pfSense Gold Subscription

Poll

Your feelings on Rack-Mounted/Internal DOCSIS hardware?

Wouldn't use either. Bad idea.
4 (33.3%)
Would use Rack-Mount with SFP.
1 (8.3%)
Would use Internal PCI Express Card.
0 (0%)
Wouldn't use either. Good idea.
0 (0%)
Would use both!
7 (58.3%)

Total Members Voted: 12

Author Topic: DOCSIS 3 Internal PCI-Express Cable Modem Card  (Read 3669 times)

0 Members and 1 Guest are viewing this topic.

Offline Harvy66

  • Hero Member
  • *****
  • Posts: 2195
  • Karma: +201/-12
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #15 on: October 14, 2017, 01:03:20 pm »
There's a good chance that the drivers for an internal DOCSIS card are going to be worse than a good Intel NIC making the latency worse. While I can appreciate the goal of minimizing latency, doing so on a DOCSIS connection that is going to have a 10ms+ ping to the first hop is quite futile. If you were talking about cut-through 10Gb switches in your network, I could get on-board.

Offline Finger79

  • Full Member
  • ***
  • Posts: 166
  • Karma: +15/-0
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #16 on: October 14, 2017, 07:17:37 pm »
if there's a way to protect coax im all ears.  Been a thorn in my side for far too long.

The only thing i've come up with is a fiber media box converter where you covert to fiber and back.  But the modem still gets toasted in that scenario.

Well I have a wall of text for you if you're looking to nerd out on this.  Credit to "Dan J." who wrote all this in an Amazon comment.  I learned so much that I just copied it into a text document. LOL.  (I've also since stopped buying MOV-based surge protectors and have bought a few Zero Surge brand series mode products.  They're like $170-250+ each but after hours of reading I'm sold on the superior technology of series mode.)

Anyways, there's a section here on how to protect coax cable coming into the house.  Apparently you find the single coax entrance to the premise and install a hybrid gas discharge thingy that's extremely well grounded, and all coax downstream should be protected.



Quote from: Dan J from Amazon
In reply to an earlier post on Jan 20, 2016 6:12:34 AM PST
Last edited by the author on Jan 25, 2016 11:06:38 AM PST
Dan J says:
Good surge protection starts with a good ground. If you are not sure if your ground at the mains electrical panel is hooked up properly, you can hire an electrician and have them verify it. Without a good ground these point of use protectors will not do much for you as they depend on a low impedance path to the mains electrical panel. You may need to hire a qualified licensed electrician for this task.

Another thing to consider is to provide surge protection at all the points of entrance where any cables/wires enter your home vs. at the point of use, as follows:

1) Whole House Surge Protection

Dollar for Dollar the most bang for the buck in that it protects everything in your home, not just computers, TV's and Home Theater gear but also refrigerators, washing machines, clothes dryers, etc ...

The protectors typically run from $50-$75 but will require a licensed electrician to install. Choose an installer that has experience installing these products. Many electricians like to make pretty right angle raceways of the wiring but for this protector to be effective you want straight point to point runs without sharp bends in the wire.

It's usually installed on a dedicated 2-pole 50-Amp circuit breaker. If you are not sure of the models, try Eaton's CHSPT2ULTRA with 20kA protection. I've been using this one for the last 2+ years.

2) Telephone (RJ-11)

Eaton makes a companion product to their Whole Surge Protection known as the CHSPTELE and has the same physical size as the ULTRA and these can be stacked onto each other if space is a premium. I have not tried this one and cannot speak to how well the product works. Then again as many of us move away from the telephone company to VOIP this may be less of a concern for many. If you don't use a telephone anymore, see if you can disconnect the wires at the NID (Network Interface Device) from the phone company on the side of your residence.

3) Coaxial Cable (VHF/UHF Over the Air Antenna, Satellite Dish, Cable TV/Internet)

Eaton makes another companion product known as the CHSPCABLE, or you can combine both Telephone and Coaxial cable protection in one product, known as the CHSP3RTELCABLE. I have not tried either of these and cannot speak to how well the product works.

