RF network AMA

00:00:01 happy Friday everyone and welcome back to our Back to Basics Series today we have another special ask me anything AMA session with the one and only Ron Hranac that’s right we’re back for round two giving you another chance to ask your questions about RF impairments troubleshooting best practices and of course what not to do in your cable plant a shout out to the Great Valley Chapter for arranging the glom so you can get credit and the many other chapters who contribute to the success of SCTE and its Mission please do write in the chat Which chapter you’re with if you’re joining from a chapter today for those new to the channel I’m Brady Volpe founder of the Volpe firm and chief product officer at open vault and as always I’m joined by the legendary Ron Hranac a 50-year veteran of the cable industry who is a trusted resource on all cable matters all right Ron let’s dive into it welcome back are you ready for another round of questions well thanks Brady and happy Friday to you too and happy Friday

00:01:06 to those who have tuned in to today’s live stream this is fun I’m looking forward to it the last time we got together we did a did a an ask me anything and um we had way more questions than we had time to answer and I was surprised at how quickly the time went by and I suspect the same thing’s going to happen today um just a a point for those who are um planning to ask questions keep them RF related I do not do Doc’s questions I’ll defer to Brady and John Downey and others for those kinds of questions but RF stuff and I think last time we had somebody who might have been a cable TV subscriber complaining about their cable service and we can’t help you with that you need to contact your cable company if you got issues there but RF stuff we’re here to address your questions and and have a good time today indeed indeed and shout out to Sean Lester in the chat room there uh Sean thank you especially in your chapter cuz uh you you guys definitely helped us with you know putting together uh the integration with the SCTE for today’s session so uh

00:02:09 definitely throw your questions in the chat room there guys and we’ll start those r i I want to start off because I you know one of the things um Ron myself a number of other people are always busy working at cable Labs putting together new content for the is industry one of the things we’ve been working on most recently Ron is um a discussion on noise and I I thought maybe we would um enter into that uh you know what what is uh what are some common sources of noise that you see Ron you know is there just

00:02:40 one source of noise that the technicians have to manage when they’re when they’re trying to troubleshoot the Upstream I thought I’d throw that out to you oh that that question is is definitely one that falls into the category of it depends and and I don’t mean that tongue and cheek I’m very serious about that because the what you know how do we Define noise and what is it typically in a cable network noise is usually meant to be thermal noise generated in say a 75 ohms impedance or in an amplifier and

00:03:14 that contributes to what we call Carrier to noise ratio it contributes to something called noise figure which is a parameter that amplifier manufacturers put in their spec sheets and and can measure in fact in the NOS working group that I chair um we working on a a draft technical monograph about noise figure and that one uh hasn’t been reviewed by the group yet but will in the very near future but it’s uh it covers some of the the stuff from way way way back so that’s the the fundamental stuff but

00:03:46 there’s an interesting type of noise that also appears in our cable networks and this started to show up as the industry migrated from all analog TV channels in our networks to a a mix of analog and digital signals and then eventually to mostly your All Digital signals and that’s basically Distortion or noise-like Distortion comp it’s it’s known by a couple different names inter modulation noise composite inter modulation noise composite inter modulation Distortion and in of course in the old days of all analog networks

00:04:19 operators were very familiar with composite triple beat and composite second order distortions which were groups of discrete beats that appeared at certain places in the Spectrum but in all digital Networks those distortions don’t go away they just take on a different appearance and they look like wideband noise so what we get in our digital networks today is a combination of thermal noise and this these these noise-like distortions um there’s no way to visually differentiate them on a spectrum analyzer and where you’ll

00:04:51 typically see this is as you look on a spectrum analyzer at say the carrier noise ratio and you start at a low carrier noise ratio and start cranking up the signal levels you’ll see the car noise ratio get better which is expected but you’ll eventually reach a point where the output levels of the active device like an amplifier um generate um a sufficient amplitude of this noise-like Distortion where it starts showing up in the noise floor and mixing together with the thermal noise and now the as you continue to increase the

00:05:20 levels the noise floor increases instead of decreases so the so-called what you thought was a carer noise ratio starts getting worse and here we’re actually looking at what’s called composite noise which is a combination of thermal noise and noise-like Distortion so we have this carrier to composite um noise ratio um and unfortunately the only way you could differentiate thermal noise and and these noise-like distortions would be to shut off all your active channels and then you’ll see the noise floor drop

00:05:49 and all that’s left is the thermal noise now underneath the umbrella of this composite inter modulation Distortion or composite inter modulation noise the noise- like distortions we find things like um common path Distortion that tends to manifest in the return path and in an all digital Network it looks like an elevated noise floor so those are the common types of noise in the network where things get really interesting with noise is when we get Ingress interference from different sources for example Powerline Gap noise

00:06:22 um maybe um interference from uh other noise sources in the over thee a environment bursty noise sources impulsive noise sources automobile ignition static from lightning um you know the list goes on and on and on and many sources well yeah and broadly speaking we tend to call these things noise I prefer to call the external stuff um Ingress interference be and to differentiate it from thermal noise and composite inter modulation Distortion um but yeah this stuff can come from all over the place inside our networks and

00:06:55 outside our networks right so the thermal noise you’re seeing that that’s like generated by the RF amplifier maybe even an equipment in the headend and these other noises the overair noises Gap noise things like that is that’s external to the RF Equipment that’s external to our networks but um in addition to amplifiers generating noise a 75 Ohm resistor at room temperature generates noise um that’s thermal noise um I I don’t want to get into the math behind that but it’s basically caused by the the thermal agitation at the atomic

00:07:26 level as the temperature increases the atomic part particles jiggle around and that generates a measurable noise and in a in a 75 ohm impedance so a 75 Ohm resistor if we measure the noise power in a 4 megahertz bandwidth uh that because that’s the that’s the bandwidth for analog ntsc TV channels it’s just one I like to keep in mind I mean if you use a quam channel it’s a different bandwidth but use that bandwidth the noise power bandwidth for an ntsc analog Channel at room temperature and we and I’ll I’ll use the

