Author Topic: 24V power supply generating higher voltages  (Read 6033 times)

0 Members and 1 Guest are viewing this topic.

Offline WojtaTopic starter

  • Contributor
  • Posts: 33
  • Country: cz
24V power supply generating higher voltages
« on: May 14, 2017, 04:19:24 pm »
Great evening guys, (at least for me  ;) )
Im having some dificulties with my power supply design. Its a pretty basic one, it takes 24VAC from a transformer and converts it to 24VDC and 12VDC, Ive attache a schematic. There is also a fet as its made for powering a load thats later regulated with the FET. The point is there is something weird going on. The 24VDC line is OK but the 12VDC isnt. The regulator keeps breaking on me and then regulates to weird voltages, now there is 34V on the 12V line... The exact regulator Im using is this one http://www.ges.cz/sheets/l/l78xx.pdf
Im guessing there could be some voltage spikes and these break the regulator but Im not sure. Cant really check for spikes as I don have a scope.
Do you see anything that could cause problems in my design, It aleready broke twice on me and I still dont know what caused it.

Thanks for any help.

Oh and the diode, it was added to block the current going back to the load and discharging the caps that smooth out the 12V line but I guess its not really necessary as the regulator shouldnt pass any current back to the load.
 

Offline Kalvin

  • Super Contributor
  • ***
  • Posts: 2145
  • Country: fi
  • Embedded SW/HW.
Re: 24V power supply generating higher voltages
« Reply #1 on: May 14, 2017, 04:25:23 pm »
 

Offline bktemp

  • Super Contributor
  • ***
  • Posts: 1616
  • Country: de
Re: 24V power supply generating higher voltages
« Reply #2 on: May 14, 2017, 04:27:24 pm »
- rectifier is wrong
- BC557 for Q1 is wrong, use BC547 or other NPN BJT instead
- freewheeling diode missing across the load
 

Offline rob77

  • Super Contributor
  • ***
  • Posts: 2085
  • Country: sk
Re: 24V power supply generating higher voltages
« Reply #3 on: May 14, 2017, 04:27:50 pm »
as mentioned by Kalvin, your rectifier is wired wrong.
D1 is pointless, C2 is in wrong place. and why you don't have a bulk smoothing capacitor after the diode bridge ?
add approx 3000uF of capacitance per ampere of current drawn right after the diode rectifier, remove C2 ad increase the ceramic caps to 100nF (you have 10nF) around the 7812.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 19962
  • Country: gb
  • 0999
Re: 24V power supply generating higher voltages
« Reply #4 on: May 14, 2017, 04:30:59 pm »
I wouldn't be surprised if the AC voltage is a bit higher than 24V which is killing the regulator. The peak voltage of 24V is 34V, which is very close the the maximum rating of 35V.
 

Offline WojtaTopic starter

  • Contributor
  • Posts: 33
  • Country: cz
Re: 24V power supply generating higher voltages
« Reply #5 on: May 14, 2017, 04:41:14 pm »
Oh, the rectifier is actually OK (not in the schematic), Ive drawn this one just as representation as I have a bridge rect as all in one solution on the board so no problem there. Still sorry for missinformation.
The BJT shouldnt be interfering with anything so Im not solving that as it works fine for me.
Flyback diode is on the load but there was no load connected when testing so that shouldnt be a problem aswell.
I dont have a smoothing cap before the regulator because the PSU is made to supply up to 10Amps so the cap would have to be gigantic to smooth the voltage. Will it be better if I add at least some capacitance? Im gonna increase the 10nF caps.

So I guess the PTP voltage of the AC is doing it. Could 24V zener save me? Or higher, 30V or so.
 

Offline retrolefty

  • Super Contributor
  • ***
  • Posts: 1648
  • Country: us
  • measurement changes behavior
Re: 24V power supply generating higher voltages
« Reply #6 on: May 14, 2017, 05:04:20 pm »
Yes, even 1,000 mfd main filter cap would help.
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5140
  • Country: ro
  • .
Re: 24V power supply generating higher voltages
« Reply #7 on: May 14, 2017, 05:28:52 pm »
A simplified formula for capacitance :

C =  Current  /  [ 2 x AC Frequency x ( Vpeak - Vmin) ] 

So for 10A and 60hz mains frequency and let's say you want minimum 14v DC at any time ...  C = 10 / [ 2 x 60 x (34-14)] = 10 / 1200 = 0.0083333 farads or ~ 8300uF ... so a 10000 uF would be enough to keep the voltage above 12v all the time.

