Antenna and Filtering front end for weak signals?

[rwgast_lowlevellogicdesin]

So if your working weak signals on a wide band radio like an SDR or scanner, things like EME, or space communications (ive seen some people have luck picking up mars rover signals!) im curious about a few things that I see commonly done, and wonder why.

First one would be antenna, so anyone doing radio astronomy/deep space uses a dish. Is this necessary? ATM I have a big 18ft C-Band, auto rotor and all, im just at a strange place in my life which prevents me from putting it up where i am currently living. I have also seen people using 4+ stack yagi arrays. When working 1ghz+ antenna elements of all sorts are a pretty manageable size. What I do not see are people using a big flat reflector full of elements like patches, biquads/double biquads, helical/double helical. Wouldn't an array of 4x4 to 16x16 elements give more gain than a bigger dish while at the same time being more compact and cheap if done right. My idea for a super cheap reflector is to use a piece of press board, or plywood that has been coated a few times in polyurethane and then wrapped with regular old non ferrous foil tape (I would imagine wrapping in copper tape would get pricey, sheet metal is cheap but ferrous and I know that causes a little bit of performance issues due to the skin effect of ferrous metals at higher frequency's). Next depending on chosen element, some cut pcb, copper taped wood or steel, 14awg or even copper tube is cheap enough. I have seen one guys site who built a 4x4 70cm patch array (pretty big, I wouldn't do this at 440 but it was a cool read) for amsat and he claims it worked well, although I can not find the site again. So why the is a dish common practice or even a 4-8 23 element loop yagi array? I have nothing against yagis except they can get cumbersome if you want to motorize your stack, I think over all 16-64 small elements used in home brew wifi set ups tuned to 1296/1440 etc is more compact. So is there a reason people don't do this? Is my idea just ridiculous here? You could even mount different bands arrays on one reflector phased properly and use a relay to switch bands. Disreguarding RX, are these sort of set ups not generally great for hi power TX, admitidally all these elements I am listing are used in small power system when they are used for TX, as far as know... don't take that as gospel but running 100-1000w im not sure if something like this would be suitable. In the end I plan on using a receiver and separate transmitter so I can optimize for the best of moth worlds. So what would work better aricebo or an array of micrwave element spread over the same size as arecibo?

Next is feeding a microwave system. I get the feeling that you are better off with out an LNA from reading some peoples texts. PHEMT MMIC are cheap they have .2 to .75 db of noise over there broad range. Maybe if you have a nice antenna set up and are running hard line you dont need the LNA but why not use if it's noise floor is lower than your receivers anyway? I would rather split the gain load by putting an mmic at the antenna and then cranking the radios gain right until the noise floor raises then stop.

Maybe this is the same as the answer above, but what is wrong with using cheaper coax for weak signal and just using LNA's to overcome it. The way I see it you always going to need a hard analog filter for weak signal if your using a scanner/SDR no matter how well it can handle channel separation and intermod. Unfourtantley I am not loaded so I cant afford good low loss filters for the most part, for instance I want and L-Band filter that isn't for gps, I want something a little more broad band like 1500-1650, I can not find any filter components even close to that i.e. SAW, ceramic etc.. I am starting to understand pcb filters, there awesome I can make it exactly how I want it, print and do a cold transfer and I have a filter in a few minutes. These filters can get up to 5db of gain on fr4 and even on rogers you may be looking at 3db so this even with a good feedline I would need an LNA. I guess im saying if you ran something stupid like 100ft of rg58 and stuck 20db worth of filtering in the feed chain what is wrong with using multiple Phemt MMIC's and not worrying about loss, at least on the RX.

Lastly if you were to mix SMT caps and make on PCB spiral inductors to make a lumped element would this work well in the microwave bands with lower loss?

Im just trying to get really weak space signals including EME, this can get really expensive, if you home brew it can be really cheap too... but all the solutions I have seen are just bulky, Im just trying to understand weather it is possible to ditch the hard line/lmr600 and make feasible antennas that can be rotated with smaller hi torque gear motors and regular tripods. I am mostly concentrated on RX right now but want to do TX too, and I know the power amps at these frequencys will be expensive weather there done DIY or not