SMPS Transformer noise and Low output voltage

Thread Starter

SMPS Lover

Joined Apr 9, 2021
19
Hello, i have an issue with a 12v30A PSU: low output of 4v instead of 12v, SMPS transformer noise like low frequency issue.
No short element in the secondary, RT/CT of tl 494 within specs, but i suspect the capacitor at VCC of TL494 : it is only 10mF and has only 7.5V. The DC output voltage increases to 6v when i cut off the mains power then dies. Any idea where to look at ? Thanks.
Here is a video of it.
 

Attachments

Irving

Joined Jan 30, 2016
2,112
Warning: Rectified mains voltages of 300+v may be present. Take care and always discharge large smoothing capacitors before working on such PSU. Be very careful using 'scopes on the primary side, you can easily exceed scope input limits or apply scope ground to something that is 100v or more above or below ground.

This is quite an old SMPS chip. Typically the low freq noise is due to the inverter trying to start then dying repeatedly. Often on SMPS of this nature there will be a dropper resistor off the rectified mains to Vcc to get the inverter started and a secondary winding on the transformer with a diode & smoothing cap to provide Vcc once the inverter is partly up to speed. If this second Vcc feed isn't established the chip shuts down and tries to restart.

It could be a duff Vcc smoothing capacitor. Consider removing and testing out of circuit. Another common fault is an open circuit secondary feedback winding.
 

Thread Starter

SMPS Lover

Joined Apr 9, 2021
19
Warning: Rectified mains voltages of 300+v may be present. Take care and always discharge large smoothing capacitors before working on such PSU. Be very careful using 'scopes on the primary side, you can easily exceed scope input limits or apply scope ground to something that is 100v or more above or below ground.

This is quite an old SMPS chip. Typically the low freq noise is due to the inverter trying to start then dying repeatedly. Often on SMPS of this nature there will be a dropper resistor off the rectified mains to Vcc to get the inverter started and a secondary winding on the transformer with a diode & smoothing cap to provide Vcc once the inverter is partly up to speed. If this second Vcc feed isn't established the chip shuts down and tries to restart.

It could be a duff Vcc smoothing capacitor. Consider removing and testing out of circuit. Another common fault is an open circuit secondary feedback winding.
Thanks for your reply.
This PSU gives VCC through a small transformer taking high voltage 320Vdc from primary but could only identify 2 resistors that are good values. The TL494 capacitor is 10 Micro Farad, is this the correct value because i suspect someone changed it before. I usually seem 47Mf around PWM chips...
 

Irving

Joined Jan 30, 2016
2,112
Hello Dick, no clue about the transformer, but it only gives 4v instead of 12v. Can capacitance of VCC capacitor have an influence on the frequency of the TL494 ?
Unlikely, but if its dried out or failed in another way it wont be acting a smoothing capacitor and you wont get 12v

This PSU gives VCC through a small transformer taking high voltage 320Vdc from primary but could only identify 2 resistors that are good values. The TL494 capacitor is 10 Micro Farad, is this the correct value because i suspect someone changed it before. I usually seem 47Mf around PWM chips...
A transformer wont work on DC so either there is a small secondary power supply from the AC input, which is unlikely if the capacitor is only 10uF as thats too small for 50/60Hz or there is a small low-power secondary inverter off the HV DC input (which I've seen on other example circuits) and again the quality and effectiveness of that capacitor will be paramount to Vcc being the right voltage. Its value won't be too critical, anything from a few tens of uF upwards to a few 100 uF.

Testing resistors in-situ will rarely give correct values.

Ideally we'd like to see waveforms on certain components, but I'd warn against trying to use a 'scope on a mains-side SMPS unless you have a High Voltage differential probe.
 
Last edited:

LesJones

Joined Jan 8, 2017
3,498
Switch mode power supplies seem to need capacitors with a low ESR. So I suggest checking the ESR and value of the 10 uF capacitor and the main reservoir capacitor. I think the low frequency is the power supply continuously trying to restart. I think you will find that there is relatively high value resistor from the +300 volt rail to the VCC pin on the TL494. This initially powers the TL494 which is enough to get things started but then it is supplied via a diode from a winding on the transformer. If the main reservoir capacitor has very high ESR then there may be so much ripple on the +300 volt rail that the power supply keeps attempting to start at twice the mains frequency. (This would happen if the 300 volt rail dropped to a very low voltage at the mains zero crossing points.) If these two capacitors test OK then I think you will need to trace out the schematic for us to be able to make any useful suggestions. DO NOT try to do any testing on the primary part of the circuit with an oscilloscope unless you are using an isolating transformer to feed the power supply.

Les.
 

Thread Starter

SMPS Lover

Joined Apr 9, 2021
19
Switch mode power supplies seem to need capacitors with a low ESR. So I suggest checking the ESR and value of the 10 uF capacitor and the main reservoir capacitor. I think the low frequency is the power supply continuously trying to restart. I think you will find that there is relatively high value resistor from the +300 volt rail to the VCC pin on the TL494. This initially powers the TL494 which is enough to get things started but then it is supplied via a diode from a winding on the transformer. If the main reservoir capacitor has very high ESR then there may be so much ripple on the +300 volt rail that the power supply keeps attempting to start at twice the mains frequency. (This would happen if the 300 volt rail dropped to a very low voltage at the mains zero crossing points.) If these two capacitors test OK then I think you will need to trace out the schematic for us to be able to make any useful suggestions. DO NOT try to do any testing on the primary part of the circuit with an oscilloscope unless you are using an isolating transformer to feed the power supply.

