coax splitter replacement

Make and model of cable modem? Most cable modems have internal diagnostics that will provide you with the line levels. If the levels are too low or vary, you will have problems.

True. For cable TV, the bandwidth has to go down to 5 MHz. The upper end is determined by the highest frequency used by your cable provider, which you also didn't bother mentioning. 1000MHz is usually sufficient unless you're piggy backing satellite TV, which requires

2000MHz.

I guess it's too much trouble to disclose the model number. If you can't manage that, perhaps the number of ports on the splitter would be helpful.

Incidentally, you overpaid. I pay about $2/ea for 2-1000MHz 2 port splitters.

It's unlikely to be the splitter, especially since your previous splitter showed similar problems. Much more likely are crappy crimps on the F connectors and RG-59a/u junk cable instead of RG-6a/u. If the connector looks and feels like it's about to fall apart, it probably has already done so internally.

Solder to what?

Don't bother. Inside your two unidentified splitters is a small torroid transformer and a resistor. You need those between the center pins. If you just connect everything in parallel, it won't work.

Don't forget to wrap the whole thing in aluminum foil for shielding and so that nobody will see the mess.

Hint: If your question is lacking numbers, you probably won't get a usable answer.

Incidentally, you're header shows a peoplepc.com email address and an Earthlink ISP, neither of which offer cable service. How did you manage that?

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann     AE6KS    831-336-2558

F connectors are not bad *provided they are properly installed on the proper type of cable*. I wouldn't be surprised if the problem isn't just marginal signal levels. Like all things "digital", the system will transition from working fine to not working at all with a level change of half a dB or less. That can happen just from the backbone cable going from full sun during the day to cooling off after dark. Especially if your system was set up without a splitter, or a while ago, losses may have increased over time.

Best solution is to get a cable company tech out to your house with a tool to measure the level, including the splitter loss.

The tech can adjust the attenuator that connects your drop to the backbone cable (it's probably on the pole), and things will be fine again.

If you can do a test by taking that splitter out of the circuit, and the CM works fine, then signal level is the culprit for sure.

Isaac

It appears that advice from The Lady from Philadelphia is again needed.

The Lady from Philadelphia uses the splitter supplied by Comcast. One of its outputs connects directly to her Motorola modem, using a double-male (hmmm...) barrel connector from MCM. The other output connects to a cable that feeds her Vizio monitor. She has no problems.

If changing the splitter doesn't fix the problem, then the problem is probably //not// with the splitter. (The Lady from Philadelphia does not understand why this is not obvious.) Furthermore, it is not clear to her how soldering the leads is going to fix //anything//. How does one //get// to the leads, anyway? F connectors were not designed for soldering to their innermost... parts.

The Lady from Philadelphia's gut feeling (pardon her vulgarity) is that your modem has a bad F connector, making intermittent contact. She suggests that you drop by the cable company and explain the problem. Of course, as you are (likely) the modem's owner, the technically ignorant people behind the desk will probably tell you that you need to buy a new modem. Perhaps a veiled threat (of your choosing) will encourage them to loan you a spare, so you can see what happens. Do not forget to have your account's MAC changed before you leave, or your Internet cable connection will be deader than Marley.

There is also the possibilty that the little bare wire sticking out of the cable is, shall we say, too short to perform its required duties? Have you experimented with a better-endowed cable?

The Lady from Philadelphia concludes by offering this rude (and unsolicited) observation... TWENTY BUCKS FOR A BLEEPING SPLITTER? She hopes it was at least gold-plated. The Lady from Philadelphia knows from personal experience that expensive items from RodentShack (polite folk do not talk of r**s in public) are not necessarily better than those from elsewhere, costing one-fifth as much.

You subscribe through TW Cable or Brighthouse like I do. They offered a choice of Road Runner, AOL or Earthlink when I signed the contract. They used a directional coupler instead of a splitter when they installed my cable modem. I can't get to it right now to tell you how many dB drop right now.

Brighthouse currently uses S.A./Cisco Webstar modems in this area.

192.168.100.1 will display the current data.

(I hid the MAC Address)

Name CISCO DPC2100R2 Modem Serial Number 224656807 Cable Modem MAC Address xx:xx:xx:xx:xx:xx Hardware Version 2.1 Software Version dpc2100rx-v202r1256-110513Uas-TWC Receive Power Level -4.9 dBmV Signal to Noise Ratio 37.7 dBmV Transmit Power Level 41.5 dBmV Cable Modem Status Operational

Thanks. I didn't know that. On the left coast, we have mostly Comcast, which offers ummmm.... Comcast. Locally, AOL and Earthlink both resell DSL. However, they can be purchased with a "bring your own ISP" contract that will work with any means of connectivity.

Locally, Comcast uses splitters. The only time I've seen directional couplers used is when the line level is to high and the installer wants to drop the receive level to the cable modem, while not introducing any additional loss to the return path.

That's the most common IP address for cable modem diagnostics. The same IP is used by Comcast modems. I haven't found it in the CableLabs specs, but outgoing traffic to 192.168.100.1 is apparently trapped by cable modem firmware, and redirected to the internal web server, instead of going out to the internet.