00:07:59 uh the room temperature um defined in the that’s used in the freeze equations and that gets into noise figure that’s 290 Kelvin so roughly 62 is degrees Fahrenheit um that generates a a noise power with an amplitude of minus 59.2 dbmv if you want to use 68 degrees Fahrenheit for room temperature which many do then it’s minus 69.6 dbmv and you use that you plug that into an amplifier and then you measure the the carer noise r at the output of the amp and you find that um you’ll find that wow the carer noise

00:08:34 ratio is worse than it is at the input and that’s because of the added noise from the amplifier that which is the noise figure right okay and um so I know sh the questions coming in from Shawn Lester but I think there’s actually a group of people in with uh in with Shawn there so the question is Ron years ago you wrote an article and referred to this as carrier to junk ratios I humbly submit we use that that moniker going forward I have used that and I still use that however in in things like the DOCSIS

00:09:08 physical layer specifications and various St standards and recommended practices and operational practices um in those types of documents the the powers of B tend to frown on um things you know terminology like Carri to junk and although I’ll sometimes use it as a with quotation marks around it to indicate that it’s basically carrier to the stuff we don’t like or the undesired stuff which could include noise and Ingress and you know any other kind of interference right but yeah I like carrier junk it’s it it’s very

00:09:40 encompassing and tends to describe tend to describe a whole lot so so you mentioned that you know RF amplifiers resistors all these make some amount of noise and and I we kind of call that like the noise floor and I I would I would argue that we would say that noise is most of the time reasonable we’re going to have that noise in the plant but what is probably more unreasonable what we’re more focused on is I I’ll use uh from from Sean’s group the carrier to junk description that becomes more problematic for the for um you know for

00:10:15 our subscribers for our equipment what what is what are we looking for when we’re trying to differentiate between our typical thermal noise floor that you know the floor that is the noise that is always there in the background and the more problematic noise that is coming in from maybe signals in the outside air or you know prob or even in in from subscribers homes when does it become a problem how does how do we see that as a problem what we what should we do oh gosh there’s more there’s more it

00:10:47 depends than the answer to this question and but it’s a really good question Brady um and you’re you’re absolutely right when you say that most of the time that thermal noise is something you know it’s there and it’s usually not a big deal but it can be a big deal if the signal levels are too low then your carrier noise ratio is out of whack um and you’re going to have degraded performance in a variety of ways one if you think think about analog TV channels where you have degraded carrier noise

00:11:16 ratio strictly because of thermal noise maybe because the signal levels are too low relative to the noise floor then you see snow or grainy pictures in analog TV channels in in um digital video you don’t see that degradation as with analog channels as the carer noise ratio degrades let’s see on a DB for DB basis as you as you reduce the the RF signal level or let’s say increase the the attenuation in an amplifier Beyond where it should be have too much attenuation um you get this gradual degradation in the picture quality but

00:11:49 in you get snowy pictures in in analog snow pictures they get they get slowly get snowier and snowier or grainier and grainier because of the the U the decrease in the carrier level relative to the noise power with digital signals you don’t see that gradual degradation in fact with digital signals because of Forward Air correction and other things um you see Crystal Clear sharp pictures I mean they’re perfect and then as the Carri noise ratio continues to degrade um the receive modulation are error

00:12:20 ratio continues to degrade and then all of a sudden you reach the cliff point or the cliff threshold where um you start to see maybe some tiling in the digital picture and then they’re G and then another half DB degradation and it’s gone y it just it just goes away um with with uh voice let’s say voice telefony you might hear roboticized voice as the carrier noise ratio degrades because um packet loss is increasing and when the packet loss gets beyond the error correction capabilities of of U the transmission format being

00:12:55 used um the digital voice goes away too but you get that roboticized voice where it can’t fix the broken bits anymore and and things go out of whack but digital gives us that advantage of of um being able to provide very good quality service um till you hit that crash point so that’s a plus but it’s a minus too because just by looking at the picture quality or listening to the sound quality you have no idea how good the performance is you don’t know how close to that Cliff how close to the cliff you

00:13:23 are you just you don’t and when you reach the cliff it’s gone it’s just plun it’s it’s gone yeah so that’s that’s where noise can be really really problematic um so we have this guy named John Downey who’s asking a question he says please touch on Noise near noise test and how the test equipment could have poor dynamic range and or signal level that is too low for a good reading so I think noise noise Spectrum analyzer I’m gonna say John first welcome I’m glad that you’re able to make um and you

00:13:59 you you bring up a good point about Noise near noise correction um Spectrum analyzers as an example have horrible noise figure um when we look at say a trunk amplifier the noise figure might be eight or 10 DB somewhere around there that’s pretty typical for the amplifiers in the outside plant but a spectrum analyzer It’s Not Unusual to see a noise figure of 25 DB to 30 DB I mean that’s the nature of the Beast um the problem is when you’re using say a spectrum analyzer to measure carrier noise ratio

00:14:28 and the true could be the same could be true of using a signal level meter measuring a carrier and then measuring noise and then doing a bandwidth correction if the noise or sorry if the carrier level is too low that is the carrier to noise ratio that you’re seeing on the on the display is too low it’s not so much the dynamic range of the spectrum analyzer that’s an issue but the carrier level is close enough to the noise floor where um when you measure the carrier level you are measuring the carrier level plus

00:15:00 noise because the noise level relative to the carrier level is is high enough that in the resolution bandwidth of the instrument and in the the detector stage of the instrument it’s making a measurement of noise and carrier Power PL they’re adding together and now it’s it’s giving us an erroneous measurement you get an erroneous measurement and realistically what you need to do is find out is is the displayed noise floor on say a spectrum analyzer the noise floor of the test equipment or the noise