Les.
Hello Les, i tested all capacitors with ESR meter, all good. TL494 has a VCC of 7.5V (with new capacitor) and a frequency of 20Khz. The output of the SMPS transformer (after diodes) is only 4.5V instead of 12V. Another clue i just discovered : when i cut the mains the output voltage goes up to 6V before diying. This is an old classical 12V30A PSU i believe used to power led lamps.
 

Thread Starter

SMPS Lover

Joined Apr 9, 2021
19
What are you testing? Is this a power supply that failed, a new power supply, or....?

Is it mains powered as many assume?
Yes, it is a PSU that works from mains 220V to give 12V30A. Those with the case fully in metal with wholes on the top. Thanks Dick for the help.
 

LesJones

Joined Jan 8, 2017
3,498
I think the next step is to look at the output of the transformer (Before the rectifier diode. I am assuming it is a single diode rather than a bridge) with an oscilloscope. (With the scope ground lead connected to the output negative.) Tracing out the secondary side would be a help so we can understand how the voltage control feedback works. There will probably be a zener diode or LM431 or similar that is used as a reference to compare with the actual output voltage to provide feedback to the TL494 via an opto coupler.

Les.
 

Irving

Joined Jan 30, 2016
2,112
I think the next step is to look at the output of the transformer (Before the rectifier diode. I am assuming it is a single diode rather than a bridge) with an oscilloscope. (With the scope ground lead connected to the output negative.) Tracing out the secondary side would be a help so we can understand how the voltage control feedback works. There will probably be a zener diode or LM431 or similar that is used as a reference to compare with the actual output voltage to provide feedback to the TL494 via an opto coupler.
Be very careful. DO NOT CONNECT a scope unless the PSU is powered through a mains isolation transformer, as what appears to be 'ground' may actually be + or - 120 or so volts wrt to earth on the primary side of the SMPS. Looking at example schematics using the TL494 there are many variations several of which it would be highly risky to connect a scope to without using a high-voltage differential probe.

Suggest you post some good hi-res photos of the board from above and a couple of other views so we can get some idea of the layout.

Do you have a part or model # and manufacturer for this PSU? Have you tried searching for schematics?
 
Last edited:

Thread Starter

SMPS Lover

Joined Apr 9, 2021
19
I think the next step is to look at the output of the transformer (Before the rectifier diode. I am assuming it is a single diode rather than a bridge) with an oscilloscope. (With the scope ground lead connected to the output negative.) Tracing out the secondary side would be a help so we can understand how the voltage control feedback works. There will probably be a zener diode or LM431 or similar that is used as a reference to compare with the actual output voltage to provide feedback to the TL494 via an opto coupler.

Les.
Hello Les,
Here are the pictures of the oscilloscope taken straight out of the SMPS transformer. I changed all caps to a higher value and resulted to an increase in output voltage of 1.5V additional, nothing magic.
 

Attachments

Thread Starter

SMPS Lover

Joined Apr 9, 2021
19
Be very careful. DO NOT CONNECT a scope unless the PSU is powered through a mains isolation transformer, as what appears to be 'ground' may actually be + or - 120 or so volts wrt to earth on the primary side of the SMPS. Looking at example schematics using the TL494 there are many variations several of which it would be highly risky to connect a scope to without using a high-voltage differential probe.

Suggest you post some good hi-res photos of the board from above and a couple of other views so we can get some idea of the layout.

Do you have a part or model # and manufacturer for this PSU? Have you tried searching for schematics?
Hello Irving,
Here are some pictures, i also put some light behind for better tracing and understanding. This PSU has no optocoupler nor TL431, i believe the TL494 takes direct feedback from the op-amp HA17358.
 

Attachments

Irving

Joined Jan 30, 2016
2,112
This board seems to follow some of the schematics found on the web, its very similar to the attached, at least on the primary side. So photo 1 shows along the top R > L, 110v in, fuse, input choke, bridge rectifier, 220v secondary input(?) or jumper for 110/220 operation, smoothing capacitors. Q1, Q2 are the main switching transistors, Q3,Q4 are the base/gate drivers via T2. What are Q1,Q2 part #s?

Can you do a high res photo of the back of the board as well...

s_atx01g.png
 

Thread Starter

SMPS Lover

Joined Apr 9, 2021
19
This board seems to follow some of the schematics found on the web, its very similar to the attached, at least on the primary side. So photo 1 shows along the top R > L, 110v in, fuse, input choke, bridge rectifier, 220v secondary input(?) or jumper for 110/220 operation, smoothing capacitors. Q1, Q2 are the main switching transistors, Q3,Q4 are the base/gate drivers via T2. What are Q1,Q2 part #s?

Can you do a high res photo of the back of the board as well...

View attachment 235238
Hello, here is a picture of the PCB. Still having hard time to figure this issue out... maybe the feedback circuitry with ha17358... big transistors are 2sc3320 and small ones ca2655. Thanks
 

Attachments

Top