Splitter/Amplifiers, Part Deux

[WA4HHG]

Have seen question and discussions which begin as questions about splitters and eventually evolve into discussions about pre-amps.

Multi-port CATV amplifier's have been discussed, these can also perform the function of a splitter. They are extremely wide band and some are 1 antenna input, with 2, 3 or 4 outputs. Mount the CATV amp very near the feedpoint of the antenna and run the appropriate number of separate feedlines, one from each port amplifier port, to your receiver(s).

The upside, this approach helps overcome both feedline and splitter port loss and in theory* improves the S/N ratio over an amp which is located in the shack in that stray feedline pickup is not amplified.

The downside, if you have 4 receivers a CATV amplifier with the same number of output ports will be needed along with 4 separate feedlines, one to each receiver.

I was very active on OSCAR 10 and this method of locating an amplifier was the preferred choice. Same with Moonbouncers and VHF/UHF terrestrial enthusiasts also. The amp may not be necessarily weatherproof, I have cut the bottom out of an empty Clorox bottle and leaving the bottle cap on, stuffed the amplifier into the bottom which faces down. It'll last a few years until the UV gets to it.

When selecting an amplifier, choose one with the lowest Noise figure possible. In another thread, I saw a 2 port PDI CATV amp model mentioned. Did a quick Google search and came up with:

http://www.pdi-eft.com/htmlandflash/proprietary/house_amplifiers.html

Look at the noise figure, 2.4db nominal within the 54-1gHz bandwidth. Is that good? As a comparison the 432 mhz VHF DX'ers look for an amp with a noise figure something under 1.0db. However those pre-amps are intentionally somewhat narrow banded and cost well in excess of $100. Both of those attributes contribute to an excellent noise figure. A 2.4 db noise figure for an amplifier costing in the $20 range is acceptable. When choosing a pre-amp, low noise figures, <3.0db always trump gain.

http://www.ebay.com/itm/PDI-2-PORT-...332&pid=100015&prg=1006&rk=1&sd=190308512393&

What about amplifier gain? Some is good, more is better, right? Well, not necessarily.

Good system design employs a pre-amplifier to compensate for certain embedded losses. For a quick and dirty, thumbnail explanation, let's consider feedline and splitter port losses. In this case, you have 2 radios and the feedline run is 100' run of RG-213 from your antenna and your band of interest is 800 mhz. For the sake of discussion, we'll say the line loss is 6db at that frequency and each splitter port has 3 db of loss.

PDI build a CATV 2 port, model available which has 11db of gain from each port. The port losses for the splitter function are already subtracted from the gain specification. With 6db of cable loss and allowing an additional 1db for connectors, the amplifier has ~4db excess gain. That's good, right? Maybe not.

Excess gain can cause front end overload and other problems.

The solution is to add a 3db inline attenuator (or a 4db if you want to/can find it) NEAR THE RADIO, INSIDE THE SHACK. So, at the end of the day, you end up with that 2 port splitter and neutral feedline losses. Why place the 3db attenuator inside the shack/near the radio? Remember the pre-amp noise figure and stray feedline pickup mentioned earlier? Those are reduced by the attenuator but to work properly, must be done at the radio end of the feedline.

There is a formula for calculating the ratio for the numerical system result. That number is dependent on a number of variables and frankly, is more work than I want to invest in this post. Just off the cuff, think of a antenna feed system noise reduction number south of 0.3db or so.

Something to consider.

 

[zz0468]

Interesting post, and technically sound.

A 2.4 db NF from 54-1000 MHz is half decent, but not spectacular. One also wants to consider the susceptibility to overload of the amplifier. Those cable tv amps are designed for controlled conditions where all the signals they see are within a specified range. Depending on the application, that can be a substantial amount of signal power hitting the amplifier, but probably not so for residential grade active splitters.

I've measured the noise figure of scanners before, and generally see from 4-6 db. So, a 2.4 dB NF preamp would be an improvement, if the improvement isn't lost to either feedline loss or overload. Cascading too much gain in front of a receiver will destroy the NF performance of the whole system, so putting pads in between stages can actually make it hear better.

The specs on a lot of cable TV drop amps can be as high as 8 db, depending on make and model, so those aren't automatically a good choice for scanner use. Some people swear by them, though. The end result is frequently strong signals make "more bars". What's often not noticed is the weak signals that just aren't there.

Optimizing receiver NF performance is all science, so far as theory is concerned. But in practice, it can border between art and black magic. Noise figure meters are expensive and hard to come by, but a signal generator and a SINADDER can be used, minus the end result in a known noise figure specification. But it's an interesting exercise to build up a multicoupler system with a preamp, distribution amp, receivers, and then add some attenuator pads in strategic places and watch the system become MORE sensitive.