If you have problems accessing the SNR values, try this work around:

Looks good.

For reference, the cable modem is designed to work with -15 to +15dBmV receive power, but it's usually best to keep it between -8 and +8dBmV. Upstream power should be at the lower end of 37 to 55dBmV. It's a measure of upstream line loss. You're fine.

SNR (signal to noise ratio) isn't measured in dBmV. Did you edit the above diagnostic output? It should be greater than 30dB SNR. I've seen the SNR vary erratically as a result of ingress and crappy connections, while the other numbers remain fairly constant. That's my guess what is happening with the OP's setup.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann     AE6KS    831-336-2558

Noooooooooooooooooooo!

I've found out the hard way that having the center wire sticking out too far from an F connector is a great way to destroy F jacks. A wire that goes too far in will spread the two inside contacts. Tightening down on the threaded part can apply a lot of force to the inside guts.

Mark Zenier snipped-for-privacy@eskimo.com Googleproofaddress(account:mzenier provider:eskimo domain:com)

Noooooooooooooooooooo!

I've found out the hard way that having the center wire sticking out too far from an F connector is a great way to destroy F jacks. A wire that goes too far in will spread the two inside contacts. Tightening down on the threaded part can apply a lot of force to the inside guts.

Mark Zenier snipped-for-privacy@eskimo.com Googleproofaddress(account:mzenier provider:eskimo domain:com)

They were forced to offer other choices, since AOL owns TWC/Brighthouse and most of the RR fiber optic backbone. AOL was the obvious choice, and since Earthlink was one of the largest ISPs they only had to add one other service to satisfy the regulators. I'm not sure that you can still get Earthlink from Brighthouse for new accounts. I haven't seen it in their ads in some time.

I only edited the MAC Address, for security reasons.

Directional couplers provide better isolation from the garbage sent upline by crappy TV tuners. I worked in CATV for years, and tested hundreds of samples of 2 to 8 way splitters for return loss in our lab.

I'm used to using a Fireberd to measure the BER on digital communication systems. FPQSK was real fun when we first added it to our telemetry products. The units are those selected by SA, long before Cisco bought that divison. At one time we built the SA Telemetry receiver that they had cloned from our 1200 series.

I won't disagree with you -- but the wire just might be a bit too short.

There is no way to lengthen the wire, so no harm can occur simply looking at it...

Veddy strange. SNR usually isn't expressed to 3 significant figures and should be in dBm, not in dBmV. The firmware scribblers seemed to have goofed. Oh well.

Huh? The FCC demands that anything connected to an antenna have very low LO radiation. That transfers to the cable TV connection when the F connector is attached to the cable instead of an OTA antenna.

This is kinda old, but covers the topic: New Measurements and Predictions of UHF Television Receiver Local Oscillator Radiation Interference

With 40dBmV (or about -9dBm @75 ohms) transmitted on the return path, even a lousy LO radiation of maybe -40 to -60dBm in the return channel isn't going to do much damage.

The Acterna (JDSU) Fireberd is a nifty gadget. I had an Acterna HST-3000 but didn't have the CATV plug in or software. I did have a BERT but it maxed out at T1 rates.

I think you mean FQPSK.

--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com               jeffl@cruzio.com
# http://www.LearnByDestroying.com               AE6KS

Some tests I did on Telemetry systems were to four places, and had to match to within .01 dB in diversity sysems.

SA probably didn't bother to review the design, since it's rarely seen by anyone. Most who would look have no clue what a Decibel is, or that there are multiple references.

I was never impressed by SA designs. Their C band receivers were noisier than even the cheapest Microdyne. They reverse engineered a Microdyne receiver for their only Telemetry product, and lost the court case for using a patented method. It was a slam dunk case. They bought a couple receivers, and a year later they release their design. The only real difference was they replaced the LED displays with a bad LCD display. Most of the units we repaired had failing or dead displays, and had to be modified to use a better panel.

I had trouble teaching some techs to use the 'relative' function on a Fluke 8920 True RMS Voltmeter, and that you had to make sure you used

50 or 75 Ohm cables and terminators that matched a particular board or module. We used them on the video amps, with up to 20 MHz video outputs.

You might be surprised at the crap that can come out of a damaged tuner, or one with a damaged F connector. The backmatch was horrible on some brands of splitters I tested. Some were over 20 dB off spec. It seemed like the fancier the label or plating the cheaper the inside were made. I peeled the back off of a lot of them to see that the soldering was substandard, and in some cases, they missed some joints.

We blacklisted multiple sources. When you buy them 1000 at a time and drop cable 50,000 feet at a time, it was worth the pre-screening. Our plant was the only one in the corporation with a lab & repair shop so we arranged the supplies fr all the locations.

I wasn't working in CATV by that time. We used them to test error rates for Telemetry systems.

Yes. We used it in the RCB2000/DCR2000 series, but that was over 10 years ago, and just before I became disabled. I'm lucky to sleep a few hours a day anymore and miss technical typos.