00:15:31 floor of the cable network and the cheap and dirty way to figure it out temporarily disconnect the cable that you’re using uh from the Spectrum analyzer and watch the noise floor that’s being displayed if the noise floor does not change that’s test equipment noise you’re measuring you’re not measuring the system noise floor if however you pull the the input from the Spectrum analyzer and the noise floor drops that’s a good thing but if it doesn’t drop at least 10 DB you’re measuring mostly instrument noise with

00:16:05 some outside plant noise in there too and when you’re in that range where the carrier noise ratio or the apparent carrier noise ratio is low um you either need to use a spectrum analyzer that can take into account that and use firmware to do a noise near noise correction or you can use on an ex look at an external chart it’s a graph that shows uh what the the measured car oise ratio is and then how much um correction you have to apply to that in terms of of DB um so you have to apply it manually to your

00:16:36 measurement to make sure you get an accurate measurement and John suggesting uh perhaps the need for a pre-amplifier in front of the analyzer in some cases to get your signal high high enough out of the thermal noise floor of the amplifier well yeah and that can work sometimes uh sometimes it may not it there’s some it depends that that’s in play there too but if you let’s say you disc connect the input from the Spectrum analyzer and the noise floor does not drop at all or it doesn’t drop at least

00:17:05 10 DB my recommendation when you face that condition is first see if you can find a hotter test point and plug into that hotter test Point repeat that measurement or that little test and if you if you get um the noise floor dropping 10 DB or more then you’re good to go as far as measuring carrier noise ratio at that point you’re going to be off by I forget the correction factor at 10 DB it’s around a half DB or so you’re going to be very close um but a preamplifier can be useful um you need to have a very low noise figure

00:17:34 preamplifier either external to the Spectrum analyzer or perhaps one that’s built in um but you need to take in take into account that that that is going to amplify noise and the signal a low noise preamplifier uh is going to get us into the world of the freeze equation where that first active stage is going to play a bigger role in in what your overall noise factor is and then the effective noise figure is so um you want to make sure you plugging in a preamp may or may not help it it could depend on a preamp

00:18:03 too well well it depends on the preamp but it depends on what the signal level is coming into that preamp even if it’s low that may not be enough but you may also have to use a band pass filter and other things with that so you don’t overload the Spectrum analyzer so John brings up a good point all right and from Sean Lester’s group he says over the past few months I’ve noticed an increase in LTE Ingress in many customers across the country particularly in bands 11 and 13 he goes on to say have you seen this uh have the

00:18:32 carriers recently increased their transmit power level to improve 5G performance other than tightening our Ingress mitigation efforts what other recourse do we have you don’t have any other recourse you need to tighten your plan um no it’s not a it’s not a function of the LTE signals increasing in power they have FCC limits like any other um over the a licensed service that’s transmitting signals into the over the a environment what you may be seeing is an increase in the number of LTE towers as as they move into the use

00:19:05 of higher and higher frequencies to get higher data rates um what happens of course is they they add more Towers to improve the coverage U while still being able to provide let’s say 5G type Services where they’re providing higher data rates so it’s not so much that they’re increasing their power as it is they’re May well be more of them um and so that means you just got to pay more attention to making sure the plant is tight and and and the only way to do that is the tried and truee methods of

00:19:33 let’s find out where leakage is getting out and then that usually indicates where Ingress is getting in and you got to fix those things and even and sometimes you may go out and measure leakage and say well I don’t see anything anything above 20 microvolts per meter it’s it’s lower than that um but the reality is with very few exceptions when Ingress gets into the plant whether it’s Downstream or Upstream it usually gets in by multiple low level points sometimes it’ll get in by what I call a high level point but

00:20:03 usually it’s a bunch of low-level points and that could be a loose housing to housing connector a loose face plate on a tap a loose F connector on the drop and then you start multiplying these in a neighborhood where you’ve got a variety of these low level points and this Ingress gets in Via multiple points and then of course inside the plant the magnitude of the Ingress increases to where it’s causing harmful interference to your signals so you’ve got to go out and find all these lowlevel points and

00:20:27 fix them and sometimes that can be real challenge tightness connectors and terminate uh Jeremiah D says can you shine a light on Annex B 6 megahertz versus 8 MHz Euro DOCSIS and other 8 megahertz quam standards the landscape that drove these standards has changed are we doing the best we can staying with six megahertz um oh six megahertz or 8 megahertz or in some countries seven megahertz right um the bandwidth for the channels used in cable networks is largely a leftover um I don’t want to call it

00:21:06 artifact but you know sort of like that it’s it’s largely left over from the channel plans used by broadcasters TV broadcasters in the over theair environment so in Europe and and many other locations around the world um the channel the over the a analog TV channels used either 7 MHz or 8 MHz bandwidth depending on the frequency range depending on the country um 8 megahertz was very common throughout Europe and in the Americas typically used 6 megahertz there are a couple other places that use six megahertz but

00:21:38 that that goes back to the development of ntsc analog TV channels in fact there’s a standard that defines the six megahertz Channel Plan used in cable networks and that that standard is under the opes of CTA the consumer technology Association so those of you that say you’ve got channels on your network you’re about 20 years out of date that that designation was deprecated two decades ago uh and was replaced when the no longer managed the channel standard and it went over to the Consumer Electronics Association the CA

00:22:12 and the ca changed its name to the CTA and then the name of the standard change but it is CTA 542 D and then there’s some other stuff that come after it that uh something like r223 or something but um that defines the six megahertz Channel plan so for digital signals in the Americas we use the same six megahertz Channel plan um for both analog TV channels and digital signals so single carrier quam channels in Europe they use 8 megahertz um that’s not going to change anywhere it’s they’re going to they’re going to

