NotoriousBRK

Is the reason for the backbone being solid core simply for a punch block connection? No, solid-core cabling also has less attenuation than stranded. It's been quite a while (10 years or so?) since I was heavily involved in all of the intricacies of full-on cabline best practices... Stranded cable is not rated for a full 100M run (as I recall), and the jacket is usually less durable because it's designed to be more flexible. It will also sag more and would require more support than solid cable bundles. Then we're also back to the bit about it not being a good idea to field-crimp RJ-45s...

I can terminate an RJ-45 plug that'll pass any test in about 30 seconds. It's not that it's a *problem* for me to do them, but it IS bad form. Your backbone cable should be solid-core, which is intended to be secured properly and not moved/flexed. An RJ-45 plug on the end of a solid-core cable violates this concept (even though I realize it is hardly ever likely to be moved/flexed after install). But the other issue is that most people don't realize that RJ-45 ends are not "universal", there are different ones for solid and stranded cables, and even across different brands of cable they don't all work 100% reliably. This is sort of the equivalent to some of the hack 110V wiring we've all probably come across. There is more to doing it right than just making a good electrical connection. You want the termination to be clean, neat, serviceable, and also conforming to standard best practices. Field-crimped RJ-45's are generally none of the above. JMHO.

This is by FAR the best and most professional approach. I stopped collecting all the "but I'm a pro and have been doing this for years" hack RJ-45s I've cut off and replaced with keystone jacks to fix cables that couldn't perform to spec.

Yup, Mike nailed it. You need to do the "bandwidth math" first. Every problem is solvable for the right price, but full remote/cloud NVR is still a pipe dream. You could store SOME video remotely, or you could store ALL video remotely at a really crappy resolution and quality level. You could possibly have something setup that offloads selected video from particular times or whatever. But, getting decent video from 5 cameras is going to be a minimum of 6GB/day (h.264, 5fps, moderate compression, D1 resolution). There is no magic after-hours upload. To transfer 6GB of data in lets say 4.5 hours, you'd need a sustained upload speed of 3Mbps (and an ISP that didn't try to impose bandwidth caps on you, that's 186GB per month of just video uploads). Also, you'd need something in a centralized site that could handle incoming video from all these sites. That same 6GB/day is about 500Kbps of continuous streaming. 500Kbps x 30 sites = 15Mbps of raw bandwidth at the central site. Wholesale committed Ethernet handoffs in data centers in the US are averaging around $10-$12 per 1 Mb per month. So, you've got $150/mo *just* in bandwidth. Dedicated pipes not at centralized locations are averaging around $80-$150 per 1 Mb per month in most of the US. So you could get that 15Mbps pipe at your remote office for around $1500/mo. Don't even think about trying to ride this on a cable modem at the main site. You're going to be moving data constantly, and exceeding their idea of what is a "fair" amount of data, so you'll end up getting throttled at unexpected times or even temporary blackouts. You're probably better off keeping the main video at the stores and architecting a way for them to do live views at times and selectively export out the stuff they want to keep long-term.

For the most part, that is going to require IP cameras. Analog is pretty much a fixed resolution across the board. Granted, a really cheap camera won't have as good of an image as a better quality unit, but overall things top out pretty quickly. Years ago there was much more of a delta between the low-end and high-end gear. You'd see things like 420 line cameras, or cameras that had really low res in black and white, etc. But now that the analog gear is so commoditized, pretty much everything is putting out a full-res signal. As long you don't use garbage for lenses, you'll see relatively little difference camera to camera. Image quality is primarily a factor of pixels on target... more pixels = better detail. The analog NTSC video standard has a fixed resolution and no way to go any higher, though you WILL see gimmicks like "700 TV Line cameras", your recorder and monitor are still going to show you the same 520 lines.

Not totally true... this depends on cam vs monitor resolution. Otherwise, good recommendations. We're probably looking at it two different ways. If your camera resolution exceeds your monitor resolution, then the image you see on the screen has to be down-sized in order to fit. When that happens you're throwing away some of the pixel data to make image fit. In that case it's helpful to have a way to select a sub-set of the camera's view to display on the monitor so that you have 1:1 pixel representation. IMO, you're not really "zooming" in when you do that, you are just no-longer throwing away data. Digital zoom is *usually* represented as a way to somehow get more detail from an image. Either post-facto from recorded data, or from within the camera itself. Both scenarios are trying to display data beyond what was captured by the camera and/or recorded, and both tend to yield equally useless results.

