ULTRA WIDEBAND POWER AMPLIFIER DESIGN

[venus]

HOW TO DESIGN A 30W 20MHZ TO 1GHZ CLASS AB AMPLIFIER?
WHICH DEVICES ARE USEFULL? BIPOLAR OR MOSFET? ANY METHOD OR SOFTWARE?

 

[flatulent]

With this wide frequency range you will probably need to use several amplifiers and diplexers on the input and output. The high end amplifier will probably be 300 MHz to 1000 MHz. The low end amplifier will cover the remainder of the range.

MOSFET devices are somewhat more convenient because their input impedances are higher and the input matching network is easier to make because of the higher ratio of impedance transformation.


Commercial amplifiers for this wide range end up with 10% efficiency.

 

[nandopg]

You can use a LDMOS to get a wideband amplifier from 30 to 600MHz. Take a look on the Polyfet web site ( http://www.polyfet.com/ ). They also have there a nice app note regarding broadband matching for LDMOS transistors.

NandoPG

 

[cross coupling]

This is a challenge work. The PAE will be low. Design the matching is the core.

 

[freddy_microwave]

Interesting company, although the webpage appears to be 3rd class
interesting products..anyone knows something about the prices compared
to motorola? They seem to cut cost where they can, so do they give it
to the coustomers?

You can use a LDMOS to get a wideband amplifier from 30 to 600MHz. Take a look on the Polyfet web site ( http://www.polyfet.com/ ). They also have there a nice app note regarding broadband matching for LDMOS transistors.

NandoPG

 

[nandopg]

Dear freddy_microwave
You are right, the polyfet web site looks terrible. As to their products at least the transistor I've been using are premium. The prices are better than Motorola and much better than Philips for a compatible part number. It's worth to try.

NandoPG

 

[venus]

PAE IS NOT THE CASE IN SUCH DESIGN. I HAVE SEEN POLYFET WEBSITE.IT IS OFFERING BETTER SOLUTIONS. BUT WHAT I AM LOOKING FOR IS UWB PA, ESPECIALLY IMPEDANCE-MATCHING STUFF. MOSST PROBLEMATIC IS OCTAVE ROLL-OFF COMPENSATION FOR FLAT RESPONSE SAY 3DB THROUGH RANGE. ANYBODY WITH AGILENT ADS EXPERIENCE?

VENUS

 

[webdog]

I know that ENI (now out of the market) have made amplifiers covering <1-1000 MHz with an output of 3-10 watt using bipolar transistors.

It seems the choice between FET or bipolar is not the most difficult but to have proper matching, finding ferrite material that will work....

The ENI seems to rely on very strong feed-back and low inter-stage amplification.

Amplifier Reserach is still in business and they offer a few broad-band amps too, but expensive.

I might find the schematic on one ENI-amp if you are interested??

------->>>I will post schematics over 1-1000 MHz 3Watt amp asap... just wait a few days :!:

Good Luck
Cheers
/WD

 

[MSH]

Have Anyone a practical sample?(Schematic , project,...)

 

[webdog]

Sorry.... it took a while longer than I first stated
But here is a service manual and schematic over an ultra wideband amplifier!

BTW ENI is not longer in the market of RF amplifiers.

Have fun!
/WD

 

[vfone]

Amplifier Research is the best company for this kind of application.
Was happened to me, to ask for a schematic and they sent to me.
You can try your luck for this one:
http://www.ar-amps.com/post/30W1000B.pdf
regards

 

[deadsoul]

broken link....

 

[webdog]

broken link....

I just checked, my link to dl and Vfone's link works!
Must have been a systems hick-up

/WD

 

[biff44]

Amplifier??? I am not sure the concept applies.

I would go make the equivalent of a CMOS buffer gate, although using a 200 volt power supply and some 500 pS rise time FETs.

 

[ttse7]

a nice document, let have a look

 

[microwavehongjie]

Hi,
i am also beginning to design a 1W 1G to 10G power amplifier, anyone here can give me some suggestions about how to design it? and which device or chip should be used in my design,and how to implement the wideband matching?
thanks!
i am vexatious now!
i am looking forward to your help!
my e-mail:microwavehongjie(at)sohu.com, and i hope we can make friends,thanks!

 

[sun2002]

a nice document, let have a look ,thank you

 

[biff44]

Hmmm, designing an UWB power amp. I would think that a feedback type design would be bad, as the feedback might have some time delay element that would result in distorting the pulse shape. There are DC coupled type amplifiers that are just a darlington pair, like this:

https://www.hittite.com/product_info/product_specs/amplifiers/hmc482st89.pdf

I suppose you could do the same thing with discrete transistors and a 40 volt power supply and get some serious output power.

I would also consider the trick that microwave designers had done for years--the distributed amplifier. However, at these low frequencies I would use lumped element L-C transmission lines between the devices.

 

[xiaohan]

a nice document, let have a look ,thank you

 

[ramin.ala]

you can use coaxial cable for matching in this range