00:22:48 continue in in different parts of the world to use those band plans that they have always used um the technology supports it the cable modem support it the cmts is support it the standards are defined and Euro DOCSIS Define um those bandwidths based on on the location so Euro DOCSIS covers stuff in Europe and elsewhere using 8 megahertz bandwidths and in North American North and South America and some other places six megahertz now is that the most efficient use of the spectrum the answer is no um spectral efficiency or bits per second

00:23:20 per Hertz yes isn’t bad as you I mean with a digital signal as you increase the modulation order you increase the number of bits per second per her that’s good you can get pretty high data rates doing that but you’re still not getting the maximum spectral efficiency where the the better spectral efficiency comes into play is with DOCSIS 3.1 in later using ofdm and say the downstream and the Upstream but in particular in the downstream um with ofdm you’re not limited to six megahertz or 8 MHz Channel

00:23:50 plans right and the the bandwidth of of a DOCSIS 3.1 ofdm signal can vary from a minimum of 24 megahertz to a maximum of 192 megahertz and you don’t have those gaps in between you know every six megahertz or every eight megahertz so the spectral efficiency is much better which supports higher data rates and then you throw in the much more powerful forward eror correction technology I was going to say  which is the better error correction over  Reed Solomon that gives you about gives you about a 3D

00:24:18 about a 3db bump over um the Reed Solomon error correction used with with Legacy single carrier quam channels so we’re never going to get away from six 8 megahertz spacing that’s that’s always going to be there but as the industry M migrates away from what I’ll call Legacy uh docus and digital video transport using those B those Channel bandwidth and moves more to ofdm with Doc’s 3.1 and 4.0 then you’ll see the improvements in spectral efficiency and and um the much higher data rates supported by that

00:24:53 y John says we’re moving to ofdm anyway so no 6 or 8 megahertz either way spacing will still be there but we also we’ve also removed EG 2 encapsulation that is in DOCSIS 1.0 to 3.0 um so well he’s got one thing he’s got one thing wrong in there um he says so no six or eight either way with the six yes that’s still there with ofdm but it what it applies to is the ofdm channel power which is defined in DOCSIS as the power per CTA channel that is exactly the way it’s written it is power per 6 megahertz um that so that

00:25:33 doesn’t go away but of course then we get into the 6 versus 8 the Europeans said well we don’t like the power per CTA Channel because we don’t use CTA channels so we want to convert the doxis channel power or the ofdm channel power to power per eight megahertz so they’ve got to do a correction to that but if you look at at modern field meters they they can display the power per six megahertz um I know on the the viavi 1 expert um it has this graph of the ofdm channel it shows these little blue horizontal lines that represent

00:26:04 graphically the power per six megahertz or the power per CTA Channel but that’s actually spelled out in in the doc’s physical layer spec so that doesn’t completely go away I know where John’s coming from and I got to give him crap yeah he says you got them on that one so I know John and I we’ve known each other for way too long yeah um Caleb says I’m seeing the LTE getting into in at the headends well-maintained and well terminated headends oh yes and that that’s getting in um and this is not new this this was

00:26:38 discovered several years ago particular particularly at headends and hubs sites that were collocated with say a tower next to the headend building where maybe the cable company was was leasing space on the tower to one of the phone companies and they had um cellular transmission equipment there and they were finding um that they were getting interference from LTE signals directly in some of the optical transport equipment they were getting it directly in their test equipment even with nothing connected to the input of those

00:27:09 devices and they put a 75 ohm Terminator on the F connector on the input to the their test equipment or their other instruments and the interference was still there the interference in that case has to do with the shielding or I won’t say lack thereof because there is some shielding but the shielding quality of those devices and I remember remember the first time I heard of this in a major operator’s headend site where they got this kind of interference um they said that the they were estimating the field strength and

00:27:38 the head end and and the field strength just sounded ludicrous um so I did a quick back of napkin calculation with one of the engineers from that company and I said no you should be more like this because you have to take into account um the the the main beam of the the antenna is going out this way and you’ve got side lobes and they’re going to be and I think I picked the number you know 10 DB down and then you got some attenuation through the cinder block wall and stuff and I so I I calculated the field strength in the

00:28:03 head end and they that same operator later brought in a I think an engineer from Rodi and Schwarz who had a very expensive piece of test equipment that that could accurately measure test equipment or measure field strength and their measurement was almost dead on with what I had calculated of course it was which is still high it was still high but but it wasn’t nearly as bad as somebody had had tried to estimate or measure and um but the gist of it was fair cage for the head it was Optical it was Optical it was some Optical

00:28:33 transport equipment you know Downstream lasers that was affected by this and of course at the time I worked for a vendor that made that that kind of stuff and I said is that our product and they said no it’s somebody else’s and the manufacturer of that equipment had to um do some tweaking to the design of the shielding on their on the the enclosures basically the modules and everything else so they had to put fingers new finger stock in and a bunch of other stuff to improve the shielding um the

00:29:01 company I worked for had done testing they tended to be so conservative on stuff they were their stuff was rated for 10 volts per meter field strength which is just an insane field strength um but that’s kind of beside the point no the issue there is if you see that um getting into equipment in the headend or your test equipment the the culprit is not the LTE signal itself it is the shielding Integrity of the affected equipment yep and that’s an expensive fix to go back and do well you go to the

00:29:33 big say with Spectrum analyzers you go to the big names like well HP now what Agilent and then what are they today um rodian Schwarz and others they they design um kyite is what became um they designed their stuff to to be usable in in high field strength environments typically to 10 volts per meter or greater which is very very high very very high level shielding and robustness in in ambient field strength all right so Sean was uh going back to the uh 6 MHz 8 MHz ofdm discussion he’s saying is is well aware

00:30:10 he firmly believes bits per second per Hertz is one of the best measurements we have on how well a DOCSIS network is performing um he also says uh he will take he will call it to my grave EI all right I see a lot of um even current documentation I see out there with eia channels still being used I think it’s going to take a long time to change old habits are hard to break I’ll I’ll Grant Sean that but it is CTA that the eia designation did it literally went away and was deprecated over 20 years ago so that terminology is

