解决方案搜寻 技术支持

TOP271EG/VG Load regulation with Universal Input

Posted by: aliaykan on

Hi,

I have designed universal input 85-265VAC input 24V/5A output power supply with TOP271EG.

My efficiency target is %99. I know this is not realistic. But I forced the design to use smallest transformer and winding.

My design working stable between 160-265VAC input under full load. Below the 160VAC input  Output voltage decreasing linearly

down to 85VAC (15V). With no load, output regulation is very stable between 85-265VAC.

How can I improve the load regulation according to input? Is it enough increasing secondary turns from 3 to 4?

I think core is saturating under 160VAC. BM=2390 BP=3970 with current design. Is it useful if I change the seconder with 4 Turns  BM=1792 BP=2978.

Is it correct approach and also do you have another advice?

 

Best regards,

 

 

 

 

https://www.power.com/community/forum/ac-dc-conversion-forum/2021/top271vg-start-problem-low-line-input-voltage

Files

Attachment 大小
EVAL-120-PQ3220-24V.uds 1.27 MB

评论

Submitted by PI-Wrench on 04/18/2022

I replied to the original query a while back. The desired efficiency (99%) was unrealistically high in the original design, and the input capacitor too small, possibly resulting in duty cycle/peak primary limiting at low line. Even using an Inno-GaN device with synchronous rectifier output, efficiency will be around 94-95% at low line. I ran a PIXLS spreadsheet to try and force the design to work with a TOP271, and the only option that gave the power output you seek was to increase the input bulk capacitor to 470 uF, which brings up the minimum DC voltage at the bottom of the B+ ripple trough, reducing the peak drain current at low line. I am not familiar with the core you were trying to use in the original, so I substituted a PQ32/20 core, which looks to yield a suitable transformer design that doesn't saturate. I used a a more realistic efficiency value of 86% in the PIXLS design. The zener clamp diode should be changed to a 180V rated device to provide more headroom for a reflected voltage of 130V. A PDF copy of the PIXLS design is appended here. Please let me know if you have any questions regarding the design spreadsheet. The power requirement combined with universal input voltage pushes the TOP271 to its absolute limit, so you will need to use extra measures to get the device to work at this voltage range and power level. Depending on the available heat sink, you may run into lower efficiency at low line due to RDSon losses as the TOPSwitch heats up at low line.

Attachment 大小
TOPSwitch-JX Flyback Design24V_120W.pdf 737.72 KB
Submitted by aliaykan on 04/21/2022

Hi Wrench,

Thank you very much for your valuable comments and help. I redesigned the transformer in both Pixls designer and PI Expert with %86 efficiency.

Each design results are very close except bulk capacitor. As we now 2-3uf per watt is sufficient for universal input flyback. Also PI expert accepts 240uf 

for this design. What is the key difference between two design? Any way to use smaller than 470uf capacitor? According to your advice I see 130V VOR (NP/NS ratio 5.33) is needed for full power delivery at low line input.

In addition to these do you advice another key points or check list for universal input design?

 

Best regards,

 

Attachment 大小
%86-24V-5A.pixls 55.5 KB
%86-24V-5A.uds 544 KB
Submitted by PI-Wrench on 04/22/2022

I don't necessarily trust the results from PI  Expert allowing a 240uF bulk capacitor, especially as the design is actually failing under real-world conditions. PIXLS says that you need 470uF for the design to work at all at 120W, universal input, with a TOP271. Thermal management will also be challenging, especially at low input line voltage.

Another option would be to use a PFS-based PFC input stage to provide 380VDC to the flyback DC-DC stage. This may actually allow you to use a smaller TOPSwitch for the same output power, as well as a ~100 uF bulk capacitor.

One other option to consider is to try running a PIXLS spreadsheet using 140V VOR. This change adds current stress to the secondary stage, but takes it away from the primary stage, which is where the problem is in the first place. If you use 140V VOR, you will need to use a 200V TVS in the primary snubber to avoid excess power dissipation in the primary snubber. .With this option, you will probably still need to use an oversized bulk capacitor  to get the design to work at low line. Another thing to consider would be to relax your input line voltage spec to 100-264 VAC rather than 85-264 VAC. This will also make a fairly substantial difference in low-line operation.

Submitted by PI-Wrench on 04/25/2022

OK, after some discussion with colleagues, here is a design in PIXLS that actually works with a 240uF bulk capacitor. I had to do two things in the design to force a solution - first, change VOR to 140V, and second, to forcibly change KP to 0.5, the lowest practical value. Both of these changes lower the peak primary current and make the design possible for a TOP271. I would suggest that you run the spreadsheet again with the parameters for your preferred transformer core to see if a design solution is possible with that core. I suspect that more turns will be required to keep the transformer out of saturation. This may not be optimal for copper loss at low line.

Just for speculation, I ran a PI Expert with your design parameters and came up with several solutions using EER35 and ETD39 transformer cores - the first one I opened up  used a 330uF bulk capacitor...

 

A PDF file of the design I ran with PIXLS is appended.

Submitted by PI-Wrench on 04/25/2022

Here is a design using a 330uF bulk capacitor - this design will be less thermally challenged at low line, as the minimum voltage at low line will be higher. 

Submitted by aliaykan on 07/31/2022

Hi Wrench,

 

  1. I redesigned the power supply according to your advice. I tried both design TOPSwitch-JX Flyback_132k_KP0R5_VOR140_120W.pdf and TOPSwitchJXFlyback_132k_KP0R5_VOR140_120W_330u.pdf
  2. With both design results and problems are nearly same. Under %100 load (resistive) start-up begins above 180VAC.

Output regulation is very tight nearly %1 up to 265VAC. Below 180VAC output regulation drastically loosing.

  1. With %50 load, output regulation and start-up ok between 90VAC-265VAC.

 

Some experiments I have done (focused on TOPSwitchJXFlyback_132k_KP0R5_VOR140_120W_330u.pdf);

 

  1. Bulk capacitor: 240,300,330,360 and 440uf
  2. With/Without Soft finish circuit
  3. W/Wo phase boost network
  4. Replacing TOP271EG with TOP262EG w/wo different current limiting resistor.
  5. Vds voltage appr. 600V incl. spike.
  6. Bias voltage is 13-14V with no load, 32V under full load. Control pin voltage is 5.7V  

Best regards,

Attachment 大小
Schematic 96.73 KB
Submitted by PI-Wrench on 08/08/2022

If the supply is not starting below 180VDC, it is either running out of regulation or is shutting down due to excessive peak primary current. I would suggest monitoring the bulk capacitor voltage to make sure that it is staying above the value needed for operation. To monitor the primary peak current, you will need to cut the drain trace and insert a current loop, for use with a current probe. Excessive peak current could occur either due to improper value of primary inductance, or due to transformer saturation.