Page 1 of 1
Matching circuit MOSFET
Posted: Thu Apr 23, 2020 2:28 pm
by ronald001
I know that mosfets have a giving input and output resistance - and that it has to be matched to 50R - for example with coax or striplines - but how do you calculate these coaxes? ( resistance and lenght )
Is there a relatively simple calculation for it?
Also, is it possible to verify the calculated lenght with for example a NanoVNA?
I already tried to measure a output matching circuit on one of my test boards ( 2 pcs of 25R coax / 13CM each ) but did not seem to get logical results..
Re: Matching circuit MOSFET
Posted: Thu Apr 23, 2020 2:35 pm
by Bton-FM
Re: Matching circuit MOSFET
Posted: Thu Apr 23, 2020 3:56 pm
by Bton-FM
Also itβs impedance when talking about AC.
Re: Matching circuit MOSFET
Posted: Thu Apr 23, 2020 8:41 pm
by ronald001
Bton-FM wrote: βThu Apr 23, 2020 3:56 pm
Also itβs impedance when talking about AC.
Thanks for the link

And you are correct, made a translating error

Re: Matching circuit MOSFET
Posted: Thu Apr 23, 2020 9:00 pm
by Bton-FM
ronald001 wrote: βThu Apr 23, 2020 8:41 pm
Bton-FM wrote: βThu Apr 23, 2020 3:56 pm
Also itβs impedance when talking about AC.
Thanks for the link

And you are correct, made a translating error
That's what I thought It was because i remembered you are dutch

Re: Matching circuit MOSFET
Posted: Sat Apr 25, 2020 1:57 pm
by nrgkits.nz
I use SimSmith do all the calculations for me, rather than messing around with complex mathematical equations involving reactive impedance.
You can get SimSimth here
https://www.w0qe.com/SimSmith.html
It's very useful for simulating matching networks involving lumped elements and coax of any impedance etc... and its also free software.
Calculating the coax lengths and impedance for push pull amplifiers is very easy with SimSmith. You first need to draw out the circuit which I've already done - see the attachment.
Next you need to know the output impedance of the dual LDMOS device you're using (as seen from drain to drain - balanced configuration) - you can refer to the datasheet for this, or if its not there you can calculate it. To calculate it, take one half of the dual LDMOS device and use the formula which Albert has posted on the forums previously. Output impedance = VCC * VCC / 2 * Power out. Also since we are using one half of the LDMOS to do the calculation, we also use half the power. So for a 300W push pull amplifier operating with 50v, the equation is 50*50 / 2 * (300/2) = 8.33 ohms output impedance for one half of the dual LDMOS. Now since we are operating it in balanced push pull, the output impedance will double = 2 * 8.33 = 16.66. Its the same as connecting two resistors in series, the resistance doubles. This is not going to be exact but its close enough.
Now back to SimSmith, enter in the impedance on the left hand side (see my example), and then use the scroller on the mouse to change the length and impedance of the coax up and down while looking at the SWR, you're aiming for an SWR of 1.5 or less. Lengths are also in feet. The coax impedance can be any impedance, however for obvious reasons you'll need to stick to known values eg 25/35/50 etc...
300w.PNG
pish pull circuit.PNG
Re: Matching circuit MOSFET
Posted: Sat Apr 25, 2020 7:01 pm
by Bton-FM
How do you calculate the input impedance on a device?
Re: Matching circuit MOSFET
Posted: Sun Apr 26, 2020 2:14 am
by thewisepranker
Nice one NRG, I was also going to suggest SimSmith. I think the best bit about it is that you can use the Smith chart, though. I really like the Smith chart because you can see very quickly where you are and the paths required to get to where you want to be, and ultimately the correct element to apply. For anyone unfamiliar with the Smith chart, it can appear quite daunting at first but with a small amount of practice it's very useful.
Re: Matching circuit MOSFET
Posted: Sun Apr 26, 2020 2:30 am
by thewisepranker
Bton-FM wrote: βSat Apr 25, 2020 7:01 pm
How do you calculate the input impedance on a device?
You don't calculate it, you either assume it or you measure it.
You either assume and interpolate the data from an
always incomplete and very granular datasheet, or you get a few samples and measure it for yourself if you've got the luxury of the equipment required to do it.
Some of the better datasheets like those from Philips have hyperlinks in them to more comprehensive data sets of the S parameters, but lots of the links are broken now.
Re: Matching circuit MOSFET
Posted: Tue Apr 28, 2020 10:21 am
by ronald001
nrgkits.nz wrote: βSat Apr 25, 2020 1:57 pm
I use SimSmith do all the calculations for me, rather than messing around with complex mathematical equations involving reactive impedance.
You can get SimSimth here
https://www.w0qe.com/SimSmith.html
It's very useful for simulating matching networks involving lumped elements and coax of any impedance etc... and its also free software.
Calculating the coax lengths and impedance for push pull amplifiers is very easy with SimSmith. You first need to draw out the circuit which I've already done - see the attachment.
Next you need to know the output impedance of the dual LDMOS device you're using (as seen from drain to drain - balanced configuration) - you can refer to the datasheet for this, or if its not there you can calculate it. To calculate it, take one half of the dual LDMOS device and use the formula which Albert has posted on the forums previously. Output impedance = VCC * VCC / 2 * Power out. Also since we are using one half of the LDMOS to do the calculation, we also use half the power. So for a 300W push pull amplifier operating with 50v, the equation is 50*50 / 2 * (300/2) = 8.33 ohms output impedance for one half of the dual LDMOS. Now since we are operating it in balanced push pull, the output impedance will double = 2 * 8.33 = 16.66. Its the same as connecting two resistors in series, the resistance doubles. This is not going to be exact but its close enough.
Now back to SimSmith, enter in the impedance on the left hand side (see my example), and then use the scroller on the mouse to change the length and impedance of the coax up and down while looking at the SWR, you're aiming for an SWR of 1.5 or less. Lengths are also in feet. The coax impedance can be any impedance, however for obvious reasons you'll need to stick to known values eg 25/35/50 etc...
300w.PNG
pish pull circuit.PNG
Very detailled explanation Mr.NRG - thanks!
Will try to follow the steps above and see if i can calculate it aswell with RFSIM

Re: Matching circuit MOSFET
Posted: Wed Apr 29, 2020 5:50 pm
by Bton-FM
thewisepranker wrote: βSun Apr 26, 2020 2:30 am
Bton-FM wrote: βSat Apr 25, 2020 7:01 pm
How do you calculate the input impedance on a device?
You don't calculate it, you either assume it or you measure it.
You either assume and interpolate the data from an
always incomplete and very granular datasheet, or you get a few samples and measure it for yourself if you've got the luxury of the equipment required to do it.
Some of the better datasheets like those from Philips have hyperlinks in them to more comprehensive data sets of the S parameters, but lots of the links are broken now.
What data can I use to Interpolate the input impedance and how?