Splitter

That could very well be water condensation, and I suspect that it's in one of the cables, since replacing the splitter didn't fix it. If you have a cable coming in from outside, it's not at all uncommon for it to get filled with water.

Nothing to do with the splitter, it's the connections to the splitter, what we termed "suck out" at the cable company when it happened on line gear. What you have is "F" connectors which are little more than nuts crimped onto a piece of coax with the copper coax center conductor acting as the center pin for the connector. If that center conductor is cut too short when things get cold and the metal contracts it will pull back into the coax and out of the connection in the splitter or whatever. The connections on the hard line coax on CATV line gear are a little different, but the same effect can occur there with the center conductor pulling out of the connection if it isn't cut long enough.

Actually, it was the opposite on hardline. The foam insulation slowed the contraction of the inner conductor and caused the shield to pull out of the connector. BTDT, and needed the propane torch to thaw out hands.

Funny that you mentioned CATV. I have a pile of Agile Modulators sitting here to convert to the old US FM band, and an agile processor. I wish that I had ahd a few of these 25 years ago when I repaired headends and designed interconnects between community loops for seperate CATV systems.

is a cheap 10 dB amp that would likely help in a marginal area.

Not in our experience in the frozen northeast. The hardline connectors grip the AL shield/jacket quite well, and the span between poles always has a slack loop to allow for contraction without pulling on the connectors. The center conductor which just connects inside the line device with a screw terminal lug is where some inexperienced techs would clip the center conductor off just past the lug, leaving only 1/4" or so before the conductor would pull free of the contact. Leaving a good 3/4" past the lug prevented the problem. Of course on line gear the problem is worse than on home gear since RF can bridge a bad connection, but the

60VAC line power for the amps can't.

I've got a couple modulators kicking around as well. Not much use any more in these ATSC days though.

An amp located near the antenna, or there are some pretty good amplified unit antennas available these days.

Raychem developed their ceramic hardline connectors which helped. Most of our suckouts were underground cables, till we switched from the traditional Gilbert connectors. The Raychem was better, but over time the aluminum skin would fatigue and break within a few inches of the connector.

60 VAC modified sine wave at up to 30A was part of what caused the hardline suckouts. The center wire stayed warmer because of current crowding.

This was all in Delhi Township, near Cincinnati, Ohio.

The real fun was getting CG&E to let use mount a large steel NEMA box on the side of one of their power poles to house an RCA Heterodyne Signal Processor to convert T-7 to Ch. 12 at the interconnect point for two community loops. Mid split to sub split. The lead tech complained that I was off by .5 dB when their community loop was powered up for the first time.

What brand(s)? Are they home units, MATV or CATV? I've worked with Holland, Blonder Tongue, Phasecom, RCA, Catel & Scientific Atlanta. I was repairing BT strip amps back in the early '70s, and some Vicoa 12 channel trunk amps in the Army. I repaired 300 & 450 MHz Sylvania line equipment, which was later sold to Texscan. I wave a Wavetek SLM, along with a portable Sadelco, a SAM and a Texscan sitting around

With digital, it's the recovered data that matters, so an inline amp at the splitter would help. Those bullet amps were fairly low noise 25 years ago in fringe locations, so if the sets work most of the time, the extra 10 dB should stop the problem for under $20. It might even get him another channel or two. With digital the final BER is what determines the picture quality. (Bit Error Rate).

Finger tight with cheap connectors? Possibly a little short on your center conductor? I am betting that you get a poor connection when the metal shrinks. Go with a quality all weather connector with an o-ring inside, enough stick-out on the center conductor, and put a wrench on it lightly and I bet its better. I don't do much anymore, but back in the mid 1990s I installed hundred and hundreds of Primestar dishes which often included splitters (upto four way) outdoors with weatherboots. Proper installation and they worked all year around just fine.

The connectors used for cable TV -- and the splitters -- are far from the best. They are as cheap as they can be made -- even using the center conductor of the cable as the center pin of the connector.

The temperature may be affecting the grip of the female connector's center pin on the coax cable center conductor. A bit of careful bending of the grip fingers (if you can get in to do it) might improve the grip. As would treating with a good contact cleaner. Or -- just unscrew the ring, and pull the cable out and push it in multiple times to clean any oxide off the center conductor. (Are you using coax down from the antenna, or is this a splitter for the old 300 Ohm ribbon antenna cable?)

