Why the LDMOS AWR simulation is not so good as the datasheet?

[tony_lth]

Hi, ALL,
For one NXP LDMOS, AFT09MS015N, in its datasheet, NXP presents a test circuit for 870M, and get good gain performance.
But I simulated the test circuit in AWR, and can't positive gain, all the S21 are negative in wide band.
So I wonder where is the fault in my simulation？
Possible wrongs:
1. The bias circuit is not given by NXP and this may cause negative S21.
2. The circuit presented by NXP is wrong.
Who can help to check??
I simulated by Fig4 and Fig 5 in the datasheet, and the AWR file is attached. Since this forum can not upload AWR *.emp file, I changed the file format to *.emp.txt. When you want to open it in AWR, just change the file extension from *.emp.txt back to *.emp.
Cheers,
Tony Liu

- - - Updated - - -

It seems the PA S params is wrong, let me check.

- - - Updated - - -

I finally download the right S-param, and the issue solved. Thanks.

 

[vfone]

Yes, the model from NXP is wrong.

 

[BigBoss]

Some manufacturers constraint their models by the version of the simulator and AWR is the one of them.Because AWR used to use 32bits models in previous versions and some DLL stuffs should have been installed but there might still be some issues because the models are " compiled and encrypted " things and this creates some issues.

- - - Updated - - -

As I have said before.Another Simulation Code should have been installed first..
https://www.nxp.com/products/rf/rf-...ion-microwave-office:RF_HIGH_POWER_MODELS_AWR

 

[tony_lth]

Thanks vfone and Bigboss.
Best Regards,
Tony Liu

 

[BigBoss]

I have checked this transistor in ADS and it's got really a model problem compare to datasheet.For instance, the operating specifications have been defined @ 100mA Idq and Vp is specified as 2.2V but Idq is very low as uA range if you apply 2.2V at Gate and Gp is very low @100mA Idq ( few dB).
It shows me that there are some modeling problems in DK.If you're really designing seriously an PA with this transistor, you should warn NXP guys and request the correct model ( preferably Modelithics model).
Or, try to design with Cree because their models are pretty consistent.

 

[tony_lth]

OK, thanks, Bigboss.
I had verified NXP recommended circuits with S-params in AWR, it works.
And I will checked with Cree, and I never know Cree has LDMOS before.
Thanks for your comment.

 

[tony_lth]

Cree seems not have LDMOS.
And finally I guessed what the package S-params means, it should be combined with S2P of FET to simulation the live situation.
NXP doesn't clearly notice that. How confused.

 

[vfone]

This is not the first model from the NXP that I found is not working. And they have a bad customer service too.
Maybe next year after Qualcomm will buy them, things will change..

 

[BigBoss]

Cree seems not have LDMOS.
And finally I guessed what the package S-params means, it should be combined with S2P of FET to simulation the live situation.
NXP doesn't clearly notice that. How confused.

Have you ever looked that company ??? I think it's old NXP guys and they took over the license of some NXP/Philips stuffs.
It's worth to try..
https://www.ampleon.com/

 

[tony_lth]

Thanks, Bigboss & Vfone.
Cheers,
Tony Liu

 

[BigBoss]

I found a compatibility problem with a DK of Ampleon, I have signaled them and they have quickly responded me with new and corrected DK..
It's impressive..