00:30:48 no longer accurate it is the channel plan falls under the CTA standard and has for a long time all right local trees good to see you in there um thanks for joining again folks get your get your questions in there I’ll ask Ron a couple more more here Ron one I want to ask you on is group delay because I always that’s a very common question I get from people what is group delay what causes it and how does it you does does it even cause an impact is it something we should worry about in our networks um well last question first the

00:31:20 answer is yes it’s something you should worry about but in some cases it’s it’s something you may not be able to do anything about um where does it occur group delay Distortion and I’ll I’ll give a kind of a high level definition of it here in a moment tends to be more prevalent or occur more more often at at the edges or near the edges of a band pass in a spectrum so think of let’s say the Upstream spectrum is the equivalent of a band pass filter so we we pass signals from 5 megahertz to say 42 MHz

00:31:53 or whatever but I’ll use 42 MHz as an example down at 5 megahertz we have that frequency response rolloff um which attenuates signals below 5 MHz doesn’t get rid of them completely but it attenuates signals down there and that’s a function of a couple things one is the frequency response of the amplifier circuitry and the other is the so-called lightning filters that are used in amplifiers to help minimize you know the effects of surges and noise from lightning that’s stuff that’s built in so there we’ve got

00:32:23 the equivalent of a band pass filter on one end at the other end we run into the diplex filter rolloff so the diplex filter the part that passes the return path is a low pass filter so there’s some rolloff at the upper end so now we have this band pass filter and near those band edges same’s true in the downstream the the we have these rolloff areas where the group delay tends to be worse all right so all that said what is group delay so think of um think of a if visual visualize this this is something

00:32:56 that that Tom Williams who’s now reti retired used to be with cable Labs um described it this way and this was such a good description I created a an animation in a in a PowerPoint slide deck that I did on group delay Distortion and he said imagine um a track where you’ve got a bunch of athletes running down the track to the finish line from the start line to the finish line and they’re you more or less they let’s say they’ve got evenly matched athletic abilities they’re all more or less going to reach the Finish

00:33:25 Line at the same time but if you have more Runners than you have track where do those Runners on the edges go well they got to run through the ditches next to the track through the weeds and the mud and stuff so they get slowed down relative to the runners on the asphalt and the runners on the asphalt arrive at the Finish Line sooner than the r getting bogged down and all that getting bogged down they’re running through the ditches so the analogy is that the ditches are like the rolloff areas in

00:33:55 the Spectrum and the group delay Distortion means that at frequencies where the rolloff occurs because of some factors I’m not going to get into because it’s going to get Way Too Deep the time delay through that part of the spectrum is greater than the time delay through the flat part of the spectrum so visualize that five roughly 5 to 42 megahertz return path and again this would apply to a that’s your nice running track where everyone can run you got a nice running track and all the frequenc arrive at the destination you know

00:34:31 whether it’s going from cable modem to headend or whatever y they all arrive more or less at the same time but the one the the frequencies in that rolloff area because of that nonlinear amplitude versus phase relationship that’s going on there they take those frequencies take a little bit longer typically measured in in tens of nanoseconds or hundreds of nanoseconds they arrive later um and if you’ve got group delay Distortion within a digital channel some parts of that digital channel arrive later than other parts of the digital

00:35:06 channel and the result is you get this we’ll call it a kind of a time smearing but you’ll you’ll get um a degradation in the receive modulation error ratio and the group delay Distortion that exists in diplex filters and amplifiers and stuff that’s you can’t do anything about that that’s there you you can’t get rid of that I mean you can reduce the Cascade depth and you can improve the group delay Distortion performance but you can’t get rid of it that’s a matter of physics and that’s the way the

00:35:37 edges of those filters work that’s where um something called adaptive Equalization or adaptive pre-equalization comes into play it can compensate for that it doesn’t get rid of that Distortion it compensates for it so right an Adaptive equalizer let’s say in the upstream or pre- equalizer the effect there would be if we go back back to the the athletes on the track the athletes on the track are slowed down by the group or by the Adaptive equalizer or pree equalizer to let the runners in the ditch catch up so that when they all

00:36:10 arrive at the destination they all get there at the same time so that’s kind of what’s going on with adaptive pre-equalization it it think of it as sort of slowing down the runners in the middle of the track and letting the other ones catch up so that all the frequencies arrive at more or less the same time that’s a great anal for group delay I got to sa R I’m going to have to use that in the future when trying to explain i’ I’ve got a I’ll try to remember send me an email or something and I’ll send you the Powerpoint because

00:36:37 it’s got I did the animation um that shows that and it it really I think it really helps to describe what’s going on yeah so um and and I mean fundamentally we know group delay if if it’s not compensated for by The pree Equalizer if it’s so high it it can cause um cause issues for the modems uncorrectable code war or eventually it could even cause the the modem to lose that channel that is uh being impacted by group delay so we know when the when group delay Distortion occurs it trashes the modulation error Ratio or you know the m

00:37:08 rxm is really what it is um and if the rxm gets low enough um you can start to suffer bid errors and packet loss and and pretty soon the fech the Ford air correction can’t fix it and now you’ve got packet loss yep absolutely all right uh so uh look says happy to be here Nicholas P says group delay is definitely a hard one to understand great explanation and and I agree that’s it’s it is not something that’s easy to explain um but we do we do see issues with it and something that um you know we uh I I first really

00:37:47 started looking at group delay when I got into P&M proctive network maintenance because it would be something that was show up there but because we’re using pre-equalization we could see really high group delay vales and if you’re familiar uh there’s a doc specification that says your group delay must be less than 200 nod per megahertz but it’s very common with pre-equalization to see group delay values at like a thousand Nan per megahertz Way Beyond the spec but the the subscriber not complaining their

