解决方案搜寻 技术支持

SMPS output fluctuating

Posted by: nishap on

Hello PI team,

We have made SMPS circuit using the reference design provided by you with all components values same as the reference design.

Please find the schematic in the attachment.

But after connecting it to mains we are getting the fluctuating output ranging from (45V to 30V) and (12V to 8V) at both 12V and 5V respectively. 

We couldn't pinpoint the exact reason for this. Request you to suggest the test conditions or process, that we could do to rectify the issue asap.

 

 

Thanks in advance.

 

Nisha Pandere

 

Files

Attachment 大小
sch_sem2.PNG 39.34 KB

评论

Submitted by PI-Wrench on 12/06/2021

The schematic attached shows separate grounds on the 5V and 12V outputs. Because of this, the circuit will not regulate properly, since the TL431 will not be able to sink current through the optocoupler to control the primary side. The top of R76 should be moved to the same output that is being sensed by voltage divider R79 and R80. This will allow the 5V output to regulate properly. However, you will still have cross-regulation issues between the two outputs because they have separate grounds, so that the usual techniques for sharing regulation between the two outputs will not work. The 12V output may require a preload (active or passive) or a separate post regulator so that the 12V output can maintain regulation under all extremes of operation. A lot of this will depend on the actual regulation requirement for the 12V output. Also, both outputs will have a large amount of high frequency noise due to the lack of  second stage filtering.

Submitted by nishap on 12/07/2021

Hello,

Thanks for your swift reply.

We were shared the working design by PI team/distributor.

There has been no changes made on our side.

Please find the reference design.

You have mentioned cross-regulation and "that the usual techniques for sharing regulation between the two outputs will not work".

What do you suggest for this??

You have also mentioned, "The 12V output may require a preload (active or passive) ".

Please do share any literature related to this.

Also we have observed ripples in our output @ few mHz.

Please find scope images for them. I request you to please suggest neutralizing techniques and the value of components.

 

Thanks in advance.

Nisha Pandere

Submitted by PI-Wrench on 12/08/2021

The big question in this case is whether you actually need a power supply with separate grounds for the 5V and 12V outputs. If this is so, then the schematic shown will result in a power supply that cannot regulate, as there is no way for the TL431 referenced to the 5V output to draw current from the 12V output in order to regulate, as the grounds for the two outputs are totally separated.

I have included a schematic that shows the proper connection of R11 to allow the supply to regulate. The top side of R11 should be connected to the 5V rail rather than the 12V. This will at least allow the 5V output to be reasonably tightly regulated. However, the 12V output will fluctuate with changes in the 5V output. This will be especially evident during the 2A peak load on the 5V. 

I have also included post filters on the outputs of the supply to filter high frequency noise.

Please try the changes shown in the marked-up schematic to allow the power supply a chance to regulate. Once you have a reasonably stable power supply, we can then discuss the options for improving cross-regulation for the 12V output.

Attachment 大小
Forum_supply.png 251.86 KB
Submitted by nishap on 12/08/2021

Hello,

The answer to the big question is YES!!!

We need a power supply with separate grounds since we are referencing one of it to 230VAC(AVSS).

And we don't want to get electrocuted.

We have tried your solution and we are getting a stable 5VDC o/p but a stable 17VDC @12v o/p without load.

What could be a reason for this happening??

Should we try to load the SMPS and check for variation in 12VDC o/p??

Submitted by PI-Wrench on 12/09/2021

The reason for this happening is that the 12V output is only regulated by virtue of its coupling to the 5V output winding - very loose regulation. At no-load, any bit of noise, overshoot or ringing in the 12V output winding waveform will cause the output to peak charge to a higher voltage. Also, since the 5V output is the sole point for output regulation, the control loop will be adjusting the primary for optimum 5V output, overcoming any losses in the 5V circuit. This operating point may not be optimal for the 12V output.

The worst case operation for your application would be a very light load on the 12V output combined with the 2A peak load on the 5V. I think it would be wise in this case to establish what the minimum load on the 12V output will be, then to check its interaction with the 5V load.

 

Submitted by nishap on 08/22/2022

Hi, we have made changes to our schematic. we are getting stable output at 5V/0.5A. but on the 12V/0.4A it is still fluctuating significantly. whenever we apply load to any of the outputs the 12V/0.4A is getting affected. please check the schematic and update us.  