Alternatively if you're somewhat handy and have several locations to protect different kinds of coaxial cables, consider a "ground block" swept to 3 GHz to ground the coaxial shield (or sheathing), then continue into a second device known as a "Cable Surge Arrestor." Morgan Manufacturing, Inc. makes the Model M311 that is a hybrid Gas Discharge Tube/Metal Oxide Varistor design that can take the center conductor to ground in the event of a surge. This is the method that I used and have not had any issues with going on almost a year now.

It goes without saying that all of these service entrance devices will require a good connection to ground; at least #6 gauge solid copper wire to the mains electrical panels ground. If any of these service entrances are not near the mains electrical panel, it may be necessary to also install additional ground rods and connect the ground rods together with #6 gauge solid copper wiring avoiding any sharp bends.

Surges can originate from the outside as well as within a home. Outside, being induced by lighting onto transmission lines (overhead or buried) or due to Power Company switching equipment. Inside surges are generated by electrical appliances when they turn on and off, especially those with a motor i.e. refrigerator, furnace fan, box fan, window fan, stand-alone fan, mixer, blender, juicer, air-conditioner, freezer, washing machine, clothes dryer, hair dryer, power tools, etc. If your lights ever dim you can bet there was also a surge generated as a result.

If lightning is a great concern to you, there are other services that you can hire to design and install a system for you that can include rooftop lightning rods, additional ground rods, etc. to protect your home, but all would be pointless without a good connection to the earth ground.

If you installed surge protections at the (point of) entrance to the house, there is no need to make use of the coaxial (Cable TV/Internet, satellite, OTA Antenna), RJ-11 (phone), or RJ-45 (network) connections on point of use Surge Protectors. Just remember to plug all the AC devices into surge protectors.

There are two kinds of MOV surge protectors.

The most common type has triple mode protection and you'll see L-N, L-G, N-G on the side panels of the packages. These represent the 3 prongs on an AC outlet.

The 2nd kind is known as single mode protection and only provides protection from the L-N prongs on a 2 or 3-prong outlet (the two vertical slots). Single mode was once very popular but in the last 20 years most products sold use triple mode except for some very expensive models for audiophiles.

You can thank marketing for the rise of triple mode protection. Why settle for "single" mode, when you can have "triple" mode.

Both modes work by diverting the surge to both the Neutral and the Ground (triple mode only) . As long as there is a low impedance path to the mains electrical panel, then all is well, however if there isn't the surge once it's at ground and neutral will continue to seek an alternate path to ground even if it's through the very electronics plugged into the surge protector and zap it in the process. The electronics may not die immediately but over time its life will be shortened. This phenomenon is known as Ground Contamination and is the reason for many unexplained hard drive crashes, fried motherboards, and premature failure of power supplies, bad USB and fried HDMI ports. This is the dirty little secret about triple mode surge protection, but it's a big one.

MOV surge protectors wait until the voltage rises to a certain level and then start diverting the surge, but here's another dirty little secret, the entire surge is not diverted. Some will be diverted to the Neutral, some to Ground and some will continue into the very electronics the protector was purchased to protect. It's because depending on the size of the surge, the MOV can be easily overwhelmed. Even when manufactures place two or more MOV's in parallel the manufacturing tolerances are such that one of the MOV's will turn on before the other.

Each time the MOV works, it takes a little less voltage for it to turn on the next time. Eventually it will be on all the time and it will smoke, explode and/or catch fire. In order to mitigate some of these problems, some manufactures encase the MOV in a fireproof material like ceramic, and/or place a thermal fuse in-line with the MOV (physically the two are wedged tightly against each other). When the MOV gets very hot and failure is imminent, the thermal fuse opens and either takes the MOV out of the circuit or kills power to the rest of the surge protector. In either case, the status light may or may not be lit informing one that the surge protector needs to be replaced.

Many manufactures of MOV surge protectors recommend that you replace the surge protector on an annual basis as its unknown when the MOV has worn out. Most manufactures will replace your MOV based surge protector but you have to wait till the protection has failed. Generally it's cheaper to buy a new one rather than pay the postage to send the one you have back to the manufacture. If you send one back each year and it hasn't failed they'll send you your old one back.