00:38:16 modem is you know is behaving quite well because of pre-equalization compensating for that now here’s a here’s an interesting thing that folks that have been in the cable industry a long time going back into the days of analog TV channels may not be aware of it but they do have experience with group delay now not only did the amplifier manufacturers specify group delay typically St channel two and channel three but in the forward path but in Channel group delay manifest is something called chrominance to

00:38:48 luminous delay Distortion and inside of an analog TV channel I’ll go to an ntsc Channel visualize you’ve got visual carrier here and you’ve got the color subcarrier here and then the oral subcarrier over here and the luminance information is carried in the visual carrier well not carried in the carrier itself but in the luminant side bands that are spaced 15.73 4 khz apart uh a a little bit on the Lower Side Band that vestigial Side Band and then going up the other direction then the chrominance

00:39:18 information is is carried in the chrominance sidebands as phase and amplitude information kind of like quam phase and amplitude but those um sidebands are spaced in between the luminant sidebands and they’re every 15.7 34 khz so now if if you have group delay inside of the 6 mahz channel the color information may arrive at the TV set slightly later in time than the luminance information and we get what was called the funny paper effect where you see the outline of the image and you see this color smearing off to the right

00:39:57 of the image ghosting right well no this is it looks it resembles ghosting but with ghosting the entire channel is offset In Time by the delay that that second path took okay the signal path took in this case within the channel because of the group delay Distortion there the color information which is three and a half megahertz roughly above the visual carrier took a little bit longer to get to the TV set than the luminance information did so the color does doesn’t register exactly with the outline of the image and you get this

00:40:31 color smearing or an offset in the color kind of like looking at the funny papers in the Sunday Funnies for when people used to get those you might see that during the printing process the color was offset from the outline of the cartoon and that’s why it was called the funny paper effect so people saw that and and may not have known what it was but that was a type of group delay Distortion where it really showed up was in character generator channels which has real fast rise time and fall time on those characters and

00:41:00 you’ll see that smearing off to the right of the lettering that’s that’s caused by that that group delay Distortion in the channel definitely saw it many years ago so Jeremiah D says can group delay affect SNR does modem traffic arrive at the wrong time and step on other modems in severe cases not not so much in the way that you’re describing I mean that’s a that’s a really good question um the some people call rxm SNR they’ll say what’s my Downstream SNR Upstream SNR when they really mean me modulation error ratio

00:41:37 group delay Distortion if we think of SNR as basically being carrier to noise ratio in the frequency domain group delay Distortion doesn’t affect that so the answer there is no it does affect me and if you think of if you like to call me SNR then yeah it would affect that but it’s it affects the rxm and what it’s doing is it’s smearing the data in time um because of that inchannel Distortion it’s not so much that you’re going to get modems stepping on one another because at that point you’re you’re getting into modems transmitting

00:42:10 at different times so no it’s not going to cause that it’s going to cause a degradation in the me or rxm specifically of that signal which which can if it’s severe enough um cause packet loss but TP hopefully the the Adaptive Equalization will compensate for that all right and Nicholas follows up follows up with how bad is group delay in the downstream I was in a discussion with a colleague recently and we didn’t quite know how it would occur or be mitigated well it’s the same thing I described a little earlier in the

00:42:44 downstream let’s let’s visualize now the downstream starting at channel two so 54 megahertz um on the amplifier manufacturers would publish a group delay specification for the diplex filter and and that’s where it’s affected by is the D Lex filter largely that separates the downstream from the Upstream so you’ve got that rolloff just below channel two so the group delay Distortion tends to be worse there so manufacturers would specify group delay in channel two maybe also channel three you can’t do anything with amplifier

00:43:14 adjustment or anything else to mitigate that that you can’t get rid of it that’s an that’s inherent to the design of the filters that are used in the amplifiers what you can do is if you’re TR well what you make you want to make sure you do if you’re transmitting digital signals near the lower band edges make sure that that you’re using adaptive Equalization in your data receivers so your your cable modems have should have it on all the time and that should compensate for the group delay Distortion um where it can occur and

00:43:45 that could be at the lower band Edge if you have a suckout in the frequency response where you have that rapid amplitude versus frequency change you can have group delay Distortion there too so uh any place where you have the let’s say amplitude Ripple or the a suckout or you’re at the band Edge you can have group delay Distortion for things like suck outs and amplitude Ripple you can fix those the problems that cause those but for the group delay Distortion that’s inherent at the band Edge you can’t do anything about that

00:44:11 but if you have digital signals there the Adaptive Equalization should be able to compensate for it now you could reduce the Cascade depth which means you’ll reduce the cumulative group delay Distortion and you’ll be in good shape but that may not be practical you may not be able to go out and reduce the Cascade depth all right John says comic book effect Channel 2 since it’s the first channel the hack was to pre- distort 170 NCS the opposite way so you could double your amp Cascades um you could but you want to be

00:44:44 careful about doing that Beyond how it’s done in ntsc video transmitters like head-end modulators they have a predistortion in there U to compensate for that and I think it is a 100 nod um delay it’s it’s done in broadcast transmitters it’s done in modulators well I’ll say some modulators because when I ran the uh corporate evaluation Lab at Jones inter cable back in the 80s I used a a ton precision tectronics demodulator and and a VM 700 automated video test set so I could measure measure all that and I was surprised to

00:45:23 find that some head-end modulators did not use that predistortion that’s that’s used in some of the better quality modulators and in broadcast transmitters and the idea there is it’s it’s designed to offset um oh there was some group delay caused by what was it the the oral carrier Notch I think in the the video circuitry so you’re not getting video so anyway so that you’re not getting that interference I think is where that where that came from but yes there was a predistortion that was used in broadcast

00:45:52 transmitters and some of the higher end modulators and where a modulator that was low cost did not include that predistortion you could see the comic book effect and say where where in the heck where the heck did that come from very good John asked a follow-up question which I think is actually a good question John can severe standing waves cause group delay or maybe even a suckout Ron it’s not a suckout but it’s well not in the strictest definition but I I answered that previously but maybe maybe it didn’t come across but the

