|
Archives
 November 1999 Issue
Columns
Hranac-Notes for the Technologist
High-Pass Filters Revisited: Lines Drawn in the Sand
By Ron Hranac
 |
| Ron Hranac |
High-pass filters are passive devices that block most or all frequencies below a so-called cutoff frequency and pass most or all frequencies above that same cutoff frequency. They physically resemble the positive and negative traps that we've used for years for premium channel security, cost about the same and, in most cases, are even made by the same manufacturers.
One common use for high-pass filters is to block reverse path noise and interference in the 5-40 MHz band while passing conventional downstream signals above 50 MHz. Installation of a high-pass filter at the tap will keep interference that may be generated in a subscriber's drop or home from entering the cable network's reverse path, while allowing that subscriber to receive the downstream TV channels.
Armed camps
In the past, I've taken a somewhat neutral position on the use of high-pass filters. These things often result in emotional debates.
One camp says to install the filters in the drops of all one-way subscribers and then remove them as the subscribers take two-way services. The theory behind this line of thinking is that the cost of the filters-$4 to $6 each, plus installation-is much lower than replacing or upgrading a lot of old or otherwise imperfect drops at $50 or more each. The other camp says high-pass filters are unnecessary in most cases if the drops were installed correctly in the first place.
The latter scenario works well under the Data Over Cable Service Interface Specification (DOCSIS) retail model. It means a subscriber can purchase a DOSCIS cable modem at the local Circuit City, connect it to the drop and happily surf the Web after a self-installation and automatic provisioning process, all without the need for the cable company to send a technician to remove a filter. After all, a two-way cable modem won't work if a high-pass filter is in the line. The filter will block the modem's upstream signal.
As you can see, both sides of this argument have merit.
Take a stand
These days, I'm leaning more toward advocating use of high-pass filters. The unfortunate reality is that the condition of the reverse spectrum in the majority of two-way systems that I've had a chance to look at over the last couple years is only so-so. There's still a lot of the attitude that "all I need to do to fire up the reverse path is install some reverse actives and maybe adjust a few amplifiers." Yeah, right, and I've got the deed for the Brooklyn Bridge.
Most reverse "junk" problems I see are drop-related, except for alignment issues, but I'm saving that for a rainy day. I also see a lot of low-level signal leakage. While nearly all systems today are pretty good about meeting the Federal Communications Commission's 20 microvolts per meter (µV/m) signal leakage limit, that's simply not good enough for two-way.
Every leak, no matter how small, is a potential ingress point. My experience suggests that most low-level leakage occurs in the drop portion of the network. This seems to confirm that up to 95 percent of the junk that gets into the reverse path comes from the drop: 25 percent or so between the pole (or pedestal) and the side of the house, and the other 70 percent between the ground block and TV set. Where there's a leak, there's ingress.
Oddly enough &.
Speaking of ingress, allow me to digress for a moment. I've spoken with some folks who think that if a given leak is low-level-that is, below the FCC's 20 µV/m limit-then ingress getting into the system via that same leak also will be low-level. While it's generally true that a minimal amount of overall leakage usually results in a minimal amount of overall ingress, individual leaks are not necessarily balanced from the perspective of "if the signal leaking out is low, then the ingress going the other way also will be low."
For one thing, downstream leakage is measured in the VHF midband, and reverse path ingress occurs in the 5-40 MHz band. The frequency response of the leak is not going to be flat across the spectrum. As well, the amplitude of the signal inside the cable likely will be much different than the level of an interfering signal outside the cable, even if both are on the same frequency.
So, the typically low drop levels of 0 dBmV to maybe as much as +15 dBmV may indeed produce a low-level leak, but a high level over-the-air signal entering the drop through the same leak may produce a high level of ingress.
Reality check
Anyway, it would be nice if every system had really good drops, but many don't. Another unfortunate reality is that cable operators willingly spend millions to upgrade their networks to state-of-the-art hybrid fiber/coax (HFC) architecture, yet do little or nothing about the drops. Granted, some will spend another 10 percent or so to take care of the drops at the same time, but from what I've seen, most don't.
So where does that leave things? Fire up the reverse, align it, and pull out your hair fighting ingress and noise problems.
An effective fix is high-pass filters. Install them throughout the system, and the reverse path spectrum will look, well, clean. Any junk that's left over most likely will be in the distribution network, and you should be able to sort that out fairly quickly. This is a heck of a lot cheaper than replacing a bunch of bad drops.
Then, as a given subscriber signs up for your new two-way cable modem service, remove the filter from that particular subscriber's drop, and upgrade or replace that drop as necessary. Even better, take the high-pass filter you just removed from the tap port and move it to the side of the house. Put it on the splitter output port that feeds the TV sets, videocassette recorders (VCRs) and FM tuners. The port that feeds the cable modem can be left unfiltered, providing a full two-way path for that portion of the drop only!
I've seen this practice done in a number of systems, and it works. The filters keep drop-related junk out of the reverse. It's possible to have a 5-40 MHz noise floor that looks like something you'd expect only in a lab environment.
But what about retail?
"Doesn't this take a lot of wind out of the DOCSIS retail model?" you ask.
Yes, it does, from the perspective of self-installation and automatic provisioning. But let me ask you this: How many of your subscribers have cable outlets near their personal computers (PCs)? Given that most probably don't, how many subscribers would you trust installing their own outlets? You know, an installation that will be done correctly and not be a source of leakage or ingress.
Hmmm, I don't see any hands out there. That tells me it'll still be necessary to send someone to the house to install the outlet, at least until the day when properly installed cable outlets are ubiquitous. While the technician is there installing the add-outlet, the high-pass filter can be moved from the tap to the side of the house and the rest of the drop cleaned up or replaced as necessary.
What about systems with addressable two-way impulse pay-per-view (IPPV) converters? How can high-pass filters be used if the filters are supposed to block reverse signals coming from the homes? Won't they also block the reverse IPPV signals? Yep! To deal with that, our friendly filter manufacturers have available what are called windowed high-pass filters. These are almost the same as regular high-pass filters, except for a narrow bandpass window in the reverse spectrum. In other words, the filter blocks all of the reverse spectrum except for a small part centered on the IPPV converter's upstream carrier frequency.
In all likelihood
If your system happens to be one that's well-designed, built and maintained, is tight, has virtually no reverse ingress, and does all of this without filters, please accept a tip of the ol' hat. But if your system is like most that I've seen, the reverse path can use all the help it can get. High-pass filters are one of the many tools available that will make your job easier and your system's reverse path more manageable.
Ron Hranac is vice president of RF engineering for Denver-based High Speed Access Corp. He also is senior technical editor for "Communications Technology." He can be reached via e-mail at .
 Back to November Issue
Access Intelligence's CABLE GROUP
Communications Technology | CableFAX Daily | CableFAX's CableWORLD | CT's Pipeline
CableFAX Magazine | CableFAX databriefs | Broadband Leaders Retreat | CableFAX Leaders Retreat
|
