So you are now saying that the reference *you* cited is not correct? And is as bad as the reference you now cite below...
Hmmmm...
Exactly the same as above!
In fact, if we do follow Section 5.2.2.1, and assume that it means to use short circuit and open circuit conditions on one port while the other is measured... we *don't* get your claimed
6 dB of loss. Instead what we get is something slightly less than the -3.5 dB that I've said is usual. See the explanation that I provided, or the one that is in your earlier cite if you want to understand why that is.
(BTW, I do *not* thing that section was meant to describe how to measure insertion loss of hybrid splitter/combiners.)
And just why would we measure the device at some impedance
*other* than the characteristic impedance of the circuit? If the load is 50 Ohms, we certainly should be using a 50 Ohm splitter...
Right. And if you are not smart enough to use a 50 Ohm splitter in a 50 Ohm circuit, what point is there in discussing the impedance match?
If you actually do find *any* credible description of the methods used to measure insertion loss, it *will* specify not only that the circuit impedances be match, but that 3 dB or 6 dB impedance matching pads be used between the signal generator and the input of the device under test as well as at the output of the device if it is connected to anything other than a resistive load.
That, for the uninitiated, is to eliminate silliness such as you are suggesting. (And I'll note that the Mil-Std you cited suggests
10 dB isolation pads! So much for your impedance mismatch!)
Why would you use a 50 Ohm splitter, or measure one, in a circuit that is not 50 Ohms?
Quotes please. It does not say that.
That may or may not be. I suspect they had it *exactly* correct, and that you simply didn't understand it at the time.
One method that really helps in understanding how hybrids work is to look at more than one type. The *principle* becomes obvious, instead of the details of any one single method of making a hybrid.
Note that in some hybrids there will be *significantly* different characteristics than those we have discussed. For example, it is relatively easy to make a resistive hybrid; there are at least to forms of transformer hybrids; and if you want to get really interesting there is a zero impedance summing junction bridge too! (That last one is not passive though, so it has very different applications than the others.)
(I'll also grant that there *is* more than one definition of "insertion loss" being commonly used. However, I'm going to leave it to you to find something that demonstrates it.)
Did you notice the similarity in that discussion and what it says about multi-circuit components in the Mil-Std documents you cited?
Both of my cites agree and both are correct - signal in the line before the insertion of the device compared to the signal in the line beyond the device after the device is put into the line.
Same definition.
yes - why you have such a hard time reading the "signal before insertion" part?
multiple outputs
1) the device's listed value is not measured at the impedance of the circuit. It is measured at a standard impedance which may or may not be the impedance of the circuit.
2) don't make the mistake of confusing measurement of a parameter with the definition of a parameter. E.g., this definition is not impedance specific (but a bench measurement and standard published value must be at a specific impedance)
3) As you know, the impedance of RF line varies with frequency and length. On many circuits, it isn't that easy to hold 50 ohms or 75 ohms. Long runs for some frequencies may require 75 ohm cable to match 50 ohm outputs. E.g., the actual impedance of a home cable system with its fairly wide frequency spectrum is not 50 ohms - it is only nominal 50 ohms. (This condition goes to the problem of defining Insertion Loss, circuit variability, and not accounting for the particular circuit's impact on a device.)
4) Insertion Loss is not an impedance specific term but it accounts for impedance, and that also means
5) that a standard measurement is not meaningful if your method of in-place input vs output is measured - and
6) what is "proper impedance" depends on what is needed from the device, not on a standard number. If I want to shift bandwith by mismatch, then that is proper for my device, even though it is not what the manufacturer had intended for its most common use. So while my device will have an insertion loss, it may or may not be the same inertion loss found by measuring using the 50 ohms used in Mil Std 220b. (The text of 220b notes that for purposes of that std, 50 ohms will be used to test in that procedure- and that 50 ohms is not part of the definition)
For example, the same 75 ohm input impedance device put into a 50 ohm impedance line will have a different insertion loss than a 75 ohm input impedance device put into a 75 ohm line. And they will likely have a different Mil Std listed insertion loss than their actual insertion loss.
Another example - a given splitter can have two insertion losses, depending on the impedance of the two circuits into which it is placed. It will function equally well in either. (If one is a 75 ohm circuit, and the other is a 50 ohm circuit, it has an internal resistor that splits the difference)
So, for Insertion Loss to mean the loss due to insertion, impedance matching it must be accounted for in an impedance-isolating manner.
The way one remove impedance and all other circuit affects in the definition reference level is to have a device-free circuit compared to a device-included circuit.
that is why the definition uses the "no device in the line" level -
don't know much about rf generators or voltmeters, do you?
to broadband on the cheap for one
You want yet another quote that says "before the device was put into the line"?