Furman sound makes some high end surge protectors that are single mode and cost a pretty penny. Their limitation is that the entire surge is dumped onto the Neutral and like the problem with ground if the return path is not a low impedance path, the surge will still enter the electronics the protector is designed to replace. Over time, damage still occurs. This company employs a number of technologies to help including over under/voltage protection that will kill power to the surge protector if the voltage gets too low or too high but until it does, the electronics plugged into it (just like the electronics plugged into a triple mode surge protector) will take the surge and protect the surge protector. It's a step in the right direction and has caused Furman to work on fewer returned models, but can we do better?

Isolating transformers (sometimes find in very high-end audio equipment) work really well but are heavy and expensive and therefore not practical. They were also common 50+ years ago in electronics and appliances.

Series Mode Power filters are heavy duty low pass filters that filter anything faster than 60 Hz, so they are great for noise reduction, EMI and RF interference, but they also filter surges. They work by slowing the surge down in real-time, storing the surge energy in capacitor banks and then bleeding the energy onto the Neutral a few volts at a time until gone. They are also known as single mode (L-N protection), but don't rely on an MOV as a protection element and therefore do not wear out or need periodic replacement. Over long periods of time they are very economical and over short periods of time provide superior surge protection since they eliminate the surge instead of just diverting and hoping for the best.

APC engineers described series-mode technology as "the best type of surge suppressors use a "series" or isolating design". Manufacturers of (series-mode) surge suppressors, such as Zero Surge, Inc., virtually eliminate ground wire contamination."

APC engineers had the integrity to acknowledge the benefits of the series-mode technology, even though they offer shunt-mode technology, and do not benefit from their candid observations.

One can buy new or used series-mode power filters. A new 2-outlet 7.5 amp model (2R7.5W) sells for $139 from Zero Surge, Inc. A new 2-outlet 15 amp model (2R15W) sells for $159 from Zero Surge, Inc. Other companies that make them include Brickwall (made by Zero Surge, Inc) and aimed at audiophiles. SurgeX licenses the technology from Zero Surge, Inc and makes a wide variety of models. Two of their most popular models are the SA-82 FlatPak (a 2-outlet 8 amp model), and the SX1115/SX1115-RT series which are 1U rack mount capable units with 9 outlets. The -RT models have a Phoenix connector for remote operation and incorporate COUVS (Catastrophic Over/Under Voltage Shutdown), and ICE (Inrush Current Elimination). ICE is very useful for Home Theater setups when you power everything up at the same time. All of these models and others are often available on the popular auction site eBay either used from $30-$75+ or new (old stock) from $60-$150. Since they don't rely on sacrificial MOV technology, they won't wear out or need periodic replacement.

If you take a long term approach and contrast the annual replacement of a $10, $20 or $25 MOV based surge protector vs. a new 2-outlet 15 amp series filter, you'd have spent $100, $200 or $250+ over a 10-year period vs. $159 for a new, $75 for a new old stock, or $50 for a used series-mode power filter. Long term economics favor series mode.

If you consider the type of protection you're buying: divert and hope for the best or true surge elimination and series-mode wins again.

Some final thoughts: Review sites

There's a big review site that uses the catchy title "The Best Surge Protector" easily found via a Google search. They claim to test a sampling of surge protectors that are all MOV based and typically sell for $25 or less. Their testing methods have to be taken with a grain of salt as they subject the surge protectors with surges packing 0.01 joules to 0.06 joules. The #1 product rotates about 2-3 times per year. Each time there is a new #1 product it's because some new information about the product has come to light from the manufacture that is not published on their website but provided to the staff by marketing for the company. Of course it's in marketing's best interest to promote the companies product. The way it works is you land on the site by a search engine, read the article and click on a link that takes you to a site where you can purchase the product. The review site gets a kickback from the purchase. Meanwhile the surge protector manufacture touts the #1 status of their product and provides a link to the review. The company selling the product makes money, the review site makes money and the consumer ends up with a product that has questionable surge protection and everybody is happy.