Digital zoom is useless. You need to capture the detail optically. If you are not capturing the information, then no matter how much digital "zoom" you apply, you are just zooming in on more pixelated non-information. You'd probably be better off relocating the LPG tank display to a location where you can put a camera within 8' or so of it. That would allow you to use one of the cheaper megapixel cameras with built-in lenses. For monitoring the relays remotely, you might want to consider something like the http://www.controlbyweb.com/ units, where you could basically put a remote status display in the house. You might also be able to monitor them remotely the same way with some Linear wireless gear.

To read characters that small you're going to want about 120 pixels per foot to minimize confusing things like a "6" with an "8", or a "3" with a "9" (I'm assuming these are standard 7-segment LED displays). You've got an area 8 feet wide, so 8 x120 = 960, you want a camera with an image sensor that is at least 960 pixels wide. A good quality 1.3 Mega pixel camera should work pretty well (they are usually 1280px wide), and there are some basic ones within your price range. At 25 feet away, you'll want a 20mm lens, give or a take a little depending on exact distance, so I'd look for something with a varifocal lens that goes to at least 22mm (the exact lens is also dependent on the image sensor size...) You should be able to find some 1/2" 1.3MP rated lenses in the 10-30mm range for $150-$200 street price.

As to Raymax, to be fair my only experience with them was the visible LED, that was disappointing and I would never waste a clients money on that again I cant speak for their Infrared Illuminators. Raytec and Iluminar are the two dominant pro illuminators right now, and for good reason. Most of the people I talk to have moved on from the Bosch units, but if they're working for you, that's great. It would have been helpful for you to share that info 2 pages ago.

You need to figure out how much total power the cameras and housings will require. Then buy a power supply rated for that amount, plus about 15-20%. Normally you'd run each camera/housing combo on a separate fused line back to the power source. A 8 or 16 device power supply is just one big-ass transformer, and 8 (or 16) fused connections off that single transformer.

Why is that? Doesn't anybody (besides me) want to buy the covert models? The 940nm IR illuminators are not that commonly used, in my experience. They are more costly (the LEDs are more expensive, and their power output is lower, so you need more of them to reach the same distance as an 850nm unit), and they are still not fully invisible. So, most people decide that paying a 30-50% premium for a unit that is still not totally "covert" isn't worthwhile.

Good luck getting any kind of stick-on material to not introduce visible distortion, especially if they're being installed by "maintenance guys".

Not true. Properly back-focused, the lenses will maintain focus while zooming in or out since they are true zoom lenses, not varifocals. That said, there is a knack to getting the back focus locked in. First, you need to have the iris fully open (an ND filter or welding glass helps). Then you zoom fully to telephoto (on a distant object), adjust lens focus, pull all the way back to wide and adjust back focus. Repeat as necessary until focus stays sharp across the entire zoom range. You may be correct. It's been forever since I worked with these, but even so the focus point may stay constant, but if you're panning and tilting (which is likely, otherwise, what's the point?) you will most likely still need/want to shift the center focus point around quite a bit.

Give me a break Rory. He will NOT cover that much area with a D1 camera and get identification level detail. You can zoom in, decrease your field of view, and get more detail, OR you can buy a better camera (eg: megapixel). It doesn't have to be "10,000 Megapixels", but there aren't many ways around this problem. If you're budget constrained, then quality D1 analog gear is certainly the most practical approach. If you're trying to cover a moderate scene AND get good detail, then there are about 1000 low-megapixel (eg: 1.3 megapixel) camera options that will yield better overall all-around results.

I built a system around one of those Rainbow lenses (or a very similar one) sometime around 1999. What a pain in the ass... As I recall the zoom and focus are DC motors and you move them by applying a voltage and/or changing the voltage polarity. Like most lenses, any time you shift the zoom, the focus needs to be adjusted. Without some kind of auto-focus circuitry this quickly becomes an effort in frustration to operate. Unless you are REALLY looking for an electronics engineering type of project challenge, you might want to not get too hung up on putting those lenses to work. You'll either need a camera that can auto-operate the focus portion (no idea where/if one exists), or you'll need a peak contrast detection circuit (the heart of most auto-focus systems).