The rectified voltage will have a peak voltage of 1.414 x 24v AC minus 2 times the forward voltage of a diode in the rectifier  = ~ 32v DC and the DC current would be around 0.62 x Iac if you have the basic bridge rectifier  ... so let's say if your transformer is rated for 300VA at 24v (12.5A) , it would be able to output around 7.5A dc current

Also keep in mind that linear regulators work by dissipating the difference between input and output voltage as heat, and without any heatsink a regulator like 7812 would probably only be capable of 1 .. 1.5w of heat  ...  (Vin - Vout)  x current = power dissipated ...   so ( 34v - 12v ) x current = 1w  .. so max current = 1w / 22 = ~ 0.05 A or around 50mA

If you need more current than that, you need to put a heatsink on the regulator and even then you're limited to a maximum of 15w for most linear regulators ... but most heatsinks you'd actually find easily would only be able to dissipate about 5-6 watts of heat.

Some regulators also have some maximum input voltage or can handle only a maximum difference between vin and vout ... you may find out the regulator can't accept up to 34v .. in which case you could use a regulator to output an intermediary voltage like 24v and then feed the 12v regulator with 24v. This will also spread the heat over a larger surface.

« Last Edit: May 14, 2017, 05:36:22 pm by mariush »
 

Offline WojtaTopic starter

  • Contributor
  • Posts: 33
  • Country: cz
Re: 24V power supply generating higher voltages
« Reply #8 on: May 14, 2017, 05:51:32 pm »
A simplified formula for capacitance :

C =  Current  /  [ 2 x AC Frequency x ( Vpeak - Vmin) ] 

So for 10A and 60hz mains frequency and let's say you want minimum 14v DC at any time ...  C = 10 / [ 2 x 60 x (34-14)] = 10 / 1200 = 0.0083333 farads or ~ 8300uF ... so a 10000 uF would be enough to keep the voltage above 12v all the time.

The rectified voltage will have a peak voltage of 1.414 x 24v AC minus 2 times the forward voltage of a diode in the rectifier  = ~ 32v DC and the DC current would be around 0.62 x Iac if you have the basic bridge rectifier  ... so let's say if your transformer is rated for 300VA at 24v (12.5A) , it would be able to output around 7.5A dc current


Oh, taht raises a few questions.
If my load doesnt require clean supply, do I really need the cap? The voltage can fluctuate from 24V to 0V and the load, being a lamp for example, shouldnt be bothered right? So the only need for stable voltage is the 12V line. When I have the cap after the regulator, it should keep the 12V steady even though the regulator isnt outputing 12V continuously. So does the voltage fluctuation possibly hurt the regulator itself or why is the cap adviced? I have no problem with adding a cap but Im curious to know. Maybe Im asking pretty stupid questions, stop me if I do please...

I didnt know about the current difference between AC and rectified DC, thats good to keep in mind, thanks. So my transformer marked "24VAC = 10A" should be able to output 6.2A max at 24VDC (I guess less in reality). Still it should be capable enough, Ive powered the load with it with no problems, so maybe its taking less than 8-10A as marked omn the label, I didnt meassure that.


Could I say than that the problem seems to be the peek to peek voltage of the rectified AC? 32V is really close to the 35V maximum of the regulator so thats the only thing that could cause problems, right?Can a simple zener (30V or so) across the rectified 24V get rid of the spikes or do I need a better solution?


Thanks guys, I really appreciate your help.


Quote
If you need more current than that, you need to put a heatsink on the regulator and even then you're limited to a maximum of 15w for most linear regulators ... but most heatsinks you'd actually find easily would only be able to dissipate about 5-6 watts of heat.

Some regulators also have some maximum input voltage or can handle only a maximum difference between vin and vout ... you may find out the regulator can't accept up to 34v .. in which case you could use a regulator to output an intermediary voltage like 24v and then feed the 12v regulator with 24v. This will also spread the heat over a larger surface.

I need a max of 100mA and the regulator is stuck to a large piece of aluminium so there should be enough surface area to dissapate that heat. Specified max input voltage on this regulator is 35V as said before, I couldnt find anything about max difference in voltage in the datasheet.
« Last Edit: May 14, 2017, 05:56:33 pm by Wojta »
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 19962
  • Country: gb
  • 0999
Re: 24V power supply generating higher voltages
« Reply #9 on: May 14, 2017, 06:21:41 pm »
Could I say than that the problem seems to be the peek to peek voltage of the rectified AC? 32V is really close to the 35V maximum of the regulator so thats the only thing that could cause problems, right?Can a simple zener (30V or so) across the rectified 24V get rid of the spikes or do I need a better solution?
If you do that, the zener diode will probably burn up.

