Check the company website. Or if you have the transistors, you can measure the S-parameters yourself. Or get the transistors and use load-pull to characterize them.


Aaron

Hi,

to be honest, you are asking questions about design using discrete transistors whereas this forum specializes in integrated designs (millions of transistors, resistors, capacitors, etc on one chip). So I doubt you will find too many experts in your field here.  

That said, if you already have the transistors, you can make a board with 50-ohm traces at the inputs and outputs to the device, and use bias-tees to power the device, and measure s-parameters using a vector network analyzer. The VNAs from agilent can directly output touchstone format s2p files which you can then use in your simulation. However, you will probably need to de-embed or calibrate away the traces used in your setup. So you will have to read into this.

Bear in mind that a PA is a non-linear circuit, so ultimately, you may need to do a load-pull measurement to get the best performance out of your device.

Another thing to keep in mind is that it might be difficult to get good measurement accuracy for such a high power setup. For instance, if you simply drive the device with a small signal and measure S-parameters, you may be bypassing any self-heating effects which could affect the performance. Likewise, if you attempt a high-power measurement, you will need high-power attenuators, and some kind of linear PA to drive the device (which will also need to be de-embedded).

There are probably ways people normally do these designs at such high power, but as I said, this is not the forum where you are likely to find such expertise.


regards,
Aaron

Hi smwk,


my experience is also all on integrated circuits. For IC design we use transistor models like BSIM or PSP. I'm not sure what circuit models you are using or how reliable they are. Some device models are not designed for RF simulation, although again, I don't know if this applies in your case.

Since you are interested in very high power, I can tell you that even though our models are very complex, they still do not offer any kind of prediction of long-term reliability. For that, you could refer to the datasheet you are working with.

As for S-parameters, I doubt they are your best choice when designing a PA. It depends how non-linearly you drive your device. Your best bet might be to try and do a load-pull and source-pull measurement to find the optimum source/load for the design.


regards,
Aaron

Since that is an unrelated question, I suggest you start a new post. Anyway since you asked, it seems to be for IMD simulation, but I've never actually used ADS for PA design.


Aaron

Hi smwk,

you may find that the transistor manufacturers will be a lot more helpful/knowledgeable in this case ...

they would have done a lot of performance characterization on their transistors, including things like reliability.

they should also be able to advise on how to use their transistors in simulation or indeed which simulation tool to use.

all the best ...

that wasn't me...