[Pafgbt] Phased Array Coordination meeting
Rick Fisher
rfisher at nrao.edu
Wed Apr 10 14:49:39 EDT 2013
Hi Anish,
I think it would be good to put this discussion on 'pafgbt' so here goes.
You are probably right that there's not much margin for adding attenuation
at the output of the Dewar. It depends on the noise figure of the
amplifiers at the input of the fiber link package or, more accurately, the
effective noise figure of the fiber link input. This will be good to
know, but for now we can probably assume that Roger designed the fiber
link with just low enough noise figure to work with the Dewar design as he
knew it. This would be good to verify, however, once the uncertainty of
the cold LNA integrity is resolved.
Keep in mind that terminating the fiber link input adds 300K of noise to
its input. Hence, if the noise figure of the fiber link input amplifiers
is better than 3 dB, you're more than doubling its noise power with the
termination.
Rick
On Tue, 9 Apr 2013, Anish Roshi wrote:
>
> Hi Rick, Bob,
>
> It looks like the disagreement is in the measured noise figure and estimate
> noise figure of BYU receiver. May be we can repeat the noise figure
> measurement of BYU receiver. Bob, would you mind measuring the output noise
> spectrum of BYU receiver after terminating its input for couple of boards
> with a spectrum analyzer.
>
>
> Regarding putting the attenuator between A and B -- won't it reduce the
> noise margin of cold sky signal ? With the current gain, the cold sky signal
> is about 20 dB above the noise level when terminated.
>
> -----------------------------------
>
> I checked the noise due to the first stage op-amp LMH6639. The voltage noise
> from the spec is 16 nV/sqrt(Hz). The source impedance I have taken as 50
> Ohms (ie the mixed output impedance) and so I neglected the noise due to the
> current source. The noise power is -143 dBm/Hz and reference at the input
> with 23 dB gain ahead of op-amp is -163 dBm/Hz -- comparable to the noise
> due to the first stage !! Does this make sense ?
>
> Anish
>
>
> On Tue, Apr 9, 2013 at 4:09 PM, Rick Fisher <rfisher at nrao.edu> wrote:
> Hi Anish,
>
> >From your short write up, the noise temperature at the output
> of the fiber link (point C) with the link input (B) terminated
> is about 3e6 Kelvin. This says that either the input noise
> figure of the fiber link is much higher than 300 K or the net RF
> gain of the link is nearly 40 dB, or some combination of the
> two. Judging by the noise figure and insertion loss specs on
> the input components of the BYU receiver it seems unlikely that
> its net input noise temperature is much greater than about 1000
> K, unless something has failed or an LO has been set to the
> wrong frequency or power level.
>
> Your noise measurement at the output of the BYU receiver (E)
> with the fiber link output connected to the BYU receiver input
> shows a net gain in the receiver of about 66 dB (133.7 dBm/Hz -
> 67.7 dBm/Hz). This would suggest that the receiver is working
> more or less correctly.
>
> For the sake of argument, let's say that the gain in the Dewar
> is 40 dB, the net gain if the fiber link is 35 db, and the BYU
> receiver gain is 66 dB for a total gain of 141 dB. With 300 K
> connected to the Dewar input (array looking at the hot load) the
> receiver output power density would be 141 - 174 = -33 dBm/Hz.
> The BYU receiver output bandwidth is about 600 kHz so its
> output power would be +24.8 dBm or about 3.9 Vrms, which is far
> more than the ADC wants or needs and more than the BYU receiver
> can deliver. Hence, there's way too much total gain in the
> system.
>
> Where to add attenuation is determined by a balance between the
> need to dominate the noise of the system on the down-stream side
> of the attenuator when looking at cold sky and the requirement
> that all stages of the system be well within their linear
> operating range when looking at the hot load. My guess is that
> this attenuation should be divided between the A-B junction and
> the C-D junction since the BYU receiver output level is
> reasonably well matched to the ADC level requirements, judging
> by previous experience.
>
> The ADC range is +/-1 V with 12-bit resolution so one ADC level
> spacing is 2.0 / 4096 = 0.49 mV, and the minimum voltage level
> presented to the ADC when looking at cold sky should be about 2
> mV rms into a 50 ohm load, and the maximum should be about 200
> mV rms when looking at the hot load. Hence, a good level for the
> ADC input is between -1 and -41 dBm
>
> Someone should check my figures.
>
> Rick
>
> On Tue, 9 Apr 2013, Anish Roshi wrote:
>
>
> Hi All,
>
> Attached is a short write up on the measurements
> done on the PAF system.
> Anish
>
>
> On Mon, Apr 8, 2013 at 4:59 PM, Bill Shillue
> <bshillue at nrao.edu> wrote:
> Meeting tomorrow Tues April 8th
>
> Phased Array Coordination
>
> 9.30--10.30 NTC-200 and GB-137 by video
>
> topics:
>
> 1. Data from last week (Anish). Discussion of
> power levels and
> other issues (see Anish email)
> 2. Testing schedule going forward
> 3. Cornell tests indefinitely postponed
> 4. Dates for testing with BYU backend (2nd
> half of May ?)
> 5. Australia travel: travel is being setup.
> Waiting for NDA.
> Presentation needed.
> 6. Summary of meeting with Tony Beasley
> 7. Summary from David on setting up Roach
> testbed
> 8. Other
>
>
>
>
>
> -----Original Message-----
> From: Bill Shillue
> Sent: Monday, March 25, 2013 4:28 PM
> To: Rick Fisher (rfisher at nrao.edu); Anish
> Roshi; Steven White;
> Robert Simon (bsimon at nrao.edu); Michael
> Shannon; Matthew Morgan
> (mmorgan2 at nrao.edu); DAVID SAROFF (RIT
> Student)
> (dps7802 at rit.edu)
> Subject: Phased Array Coordination meeting
>
> Our usual meeting time of 9.30--10.30 is
> impacted tomorrow by a
> Division head meeting starting at 10 am GB137
>
> So we'll cover as much ground as we can before
> 10 am.
>
> We'll start with testing status (Anish) and
> Cornell/BYU tests
>
> Bill
>
>
>
>
>
>
>
>
More information about the Pafgbt
mailing list