[time-nuts] low phase noise, noise floor and noise figure amplifier at 400MHz

[Gerhard Hoffmann]

Am 10.01.20 um 03:01 schrieb jimlux:
> On 1/9/20 5:38 PM, Lifespeed wrote:
>> Hi Time Nuts,
>>
>>
>> I have a need for a low phase noise, noise floor and noise figure 
>> amplifier
>> at 400MHz.  I have tried some off-the-shelf 50 ohm amplifiers, the 
>> best of
>> which degrades phase noise by a couple dB.  I'm working with a signal 
>> with
>> -172dBc/Hz PN, so not much tolerance for degradation here.  The input 
>> signal
>> level is only 3dBm, so noise figure still matters as well.  I'm 
>> looking for
>> 15dB gain, 16dBm P1.  The residual phase noise would have to be 
>> better than
>> -180dBc/Hz, and I would probably operate the amplifier slightly 
>> compressed.

A friend of mine measured my SKY 75150 demo board at F= 0.35 dB last weekend

on 432 MHz for EME purposes. More careful measurements to follow.

chip    = < https://www.digikey.de/products/de?keywords=863-1559-1-nd     >

board = < https://www.digikey.de/products/de?keywords=863-1562     >

When I ever get it back  :-) , I might characterize it for added phase 
noise at 400 MHz,

but he should have his fair share of time to play with it.

Given the extra tiny SMD input inductor, an even better F must be possible.


The Mini Circuits SAV-551+ has a Fmin of  ~ 0.08 dB at 450 MHz according 
to the

data sheet IIRC. When I add 0.1 dB of losses, that is still outstanding.

Does anybody know of an ADS design kit for these parts? Spice is 
probably hopeless.

ADS and MCL do not seem to go together well.


If you want to allow compression, there will be 1/f up conversion. 1/f 
is completely

unspecified for these parts.

> If you need lower than that, I can ask around about discrete designs. 
> There's some group in Norway that is using FETs from Switzerland and 
> getting L band noise temps in the single digit Kelvin range. (without 
> cryocooling).  I don't know if it would go down to UHF and hold that 
> performance
>
Interesting. What are these FETs from Switzerland? Any pointers?

>>
>>
>> Any suggestions on topologies, transistors, white papers, etc? I'm
>> considering the NXP BFU590Q silicon bipolar transistor, which I have 
>> used in
>> a transformer feedback configuration at 100MHz with less than 
>> -180dBc/Hz PN.
>> But this topology doesn't appear practical for 400MHz due to the 
>> difficulty
>> maintaining a high collector impedance at that frequency with a 
>> transformer.

Have you tried Infineon BFQ19S ? I just have bought quite a lot of them.


>> Nor do I need to control the gain, which is one of the features of the
>> transformer-feedback topology.  I was thinking about common emitter with
>> inductive emitter degeneration.  Not sure cascode is right for this UHF
>> application.
>>
Cascodes do not add noise in real life. p-HEMTs have ridiculously low output

impedance, so cascoding with SiGe BJT might be a good thing. The output

could be a low-Q tuned tank if the impedance needs to be high.


regards, Gerhard