While possible, only remotely so, as I had been a tech for 6 years before I went to the University to get my degree, and I had used a lot of splitters and combiners and rf power as reflectors to check levels and points of degradation to find the location of cable problems, and that set of equations showing that there had to be a 3db loss had stuck in my mind.
It was a straightforward set of equations, and it was news to me, but his equations were accurate and my recheck bore out his statement.
In the next few days, when I get the time, I will get the RF equipment out and measure a splitter.
No, I don't. From my cited article, page 5:
"Mismatch effect on insertion loss
For the power splitter of Fig. 4, consider the situation where one output port, B, is shorted. A signal applied to port S would result in half the power appearing at each port, A and B. Since port B is shorted, all the power appearing there would be reflected back into the power splitter. Half of this power reflected would be dissipated in the internal resistor and the other half would appear at port S. The power loss within the power splitter would therefore be 1/4 or 6dB below the signal power originally applied to port."
MilStd requires shorted or open ports, not impedance amtched ports.
The Mil Std procedure does not terminate multi-ported devices with pads of the devices rated resistance. It requires opens and shorts.
The cold one, on the other side of the world from Point Barrow.
BTW, you don't get the info I noted by dropping in. Much of the stuff I listed are areas reserved for the several senior technical personnel, and one wouldn't know the names of much of it unless you actually worked on them and had been on site for several months. Their existence isn't secret, btw, just details were.
FWIW, there weren't ten people in the world that worked on the site's "special stuff" in a 20 year span - and in a couple year span, there was just one person for all sites for two particular rather special pieces of equipment. I know this from a couple of experiences with them, but the best one was from calling Point Barrow on a secure line to check with their seniormost tech on an arduous aligning of one very very access-restricted piece of equipment (as an aside - only I and one other person were even allowed in the room, with dual log-ins, guards brought in to guard the door when I went in, etc, and not even the site commander was allowed in there- not as unusual as it sounds, btw. Had a similar setup in the marines w/o the guards-the CO stopped at the door and said he was not allowed in but wanted to see how it was going. ) and the Point Barrow tech told me that "X" on our site was the leading expert on that piece and area and was the one who would be doing theirs, and "X" should be able to help - that was news to me, since I am "X". Pretty much a clue that I was on my own with that equipment.
You know that at the end of this thread, if we are men about it, one of us will be wrong and may even have to admit it.
Bright side is that if we didn't care. we wouldn't hold a position so vigoruously.
That statement is absurd. Learn how to articulate, because the words you used do *not* mean what you were should have been saying.
What you meant was that a sine wave can be analyzed as if it were an equivalent series of pulses with linear rise times.
Hee hee, that *is* funny. It doesn't "create an impedance" during the first 10% or during the last 10%?
So you are blowing blue smoke, trying to obscure some fairly simple principles. Input circuits are frequency sensitive, which manifests itself as an input impedance that varies with frequency, and to whatever degree that occurs and is significant, the amplifier will exhibit amplitude distortion.
Really? Who'd a thunk, eh?
Of course in practice, if broadband characteristics are desired, the circuit is designed to make a short enough rise time to cause that effect shorter than will in fact exist during operation. Generally speaking, with modern technologies that is fairly easy to do too!
That statement is meaningless. In particular it is invalid in an argument claiming that rise time has some significance in a digital circuit in regard to decoding the symbol values in a digital signal.
The fastest possible rise time might well *limit* the symbols rate that a particular channel can use, but it is *not* necessarily used to decode symbol values. Which is to say, if the transient response is at least sufficient to allow a rise time higher than the requirements for any specified symbols rate, it makes *no* difference how much faster than that the actual rise time is, or how much slower the actual symbols rate is.
For any given symbol to symbol change in values, the fact that rise time might be virtually non-existent (they may have the same pulse amplitudes) means that lower rise times equally do not affect decoding values of the symbols.
As I've said, rise time is not a factor in decoding symbol values for a T1 system, nor is it necessarily the case for other digital systems.
Insignificant. A T1 system has a broadband input, it does *not* have an impedance mismatch sufficient to cause amplitude distortion and *any* of the requisite frequencies (DC to
775KHz).
Unless of course the input is broadband, compared to the signal frequency spectrum.
And if the slope does *not* "create" such a mismatched impedance, then everything you are saying can be ignored!
All you have demonstrated is that a 1.544 Mbps T1 circuit might not work if clocked at 10 Mbps. Whoop dee doo!
That range might be from DC to 10 times the fastest rise time necessary. And in fact, with a T1 system that is true. It has essentially a broadband input. The rise time is *not* affected by the amplifier input circuitry *at all*, if for no other reason than the fact that all practical transmission lines (i.e., cable pairs) have a dramatically greater affect on the slope of the signal.
You are mighty confused about just what is and what is not BMEWS!
There is no BMEWS facility at Barrow, and never has been. The original DEWLINE had the POW MAIN site at Point Barrow, just north of Barrow. Today that is a Long Range Radar Site (LRRS).
The Alaska BMEWS site is located on Clear AFB, south of Fairbanks. (Nearly 600 miles due south of Barrow.)
The original communications connection between the DEWLINE and BMEWS was a 50 KW tropo link between Barter Island (BAR MAIN) and the Fort Yukon (FYU) White Alice Communications System site.
Oddly, voice circuits between the DEWLINE and BMEWS were routed through the Eielson AFB (EAF) switchboard located near Fairbanks.
You should try being honest about where you have no background too.
Thule Greenland. Didn't you even know the name of the place?
Sure.
Sure.
Sure.
Hence you are admitting no direct experience with any of that, and are repeating scuttle butt from the DEWLINE about BMEWS. :-)
I'm afraid you won't impress me much with that sort of thing. Base Commanders were not allowed into the areas where I worked (for decades), and I had access to areas where no *guards* were allowed either.
Personally, I thought it was hilarious! I had occasional access to the room where seismic data was analyzed in an attempt to detect Soviet nuclear testing. The kick was that on any given visit, the majority of questions one might ask would get a response similar to "For security reasons we cannot discuss that." Of course what they really meant was that they didn't know how to answer my question, or didn't want to. But since those people did one year tours of duty, and I was a permanent local resident... it didn't take long before I know more about what was there than all but the absolute sharpest of the military personnel.
That had some strange effects... one Major, who was personally in charge of security, got upset with me one time and actually turned in a security alert claiming there was a violation. He didn't think it out very well before doing this though, because in effect he turned himself in (directly to the Pentagon no less) for allowing a security violation to exist. He was gone...
However, the one that provided continued fun was actually something very simple. There was a "secure, failsafe" scramble circuit at Eielson AFB, which was intended to allow the Command Post at Elmendorf AFB to scramble fighter jets from Eielson. It has three lights on it (green, red, and yellow) which indicated the status of the voice circuit. I had circuit diagram of the equipment, and if they opened a trouble ticket on that circuit I would dial up the telephone next to where this equipment was, and ask whoever answered the phone to tell me what the equipment was doing.
Usually, they would refuse to do so! I'd tell them it was *not* secret in any way shape or form. They wouldn't budge. So then I'd plug a test set into the circuit at my location, and switch the signaling supervision on the circuit back and forth, causing the lights to blink. I'd say something like, "Okay, just tell me if the red light blinks three times." And then do that with the other lights too. It was usually possible to *hear* the awe in the voice of the poor fellow. They'd go from total control, being able to refuse information to someone... to absolute confusion.
We already know that you are blowing blue smoke on most of this, and haven't got a clue about many technical things. Coax cable... what a joke! Insertion loss is not far behind. And your confusion trying to read a technical paper one hybrid splitter/combiners was an outstanding indication.
Take a pill. Don't try to impress people by talking about things you don't actually understand.
So why did you say they were ignorant? Your cites are not in agreement with what *you* have claimed! You simply don't understand what they are saying. Look at the quoted text above, which specifically says that what you have said is *wrong*. That is quoted from your first cite. Nothing in your second one contradicts it.
Read the "How to measure insertion loss" from your cited reference, and tell me how that does anything other than tell you how you've misunderstood what the definition is.
Perhaps. Regardless, it provides 1) a useless measurement because it does not measure the actual effect of proper use (where the ports would all be correctly terminated), and 2) an absolute denial of what you have been claiming (the results of measuring as described will not be those you claimed to be correct).
That is a stupid statement. What is "a standard impedance"??? And why would anyone design or specify or measure the device at other than the intended operating impedance?
At the impedance specified! And of course if the device is specified as a 50 Ohm device, it will *not* have a mismatch at 50 Ohms.
I know for a fact that is not true.
I know for a fact that is not true.
Wrong.
What you have said comes from not understanding circuit impedance and/or transmission line theory.
Insertion loss is *most certainly* an impedance specific term. It is specified at a specific impedance.
Wrong. (Do you understand what the Mil Standard document meant by "isolation" pads?)
You just said above that it was a "standard impedance". Make up your mind!
Nobody wants to "shift bandwidth by mismatch" with a hybrid splitter.
It won't be the way *you* want to do things! But that is why it isn't done that way, because the rest of the world actually does want to have measurements that have meaning in relationship to operating circuits.
