NRAO Home  >  Green Bank  |  Wiki Topic:    GB > Projects > TWikiUsers > RonMaddalena > QBandUpgrade

Changes | Index | Contents | Search | Statistics | Jump to Topic:

Q-Band Receiver Upgrade -- Scientific Drivers vs. Cost

Why not 4-feeds?

Of the 24 or so accepted Q-band proposals, only one requested 4 feeds (That proposals wants to map a 1 x 4 arcmin object where our 1 arcmin feed separation wouldn't help significantly anyway). I cannot see any advantage where having more feeds would have improved any of the other proposals. Since we never advertised widely that the receiver has only 2 feeds, the lack of 4-feed proposals is probably not a chicken-and-egg thing. The scientific driver for 4 feeds isn't there so the science-to-cost ratio is very, very low.

That leaves only operational reasons for having 4 feeds -- namely, to have some redundancy in case one of the feeds breaks or has sub-par performance. Deciding for 4-feeds, then, is a risk-management decision for the project manager and not one for the project scientist.

High-end of the band

The more scientifically important question is what to do about extending the upper end of the Q-band receiver. There, the scientific drivers are more obvious by the number of accepted proposals that want to observe above the current receiver's upper end. The upper end can be extended in three ways:

(1) Those components that Gary and Roger say are cost effective to change next summer. These components will also improve performance for two of the channels below 42 GHz. Changing these components will allow us to schedule many more Q-band proposals. I'd say, this it's an easy 'scientific' and cost judgement to back the decision to change these components.

(2) the current polarizer probably has little or no performance above ~49 GHz. The preliminary 'reflectivity' graphs I saw suggest that we'll need to redesign and build a new polarizer if we want good performance above ~49 GHz. That will require, according to Roger and Gary, a redesign of the dewer, rearranging plumbing, etc. The work is almost comparable to that required to build a receiver. Gary is currently modeling the performance of the polarizer so, stay tuned.

(3) the frequency range covered by the amplifiers. At 50 GHz, their gains are only 0.8 - 1.5 dB, down from 8-13 dB at band center, suggestive that the amps are running out of umpf at about 50 GHz. The amplifiers need to be measured above 50 GHz but that is probably impractical with the receiver mounted on the telescope. It probably won't be clear what the receiver's performance will be above 50 GHz until the receiver is back in the lab and the (1) modifications are made.

I've come up with estimates of what fraction of the 16 accepted proposals are being limited by the current performance and how each of the three improvements will increase our science. (I'm ignoring any bad estimates the proposers made with their noise/Tsys estimates since I'm looking at science drivers.)

• When the (1) problems are fixed and we can cover 40-49 GHz, we'll get a total of 3/16 more science over what we currently can do.
• Assuming the amplifiers have no performance above 50 GHz, building a new polarizer that gets us to 50 GHz, and doing the (1) modifications, we'll get a total of 4/16 more science over what we currently can do. 1/16 more science doesn't sound like a huge enough step.
• If the amplifiers are good to 51.5 GHz (or we get some that are), and we redesign the polarizer, and we perform the (1) modifications, we'll get a total of 7/16 more science over what we currently can do. This is significantly better than just doing the (1) modifications.

So, science drives us to either 49 or 51.4 GHz, and not some intermediate limit. Unless we take the receiver off of the telescope, until next summer we're probably stuck not knowing the amplifier's performance above 50 GHz. Gary is planning to get us more information about the polarizer but I'm not too hopeful we see anything good above 49 GHz.

Thus a compromise science/cost solution would be to just make the (1) modifications and hope (but not advertise) for performance above about 49 GHz.

-- RonMaddalena - 19 Sep 2005