Flyback PQ20/16 and TOPSwitch-GX family

Hi,

For the PQ cores, compared to the EE/EF cores, it has shorter core length, which tends to have higher leakage inductance. But there are many existing applications that requires low profile, using this PQ core.

If you want to use the custmized the core, like PQ20/16 that you mentioned, you will need to use the PIxls software, which is part of the PI-expert software kit. It is allowed to choose the customized core. Please select the core as "custom" from the list, and input all the core parameters into the spread sheet munually, then the spreadsheet will do the calculation to design the correct transformer for you.  

Hi Terry.

Thank you for you respond especially comparison PQ and EE/EF cores (PI Expert default).

We tried to simulate our supply by custom core based on PQ20/16 (PI Expert).

Simulation worked with no errors but only with 3(!) main secondary turns (and calculated other secondary turns).

With 2 turns as is in a real transformer (and calculated secondary turns for main 2 turns that are close situation on the real transformer) program publish error "Peak flux density limit exceeded"!

Unfortunately solution with 3 turns doesn't fit in the bobbin.

So according Pi Expert it seems that we use to small core.

Best regards,

Zeljko.

Hi:

Would you please attach your PI expert design using PQ20/16? I can have a quick check for you.  

Hi.

Thank you for your support.

Our design/simulation attempts are attached (hope that zip compression won't be problem).

File Nr. 1 -  design with edited components according to real device except the transformer offered by PI Expert.

Nr. 2 - changed default to custom transformer PQ26/20 with NS=2 (failed).

Nr. 3 - custom transformer PQ26/20 with NS=3 (passed).

Nr.4 - PDF snapshots of custom transformer parameters.

If by your opinion there is any problem with our design it could be good starting point for the future development of our power supplies.

Best Regards,

Zeljko.

Hi,

Thanks for the file. I checked the design with NS=3 turns, it looks good.

Would you please let me know what is the difficulty to fit all the windings into the bobbin? The difficulty is coming from the Np, Nb, or Ns? If we can further decrease the Np by adjusting the design, is it okay to fit them into the bobbin? 

Hi,

thank your for your response and time again.

I've notice yesterday PI Expert default design use triple Insulated wire and transformer with

no safe margin => more available space on the bobbin => the received file isn't adequate for our real device with magnetic wire/margin type >1.7mm x 2 (product is certified by local authorities).

Fact is with NS=2 the bobbin is almost filled => no space for more turns/all coils (NS=3).

If we neglect type of wire/space problems on the bobbin you simulation produce smooth design with NS=3!

What is with NS=2 (real device)? Do you agree and according PI Expert this solution can't be recommended?

Problem is high core flux for NS=2 compared NS=3 (correct me please if I'm wrong)?

This moment with few assumptions it seems to me that core PQ20/16 wasn't best choice for our power supply.

Bigger core like PQ26/20 would be better (like PI Expert offers) but problem is limited space on the printed board.

The transformer is designed before 10 years without assistance of great tolls like PI Expert today so it might be things that aren't good enough.

By your opinion and experience can we state design with NS=2 is acceptable or requires redesign?

Hi,

Thanks for sharing the information, it is pretty clear now.

If the secondary NS=3 turns, and triple insulation wires is used, normally there is no requirement to add the margin to pass the safety requirement. This is something that you may consider.

The problem with NS=2 is the working flux density is pretty higher, over the satuation level the core. So the transfromer will enter into satuation especially during the fault conditions (like over load, or output short circuit) or high line startup, this may cause catastrophic failure of the power supply. This is definitely something that you need need to avoid.

Do you think Ns=3, and triple insualtion wires withouth maring could be a option for you? 

Hi.
Terry.
I'm sorry I did not respond earlier because I had a problem with logging as Boomer5999.

According to the information you gave me, I do not know whether the design of the transformer to be changed. It depends on the designers. But you clearly pointed out that the problem is too few turns of the transformer!
With this information future development must be seriously reckoned with.

As we are not experts for the transformer discussion was very valuable.
Thank you very much once more. Best regards from Croatia.
Zeljko.

Great, thanks for sharing the information.
Please let me know if you need further assistance,