voltage output slowly decreasing by .01 why?

[Eldon Jake Enerio]

guys i need help.. my DIY DC linear power supply's maximum voltage output slowly decreases by. 01 every minute. i do not know what the real cause is my hunch is in the capacitors and i tried changing a much higher capacitance value. but it still didnt work.

im using an input of 12volts and i just made a full wave voltage doubler so i would get a maximum output voltage of 24volts. the circuit diagram is down below

 

[Harald Kapp]

the circuit diagram is down below

Sorry, it's not showing. Please use the upload button on the bottom right to upload as an attachment. If you want, you can then copy the attachment's URL (hover with mouse over the attachment) and insert the image using the "image" button in the menu bar of the edit window,right next to the smiley.

 

[Eldon Jake Enerio]

Sorry, it's not showing. Please use the upload button on the bottom right to upload as an attachment. If you want, you can then copy the attachment's URL (hover with mouse over the attachment) and insert the image using the "image" button in the menu bar of the edit window,right next to the smiley.

sorry about that.. i thought the picture was uploaded when i pressed the post button. XD. well i uploaded the picture now

 

[Harald Kapp]

i do not know what the real cause is my hunch is in the capacitors

Unlikely. What are your output current and your output voltage? Do you have a suitable heat sink on the regulator? I'd guess the LM317 is getting hot and shutting down to protect itself from destruction.

 

[Eldon Jake Enerio]

Unlikely. What are your output current and your output voltage? Do you have a suitable heat sink on the regulator? I'd guess the LM317 is getting hot and shutting down to protect itself from destruction.

my output voltage is 24volts... and yes i do have a heatsink on the regulator... did u examine the diagram?.... i thinm there might be something wrong with my design. :/

 

[davenn]

guys i need help.. my DIY DC linear power supply's maximum voltage output slowly decreases by. 01 every minute. i do not know what the real cause is my hunch is in the capacitors and i tried changing a much higher capacitance value. but it still didnt work.

im using an input of 12volts and i just made a full wave voltage doubler so i would get a maximum output voltage of 24volts. the circuit diagram is down below

OK lets back up a bit and get some more info from you ... I want NO LOAD voltages please ....

1) what is the DC voltage across C3 with and without the regulator in circuit ?

2) what is the DC voltage across the output ?

3) then the voltage starts dropping what does it start at ? ( presumably the answer to Q2)
unless you were initially measuring with a load

4) what does the voltage drop to, say after 10 minutes ?


Dave

 

[Eldon Jake Enerio]

OK lets back up a bit and get some more info from you ... I want NO LOAD voltages please ....

1) what is the DC voltage across C3 with and without the regulator in circuit ?

2) what is the DC voltage across the output ?

3) then the voltage starts dropping what does it start at ? ( presumably the answer to Q2)
unless you were initially measuring with a load

4) what does the voltage drop to, say after 10 minutes ?


Dave

1) what is the DC voltage across C3 with and without the regulator in circuit ?
-i think it's around 32volts cuz it's in parallel with the voltage doubler the output of the voltage doubler btw is 32volts DC

2) what is the DC voltage across the output ?
-I Tuned it to get 24volts but it's not exactly 24 it's at around 24.1 volts

3) then the voltage starts dropping what does it start at ? ( presumably the answer to Q2)
unless you were initially measuring with a load
-no im not measuring with a load on and it starts dropping at 24.1volts by. 01 per minute i think

4) what does the voltage drop to, say after 10 minutes ?
-hmm. i noticed that once it droped by around .03volt the drop rate decreases too. by about .01volt per 3minutes i think. So after 10minutes i think it would have droped around .05 volts or less....

 

[Harald Kapp]

C4 may be a bit high. Also an electrolytic type is not well suited in this position due to leackage current. Any leackage current will influence the voltage at the reference input of the regulator and thus change the output voltage until it has stabilized.

Try a 100nF ceramic or foil capacitor for C4.

 

[Eldon Jake Enerio]

C4 may be a bit high. Also an electrolytic type is not well suited in this position due to leackage current. Any leackage current will influence the voltage at the reference input of the regulator and thus change the output voltage until it has stabilized.

Try a 100nF ceramic or foil capacitor for C4.

hmm.. that does make a bit of sense cuz i noticed that when the output voltage has droped by a substantial amount the output voltage drop suddenly stops or it's drop rate slows down... maybe i'll try that.. thanks! ... are u an electronics engineer?

 

[AnalogKid]

0.01 V out of 24.1 V is 0.0415%. That's not much. The LM317 is a great part, but it is not a millivolt or microvolt-precision device. If you combine the line and load regulation and tempco, I think you'll see that you are within the specs for the part. What is it about your application that can not handle this small drift?

Also, anytime something drifts, but drifts more slowly as time increases, thermal stability is the number one suspect. It might be internal to the 317, or the peak voltage across the input caps might be changing as they warm up with ripple current, or the rectifiers Vf is changing as they warm up, or or or. Whatever, probably something or a combination of things is approaching its thermal equilibrium.

ak

 

[shrtrnd]

I'm thinking like Harald in a potential leaking electrolytic cap, but what's your load?
Is the load heating-up, or vaying in some way under voltage and affecting power out?

 

[BobK]

I am with Analog Kid. I think the part is heating up and the reference voltage inside the LM317 is being affected. It will stabilize eventually when it gets to thermal equilibrium. This is well within the expected behavior.

Bob

 

[hevans1944]

im using an input of 12volts and i just made a full wave voltage doubler

Not that it's relevant to your "problem" but D1, D2, C1 and C2 form a half-wave voltage doubler; the diodes conduct on alternate half-cycles.

Other than that, I agree with everyone else: you are seeing an inconsequential thermal drift. Live with it, or use a more sophisticated design.

 

[davenn]

but what's your load?
Is the load heating-up, or vaying in some way under voltage and affecting power out?

I am with Analog Kid. I think the part is heating up and the reference voltage inside the LM317 is being affected. It

note this is with NO LOAD

I also like the C4 value problem, Harald alluded to

 

[Eldon Jake Enerio]

0.01 V out of 24.1 V is 0.0415%. That's not much. The LM317 is a great part, but it is not a millivolt or microvolt-precision device. If you combine the line and load regulation and tempco, I think you'll see that you are within the specs for the part. What is it about your application that can not handle this small drift?

Also, anytime something drifts, but drifts more slowly as time increases, thermal stability is the number one suspect. It might be internal to the 317, or the peak voltage across the input caps might be changing as they warm up with ripple current, or the rectifiers Vf is changing as they warm up, or or or. Whatever, probably something or a combination of things is approaching its thermal equilibrium.

ak

hmmm that could be one too .. thermal equilibrium.. hmm great idea...

 

[Eldon Jake Enerio]

Not that it's relevant to your "problem" but D1, D2, C1 and C2 form a half-wave voltage doubler; the diodes conduct on alternate half-cycles.

Other than that, I agree with everyone else: you are seeing an inconsequential thermal drift. Live with it, or use a more sophisticated design.

oh yeeeah .. sorry about that.. i just realize d that... haha.. sorry... and.. yeah i think it maybe a thermal drift too... but im still gonna try what harald just suggested..