Connected Equipment Warranties:

They are designed with so many escape clauses and loop-holes that it's very difficult to collect on them. Sometimes you get lucky but more often than not it's very time consuming. If a computer is involved you could have lost data and of course the use of the computer. Replacement is generally the lesser of the fair market value based on recent "completed" eBay or Craigslist "completed" sales.

Safety Compliance Issues:

It's not recommend that you daisy chain MOV based surge protectors first because it's much harder to keep track of loads for the circuit breaker and second not all MOV's turn on the same voltage. For those of you who use a UPS (Uninterruptable Power Supply) many include a surge protector. When the surge protector in the UPS fails the whole UPS needs to be replaced.

You can safely plug an MOV based surge protector or a UPS into a series-mode power filter. In the case of an MOV surge protector you can use it to expand the number of outlets from 2 to N. In the case of a UPS, you'll end up protecting the UPS by ensuring that it never sees a surge.

NOTE 1: As previously mentioned, the MOV will (in a triple mode circuit) divert the surge to Neutral and Ground but some of it will continue into the load you're trying to protect. It's just how MOV's work.

NOTE 2: If you decide to have a mixture of both MOV and Series Mode technologies you can do so as long as you always plug an MOV based device into a Series Mode device but you don't ever want to do it the other way around or have them as first devices on the same branch circuit (circuit breaker)

NOTE 3: It's generally not a good idea to mix different brands or models of MOV surge protectors throughout your home as each will respond to a surge in a different manner. No two devices will turn on at the same time, so some devices will start diverting the surge before the others and wreak havoc on the other surge protectors. You can't even rely on two or more identical models purchased at the same time from the same seller as the manufactures does not sell matched MOV's (not even in the same unit, let alone in different units) especially at this price point. This would be the biggest reason that I would avoid this technology entirely as it's truly a false sense of security.

Older Homes and Buildings

Many older homes and buildings in the US still only have 2-prong outlets (no ground). You can't use an MOV based surge protector in these environments since there is no ground and no place for the surge to go, but since the series-mode power filters don't rely on ground you can safely plug them into the 2-prongoutlets; either install a 3-prong GFI outlet with a label stating that no ground is present or use a 3-prong to 2-pround adapter on the series mode plug.

Summary

There are many advantages to using a series-mode power filter of which total surge elimination is the strongest.

Offline belt9

  • Full Member
  • ***
  • Posts: 233
  • Karma: +24/-6
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #17 on: October 14, 2017, 10:48:36 pm »
Holy crap, that's a lot of info about surge protection.

I haven't had a piece of equipment damaged by a power surge or any power irregularity since I was a child living in the literal boonies, and that was a lightening strike.

Since then I've lived and traveled to a whole lot of places and countries - many of which have very primitive power, and nothing.

Where is it that people are living that power surges are a real concern. Am I just lucky? I've never even tlaked to anyone who has lost equipment from a power surge, but then again that isn't exactly the topic of normal conversation.

Offline kklouzal

  • Jr. Member
  • **
  • Posts: 59
  • Karma: +1/-1
  • Maximum Performance!!
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #18 on: November 05, 2017, 09:17:48 am »
For instance if the drop is fed from a pole and improperly grounded the pole can be struck by lightning and surge the coax down the drop into your modem (or in our scenario the internal card) and fry it (or potentially the entire machine) leaving you to purchase a replacement. There are options for grounding:
http://www.citel.us/coaxial_types_gas_tubes.html
http://www.citel.us/coaxial_types_hybrid.html
This would introduce a resistance on the line which would be compensated by the modems ability to adjust output power on-the-fly.
Interference? Maybe, and if so I highly doubt it would be enough to cause transmission errors.

As far as Software/ASIC I would highly encourage a manufacturer to use an ASIC. This would offer the highest performance available.

Companies like Motorola/Netgear, which ISP's already widely accept into their networks, would take their existing designs and modify them to be implemented onto a PCIe card and directly communicate over the PCI bus. Yes they would need to provide firmware for the machine using the card but they already provide firmware to the ISP's so this again wouldn't be a new process. If you want to argue Linux/FreeBSD support that might be the appropriate path here.. I don't think the big manufacturers would support anything outside Windows, possibly Mac devices, and less likely Linux. However a released Linux driver would mean a community based FreeBSD port could be doable.