It would be better to put the zener in series with the LM7812. Alternatively you could use a potential divider followed by an emitter follower, to reduce the voltage to a safe level before the LM7812.

« Last Edit: May 14, 2017, 06:24:20 pm by Hero999 »
 

Offline WojtaTopic starter

  • Contributor
  • Posts: 33
  • Country: cz
Re: 24V power supply generating higher voltages
« Reply #10 on: May 14, 2017, 06:37:53 pm »
Thanks, the  potential divider followed by an emitter follower is the correct circuit for the job i guess, but I would love to get away with the zener method. I have my PCB aleready made and merging that circuit into it wouldnt be easy. Could I get away with couple 24V 5W zeners in parallel on the rails? I guess not really if the Imax(zener) = 5W/24V = 0.2A but its peek current, it should be on about 3/10 of the rectified AC period if my calculations are correct,  so that might work.
How is the Zener in series with the LM7812 meant? I cant think of it working. There is also the solution of putting a couple of diodes in series with the regulator and get the voltage down by dropping the 0.7V on each of them but thats pretty awful.
« Last Edit: May 14, 2017, 06:51:06 pm by Wojta »
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 19962
  • Country: gb
  • 0999
Re: 24V power supply generating higher voltages
« Reply #11 on: May 14, 2017, 07:14:30 pm »
Thanks, the  potential divider followed by an emitter follower is the correct circuit for the job i guess, but I would love to get away with the zener method. I have my PCB aleready made and merging that circuit into it wouldnt be easy. Could I get away with couple 24V 5W zeners in parallel on the rails? I guess not really if the Imax(zener) = 5W/24V = 0.2A but its peek current, it should be on about 3/10 of the rectified AC period if my calculations are correct,  so that might work.
The trouble is, there's nothing to limit the zener's current. If you go with that idea, add a 100R current limiting resistor with the zener.

Quote
How is the Zener in series with the LM7812 meant? I cant think of it working. There is also the solution of putting a couple of diodes in series with the regulator and get the voltage down by dropping the 0.7V on each of them but thats pretty awful.
A zener in series will drop the voltage before the regulator. Don't use a 24V zener, that won't leave enough headroom, go for 10V instead. If the input voltage is 40V, the zener will drop around 10V, leaving 30V for the LM7812.
 

Offline WojtaTopic starter

  • Contributor
  • Posts: 33
  • Country: cz
Re: 24V power supply generating higher voltages
« Reply #12 on: May 14, 2017, 07:29:31 pm »
Quote
A zener in series will drop the voltage before the regulator. Don't use a 24V zener, that won't leave enough headroom, go for 10V instead. If the input voltage is 40V, the zener will drop around 10V, leaving 30V for the LM7812.
Oh like that, I didnt even realize that, thats cool. That will be the best solution for me I hope.
Thank you very much guys, Ill try it and let you know how it went, hopefully itll be OK after that.
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5140
  • Country: ro
  • .
Re: 24V power supply generating higher voltages
« Reply #13 on: May 14, 2017, 07:39:30 pm »

Oh, taht raises a few questions.
If my load doesnt require clean supply, do I really need the cap? The voltage can fluctuate from 24V to 0V and the load, being a lamp for example, shouldnt be bothered right? So the only need for stable voltage is the 12V line.

Correct. If you have an incandescent light or a bulb designed for 24v AC you don't need to use a bridge rectifier to power that thing. However, if it's 24v DC light, then you should not power it directly from the bridge rectifier, because the voltage will vary about 120 times a second between 0v and a peak voltage of around  32-34v ... so your 24v DC bulb could potentially die due to over-voltage.

If you have a 24v DC bulb or led lamp, you would probably want to use a 15v AC or 18v AC transformer, which will give you around 22-25v peak DC voltage after you rectify it using a bridge rectifier, and you may power that lamp without any capacitors and it would work, but you would probably see the flicker of the lamp about 120 times a second.

Capacitors after a bridge rectifier raise the minimum voltage to some value. The minimum voltage will depend on the capacitance and the current used - if more current is used, the capacitors will discharge faster and have less time to charge up so the minimum voltage will be lower. The more capacitance you add, the higher the minimum voltage would be but up to a point..