00:46:26 answer is yes where you have uh let’s say standing waves in the transmission line the coaxial cable that manifests as amplitude Ripple in the frequency domain yeah and where you have these um rapid variations in amplitude um versus frequency you can have group delay Distortion form there so you what you get is group delay Ripple and the the Adaptive equalizer should be able to take care of that in most cases unless it’s really really severe but then you’ve got to go out and find out um what go out and find and fix the impedance

00:46:59 mismatches that’s causing the uh the standing waves and the subsequent amplitude Ripple but yes severe amplitude Ripple can cause um group delay Distortion it’s group group delay Ripple y yeah I think that’s important to understand Jeremiah D says I’d hate to do the by anual proofs on that Cascade LOL and I think it’d be good uh you know what what do we mean by biannual and what are what are proofs or why do we do them twice a year biannual is twice a year what what is a proof Ron can you um well the the concept of

00:47:31 what’s a called a proof or proof of performance has been required in part 76 of the FCC rules for pretty much as long as I’ve been in the cable industry U that goes back to the early 70s um back then and into the 80s I don’t remember when the FCC rules Chang to twice a year but it used to be once a year and I would go out to Cable Systems in in the region where I was the regional engineer for Jones and do do or help with theof for performance tests and the the uh FCC rules in part 76 spelled out a number of

00:48:02 technical performance parameters in 76605 with measurement techniques described I think in 76 609 if I remember right for the analog ntsc TV channels carried in the system you had to do all kinds of of measurements as part of that proof of performance testing the FCC bumped that up to twice a year with the idea being that um you get an indication of what the cable network performance looks like in let’s say the summer when you do one of the proofs and the other one in the winter time when you do you do the second proof

00:48:35 so the idea was to make sure that the the analog TV channels and I think the intent of the rules was to make sure that the cable network wasn’t degrading the performance of the analog TV channels Beyond a certain amount and those proof of performance tests were done and were required to be done um to ensure that and you had to record all kinds of things like the can model last calibration date serial numbers and so on of the test equipment being used and also the the name of the person or person doing the proof along with th

00:49:07 those the the qualifications of the person or persons and then that had to be put in U put in writing and logged and and put in a file cabinet so if the FCC came to visit your cable system they might request a copy of your latest proof of performance test and measure your and your latest signal leakage logs and you had to pull that stuff out of the file and be prepared to hand it over for review Y and so that that does just apply to our um you know the our operators in the United States and I know that’s something Ron you strongly

00:49:36 encourage that people continue to do um whether they’re carrying analog channels or not uh for that so um Nicholas P taking us back to group delay how bad is group delay when going into the rolloff of an 862 MHz amp we are thinking of deploying an ofdm Channel around 862 MHz to 1 GHz should I be worried this is only for a brief time as am as amplifiers are going to be changed um the nice thing is at least in the downstream the roll off above the upper frequency limit of the C of the downstream Spectrum tends to be more

00:50:13 gradual than you would see it at the lower end so the group delay Distortion and there will be some but it’s not going to be as severe as it is at the lower frequencies um where you can see some issues might be in the feeder part of the plant where you’ve got a lot of and I I’ll just pick on a 750 MHz plant let’s say you’ve got a 750 MHz plant you’ve got the gradual rolloff in the the uh the amplifiers at 750 megahertz you’ve got the gradual rollof through the passives the Taps and you know Splitters directional cupers and stuff

00:50:43 and that might depending on the layout of the plant might get a little nasty looking at the upper end but surprisingly um I’ve seen cable operators carry DOCSIS ofdm signals in the rolloff region in a 750 plant in fact my neighborhood used to be served by a 750 plant for the longest time and there was a an ofdm signal when the local cable operator first rolled out docus 3.1 a few years back that operated in the rolloff and it worked fine the Adaptive um Equalization took care of that rolloff and group delay now as

00:51:15 you’re migrating to a a greater bandwidth um I don’t know that the group delay is going to be so much of an issue is just just managing signal levels and and other things as you as you make the the swap from let’s say the 62 amps to 1 gz amps yeah yeah I definitely agree with you at rolloff is a lot is is not so sharp as we would see with filters and things like that be a lot gentler so Ron I have uh I have one crystal ball question for you as we’re wrapping up the episode here um with the industry

00:51:44 moving toward dox’s 40 and fiber deep architectures lots of you know lots of more lots of talk of fiber do you think hfc will still be a dominant technology you know three years five years maybe even 10 years from now uh you know my favorite answer to so many questions is it depends um I think hfc is going to be here for the foreseeable future it’ll certainly be here in three years um it’ll be here in five years probably 10 years what you’ll find is that as as more operators roll out fiber to the home and I think the

00:52:19 European cable operators tend to be more aggressive about rolling out fiber to the home than we are in the Americas um but looking at even the major cable operators here in in the US I think all of them have at least some percentage of their homes pass served by fiber to the home now for the most part near as I can tell it’s in single digigit percentages but over time that’s going to increase and it’ll be over time and the reason is it first of all it still costs a lot of money to rip out an existing hfc Network

00:52:52 and replace it completely with fiber to the home if you’re doing a green field build that is BR new from the get-go there is no hfc there um fiber to the home makes more sense um but in where you’ve got existing hfc we can still do upgrades to the hfc part of the network and take advantage of the incredible capacity of coaxial cable a lot of people think well where does coax run out of gas I don’t know maybe a couple gigahertz um docus 4.0 supports Downstream operation to 1.8 GHz or higher how much higher can you go that’s

00:53:26 another it depends but the connectors call it the Hardline connectors the Hardline coax are all if you look on the spec sheets for the manufacturers are for the most part specified out to three gigahertz um so that’s going to give us a lot more expansion room for the foreseeable future but the other thing is well what about just the capacity of coax how how high can we go and the how high can we go in terms of the capacity I’ll start with the bandwidth for the upper frequency limit depends on the

