[time-nuts] HP 11729C versus 11848A

Adrian rfnuts at arcor.de
Sat May 8 21:16:50 UTC 2010


Thanks to Rick's great hint, I'm now awaiting some AM123 amps to arrive.

Why is it that I just don't get better than -165...-170 dBc/Hz?

Let me try some simple math:

At L and R levels of +7.5 dBm (that's what come out of my 10811's),
subtract some 7.5 dB mixer conversion loss to see that 0 dBc would be 
exactly 0 dBm at the mixer output.
(The calibration is performed with the R signal decreased by 40 db. At 
the 40 dB LNA output, I'm measuring pretty exactly 0 dBm.)

At room temperature, the thermal noise is -174 dBm/Hz.
Add some 4 dB for the LNA noise figure and LPF insertion loss, so the 
system noise floor is at -170 dBm/Hz.
Give or take a dB, but that's pretty much about it.

Now, with the calibrated carrier level of 0 dBm at the LNA input, the 
residual noise floor is -170 dBm/Hz -(+0 dBm) = -170 dBc/Hz

A modern RF spectrum anylyzer has a noise figure in the 15 dB range, add 
10 dB of 'safety' attenuation, so the analyzer noise floor is -174 
dBm/Hz -(+25 dB) = -149 dBm/Hz. Remember that the mixer output / LNA 
input of -170 dBm/Hz is amplified by the LNA by 40 dB, so the analyzer 
input 'noise signal' from the test set is -170 dBm/Hz + 40 dB = -130 
dBm/Hz, which is 19 dB above the analyzer's noise.

If we increase the L and R levels by 15 dB, we need a +23 dBm mixer, but 
we will add 15 dB to the system dynamic range.
The mixer conversion loss, LNA noise figure and thermal noise floor 
haven't changed, so the noise floor in dBm is still the same -170 
dBm/Hz, but we're now refering to a carrier level of +22.5 dBm instead 
of the above +7.5 dBm (+15 dBm versus 0 dBm at the mixer output).
That's why we can now measure down to -170 dBm/Hz -(+15 dBm) = -185 dBc/Hz.

Adrian


John Miles schrieb:
> That sounds about right to me.  I was guessing you meant 40 dB and not 30 dB
> in the previous message, or there was something else causing about 10 dB of
> loss.  Lots of things to go wrong in this process!
>
> -- john, KE5FX
>
>   
>> -----Original Message-----
>> From: time-nuts-bounces at febo.com [mailto:time-nuts-bounces at febo.com]On
>> Behalf Of Adrian
>> Sent: Friday, May 07, 2010 10:38 PM
>> To: Discussion of precise time and frequency measurement
>> Subject: Re: [time-nuts] HP 11729C versus 11848A
>>
>>
>> Just the roughly 16 dB of insertion loss caused by the 562 ohm resistor
>> at the input don't make that filter such a great solution ;-)
>> So, I brewed something better together...
>> The 10 kHz beat note is now near 0 dBm when the R input signal is
>> decreased by 40 dB, which makes a lot more sense than before...
>> And, I have a noise floor of some -165 to -169 dBc/Hz at +7.5 dBm input,
>> and I'm seeing the 10811A's some 5 dB above that.
>>
>> Adrian
>>
>>
>> John Miles schrieb:
>>     
>>>> John,
>>>>
>>>> I'd say you nailed it.
>>>>
>>>> After some more testing, I can confirm that the limiter amp and the LPF
>>>> are the culprit.
>>>> I opened the box and plugged directly into the mixer LO port.
>>>> And, for the LPF, as a quick 'n dirty solution, I connected the <1 MHz
>>>> front panel output with the LNA input.
>>>> Now, at 10 dBm each into the mixer ports, I'm getting a noise floor of
>>>> <-145 dBc/Hz at 100 Hz and about -170 dBc/Hz at 10 kHz and above.
>>>>
>>>>         
>>> That's pretty aggressive for 10811s.  The floor on those is usually
>>> around -165 dBc/Hz.  How's your calibration process -- are you
>>>       
>> accounting
>>     
>>> for the 600 ohm output Z of the <1 MHz output port?  It'll lose
>>>       
>> a few dB if
>>     
>>> you try to drive 50 ohms with it, and/or the filter response won't be
>>> correct.
>>>
>>> -- john, KE5FX
>>>
>>>
>>> _______________________________________________
>>> time-nuts mailing list -- time-nuts at febo.com
>>> To unsubscribe, go to
>>>       
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>     
>>> and follow the instructions there.
>>>
>>>
>>>       
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at febo.com
>> To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>>
>>     
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
>   




More information about the time-nuts mailing list