You need JUST ENOUGH gain to overcome the slightly worse noise figure at the next device downstream toward the receiver. So, it's not unusual to see a preamp-multicoupler system with, maybe, 4 or 5 db gain from end to end.

 

[gewecke]

I don't always use a preamp ... but when I do,I prefer a gaas-fet preamp!

I have 2 of these and they work great, and rarely fizzle out, very low s/n ratio.

gaasfet.html


73,
n9zas

 

[WA4HHG]

Appreciate an expansion of the essence of my point. "More bars" is never the goal for, any pre-amp application. Scanner front ends are usually not lacking for raw gain, adding 10+ db of additional gain can certainly have unintended consequences. If there is a choice between 2 amplifiers, one having 15 db of gain with a 3.5db noise figure and another with 10db gain but a 2db noise figure, the lower noise figure always trumps gain. Of course, there is the issue of dynamic range in the pre-amp.

Your cautions IRT CATV amps are well noted. A good friend is Network Operations Manager for Cox Cable here in town and we discussed this Friday evening, the 5th at dinner. Ideally, CATV amps like to "see" their rated input at stable levels. Excessive input can saturate them or even worse, cause oscillation, amplification of mixing products and spurs. That all said, I'd risk $20 and some time installing the model linked in the first post for the sake of experimentation. A more ideal approach would be to persuade Andy Ilkin of Wellbrook fame to design and market say a 10db 2 output pre-amp with an economically attainable low noise figure (say, < 1.5db from 100mhz to 1 ghz) that would stand up to +32db input. His MW and SW loop arrays have amplifiers with those kin of specs. Of course, its quite another thing to affordably replicate that performance across the frequency's and bandwidth above.

Rather, the point here was to illustrate signal distribution, compensating for feedline losses and port loss when using passive splitters. Ideally, the feed and distribution system should be gain/loss neutral with the pre-amp mounted physically close (say, within 1 ft) of the antenna and any excessive gain be compensated for by a passive attenuator at the radio.

Physical location of pre-amps and attenuators are some of the key considerations when making a well engineered, effective signal distribution system. Appreciate zz0468's comments on the contribution (or degradation) pre-amp IP3 specs can have on performance.

 

[WA4HHG]

Nice amps, +26db IP too

I don't always use a preamp ... but when I do,I prefer a gaas-fet preamp!

I have 2 of these and they work great, and rarely fizzle out, very low s/n ratio.

gaasfet.html


73,
n9zas

But they have a LOT of gain. Here is where (IMHO) you calculate all your feedline and splitter port (if any) losses then pad down what is excess near the receiver. That will yield the best overall system noise figure while compensating for our feed losses.

Just my $.02

 

[prcguy]

There are a couple of points that have not been mentioned like IMD generated in the preamp and local "urban" noise floor.

CATV preamps are designed with a known and fixed system RF level and not for connection to an antenna, which has varying amounts of carriers at very unpredictable levels. A lot of CATV and low cost scanner type preamps have 1dB compression points of +5 or 10dBm at best with a correspondingly low IP3 rating.

What this means is you start with a noise floor in your area that is hopefully low and are able to pick up weak distant signals fairly well. Then add the preamp which gets hit with lots of high level signals like local TV and FM broadcast, nearby cell towers, high power paging transmitters, the cop or firetruck transmitting down the street, etc, and the preamp is now generating lots of IMD.

What this does is create mixes and new frequencies that don't belong in the spectrum, sometimes hundreds of them that you can actually measure at surprisingly high levels but most just raise the noise floor to a point it covers up the weak signals you used to pick up.

Sure the overall spectrum to your scanner is much higher and the meter on your radio might read higher on some desired signals telling you you made a good decision to add the preamp, but you cant pick up the weak ones any more because your signal to noise ratio is much worse because of the high noise floor you just created.

One way to avoid the problem is to use a very high level preamp like 1dB compression of 30dBm (1 watt) or more and or filter the spectrum to be amplified to keep as much unwanted stuff from saturating the preamp and creating IMD. Super high level preamps with low noise figure are quite expensive and filtering is probably the best choice for those that pursue certain limited frequency ranges like VHF or UHF air band. One exception to this might be those that have an 800MHz Yagi which will severely limit the amount of VHF/UHF input to the preamp, possibly avoiding overload problems.

The other point I would like to make is going for the really low noise figure for a wide band omni directional antenna in an urban environment is probably a waist of money. The local noise floor is probably polluted from RFI generating sources in your house and local businesses that you will never see the difference between a .5dB noise figure and 2.5dB noise figure with all other specs the same.

If you live out in the sticks away from anything that generates electrical interference, or are a hamster doing moon bounce or terrestrial work with very directional antennas you will benefit from the lower noise figure because you can limit the amount of interference to your receive system with the directional antenna.
prcguy

 

[gewecke]

I use my AL gaas fet's for EME work off a quad in the back yard, but not much else.

73,
n9zas