 opamp that has the problem called "Opamp Phase Inversion" which causes its output to suddenly go high when an input voltage goes outside its allowed negative input common mode voltage range. The allowed negative input voltage range for the TL081 is a maximum of 4V more positive than its negative power supply voltage which was -5.6V. Then when the main power is turned off the negative -5.6V supply disappears quickly and causes the input of the voltage error amplifier opamp U2 to have a voltage outside its allowed input common mode voltage range. Then it would cause the output voltage of the project to go as positive as it can while the huge main positive filter capacitor was discharging. They added Q1 to short the output of U2 to ground when the main power was turned off.

The new MC33071 and TLE2141 opamps do not have this problem so transistor Q1 was removed.

Adding Q1 back might not fix the existing voltage spike because it might not activate quickly enough.
Liquibyte, can you try it please? 

post-1706-1427914472876_thumb.png

Link to comment
Share on other sites

Guest liquibyte

I'm starting to have trouble keeping up with the changes between how the original circuit was tied to ground vs. the revised version.  In the original schematic, U2 was at the same ground potential as U3 with Q1's voltage divider being tied to that point.  In the revised version, U2 grounds out to the common ground.  I tried to tie in Q1 like in the original but left U2 alone.  I also didn't add R15 back in as I wasn't sure of its purpose.  After doing all that, I was only able to get 200-250mV out regardless of what I had the voltage pot set at so I suspect that Q1 was shunting to ground regardless.  I only have a few basic NPN transistors such as the 2N2222A so I'm not sure if the differences between that and the original BC548 is an issue or not.  I'm also not sure if I'd need to modify the values of R14 and R15 from 1.5K and 10K respectively.  I've also been told that D10 may need to be added back to ensure reverse Vbe protection.

I was also reading another thread where a user had an inductance issue and this app note was posted as a solution.  His issue was with inductance on long battery leads being connected to the device versus a power supply but I'm starting to think that several things could be at play here and I found this app note interesting though it goes above my pay grade quite a bit.  I don't have a scope that can sample and hold so I can only go by the pics redwire posted and the brief glance I can get from the old one I have and my meter but there seems to be a correlation here.  I'm especially seeing what you mentioned with "the output voltage of the project to go as positive as it can while the huge main positive filter capacitor was discharging" on redwire's version that I've been testing so I have a feeling that the inversion is happening there but not necessarily with my versions where I just get the spikes.

Link to comment
Share on other sites

The original circuit used R13 and R14 to keep Q1 turned off when the positive unregulated supply was about +28V and the negative supply was -5.6V.
When the mains power was turned off then the negative supply disappeared first allowing R13 to turn on Q1.

To add Q1 to the latest circuit then the value of R14 must be decreased a lot so that the base voltage of Q1 is at a negative voltage when the circuit is working.

The 0V (ground) of both circuits is the output 0V, except the adjustable reference voltage for U3 from R17. The original circuit had -5.6V for the same negative supply for U2 and U3 (R10 was connected wrongly to 0V instead of to the negative supply).

The latest circuit uses an opamp for U2 that works perfectly without a negative supply so it does not have a supply voltage more than its maximum rating of 44V.
U3 needs to have its output go negative one diode voltage drop so that D9 can pull down the drive voltage during current regulation. Then the negative supply for U3 was made -1.3V.

R15 was removed because it didn't do anything useful and it wasted valuable output drive voltage. D10 also never did anything and still doesn't do anything. 

Link to comment
Share on other sites

Guest liquibyte

To add Q1 to the latest circuit then the value of R14 must be decreased a lot so that the base voltage of Q1 is at a negative voltage when the circuit is working.

Something like a standard 270R or so or am I way off here?
Link to comment
Share on other sites

Guest liquibyte

There's good news and bad news.  Adding Q1 back in with a 12K 160R voltage divider and moving U2's pin 4 to the D7/R3 junction stopped the spikes.  The bad news is that I can now only get up to around 7.6V at the output.  More than half of the voltage pots travel on the upper end does nothing.  I think we're getting close, but I wonder what may have caused this.

Just a note, I replaced redwire's original single turn pots with 10 turn versions so I could have greater resolution while testing but that shouldn't make a difference.

Edit:  Hang on, I just realized I didn't move R10 from where it is to where it needs to be.

Link to comment
Share on other sites

Adding Q1 back in with a 12K 160R voltage divider and moving U2's pin 4 to the D7/R3 junction stopped the spikes.  The bad news is that I can now only get up to around 7.6V at the output.

Sorry, I made a mistake converting the original schematic to the fixed and improved one.

You do not want the negative supply pin 4 of U2 to be negative -1.3V because then U2 will have its total supply voltage very close to its maximum rating of 44V (+42.6V plus -1.3V= 43.9V).
U2 does not need a negative supply so connect its pin 4 and R10 to the 0V from the unregulated power source.

Most little NPN silicon transistors have avalanche breakdown (like a zener diode) of their base-emitter junction when it is reverse biased more than about 6V so maybe you have the emitter and collector pins of Q1 reversed?

Here is my corrected schematic:

post-1706-14279144729486_thumb.png

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...