That's about what I'd expect of a D1 analog camera. A 4mm lens at 16 feet has a FOV about 18 feet wide. For an analog camera that is 39 pixels per foot, well below what you would need to get recognition (about 50 ppf) or identification (about 80 ppf).

You're doing it wrong. You specify that the customer picked some arbitrary budget number ($4000) and number of cameras they want to get for that number (. This is like saying "I'd like to get 8 new laptops for $1000". Sure, you may be able to fulfill this request, but not necessarily to the degree of performance the customer was expecting. Personally, I've never seen a CCTV "kit" that was anything more than low to mid-grade gear targeted towards those generally lacking a clue. Ideally you would assess what problem(s) your customer hopes to solve with their CCTV system, determine correct number of cameras, placement, resolution, recording duration, etc. Then present a "good, better, best" proposal, or at least a "barely compliant" and "more robust" solution. More than likely a solution that uses modern technology and provides usable video on par with their expectations will look very different from this kit. At the *very* least you should discuss all this up front so that when you are done with the install and they are disappointed at not being able to read a license plate across the parking lot in the dark you have this conversation to fall back on for "I told you so" purposes.

Yes.

Well, you don't give us much to go on, but I'm going to guess that you've got low to mid grade analog cameras covering too much area. Digital zoom expectations are mostly a factor of pixels per foot (assuming decent lighting, focus, etc.). You'll want 50 pixels per foot minimum in most cases. This means an analog camera can cover an area about 14ft x 10ft max and give you good detail on zoom. Larger areas will yield less zoomable detail.

The solution is to not use junk cameras with IR LEDs sprinkled around the lens. Use a quality camera *and* a quality external illuminator.

They will be most likely *barely* better than the 420 line cameras. You could get a 1000 TVL camera, but the image you see is still going to be D1 (~704 x 480 pixels). To get good facial details you're going to need 50 pixels per foot minimum. 704/50 = 14.8. If you try to cover an area more than 15 feet wide your detail is going to suffer. With *ANY* D1 camera. IP cameras come in D1 resolution, but also in much higher resolutions. D1, BTW, is about .3 Mega-pixels. A cheap 2 mega-pixel camera has 7x the total pixel count, and can cover an area about 38 feet wide and get strong detail. On your internal network it doesn't matter if you have to assign an IP address to each camera. You're using a reserved subnet (eg: 192.168.x.x) so the IPs are free. Type in a couple dozen numbers (IP address, subnet mask, gateway) to the camera and go. You can fight with the D1 gear as long as you want but you'll never get the image quality that I think you are expecting to get. It's not the IP address that gives the camera the higher resolution, it's the fact that IP cameras are not bound by the signal constraints of D1.

There are two issues with most digital cable systems: 1) The frequencies allocated to channels are remapped at times. So, even though MTV might always be channel "50" or whatever on your box, the specific frequency (and compression) used to deliver it may change as channels are added to the system or other upgrades are done. Your STB (set top box) is given instructions on what in the spectrum is "channel 50" to you. This makes it difficult to find a free channel like with analog systems because things may move around. 2) Your STB is programmed by the cable company, and checks for authorization for each channel. So, you might find a free bit of spectrum, purchase a $1000 QAM modulator (yeah, they ain't cheap), and hook it all up and the STB still wouldn't tune the channel anyway. If you're just getting basic digital cable in clear QAM (meaning that you can use the tuner in a newer TV to watch your TV without needing a STB), then you can get a QAM modulator and usually make this work. But, it's going to be around $1000... Or, if your TV has PIP you could try using that and an analog modulator.

I hide mine in the cameras...

In all seriousness, there is no "future" for analog. What you see today is what you get. Those cameras may continue to sell, but there is nothing new coming.

Who is talking about analog cameras? You stated it as an advantage IP cameras have, just saying its not. Go back and read slower this time. The discussion was about HD-SDI vs. IP cameras. I never mentioned analog cameras. I'm aware that analog cameras can be backhauled over fiber (though even then you have to watch for high frequency rolloff and image degradation with *some* fiber systems). However, HD-SDI currently cannot. So, if you need to go a very long distance AND want a higher-res image (eg:NOT D1), then you will have major problems implementing that with HD-SDI.