Innovative Or Pointless: PoE Chain IP Cameras

A manufacturer is promoting a different architecture for networking IP cameras. Instead of all cameras connecting individually to a network switch, each camera is connected to the next closest one in a daisy chain, like so:

The manufacturer argues that:

"It saves cost of cables and PoE equipments. Suppose there are 12 IP cameras, and average distance between camera and switch is 200 meters. In a normal PoE connection, it requires a 12 ports PoE switch and 12 sets of PoE midspan, and total length of 2400 meters of CAT5e cables. With chain connection, it only needs a one port switch, a PoE injector, a total length of 500 meters of CAT5e cables."

One obvious limitation is that it is proprietary, and cannot support 3rd party cameras on the daisy chain. Also, it is from a small manufacturer, Sigrand, that most of you have likely never heard of nor used.

What do you think? Interested enough for us to test this or look into further?


The main issue I see is that it's very very non-standard. You're going to be forever limited to this hardware. One of the benefits of IP cameras (IMO) is that there are in fact "standards". You have standards for cabling, power, etc. If you install a Cat5(e) certified cable plant you can hang all kinds of equipment off it. Doing this makes your whole system unique in a very bad way. You also appear to be dependant on that funky NIC. Additionally, I'd be concerned/curious about latency on the cameras that were more than 2 hops down that chain.

I think that is a fantastic idea. I don't know if I would put 12 cameras, but certainly it could be very cost savings to put a few of them on there. I am thinking especially in cases where you have to do an outdoor run with a POE extender, instead of doing two extenders or an outdoor switch, you can just link the two cameras together. Save a thousand dollars right there.

I am thinking that Axis will come out with their version in the near future. I'll use it!

Not worth it. If every camera manufacturer had two network ports in their cameras that were PoE powered and did PoE pass through, fine, but that's too much an added cost to justify.

There are third part products already that allow you to daisy chain connect cameras down the line like NVT's 1701 Ethernet over Coax and new T-Bus line. There you can use any camera you want.

true, but if you are adding another piece of hardware to do what the camera could already do, you are adding unnecessary costs to a project. Think: three cameras on a pole 700 feet from the nearest IDF...

Then it's a tradeoff, depending if the cameras themselves are cost effective or not. If they are, and they have the form factor and options and reliability you need, then they work. Otherwise if they don't have the camera you want or need, there are other ways to do it.

personally i can't say it would be a benefit... what is troubleshooting like in the event that one camera dies, does the poe passthrough? this reminds me of a similar product viscount had years back in daisy chaining access control readers the same way...

Exactly this is a troubleshooting / service nightmare waiting to happen.

I agree. Complicates troubleshooting significantly. As far as cost savings, you might be able to save some cabling. However, running cable from camera to camera does not necessarily use less cabling than from a central switch, and may even consume more.

Doing some math: if 802.3af delivers a max of 15.4 W per port, and each camera uses 4.5W, I can only chain a max of three cameras together in one line (not the 5 they show in the image.)

Not enough of a benefit to risk an exotic implementation.

Ethernet switches are fairly reliable, and are frequently kept in nice temperature controlled, secured racks.

On the other hand, security cameras get brutalized by weather, corrosion, and even vandals. It's shortsighted to 'be pennywise but pound-foolish' to save money on cable cost but rely on a device that runs the risk of destruction to power multiple cameras in the system.

If I'm lucky enough, I take one camera down, and I potentially open up a huge blindspot when the other cameras fall offline.

The concept is similar to the power as you can run the power cable from the power source to the camera or have it like chain connection. The main issue with the chain connection is what will happen if one of the cameras is down or what if one of the chain cables is failed, then the complete chain will be down unless if there is a sort of double chain connections.

Solutions like these should only be needed on those few occasions where the benefits outweigh expected increases in maintenance time. I don't think anyone would be unwise enough to use this in every application.

At least with the bus type connection with 3rd party products, a camera going down or offline probably won't affect the rest down the line. In the connection method shown for the Sigrand cameras, any camera hanging or malfunctioning would likely break the chain and definintley taking it offline would, if the next camera down the line was too far to use a coupler.

I would not even entertain the idea for all the reasons mentioned. Now, if this were an AXIS product I would at least give it a try..........

Ethernet Daisy chain This is a crazy solution in term of maintenance.

When you have have a ring or spanning tree you have redundancy, but nothing here

So niche market for small installations but with proprietary devices... so what happen in 5 years .. in 10 years ...

The same way, some large transportation company here migrating from coax analog to IP have chosen a very specific PLC over coax hybrid solution, with a single provider on the market for 10 thousands cameras.. so you can imagine what will happen after 5 years maintenance if the supplier collapse

I think this would be a great feature to have available. I certainly wouldn't use it in every application, but it would be a nice option to have when it was needed. For example, as we speak, I am designing a system that has service counter with 16 registers all in a row. I have a very narrow canopy over the counter where I need to fit in 16 cameras. I would love to be able to loop from camera to camera instead of running 16 cables back to the switch.

