I was speaking with a relative EE who manufactures CCTV systems for the medical industry. He was telling me how they measure video cables from various manufacturers. They found that Belden was 8 IRE down at 100 ft but Mogami cable was only 1 IRE down at the same distance. He asked me what cable I use and if I had any idea how the cable measured? I could site the 3db down specification but I've never bothered to put an IRE pattern on a 100 ft of cable and see what happens. I will shortly though. The difference between theoretical performance and real world.

Alan

Must be nice to have time to do such things. We don't have the time. Most importantly, how the hell does this affect what we do?

Um! Video of any variety.

Belden what (i.e. which model)? What was the test method? I'm extremely skeptical the results are accurate.

I feel the same way.

BTW Alan, I was surprised you never commented on this thread:

[Link: remotecentral.com]

If you have the time, I was interested in hearing your opinion.

I will even go further and say that short of seeing proof it's pure hooey.

I didn't ask which model of Mogami either but I'll email him. Method was to put an IRE pattern down 100 feet of wire and measure the result. Very simple, very accurate and very much real world. I began the conversation with the oscilliscope questions on the CEDIA Designer test and was initially surprised at his agreeing with the practice of putting an osciliscope on every wire pulled. On the other hand he agreed with me that while such tests are important they oughtn't to be done if they are paid for by the client. As they often need ultra-high resolution on their medical CCTV systems he revealed his test findings on cables. I was also surprised at his pronouncement about Belden. Hence the reason I began this thread. I'll email him for more information. He's sending me Techtronix's manual so I can see what the various readings for damaged cables will be. He also told me that his software verson of an oscilosope costs 5k, nieigh as much as the real thing. Supposedly some lesser lights exist which he promised to look up for me. Empirical evidence vs beliefs? He is a traditional EE. If it cannot be measured with a test instrument or calculated mathematically it is bogus. Give me the measurements.

And Daniel fixed the spelling on my tag. Thank you.

Alan

IRE is a level, not a pattern per say. You can have a test pattern gen show a white field of 100 IRE, or a black field of 0 IRE, or a scale from 0 to 100 in 10 IRE steps, amoung other patterns.

The loss per 100' of cable is related to the frequency, so the loss in IRE units would have to specify some sort of standard video pattern that has a high frequency pattern in order to mean anything. So if by example I use a flat white field of 100 IRE, the loss I measure would basically be defined by DC resistance of the cable, since the white field is a DC pattern. On the other hand a needle pulse pattern would have a high frequency pulse, maybe 25 MHz, and so could possibly be attenuated and never get to the full level.

I stuck some numbers in Excel, based on some popular Belden cables. I checked the Mogami web site, and I can't find a cable that is equvilant to RG-6 or RG-59, only miniature video cables. If anyone has pointers to other Mogami cables I might be missing, let me know and I can compare them. I compared DC resistance, and spec'ed (some extrapolated) losses at 1 MHz, 10MHz, and 75 MHz for Belden 1694a, 1505a, and banana peel 1282S5, and boiled that down to loss in IRE units for a 100' cable.

Type, IRE loss at DC, 1M, 10M, 75M
1694a, -0.4, -2.7, -7.9, -16.8
1505a, -0.7, -3.4, -9.8, -21.5
1282S5,-2.1, -5.6, -15.9, -40.4

As you can see at DC there is not much loss, and I doubt anyone would notice even on the banana peel. However at 75MHz, it's quite a bit more. In reality most of your sharp detail in HDTV is going to be in the 10-30MHz range. High frequency helps with edge sharpness, and fine vertical lines. Images like a guy in a fine striped shirt can show this, where the stripes will be more contrasty with better high frequency response. Images that have fine dust particles floating around will benefit from this as well.

Alan,

EE or not, it would not be the first time I saw one make an error in measurement. Unless my understanding is seriously flawed a difference of 8 IRE is ***EXTREME***. It flies in the face of common sense (and every measurement I've ever taken and looked at) that two properly designed AND COMPARABLE cables that are being used for their intended application could vary that much. But again, it would depend on so many factors. What was the cable? What was the source/resolution/frequency etc?

I see after posting Late Night Bill already posted much more elegantly than I did...

[Link: mogami-wire.co.jp]

pps 40-41, Please note the specification of .018 db/ft @10MHz and the 18.3Pf/ft capacitance specifications for the high frequancy 75 ohm coaxial cable. That works out to 1.8db for 100 ft @ 10MHz and 1830Pf @ 100 ft

Let's look at the Beldon 1694 as it specs out the best of the Belden cables. It's capacitance spec is 16.2Pf/ft and loss @ 10MHz .s .72 At 100 ft of cable that's 72 db
and 1620Pf

Assuming the specifications are accurate, damned right the Mogami will perform better than the Beldon 1694A.

Alan

Alan, some of the things you say don't pass the stink test. And I mean the one where you don't even have to lift your arm:


Copy and paste, I swear. Go back and read it. This guy said that if a client DOES pay for a test, you SHOULD NOT do it? That is what you said, right?


I will bet that one hundred feet of 18 gauge SPEAKER WIRE won't be down 72 dB at 100 feet at 10 mHz. Are you sure you read or copied this correctly? 72 dB is a damn effective spec for a rejection filter! And if we figure a probably excessively BAD response curve for the wire of, say, - 6 dB per octave, which is a first-order filter, not a hunk of cable, then this wire would be down, let's see
66 dB at 5 mHz, 60 dB at 2.5 mHz, 54 dB at 1.25 mHz, let's say 52 dB at 1 mHz, 46 dB at 500 kHz, 40 dB at 250 kHz, 36 dB at 125 kHz, 30 dB at 62.5 kHz, 24 dB at 31.25 kHz, 18 dB at 15.625 mHz, 12 dB at 7.8125 mHz, 6 dB at 3.9 mHz, 0 dB at 1.95 mHz

In other words, if this wire were so crappily made that it performed as a first-order filter, we could expect it to be flat to just under 2 mHz. Now, if its high frequency attenuation were any less than 6 dB per octave, we would soon have it performing with noticeable losses in the audio spectrum. I don't think this passes the stink test.
No, this spec does not pass the stink test.

The belden spec is loss per hundred, whereas the Mogami spec is loss per foot.
Considering all that, I plugged this into Excel again:

type,DC,1M,10M,75M
1694a,-0.4229,-2.7253,-7.9550,-16.8236
1505a,-0.6721,-3.3949,-9.8429,-21.4764
1282S5,-2.1085,-5.5939,-15.8605,-40.4338
Mogami1,-2.0888,nodata,-18.7169,nodata
Mogami3,-0.5635,nodata,-10.6695,nodata

Mogami1 is the first column front thier datasheet, the miniature, and Mogami3 is the large multiconductor. So it looks like the M-1 is comparable to BananaPeel, and the M-3 cable is comparable to Belden 1505A, which is a RG-59 clone. It would be nice if thier loss charts went above 10MHz.

I'm thinking this is going to be a web page soon. Maybe tomorrow, time to go watch a movie. :)

Shock of shocks Ernie. The man agrees with the CEDIA Designer Test writer that testing cables after they have been pulled is essential but also agrees that performing work without compensation is foolish. He did not say that is should not be performed if the client was paying for it. I mistyped and should have included a "not" between "are" and "paid."

Mind clearing up what "I will bet that one hundred feet of 18 gauge SPEAKER WIRE" is supposed to mean? Belen 1694A is most assuredly a "low loss coaxial cable,"
[Link: bwccat.belden.com]

Alan