Transistors for 100 Mhz - 6 Ghz LNA

Hey there! I'm about to design a 100 Mhz - 6 Ghz LNA, but i'm still not sure which transistors to use. Any recommendations? Would be greatly appreciated. Thanks in advance!

Hello db12,

nowadays it is much easier, instead of using "bare transistors" as LNAs for such wide band like yours "100 MHz up to 6 GHz", there are MMICs available as "gain blocks" or LNAs.
What is more important, achieving a low noise figure in a LNA over such wide bandwidth is generally difficult. This is why MMIC designers developed their products, to offer you the best performance for a low cost, even without the need of a tricky noise figure adjustment.
For instance, I have used Agilent MSA2743 as a wideband "gain block" with 13 dB gain over 0.1- 6 GHz, NF <4 dB, linear up to +16 dBm.
As an excellent LNA I have used Agilent MGA 86563 for 0.5 - 3 GHz, 23 dB gain, NF <2 dB. RF Micro Devices has RF2312, usable from DC to 1.0-1.5 GHz, NF <2 dB, ~16 dB gain.
Recently, Mini-Circuits has introduced many new wideband LNAs covering up to 10 GHz, as MMICs as well as connectorized, for very reasonable prices.

Thank you very much for the good advice...it definitely helped. More suggestions welcome!

Added after 5 hours 2 minutes:

jiripolivka said:

Hello db12,

For instance, I have used Agilent MSA2743 as a wideband "gain block" with 13 dB gain over 0.1- 6 GHz, NF <4 dB, linear up to +16 dBm.
As an excellent LNA I have used Agilent MGA 86563 for 0.5 - 3 GHz, 23 dB gain, NF <2 dB. RF Micro Devices has RF2312, usable from DC to 1.0-1.5 GHz, NF <2 dB, ~16 dB gain.

Click to expand...

Would it be possible for you to post your designs? Would be of great help and tremendously appreciated. Thank you very much in advance.

I do not think too high about "my designs"; in principle, I followed the recommended designs by MMIC makers.
I have often achieved better than specified parameters, by using not the printed-circuit boards where ground vias introduce unwanted feedback. Instead I use small brass plates as carriers, only RF 50-Ohm lines are used for rf input and output.
By directly grounding the appropriate MMIC pins, I observed a lower NF as well as a higher power capability than specified.
I would recommend some creativity like that to anyone. Modern MMICs are designed extremely well, they require no adjustment other than DC bias current.

what do you think about splitting the band and taking a two amp approach to the task? maybe from 0.1-2.5 and 2.5-6 Ghz?

Yes, you can do it! Everything depend upon what you need to do!

There are MMICs selectively designed for specific frequency band (like cellular band) which are optimized just for that.
Then you can find MMIC s or ready-made connectorized amplifiers capable to operate over a wider band, also with a good NF and gain, etc.

As an engineer you should define your specifications, then search for available components.

I have used many MMICs to design my amplifiers for my needs. Recently it is becoming easier and cheaper to buy them. Those who make amplifiers in a large quantity can offer them for a lower price than the cost of your time.

Good luck!

well, ready-made amplifiers aren't an option for me...i need to come up with my own design. as regarding specifications my main concerns are:

gain: 15-20 dB
NF < 3 dB

getting close now...what do you think about these characteristics?...it's two stage...

If you use pHEMT in cascode configuration, you may obtain wideband LNA with relatively low noise and high gain.
Look at www.avago.com for discrete pHEMTs..

i obtained the aforementioned characteristics using a depletion mode pHemt (from RFMD)...NOT in cascode configuration, but two stage...all i actually wanted to know is, whether the characteristics "look" good or is there something that may be improved...i'm basically pretty satisfied with the caracteristics, except for the noise figure...i'd like to get it below 2 dB...is there any way?
...my lna is a two stage with shunt (RC) feedback on both stages...in-/output matched with LC ladders & and ladder interstage match...

How about the stability ??? Have you ever checked that ?? then What is the linearity specification of your LNA ?? IIP3 ??
When you design something, you should consider all of those things ..
When you start to implement the actual circuit, you'll see that the performance of the amplifier will look like quite different than "idealized world"..

I recommend you again that use cascode configuration for wideband, good linearity..

Maybe your LNA is not stable in some frequency.
Just Stabfact >1 is not enough, there is another para must >0, see the Stabfact help of ADS.

I ALWAYS check stability!!!....stabfact>1, delta<1, mu_load & mu_source both > 1....that should cover it, i think...i can't specify the IP3, because i only have s-params with noise to for the transistors...they don't alow a IP3 check...or do they?

perhaps you can consider splitting the band in two using Avago MGA-30889 and MGA-30789...besides having the target of NF/Gain/IRL/ORL/OIP3/Power budget, for such broad band you ll need to know more a less how flat do you need the response for example on the gain...many MMICs have negative slope on the gain for this band and you may see differences of ~5 db between 100 MHz and 6 MHz.

Cheers