hemt vs mesfet
equivalent circuit is as shown below
Hi, that means that your small signal model is false, Cds , Ri and T parameters have to be positive, your extraction is not correct.
You should try another extraction with Z11, Z22, Z21 & Z12 checking between measurement and simulation.
Aids of these Z parameters you can get an idea of parameters to move in order to match the measured data.
which type of algorithm you are using?
do you extract 15 parameters small signal model?
all positive? what confused me is that how can i know that the extracted parameters are right,even if they are all positive,is there any meaning to explain the negative parameter.
reference paper:ieee paper
1.A new method for determining the FET small-signal equivalent circuit
2.New method to measure the source and drain resistance of the GaAs MESFET
3.Improved equivalent circuit for determination of MESFET and HEMT parasitic capacitances from ?Coldfet? measurements
at last ,thank you!
yes it's my job.
All of them have to be positive, if they are negative that means the result doesn't match the physical reality.
then,which way are you using?which software?
sometimes Cpg is negative,does the Cpg influence the intrinsic parameters.
as you say,maybe the method I used is wrong,what is the right way?
plz give me some advices.
We are using personnal lab. developped softwares, I told you that all parameters have to be positive, including Cpg.
If only one of them is negative, your model extraction is not correct.
In my point of view, it's very hazardous to think that you could get a HEMT model from [S] parameter data by using ADS optimization.
In some publication you can find equations, it is based on [Y] matrix conversion.
I don't remember the papers name.
I am using the [Y] matrix conversion too.just as the papers described,we got the 3 device [S] parameters after the Cold condition and Hot condition.
1.deembed the Cpg,Cpd
2.deembed the R and L
3.then we got the intrisic [S] parameter,all the intrinsic parameters are all related to the [Y] parameter,so we can got all the 15 parameters
is there anything wrong?personally, I think the method is right,but don't understand why the result is wrong!what do you think about the method?
Hi,
Yes you are right, the intrinsics parameters are related to the [Y] parameters, but in contrast with you I never (or almost) use the cold measurement (from my experience that always failed) to extract the extrinsics, I use a personal software .
It performs an optimization process by finding extrinsics assuming that all internal parameters are not frequency dependant.
when you try to get your extrinsics with cold measurement, your device is composed of capacitances (almost) but how can you decorrelate Cgs & Cpg for example?
It's a problem.
do you have any papers or datas to share?
if don't use the cold conditions,how can i deembed the Cpg,Cpd,also the Ls,Ld,Lg,Rg,Rd,Rs,then get the intrinsic [S] parameters.
I read some ieee papers,they all used the cold and hot measments to get the extrinsic parameters,so I did the same way.
what's the method your personal software used?or the algorithm?
thank you!
Hi All,
I am writing my own parameter extraction "optimization algorithm" for extracting small signal MESFET parameters.
After extraction when i am plotting real and imaginary Y paraemetra with frequency, it fis upto 14 GHz and beyond 14GHz the fits are very poor.
Can anybody help me to identify where the problem may be. I am using standard equivalent circuit 16 element.
When i asked the MESFET fabricator, he saya this MESFET works upto 12GHz only.
If a MESFET works upto 12GHz then is it possible that after 12GHz, it will not follow the equivalent circuit?
please help me
some parameters like resistances can have a physical meaning even if they are negative....in between gate and drain there is a region which a negative resistance region(refer ladbrooke for reference)....now if the parameters before are small the resistance of channel can come negative..........