Submitted by PI-Wrench on 08/23/2022

Unfortunately, since the 12V output does not share a ground with the 5V output, there is not much that can be done to improve regulation other than optimizing the value of the 12V preload resistor R84. !k is probably too high of a value for R84 to help maintain regulation, especially with the 2A pulsed load on the 5V output, which will likely cause a significant rise in voltage at the 12V output. Is the 5V output load relatively static other than the pulsed load, or does it vary significantly?  This will be a factor in the 12V regulation. What is the actual variation in the 12V output vs. 5V load? 

 

Also, what is the regulation specification for the 12V output?  If both the 5V and 12V outputs have a wide range of load conditions, you may need to add a post regulator to the 12V output, depending on the desired regulation spec.

I also noticed something on the primary side of the supply - bulk capacitors C46 and C49 will need equalizing resistors across them to overwhelm the effect of differing leakage currents between the two capacitors. If one cap has a leakage current higher than the other, the cap with lower leakage has the possibility of getting overvolted.

 

Submitted by nishap on 08/28/2022

Okay, What is the perfect value for the R84 resistor? 

Yes, the 5V output is stable irrespective of the Load on both ends. 

whenever we use a load of 10E/25W at 12V the voltage drops to 7.81V whereas the 5V remains unchanged.

on the other hand, when we use a load of 10E/25W at 5V it remains unchanged but the 12V will get raised to 15.60V.

I think it is happening because of cross-regulation. These are my observation after testing out the circuit physically. 

need your suggestions to clear out the issues from the circuit. 

 

Thanks,

Vinayak P 

Submitted by PI-Wrench on 08/29/2022

The optimum value for the preload resistor will depend on your particular setup, and needs to be determined through experiment. There are some measures that can be taken to improve cross-regulation by optimizing the coupling between the 5V and 12V output windings on the transformer. One measure that might help would be the split the 12V winding into two halves and sandwich the 5V winding in between the two halves, using multifilar (bi-or trifilar) windings on the 12V output to improve coupling. Another measure that might help would be to use an active preload circuit that will increase loading on the 12V output in response to a rise in 12V output voltage. I have included several possible preload circuits in the attached PDF file.

I must point out that in a setup like this, cross regulation will never be perfect, and the 12V output voltage will depend largely on the minimum and maximum loading on the 5V output. If you have an well-defined and substantial minimum load on the 5V output, it will be easier to optimize regulation for the 12V output. In a worst-case situation, it may be necessary to add a post-regulator on the 12V output. 

Attachment 大小
ACTIVE_LOADS.pdf 51.21 KB
Submitted by nishap on 08/30/2022

Thank you for the solution.

I have something to say.

I am applying a load of 10E/25W at 5V/500mA.  Will this resistor affects? shall I need to opt different resistive load? need your feedback on this. 

I have attached the Load sheet. have a look

 

Thanks in advance. 

Attachment 大小
different loads with different wattages 40.19 KB
Submitted by PI-Wrench on 08/30/2022

It is not entirely clear from your chart what is happening on one output vs. the other. What was the load on the 5V output when you were varying the 12V load?. Also what was the load on the 12V output when you were varying the 5V load.? 

In order to get any semblance of regulation, there need to be well-defined minimum and maximum loads for both outputs. The 12V output will require some loading on the 5V output to maintain regulation under varying load, and the 12V output will require a well-defined minimum load to maintain regulation when the 5V output is varied.

Submitted by nishap on 09/05/2022

When I apply Load at 5V i.e. 20E/25W I get 5V at the 5V output end whereas on 12V I get 14.4V, when I use Load at 12V i.e. 40E/25W I get 5V at the 5V output and on 12V side the voltage is dropped to 10.8V.

Submitted by PI-Wrench on 09/06/2022

Again, when you mention applying a given load to one output, it is not clear what the loading is on the other output. Clearly for this output configuration, a no-load situation on either output is not desirable. 

In the case of the 5V output, the controller will adjust for optimum 5V output, and the only thing controlling the 12V output will be the coupling between the 5V and 12V output windings. This means that the regulation will be far less than perfect, as the 5V output will regulate to suit itself without any input/feedback from the 12V output.

On your schematic, a 2A peak load is mentioned for the 5V output. I would suggest applying  that maximum load to the 5V output and adjusting/increasing the 12V load until it comes into regulation. This will define the minimum load for the 12V output for your setup. Conversely, with that minimum load in place for the 12V output, then adjust the 5V output load lower to define the minimum load for the 5V output that keeps the 12V output in regulation.

In a situation like this, I would not expect any better than around 10% regulation for the 12V output.

If this is not acceptable, you need to consider adding some turns to the 12V output to provide adequate voltage headroom  for a post-regulator of some sort. Even a post-regulator will not work in this situation without an adequate minimum load on the 5V output to keep the 12V output alive.