I agree it is a niche market but I think there is money to be made with hardware like this. Plenty of manufacturers create products that only sell 100's of units, albeit the price is high, but I would be willing to pay the price for a device like this and i'm willing to bet anyone else who wants one would be willing too. If there did happen to be an improvement in latency/features I bet others would start to desire such hardware and then laws of supply/demand would start to take over.

I have yet to see a good reason why this shouldn't be available.

As a side note, this 'card' could directly interface with the machine as if it were just another gigabit network adapter in which there were no external port. Instead the port would be directly connected to the modem side of things via traces on the board. This is a bit of a cop-out however it would mean existing designs could be almost 1:1 printed onto the PCB and put on the market as a sort of 'test run' to see exactly how much of a demand there really was. This would require no engineer to reinvent the wheel costing the manufacturer time and money coming up with a whole new design.

Someone at home with enough electrical engineering knowledge could source the chips and print their own board.
https://www.maximintegrated.com/en/products/comms/wireless-rf/MAX3518.html
https://www.eetimes.com/document.asp?doc_id=1126820
https://www.altera.com/products/fpga/arria-series/arria-10/applications.html
http://www.lightreading.com/cable-video/docsis/the-chips-fall-for-docsis-31-/d/d-id/712922

Modem-On-Chip
http://www.ti.com/lit/ml/spat152/spat152.pdf

The technology is there.

Offline johnkeates

  • Sr. Member
  • ****
  • Posts: 572
  • Karma: +51/-1
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #19 on: November 05, 2017, 10:11:23 am »
As already posted plenty of times: there is no benefit. The modem parts are always done with an RTOS and almost always with an ASIC as well. There is no direct modem interface that would work with a time sharing / non-real-time OS. It always comes some sort of bus between the actual modem and the router/firewall/gateway, and it doesn't matter if that is external or internet, nor does it matter if the interface is 10cm of PCB tracks or 100cm of twisted pair ethernet wires.

The only thing you stand to gain is physical integration, and if that is the goal, then by all means, an embedded modem on a PCI(e) card could be useful. There is no performance benefit, no interface benefit and no software benefit.

Offline Stan464

  • Jr. Member
  • **
  • Posts: 48
  • Karma: +0/-0
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #20 on: November 06, 2017, 03:45:41 am »
I do agree with John.


I use an old ISP Provided "PPPOE" over FTTC, and not the Box they provided me, which would of been a cheap as shit Wireless Router, i use an White BT Box (Sorry for the flippant term, cant for the life of me remember the Connection Tech it uses or the Hardware Manu, But i use that paired with PFSense.

Does the job perfectly, dont notice any problems at all.

Offline kklouzal

  • Jr. Member
  • **
  • Posts: 59
  • Karma: +1/-1
  • Maximum Performance!!
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #21 on: November 08, 2017, 03:50:27 pm »
Fiber is a completely different story since FTTP (Fiber To The Prem) brings a standard fiber line into your home in which you can simply plug into an existing fiber channel card, get the PPPoE instructions from your ISP and away you go.  8)

FTTN (Fiber To The Node) typically uses copper for the 'Last Mile' going from the distribution box. That last mile is either Coaxial or RJ11. In the case of RJ11 you can still use a PPPoE setup but you're already using RJ11 and that's going to be horrible in terms of speed and latency especially the further you get from the DSLAM.  >:(

In a FTTN situation where Coax is used for the last mile you typically have an extremely short distance of copper (in terms of the entire connection length). The consumer has no control of anything past the drop towards their ISP. Basically you have a direct fiber connection except for this 'Last Mile'. Why not directly convert back to fiber, or as close to it as possible once it passes the drop back into our home/office?

