AWG calculations with PIXIs for primary/secondary windings

10 posts / 0 new


I do not understand well how PIXLs performs the automated calculations of the AWG wires in the spreadsheet.

I.e: If I have got a Isrms = 1.735A and for a current density of 9A/mm2 (219 CMA), the mínimum bare copper diammeter for the output will be AWG24 (d=0,51mm, a=0.205mm^2).

That is Ok in the spreadsheet and matches with this criteria extracted from Designing Multiple Output Flyback Power Supplies with TOPSwitch® Application Note AN-22.

So, as it is not recommended to use wires thicker (less) than AWG26 it is needed to use several parallel strands. If I am not wrong with 2 parallel strands of AWG27 (d=0.36mm, a=0.102mm^2) would be enough for that.

But, Why is the PIXIs program reporting an error after filling in the boxes with this information?. The report suggests to use thicker (less) AWG wire or filar.

It can be solved with 6-filar AWG24, but I cannot understand well the reason.

Could you please help me to clarify this? I am interested in understanding well this because it is the key for customizing the transformer designs.
Otherwise I could not develop my design properly.



Thank you for considering Power Integrations in your design. To have a better understanding, could you help provide the following details:

1) What IC do you use?
2) What topology/circuit?
3) What is the input voltage range?
4) What is the Output Voltage and Current?
5) What is the transformer core selected when the problem happens?


Hello Bombadil,

First off, the explanation I showed in the previous email was extrated from Designing Multiple Output Flyback Power Supplies with TOPSwitch® Application Note AN-22.

So, if that criteria is according to an application note from Power Technologies, I asume that this criteria is valid and followed in other designs or spreadsheets as well. In fact, it is quite logical, is not it?

Regarding to your details, I am providing you an spreadsheet design performed with PIXIs and you can check what I am talking about.

For instance, if you check the design you will see the 5V/1A output. Here, this ouptut requires an AWG24 (0.205mm diameter) to be winded in the coil. As wires less (thicker) than AWG26 are not recommended due to skin effect and following the criteria mentioned before in AN-22 note, 2 parallel strands or bifilar of AWG27 (0.102mm diameter) would be enough for satisfying the CMA mils per ampere needed.

But if you filled the boxes with that specifications in the 5V/1A secondary, the PIXIs suggests to use thicker (less) AWG wire or filar. HOW CAN THIS BE?. For fixing the warning reported, it can only be removed by using 6-filar of AWG24 which is quite enough and over-dimmensioned wire, I think.

The same occurs with the other output in the design. So, what is the criteria for winding the secondary wires used by the PIXIs?? Could you help me please? My design depends directly from this spreadsheet and I need technical support, please.



Hi emilzatopec,

Unfortunately, I cannot open the file that you attached. Did you use the stand alone PI Xls Designer? Or have you used the PI Expert Online (

I believe I understand your concern but I need to duplicate it on my side to pinpoint specifically where the problem might be. If in case you're using the stand-alone version, you may need to regularly update it as the calculations for all ICs and Topologies are continuously improved as bugs are found and reported. Fortunately, if you use the Online Tool, the version of the calculator there is updated and we can see if the issue you've reported will still be present there.


Hello Bombadil,

I am a bit lost right now because I do not understand anything at all.
Let me explain you all:

The file was generated witn PIXls Designer 10, which I updated from a previous version. The file attached can be openned with the latest version (PIXls Designer 10, as you suggested) and with previous versions and there were no reported errors when it was compiled.

But as if that is not enough, you can also open the file with PI Expert Online, and you will see again that there is produced the same result: NO ERRORS when AUDIT REPORT is ticked. IS THAT SOFTWARE OK?

I didn't either see that there were any problem with the previous versions or bug reported with this problem in the log changes that is described in "".

Anyway, is there any possibility we can improve the communication? I mean, Could I write you directly by private mail? It would be helpful a more fluent feedback to me. If you need my private email I will give you immediately.

I really need technical support because I have got the evidence of two transformer manufacturers that the TRANSFORMER as described in the attached document CANNOT be wired because it does not fit.

I attach you the evidences that proves the initial file attached can be openned as detailed before and I also attach a pdf file exported so that you can type it and review in PIXI software.

Looking forward to hearing from you.
Thank you all in advance.


Hi emilzatopec,
Thank you very much for providing your feedback. We spent some time investigating your findings and are currently in process of providing a fix. We will provide you an update shortly. Thank you very much for your patience.

Hi emilzatopec,
Please upload your PIXls design (provided to us recently) to
Once uploaded, go to Winding Construction -> Secondaries and de-select the user-input button next to the Wire Size input for both outputs. This will force PIXls to re-compute the wire size. The correct wire size should now be selected. Please take a look. Thank you.

Hello Kenobi,

I have checked what you said in your email and the secondary wires size for both outputs changed once you de-select the user-input button. So, it seems that it is been updated according to AWG conductor size table, properly.

I agree that for 5V/1A, bifilar (x2) wire size AWG25 is far enough to get the CMA mils per ampere needed. And also for the AWG26 wire size selected for the 16V/0.7A.

I have also filled in the boxes with other combinations like quadfilar (x4) wire size AWG27 for 5V/1A and bifilar (x2) wire size AWG28 and seems to be ok, too. There were no Reported errors after auditing the design.

On the other hand, I see that for the primary winding the AWG it's still the same (AWG26) and I think it is very thick and lead to over heating due to the copper losses. I mean, the AWG selected for the spreadsheet in cell is AWG29 and then it is finally changed to AWG26.

I guess that it is performed to fill in the whole window but I think that this wire size has not been optimized as the secondary has done, hasn't it?. As per 9A/mm2 criteria described above, AWG 29 would be enough for the primary winding. Please, confirm me whether this updating has been performed for the primary winding or not.

My final question is quite obvious after all: Can you confirm me that the wires fits in the transformer after proceeding as you detailed before? Is that checked by the software now?.

Thank you all for the effort done for fixing the problem.
Looking forward to hearing from you.

P.D.:It would be helpful that PIXIs designer 10 were updated and aligned with PI Expert Online in order to design in the same way.


Hi emilzatopec,
In case the user prefers a split primary construction, the tool will create an even split of the primary winding and find a wire suitable for integer number of layers (Eg.: 3 layers will be split at 2-2 at the expense of wire size such that each layer is fully filled). De-selecting the split primary flag (Primary and Bias -> Split Primary), will fill the primary winding in 3 layers. This is why the split construction selects AWG #26 and the non-split construction selects AWG #29. Fit factor is currently only available in PI Expert and not PI Xls for LinkSwitch-HP. Thank you very much for your valuable feedback.