00:53:56 phys physical dimensions of the cable and you get into something called the transverse electromagnetic mode of the signals passing through it and what that means is that larger diameter cables have a we’ll call it a a tem mode cut off let’s say 750 cable cuts off at about oh what 750 is or Sor sorry about 7.5 gigahertz half inch Hardline cable is good to about 11 A5 gigahertz Series 6 drop C good to I think it’s the math I did was about 26 GHz now the in the smaller diameter cables while they can handle higher

00:54:35 frequencies um at least from the perspective of of the of the capacity the attenuation is worse with lower or smaller diameter cables I wrote an article in Broadband Library a few issues back about the talking about the capacity of coaxial cable and I went through the math and showed the examples of what you could get and I and I figured that conservatively um you could easily get to 6 gigahertz for an upper frequency limit in a cable network and I did made some assumptions and said let’s just divide the forward

00:55:06 and return up in half at you know three gigahertz is the cut off for the return and six gigahertz for the the downstream with it starting at three and I said assume no diplex filter rolloff or or crossover losses and there are some amplifiers that can do that without diplex filters and I I’m trying to remember the the data rate I came up with it and I don’t remember if I used 4K quam or if I went higher but the data rates were something like 25 gigabits per second symmetrical it’s amazing with Co with coax so the capacity is there

00:55:39 that you know at that point we may say all right we don’t want to keep upgrading the the coax plant at that point we’ll just go to go to fiber but we don’t have to rip it all out that fiber’s got a or sorry the coax has got a long life ahead of it it it’s it’s it’s really amazing I mean I I always tell the story 30 30 plus years ago when I was starting out my career in cable I was telling my boss that I I don’t think cable has a future I need to go find a new job and he says Brady you know give

00:56:07 it give it a couple years there’s probably some future in cable here we are more than 30 years later saying yeah there’s still more more future in cable and I and I definitely think there is um so Nicholas says I did not do a Cascade test yet but with two amps I was getting a tilt of-7 DB this was impressive as it did not work perfectly it did work perfectly it did work perfectly yeah I mean that’s that’s the amazing thing that we’re we’re seeing with ofdm uh deployments Sean Lester says the engineers Mantra it depends yes it

00:56:40 depends there’s some black and white stuff in there where you know it’s got to be this answer or that answer but there’s a lot of it depends there’s a lot of gray in between Dominic welcome he says what frequency should you put your regular digital cable linear TV above 860 below and where should you put the doctor’s internet or the doc’s internet I believe yeah he says he’s at at the doctor’s office doing voice chat so he’s having so you know put your video I think we normally say keep your video as low as you can and put your

00:57:08 docs as high as your as high as you can you can do that but you don’t have to the nice thing is the doc’s spec supports um Downstream operation with dox’s ofdm channels all the way down to 108 mahz so you can put it there in fact I know of at least one cable operator that that did just that put their DOCSIS ofdm signal on the low end of the spectrum and and then their legacy stuff at the high end of the spectrum and they cranked up the ofdm channel 3db relative to what Legacy digital video would have

00:57:39 been I remember doing some some between midnight and 6 a. field test with a cable operator in one of their markets and then I said you know you guys don’t have to do this in the day night time you can do it in the daytime I we did one another field test in another Market in the daytime and they had done a lot of lab testing ahead of time um on the bench to make sure the equipment could support this configuration and it worked fine by putting the docks of stuff down low and the Legacy stuff the Legacy

00:58:05 digital video up high put anywhere you want uh but as Brady said most people put the this the uh Squam stuff down at lower frequencies and the dox stuff at higher frequencies but you don’t have to yeah hope things go well at your doctor’s appointment there Dominic uh Sean says Ron that’s exactly what I tell customers if you have hfc already it’s almost always better to continue using it usually cheaper usually cheaper and Ron John you like this one Nicholas P says for us as German dox’s providers

00:58:35 with 100K Subs we are migrating from a ubr 10K with MC 20 X20 cards using hfc to harmonic daa with doxes 3.1 we are also heavily focusing on fiber to the home with xgs Pawn those UV 10ks are still out there running still out there that’s or something yeah and and that’s fine keep using the technology get your money’s worth out of that um and doing what a lot of operators are doing they’re using the their legacy Big Iron chassis and they still use it and they’re they’re migrating to some type of or form of fiber to the home uh

00:59:14 usually a pond-based architecture and that works fine it’s a that’s ultimately I think that’s the end goal for the industry is to get the fiber to the home it’s just absolutely I remember people saying in years past well it’s another five years and we’ll be it’ll be more white you know there’s always another five years and we’re still not there yet but gra sandwich says remember when Clinton Gore said we’d all have F of the home by what was it year 2K y 2K here we are a quarter of a century later good

00:59:42 memory Sean Lester says he’s been recommending ofdm channels to be moved to the lower frequencies in all video and the higher frequencies so works fine yeah you can do it if you want you just said that so yes great chat everyone great conversations and great information Ron thank you so much that that was absolutely great knowledge that You’ shared with everyone we really really appreciate your insights they’re always valuable and truly appreciated um sraga just dropped in our ofdm is below 100 to 500 megahertz with a few at

01:00:19 350 until they swap out their old modems so um again thank you everyone who joined us live live your questions keep the discussions going um without you itd be pretty quiet on here definitely in an ama if you didn’t get your questions in do drop them in the comments below we’ll try to cover them in a future episode or answer them directly if you enjoyed today’s live stream please do hit that like button and subscribe turn on notifications so you don’t miss an episode and don’t forget you can catch

01:00:52 up on past episodes of our back to basic series on YouTube Spotify Apple and Google podcast thanks again Ron uh we still have questions coming in but we’ll catch up on those later so thanks and thank you Mia for moderating everything take care everyone my pleasure thanks all thanks so long