Linear regulators like 7812 need to have an input voltage slightly higher than the output voltage in order to output always 12v. That amount is called dropout voltage in the datasheet and for the 7812 it's equal to around 1.. 1.5v (depending on current and chip temperature and other things).  So basically, it would be a good idea to always have at least 13v at the input of that regulator.

The linear regulator will output a lower than 12v if the input voltage is smaller than the required one (around 13v) .. for example with 10v on the input , it will output 9v and so on.. and if the input voltage is too low, it would probably stop working altogether. For example, with just 2v input, the regulator may not output anything.

So If you don't use any capacitors between the bridge rectifier and the regulator,  100-120 times a second the voltage will slowly go up from 0v to the peak voltage of around 34v and then go down again towards 0v.  Between that period where the transformer outputs less than around 2-3v, the regulator will be completely off  and then it will gradually output higher and higher voltage until it gets more than around 13v and from that point it keeps the output to 12v and then when the voltage goes below 13v it again starts to output less voltage.  So depending on how big the output capacitor after the linear regulator is, you may be able to keep the voltage very close to 12v.

But probably the simplest solution would be to add a diode so that input capacitor won't be discharged into the lamp before it, then add a capacitor big enough to always have 13v or more at 100mA  (so about 0.1/ 120* ~ 20v = 0.1/2400 = ~ 416uF .. so a cheap 470uF capacitor would probably be enough ... and then for protection you could probably add a 15v / 24v zener diode so that the regulator won't see more than that voltage at any point.


Quote

I didnt know about the current difference between AC and rectified DC, thats good to keep in mind, thanks. So my transformer marked "24VAC = 10A" should be able to output 6.2A max at 24VDC (I guess less in reality). Still it should be capable enough, Ive powered the load with it with no problems, so maybe its taking less than 8-10A as marked omn the label, I didnt meassure that.

No, it should output about 6A at a voltage between 0v and 34v DC maximum ... the bridge rectifier just "flips" the periods where the voltage goes below 0v and makes them periods going to +v .. so instead of having a period where transformer goes from 0v to peak 34v and then back to 0v and then a period where transformer goes from 0v down to -34v and back to 0, you now have two of the first periods, both from 0 to a peak dc and back down to 0.  ... well not quite 34v, because you'll lose some voltage equal to the voltage drop on two of the diodes in the rectifier bridge.

Like I said, capacitors raise the minimum voltage, that's all. Linear regulators then make sure you get a fixed output voltage and the excess between the output voltage and peak dc input is dissipated in the form of heat.

Quote
Could I say than that the problem seems to be the peek to peek voltage of the rectified AC? 32V is really close to the 35V maximum of the regulator so thats the only thing that could cause problems, right?Can a simple zener (30V or so) across the rectified 24V get rid of the spikes or do I need a better solution?

Yes, you could say that. Remember that the output of your transformer won't always be exactly 24v AC .. your mains input may vary between let's say 210v and 240v while the transformer may be designed for 220v or 230v , so the output will also vary proportionally with the input voltage. The voltage of the secondary side of the transformer will also vary due to other reasons. Expect to have up to 38-40v in some extreme cases, and as little as around 28-30v ...

Quote
I need a max of 100mA and the regulator is stuck to a large piece of aluminium so there should be enough surface area to dissapate that heat. Specified max input voltage on this regulator is 35V as said before, I couldnt find anything about max difference in voltage in the datasheet.
Yes, for 100mA it should be easy to do ... just use a capacitor at the input enough for around 110mA (i would say go with 470uF 50v rated cap or a 35v rating if you put the zener diode in front), and some capacitor after the linear regulator, maybe 47uF to 100uF and 16v or 25v rating
 

Offline WojtaTopic starter

  • Contributor
  • Posts: 33
  • Country: cz
Re: 24V power supply generating higher voltages
« Reply #14 on: May 18, 2017, 05:18:41 pm »
Thank you, its much clearer now. I installed the zeners (2x 5V 5W) so the voltage going to the regulator is now +-15V but its still not working as it should. If I apply DC from a few 9V batteries to the circuit, lets say 28V, I get a clean 11.3V on the output (-0.7V is caused by the diode in series with the regulators output) but when I apply my non smoothed half wave, the output voltage of the regulator is the same as voltage after the zeners. So with roughly 15V at the input of the regulator I get almost the same at the output. I tried capacitor values mentioned in the datasheet but that still didnt fix the problem. I guess Ill have to go for the big cap on the 24V line next and see what that does. Or maybe its just that cheep regulator, Ill also try a better one, maybe that will work, who knows.

Have a nice day guys.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf