POWER LEVELS & SNR: A TECHNICAL PRIMER (Updated October 2016)
PART B - UPSTREAM TRANSMISSION
The so called "fibre-optic broadband" is more accurately referred to as "cable broadband" because at both ends (home and head end), termination is by means of coaxial copper cable. It is a nominal requirement that each active node (e.g. Optical Node), on the upstream path receives a minimum of 15 dBmv. This 15 dBmv is not rigidly applied because it is obviously impossible for all modems to hit the node at the same power. There is intermediate upstream amplification.
So let's look at two theoretical (but practically founded) scenarios. We will use the nominal 15 dBmv for illustration purpose. In Figure-1, the home being 150m from the street cabinet, the path to the optical node might comprise the following elements, each contributing to attenuation:
Figure 1: Upstream power from 150m to the cabinet
In Figure-1, upstream power would be 43.25 dBmv. It is assumed that the cable is connected to the 18dB tap point.
Figure-2 illustrates a home that is 50m from the cabinet and connected to the 22dB tap point. The upstream power would be 44.14 dBmv. If the coax cable goes to a 15 dB tap point instead of 22 dB, then the upstream power would be 37.14 dBmv, making 15 dBmv at the optical node. Coax attenuation varies according to upstream frequency (the higher it is, the greater the attenuation but for the small upstream range that isn't a huge problem) and temperature. Doubling the coax length doubles the attenuation.
Figure 2: Upstream power from 50m to the cabinet
Any impairments outside of these "known" attenuations requires extra power to assure 15 dBmv to the optical node. The CMTS manages this by simply deciding whether or not the CM has been able to communicate properly with the CMTS. If not, during initial ranging, the CMTS commands a 0.25 dBmv increment until communication is established. During a session, periodic maintenance opportunities ("keep-alives") occur. If impairments have arisen that disrupt keep-alive communication, the CMTS offers a series of further opportunities, each failure resulting in a T3 message. After 16 such T3 messages, the CMTS removes the modem from its polling list and a T4 timeout occurs with a subsequent modem reset and possible upstream power ramping as per initial ranging.
Modems to a single street cabinet are arranged (through differently attenuated tap points) to have as similar a transmit power as possible so as to preserve SNR; otherwise there would be wide SNR variations between modems close to the street cabinet and those further away. That said, and as already mentioned, the CMTS will command upstream power according to the attenuation (and SNR) seen for a given modem. So you can see why a ramping upstream power number is potentially something to worry about, whereas a stable, but high number without problems is nothing to worry about.
UPSTREAM POWER LEVELS AT THE MODEM
|
Per channel Upstream |
Acceptable 2 US channels |
Acceptable 3 or 4 US channels |
|
VM Recommended |
31 dBmv to 55 dBmv STABLE |
31 dBmv to 55 dBmv STABLE |
|
DOCSIS 3.0 Table B-15 |
23 dBmv to 55 dBmv 16QAM |
23 dBmv to 52 dBmv 16QAM |
|
DOCSIS 3.0 Table B-15 |
23 dBmv to 54 dBmv 32/64QAM |
23 dBmv to 51 dBmv 32/64QAM |
|
Seph’s Reasoned Power ** |
Lower than 53 dBmv STABLE |
Lower than 50 dBmv STABLE |
VM instruct their technicians that there is an upper power limit of 51 dBmv when the upstream comprises only 2 channels. Because in future everyone will have 3 or 4 channels, VM take the cautious approach; however the DOCSIS 3.0 spec defines power levels at which the modem MUST be able to operate.
** The “Seph’s Reasoned Power” level given above is based on the earlier discussion of the power level necessary to reach the first active node, having regard for the DOCSIS 3.0 spec. In particular, the reasoned user’s upstream power ceiling should not be at the DOCSIS 3.0 upper boundary, especially in summertime when an expanded coax offers greater resistance and upstream power can rise by as much as 3 dBmv. So we have to allow some margin here. “STABLE” has the same meaning as for the Downstream.
The CMTS at the VM end is a big fat modem and, depending on settings at the local hub node, power entering the line card needs to be similar to the downstream entering a cable modem. Once the upstream goes past the optical node and along the fibre backbone, it has to be re-modulated to RF so that it enters the line card from a coax connexion. There should be no noise ingress at the CMTS end – and sometimes there is, which we don’t see (so we blame the SuperHub/Hub 3). Upstream noise (SNR) is not reported to the user and is not further dealt with here.
The CMTS sets the Cable Modem’s upstream output power to whatever level is necessary for the CMTS to make sense of what it receives. If the Cable Modem is close to the street cabinet, then you wouldn't expect to see a high upstream power level if a neighbour had a lower level unless you had been placed on a different tap point (see Figure-3 below). The numbered roundels show the attenuation value in Db.
Figure 3: Street Cabinet coax tap points
It should be noted that not all street cabinets offer a full range of attenuated tap points. In low density housing areas, the cabinet will have a single block of 16 tap points, all at 15 dB attenuation.
If the upstream power is fluctuating during initial ranging (including after a T4 event), then the reported (e.g.) 55 dBmv (the point at which the upstream is maxed out) is not enough to establish communication with the CMTS. The problem could lie on the CMTS side of the network, although most cases are local with a faulty amplifier in the street cabinet or some attenuation on the line from home to the street cabinet.
Further technical reading on the Upstream can be found here.
THE EFFECTS OF LOW UPSTREAM POWER
Basically this is not a problem unless the modem is faulty. A house that is close to the street cabinet needs less power than one that is further away, subject to the tap point used at the street cabinet (see Figure-3). This is why upstream values reported on the modem such as 29 dBmv can work OK. But it could mean that the home is connected to a low attenuation tap point and thus the downstream power might be too high and would need an attenuator fitted to the back of the modem.
If you are only 20m from the street cabinet instead of 100, you immediately gain 3 dB and if you are on the 15 dB tap, you only need 32.5 dBmv to make 15 dBmv at the optical node using Figure-1 as the starting point.
So, if you are some distance from the street cabinet AND your upstream power is low (with possible attendant SNR issues) AND you're having problems, then you could have a faulty modem.
.... PART C is in the next post.
POWER LEVELS & SNR: A TECHNICAL PRIMER (Updated October 2016)
PART C - DOWNSTREAM TRANSMISSION
In PART B, we examined the upstream leading to a rationale for recommended power values in the various circumstances that pertain.
Here in PART C, we examine the downstream. It's frequency range is much higher than the upstream and so, is less susceptible to (but not immune from) noise. The downstream, therefore, works at a much higher modulation (bit packing density) than the upstream.
The CMTS at the VM end puts out upwards of 42 dBmv per channel. The level arriving at the first downstream node is variable depending on plant condition; the output of a particular line card matches those dependencies. This approach ensures (in a correctly maintained system) that the signal reaches the optical node at 15 dBmv or higher. When the signal emerges at the optical node in the relevant street cabinet, it has to overcome the 34.8 dB built in attenuation (Figure-1) or 38.8 dB (Figure-2). PLUS another 10 dB due to the much higher downstream frequency. So that's 44.8 dB or 48.8 dB attenuation for downstream in the two tap point cases. But 100m further down the road, the coax component of attenuation doubles adding 14 dB (more for higher frequencies), making the example values 58.8 dB and 44.5 dB for the circuits shown in Figure-1 and Figure-2 respectively.
So now you can see why downstream power levels can go negative when the target output of the optical node is 46 dBmv.
HOW STABLE SHOULD DOWNSTREAM POWER LEVELS BE?
By stability, we mean variance in levels, not the difference in levels among the various channels. From minute to minute, ½ dBmv variance is quite normal. In summer, downstream power levels fall because the coax copper expands with heat offering greater resistance. The fall in summer from the winter low temperature can be as much as 3 dBmv. The same the other way in winter. If downstream power level in winter rises above the maximum shown in the tables below, then a forward path attenuator would need to be fitted and can be left on in the summer.
THE EFFECTS OF HIGH DOWNSTREAM POWER
High downstream power (say in double figures), occurs if:
High downstream power amplifies noise (but if the CM is close to the street cabinet, SNR will normally be high and thus noise will be low). High downstream power can also overdrive the CM amplifier, but not all CMs are as sensitive as each other. So this can be a moving target.
The SuperHub 2 modem tunes the RF channels it’s looking for - as does the Hub 3. The incoming signal is QAM modulated (which you can read about here). Power levels higher than the DOCSIS single channel upper limit can cause the adjacent QAMs to break through the tuner's isolation and interfere with the intended signal.
The SuperHub 1 modem digitally samples 100 MHz of the RF. This process is not without loss of signal quality as it takes an analogue stream and turns it into 1s and 0s. There is a good technical read on some of this here.
DOWNSTREAM POWER & SNR/RxMER LEVELS AT CABLE MODEM
Bearing in mind the above explanations, it should be clear why the following levels are recommended. They take into account:
|
Downstream Modulation |
VM Acceptable Power |
Seph’s Reasoned Optimal Power |
VM Acceptable SNR |
Seph’s Reasoned Acceptable SNR |
|
256QAM |
-6 to +10 dBmv |
-3 to + 8 dBmv |
≥ 34 dB |
≥ 34.5 dB |
Looking at the stats on the VM forums, it is clear that power levels below -4 dBmv tend to increase the codeword error count. For DOCSIS 3, power levels > 8 dBmv have been known to cause synchronisation issues. On the other hand, power levels up to 10 dBmv are known not to have cause problems although such levels are infrequent because the installation technician usually takes care of this aspect. With the upstream, the ‘Seph’s Reasoned Optimal Power’ makes allowance for temperature variations which, when in effect, bring the range into congruency with the ‘VM Acceptable Power’. Put another way, if you have 8 dBmv downstream in summer, then in the cold portion of winter you will have 10 dBmv due to reduced resistance.
Regarding downstream SNR, don't let anyone tell you that for 256QAM modulation that anything less than 34.5 dB RxMER/SNR is acceptable. There is considerable evidence in the forums that data corruption occurs when the SNR falls to 34 dB. This is shown in the Post-RS count on the downstream status page. The total number of corrupted codewords is shown in the Pre-RS count. The modem hardware then attempts to correct the corrupt data and anything that remains corrupt increments the Post-RS Count. The proportion of Post-RS errors rises significantly when the SNR falls and, of course, we shouldn’t be experiencing any lost data packets.
Please note that at this time of writing, the RS counts are not available on the HUB 3 except via a URL:
http://192.168.0.1/walk?oids=1.3.6.1.2.1.10.127.1.1.4.1.3 for Pre-RS errors
http://192.168.0.1/walk?oids=1.3.6.1.2.1.10.127.1.1.4.1.4 for Post-RS errors
SNR/RxMER Thresholds
There are official tables that set out the lower threshold for each QAM level. Additionally, good engineering practice adds 3 dB for local conditions and best practice adds 6 dB which would be an ideal design goal. So at minimum we’d be wanting to see better than 34 dB at 256QAM and better than 29 dB on 64QAM.
The upstream is most noise sensitive due to its frequency range. VM have improved their infrastructure so that the noise tolerant QPSK modulation (2 bits/symbol) is no longer used for DOCSIS 3.0. 16QAM modulation (4 bits/symbol) has been the norm until 2013, since which VM have been progressively enhancing the upstream to 64QAM modulation (6 bits/symbol), reverting to 32 or 16QAM in poor conditions. If your bonded upstream stats show a mixture of 64QAM and 32/16QAM modulation, then your modem will only be able to put out at the slower rate. That sort of partial impairment has to be sorted out in the network. If both upstream channels have reverted from 64QAM to 32/16QAM by a process known as Dynamic Upstream Modulation, then you should look for rising Post-RS errors on the downstream to see if you have a common cause fault, which means that a component that handles both your upstream and downstream (e.g. coax cable, cable connector, modem) is electrically noisy.
Note that for the upstream a SNR threshold is set in the CMTS for each cable, below which modulation is dynamically reduced from 64QAM to 32QAM or16QAM (read about that here). You are unaware of the upstream SNR for your circuit.
The table below is adapted from learned Cisco documentation and illustrates the SNR at the CMTS that will support readable data issued by the cable modem. Upstream should always benefit from at least 6 dB headroom when designing the circuit. The numbers are configurable according to spectrum analysis.
|
Modulation |
Lower SNR Threshold |
+ 3dB Headroom |
+6 dB Headroom |
|
QPSK |
07 to 10 dB |
13 dB |
18 dB |
|
16QAM |
15 to 18 dB |
21 dB |
24 dB |
|
64QAM |
22 to 24 dB |
27 dB |
30 dB |
|
256QAM |
28 to 30 dB |
33 dB |
Making your own changes at home
VM state in their T&Cs that you must not interfere with their equipment, which is deemed to be the Superhub/Hub 3, all coax cabling they have installed and any splitters/attenuators that they have installed. That said, VM do allow Superhub/Hub 3 self-installs which must thus be performed very carefully. The point of note here is that if you do anything yourself, like adding a Forward Path Attenuator to deal with high downstream power, or extend a coax cable, you must make sure that nothing is “Heath Robinson” about the addition that would cause phenomena such as micro-reflections to occur that would then induce noise on the same cable segment.
END NOTE
The Virgin cable network is a collection of different network builds by different companies in the 1990s. These builds vary in quality. Some (at this time of writing) are still on DOCSIS (5 - 42 MHz upstream); others (most and soon all) on EURODOCSIS (5 - 65 MHz, further reading here). The upstream frequencies are a noisy part of the spectrum, particularly the lower. VM are in the process of improving plant, installing higher capacity line cards, performing software based re-segmentation as well as physical node splitting (Cat-C resegmentation), allowing better than 16QAM to be achieved with consequently higher upstream data rates.
For the downstream, everyone will eventually have 16 downstream channels (and later 24 channels), aggregating to 880 mbps shared with everyone else on the node/segment. VM are now allowing 200 mbps and 300 downstream and to avoid “unacceptable” contention, resegmentation is taking place to regulate the number of connections per optical node. Don’t get too excited, though. Just four users able to download at 200 Mbps will nearly saturate a 16 channel downstream bonding group; usually this capability is delivered by p2p/torrenting mechanism which are throttled by VM in order to minimise this form of congestion.
This article is merely a primer on Power Levels & SNR. If you are having speed problems and your power levels and SNR that you can see fit the optimum range, then it does not automatically mean that there is a utilisation issue in your area. The problem may well lie in the upstream under noise conditions that can only be read from the CMTS end. Furthermore, it has been known for the CMTS end to be misconfigured for an individual user – sometimes visible on the modem stats (e.g. Mini-slot size). So there is more to this than just Power Levels & SNR.
Further down the road, is DOCSIS 3.1 which allows for higher performance on the same plant, but with different modems at each end to handle the new protocols. DOCSIS 3.1 is out of scope for further discussion in this article.
.... PART D (Bibliography) is in the next post.
