LNA for DBS LNB design?

[patriot]

lnb design

hello,I want to know how much gain does the LNA have? On the market, most of the LNBs have a NF of lower than 0.7dB, which transistor or MMIC are usually used? NEC3210 series or any other?
If the discret low noise transistor is used, maybe we need 2 or 3 stages to design the LNA, so why don't some IC venders provide the LNA MMIC for DBS?

 

[patriot]

dbs lnb

hello,I want to know how much gain does the LNA have? On the market, most of the LNBs have a NF of lower than 0.7dB, which transistor or MMIC are usually used? NEC3210 series or any other?
If the discret low noise transistor is used, maybe we need 2 or 3 stages to design the LNA, so why don't some IC venders provide the LNA MMIC for DBS?

Can anyone give some suggestions?

 

[vfone]

dbs lna

http://www.hp.woodshot.com/hprfhelp/products/xrs/atf36xxx.htm
http://www.mitsubishichips.com/Global/common/cfm/eLineUp.cfm?FOLDER=/product/hf/gaastransistor

 

[patriot]

lnb transistors

http://www.hp.woodshot.com/hprfhelp/products/xrs/atf36xxx.htm
https://www.mitsubishichips.com/Global/common/cfm/eLineUp.cfm?FOLDER=/product/hf/gaastransistor

yes, your links are LNA products info., Since we must use two or three stages of the low noise transistors to construct the LNA, why don't they provide the MMIC with sufficient gain for DBS-LNB? hence it is more convenient.
BTW, do the transistors such as NE3210, have stability-enhancement matching? I find that the stability factor K is >1, however, the ATF36077 is not. In the 0.25um pHEMT Design kit, the phemt itself has K<1, so if we want to design a transistor like NE3210, do we need to add some stability-enhancement matching? such as source degeneration,series resistor or parallel resistor?

 

[vfone]

mmic lnb

I didn’t see any multi-stage low-noise MMICs for DBS-LNB in production.
Actually both NE3210 and ATF36077 are potentially unstable at some frequencies.
The recommendation is to plot on the Smith chart (during tuning of the matching networks), NF, gain and stability circles, to get the optimum for all of them. Providing the right input and output impedance should be fine, and don’t need any stability enhancement circuits.

 

[patriot]

ne3210 amplifier

I didn’t see any multi-stage low-noise MMICs for DBS-LNB in production.
Actually both NE3210 and ATF36077 are potentially unstable at some frequencies.
The recommendation is to plot on the Smith chart (during tuning of the matching networks), NF, gain and stability circles, to get the optimum for all of them. Providing the right input and output impedance should be fine, and don’t need any stability enhancement circuits.

Thanks for your reply. I don't know whether the LNB vender use some LNA MMIC or not, By far, the LNA MMIC with lowest NF which I can find in productions is tga2600 from Triquint. It has NF<0.7.
Here, I want to disscuss how to design a low noise transistor like NE3210 or ATF36077 with pHEMT technology, I simulated a pHEMT transistor with width of 50um×4, but the stability factor is less than 0.3, but how can NE3210 has a k factor of >1( may be <1 in some frequency points, nevertheless, it is larger than my simulation) Isn't there some stability enhancement circuit in NE3210?
should I do some efforts to make it approach 1 or let it alone.

 

[vfone]

low noise block design

Potential unstable doesn’t mean it will oscillate. Especially at Ku frequencies is hard to design an amplifier with a “rock solid” K factor greater than 1, all over the band.
I never used NE3210 but I used a lot the old ATF36077, which gives stable LNA’s whatever frequencies range was coverd, without needs of using stability enhancement circuits. Bias network decoupling and layout can make the difference designing an LNA like this.
Another rule that should be folowed is, never use K-factor in a multistage LNA. Use just for each stage alone.

 

[radiohead]

multi-stage lna

Another rule that should be folowed is, never use K-factor in a multistage LNA. Use just for each stage alone.

Vfone, could you please comment on this last statement? I heard different opinions on this issue, without real clarity. In fact if the first amplifier is conditionally stable, but is stable with the input impedance that the second stage offers (over all frequencies), no oscillation condition occurs. The same accounts for the input stability of the second stage. So you could say this is a sufficient condition for stability.

Why do you recommend to analyze both stages separately?

 

[patriot]

related:www.earf.co.uk/sat.pdf

Potential unstable doesn’t mean it will oscillate. Especially at Ku frequencies is hard to design an amplifier with a “rock solid” K factor greater than 1, all over the band.
I never used NE3210 but I used a lot the old ATF36077, which gives stable LNA’s whatever frequencies range was coverd, without needs of using stability enhancement circuits. Bias network decoupling and layout can make the difference designing an LNA like this.
Another rule that should be folowed is, never use K-factor in a multistage LNA. Use just for each stage alone.

Maybe I didn't make myself been understood clearly, I want to design a Low noise transistor to take the place of ATF36077 or NE3210, So i chose a pHEMT with width of 50um×4, however, the K factor is less than 0.3 ( but the ATF36077 or NE3210 has a much larger K). Do you mean i need not care about the K factor, directly add G,D,S pads, and package it?

 

[vfone]

atf36077 lna

@radiohead
When the amplifier contains more than one active device the Rollet K-factor measurement may lead to incorrect results.

@patriot
All right, you want to design a transistor and not to use one. In this case K=0.3 seems to be too low, and internally, definitely you have to do some stability enhancements.