Or -- the thermal stresses might be opening a connection inside the splitter. They also are made as cheaply as possible.

Out of curiosity -- did you try exchanging the output feeds to see whether it made things better?

Why the splitter anyway? Feeding the antenna to two different locations in the house -- or to the TV and a VCR?

Good Luck, DoN.

There is something that does not make sense about this. The thermal expansion of the plastic insulation in coax is about 10 times that of the center copper wire. The plastic basically determines the length of the cable between the connector bodies, so I would think the wire would actually protrude more as the temperature drops. This still could be a problem since that might make cause the center connector to make contact with a grounded shield in the splitter.

Coax all the way.

That was my original thought. Bought a new splitter, which improved performance, but did not eliminate the problem. The original splitter was purchased back in the early 80's.

We feed two different (and identical) TV sets. Both behave the same way, although not necessarily in lock-step. It's clearly the loss of signal to each set, and, in our case, most likely due to weak signal to begin with, as we're a good distance away from the transmitter (tower). When the switch to digital came along, we lost many stations. Before the change, we could get both Portland Oregon and Seattle stations (although not well). We lost all of the Portland stations and almost all of the Seattle stations with the change.

Thanks, Don, and all the others that provided an opinion. I have yet to explore further, as I am short of time right now. The light bulb has provided excellent service---the problem, what ever it may be, has been temporarily addressed, so we're happy for the moment, but I hope to discover which of these things are responsible. I'll report if I happen to solve the mystery.

Amazing. This group is exceedingly knowledgeable. Too damned bad it doesn't stick to productive things instead of all the bad-mouthing it does with trolls and those who don't have a life.

Again, thanks to all. Your comments are appreciated.

Harold

Come on guys - internal components - even splitters have them - resistors - they change values over temperature. In sensitive circuits there are designs that compensate for this effect.

You are simply changing the impedance of the circuit and once that is done, reflections and shunts occur.

Mart>

I saw a lot of damaged splitters where people left the center conductor too long. It would distort the spring in the 'F' connector and make it unreliable. Small TV coax has a copper plated steel center conductor & a steel braid over the foil shield. The foam insulation is too thin to provide much insulation, so the expansion factor is very small.

I ran the service shop for United Video Cablevision in Cincinnati, Ohio back in the '80s.

Good splitters have a multi-port transformer, not resistors. Resistors waste too much power, You would lose 7 dB per output on a two port splitter instead of 3.5 dB.

Balms maybe - RF transformers. But impedance matching is still going on with various T, L, H, patterns.

Many have LNA's in them as well. I have a dual myself. It is an active RF circuit you have to plug in. It downshifts the frequency, combs out the noise, boosts the signal and then impedance matches two outputs from a single input.

Many versions. Many of them are etch on substrates.

Mart>

Martin, I've spent my life working in RF. From AM Broadcast to well past 11 GHz. First of all, a BalUn coverts from balanced to unbalanced. In TV that also included an impedance transform from 300 Ohm balanced to 75 Ohm unbalanced. A simple broadband transformer, and often on a plastic core in the early days. TV splitters with equal outputs are all two port directional couplers to provide at least a 20 dB backmatch, and that is considered a low grade device. At one time I was testing samples from dozens of vendors to either qualify them for our CATV systems to purchase, or remove them from our AVL. Most two way splitters were a simple ferrite core with three identical windings, one to each port, with a few PF cap between the two outputs in an attempt to improve the backmatch. What Harold described wasn't an active device, so there was no reason to bring them up. A MiniCircuits MAR, ERA or any of their other 1 GHz MMICs can give you 10 to 20 dB of gain with a decent match for under a buck. Avango (HP) and several other companies make similar parts, in the same price range. If you have plenty of excess gain to burn, you can use resistive dividers. OTOH, you add more noise to the overall system gain.

A LNA is a 'Low Noise Amplifier' and it does not 'downshift the frquency'. A LNB or LNC dows convert the frequency, for Sat TV. The LNC is used with an LNA to form a LNB and there are multiple types of converters. Block, tunable or phaselocked block where a pilot carrier is sent up the coax with the DC power, ant the output comes back on the same coax.

My RF gear is in use by NASA, NOAA, the ESA and on board the ISS.