The Chelsio line of ASIC provides EXCELLENT latency, sub millisecond down to around 25 microseconds. The optimal solution would be to bridge the Ethernet PHY to the Ethernet PHY of a 'modem on chip' ASIC or something a little deeper and more integrated if you're not using a modem on chip solution. I guarantee you would receive at minimum an order of magnitude latency reduction as well as maxing out the line rate supplied from the your purchased plan.  8) :-*

https://www.chelsio.com/terminator-6-asic/



In the crude design above we have our Chelsio T6 ASIC interfacing with the 'Modem On Chip ASIC' via one of its two provided ethernet interfaces. The secondary free interface is not wasted and we provide a SFP+ port to be made available for the user to do with as they see fit. The T6 ASIC then interfaces with the host machine and presents itself as two available NIC, one dedicated to WAN which has been coupled with the modem on chip. Throw a huge heatsync over the entire thing with a dual 40mm fans and away we go!

Again, I think a card of this particular type design, for at least a first wave of exploring the real user demand for something like this, would be extremely attractive.

More research is needed as the bottleneck in this scenario would be the Modem on Chip.

Offline VAMike

  • Sr. Member
  • ****
  • Posts: 386
  • Karma: +64/-11
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #22 on: November 08, 2017, 04:25:09 pm »
Fiber is a completely different story since FTTP (Fiber To The Prem) brings a standard fiber line into your home in which you can simply plug into an existing fiber channel card, get the PPPoE instructions from your ISP and away you go.
...assuming you have a terrible ISP that uses PPPoE. otherwise, you just get IP packets.

Offline kejianshi

  • Hero Member
  • *****
  • Posts: 4915
  • Karma: +196/-40
  • Debugging...
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #23 on: November 08, 2017, 04:34:06 pm »
I think pfsense is really great at being a firewall.  I think you shouldn't drop in modems and wifi cards.  Likely to cause you WAY more problems than some infinitesimal modicum of latency a separate modem might cause. 

Offline johnkeates

  • Sr. Member
  • ****
  • Posts: 572
  • Karma: +51/-1
    • View Profile
Re: DOCSIS 3 Internal PCI-Express Cable Modem Card
« Reply #24 on: November 08, 2017, 04:55:18 pm »
Fiber is a completely different story since FTTP (Fiber To The Prem) brings a standard fiber line into your home in which you can simply plug into an existing fiber channel card, get the PPPoE instructions from your ISP and away you go.  8)

FTTN (Fiber To The Node) typically uses copper for the 'Last Mile' going from the distribution box. That last mile is either Coaxial or RJ11. In the case of RJ11 you can still use a PPPoE setup but you're already using RJ11 and that's going to be horrible in terms of speed and latency especially the further you get from the DSLAM.  >:(

In a FTTN situation where Coax is used for the last mile you typically have an extremely short distance of copper (in terms of the entire connection length). The consumer has no control of anything past the drop towards their ISP. Basically you have a direct fiber connection except for this 'Last Mile'. Why not directly convert back to fiber, or as close to it as possible once it passes the drop back into our home/office?

The Chelsio line of ASIC provides EXCELLENT latency, sub millisecond down to around 25 microseconds. The optimal solution would be to bridge the Ethernet PHY to the Ethernet PHY of a 'modem on chip' ASIC or something a little deeper and more integrated if you're not using a modem on chip solution. I guarantee you would receive at minimum an order of magnitude latency reduction as well as maxing out the line rate supplied from the your purchased plan.  8) :-*

https://www.chelsio.com/terminator-6-asic/



In the crude design above we have our Chelsio T6 ASIC interfacing with the 'Modem On Chip ASIC' via one of its two provided ethernet interfaces. The secondary free interface is not wasted and we provide a SFP+ port to be made available for the user to do with as they see fit. The T6 ASIC then interfaces with the host machine and presents itself as two available NIC, one dedicated to WAN which has been coupled with the modem on chip. Throw a huge heatsync over the entire thing with a dual 40mm fans and away we go!

Again, I think a card of this particular type design, for at least a first wave of exploring the real user demand for something like this, would be extremely attractive.


This doesn't actually change anything compared to a bridge mode cable modem. In fact, it pretty much *is* a cable modem. That is just the issue here: all of the proposed "solutions" are just normal modems crammed on a PCI(e) card with a ethernet interface to the host. It doesn't change a thing, except physical form factor.