I understand the reliability/servicability concerns, but I think with some competent engineering, these could be easily addressed. For example, a small version of a standard network switch could simply be incorporated into each camera. One port would connect internally, and another port would be available to loop on to the next camera or other network device.

Well why not just put a switch by the cameras then? You would need 15 of these things and power for each? What a hodgepodge.

Also without some form of additional power input you cannot just keep looping out poe - it's doesn't work like that.

I know that's true with today's equipment - I'm talking about developing something specifically designed for this type of application.

I agree with Mr Silva. If the technician has some skills, he could easily pinpoint which camera is the issue and replace THAT one. I think we are forgetting that in a traditional setup, if your switch goes down, all your cameras go down, too. If it is a customer that needs a star network configuration and redundancy, then this kind of setup is out of the question. However, after explaining the pros and cons sufficiently, most customers we run into will opt for saving thousands of dollars and take that slight increased risk.

This " innovation" has right now 20 years. Farallon a US network manufacturer in the 90's launched the first Ethernet daisy chain which never really found his market.

Always choose an open solution with several potential provider and IEEC .. you will do the deal, but your customer will cry 5 years later when the spares won't be available anymore.

Remember you can have 30W and more on twisted pair and put an auto powered switch to supply 3 or 4 cameras with a single CAT5or CAT6...

Well, the 'innovation' here is that it's embedded in a camera. One can argue whether if or where it's useful but it's not simply daisy chaining Ethernet.

Of course John, daisy chain with POE within the camera. Innovation is always to try to create a differenciation

Some others focus on coax PLC cameras, why not ? PLC is IEEE standard P1901 but PLC Hybrid is .... nothing, no t a chance to find it in 5 years

Serial bus are a pain in maintenance (you know I train guys here) If somebody invents a real cheap POE ring with redundancy it could be different. Think long or Mid term should be the main idea for professionnal. (according to me)

When you don't care mid term you can play strange innovations :-)

This looks like a problem just waiting to happen to me.

Man - I have a reputation for being cynical about new technology and never wanting to be an early adopter, but some of you guys make me look like Tony Robbins.... :)

A number of you asked about or commented on how the system could support up to 12 PoE cameras on a daisy chain. Here is the manufacturer's claim:

"If module MS-17E8PP [for Sigrand switch] 60W per channel is used as a PoE injector, you can connect up to 12 cameras, with the distance between them is less than 100 meters, and the total chain length is no more than 500 meters.... This is possible because cameras can work at a reduced PoE voltage (around 30V). We call this connection PoE-through."

When I first heard of this idea, I thought it was useless. But thinking about it, I agree with some of the above, like the multi-camera pole example, or the row of registers. Those actually aren't terrible applications for it, and it would save some hardware and cabling. I wouldn't do an entire building full of daisy-chained cameras, to be sure, but avoiding some complexity in specific spots, sure.

I gotta go along with Michael, Ethan and Luis: it's a niche sort of concept, but it's one that COULD be very useful/helpful/efficient in certain circumstances, to address some particular design challenges.

I don't know that it's all THAT "innovative" - it's basically just a camera with built-in switch and PoE pass-through - but it's far from "pointless".

Well this is not innovative for sure. Novel - not really. Pointless - not completely. I would never design a system from the beginning to utilize these devices but I can see where they might come in handy in some situations (those that require the redneck rigging for one reason or another).

Ok, I think the point is that we are a little split in opinion, but many are vehemently opposed to them, like they are sac religious and detestable. However if some of these same people read a glowing review from John, I believe they would shift their opinion a little bit. Which brings us back to the original question John proposed, should they look into it and do a review? An even better question is, if a major manufacture came out with such technology or if this technology gets a great review, would we actually use it. I think the number of responses on this discussion alone shows that John should look into it a little more. We need to know if there is no way in hell we would use it, or if there might be certain situations where it would be useful. I vote for the latter, but I can be open minded.

If Avigilon released this, it would be the most innovative, awe inspiring achievement of the millennium :)

That said, I am with those who point out that the manufacturer does count. If any 'major' manufacturer released this, it's a no brainer for a test. It's definitely different and would find a niche. The problem is that it's from a company hardly anyone uses and who has no real presence in North America or Europe.

Can't we just wait until Honeywell OEMs it? ;)

Well the question posed was Innovative or Pointless. The answer is NEITHER. While I'm opposed to habitually using these or planning systems based on utilizing these I am not 100% opposed to them entirely. I can think of a few applications for them (all have to do with budget limits) but that does not make these Innovative. Nor are they pointless.

I think time be better spent tetsing third party daisy chaining connection products that are not camera manufacturer specific and more video analytics comparisons.

However, if IPVM does have the time on their hands, I don't see the harm, either, and would be interesting to see.

I am not sure I buy the "3 cameras on a distant pole" argument. That can be achieved with a Veracity Highwire Quad or similar device at the pole. At that point it is a question of whether or not cost savings (if any) justify the inherent risk of locking yourself into such a small vendor who may not be around in five years. I would also be concerned about image quality (especially in "distant pole" scenarios, which implies sub-optimal lighting).

It would be better if it had an analog output too for focusing the camera.

it would be better if it was an April Fools joke