You are not reading that correctly. They are merely saying that the example they provide happens to be 50 Ohms. But in any real measurement, the impedances of *all* devices must match the device that is to be measured. You do *not* use 50 Ohm devices to measure devices that are intended to be use in 135 Ohm circuits, for example.
Wrong. It will *not* be used in a 50 Ohm line, it will *not* be measured in a 50 Ohm line. If it is designed for 75 Ohms, it will be used in 75 Ohm circuits, it will be measured using 75 Ohm equipment, and the values will be useful.
(I'll point out here that the RF splitter/combiners that we have been discussing are not particularly impedance sensitive, and virtually all of them will measure and function precisely the same in either 50 Ohm or 75 Ohm circuits. The biggest difference is usually the type of connectors used...)
That is total bullshit. As I have repeatedly pointed out to you, the internal resistor determines the impedance of the
*input* port. It has no direct effect on the output impedance.
Is that supposed to mean something?
Tell us why the Mil-Std document suggested 10 dB isolation pads when measuring Insertion Loss!
Tell us why the Mil-Std document suggested 10 dB isolation pads when measuring Insertion Loss!
Tell us why the Mil-Std document suggested 10 dB isolation pads when measuring Insertion Loss!
Tell us why the Mil-Std document suggested 10 dB isolation pads when measuring Insertion Loss!
So, I'll be very interested in the fantasy land description you give!
(And to answer your question: I know a great deal about accurate measurements using RF generators and voltmeters.)
Above you said it was to "shift bandwidth", now you claim it is to broaden bandwidth. Can you explain the contradiction?
Please cite where it says what you claimed above. First you claim one thing, then you quote something else.
Which of course has *nothing* to do with that resistor.
Not if you include loss in that resistor.
Whaa hooo!
Which
1) makes the measurement useless for design purposes, because it does not indicate what levels in the circuit will be in a proper configuration.
2) demonstrates that your claims about Insertion Loss levels are false. Instead of being close to the 6 dB you claim, that measurement will result in a loss value of *less* than my stated typical 3.5 dB.
Note that the circuit is totally unbalanced, half of the power is no longer dissipated in one of the ports, but instead is split between the resistor and the input port. That power (1/4 of that applied to the input) is not all "dissipated" in the input port, but it will actually be reflected back into the balun, and half of it will appear at the measured port, while the other half will again be split between the resistor and the input port!
The result is that output power at the measured port will be
*greater* when the other output port is not properly terminated than it will be when it is terminated. Thus, measured that way the "Insertion Loss" will be less than the levels found when the device is used in a correctly configured circuit.
It is a useless measurement.
Both require isolation pads between all external devices and the device under test. The Mil Std procedure is interesting as a diagnostic tool, but provides a meaningless measurement for design specification purposes, which is what Insertion Loss is generally used for. Indeed, for that reason I do *not* believe that the section you cite was ever intended to be used to measure insertion loss of slitter/combiners.
Regardless, you two cited texts do *not* agree with you. They are correct, but just do not mean what you seem to thing they mean.
the measurement of insertion loss is NOT the definition of insertion loss. Stop mixing the two.
proper use is the intended use, not the use you determine for another
2) an
What? It is EXACTLY what I have been saying.
The one specified in the measurement used to derive the listed value. 50 ohms.
Please - all devices with values derived from testing are from tests done at a standard impedance.
No, but who says I will use it at 50 ohms?
I know the common belief, but if it were not frequency dependent, loss tables would not need to list loss by frequency (e.g., Belden's tables for Rg-59, 6, 100). - unless ohm's law has been rescinded for cables.
Same length, same cable, same input level, different frequency - different db loss.
Z is E-field/ H-field. How do you get reduced/increased output of E and H if Z is supposedly constant? The only way that happens with a constant Z is if E goes up the same amount that H goes down. And if that happens, that means E and H varies by frequency.
And the lower frequencies are then all H and the upper frequencies are all E?
_ It varies by frequency because when measured, the losses in a given cable type varies by frequency_ (Unless those cable loss tables have been revised to remove frequency)
As to length: I only have knowledge of that as it was the military's practice, implemented due to output loading. Longer cables loaded the amplifier more than shorter cables NOT more loss at the other end, less out with longer cables in place.
An amplifer output varies according to output impedance. Variation in output means variation in what other load parameter of the amplifier? It only has two other basic parameters if impedance is constant - voltage and current. VSWR you say? What does the reflected wave actually create at the amplifer output such that less signal leaves the amplifier? Higher fields at the output point compared to fields farther down the cable? So the cable's field varies according to position along the cable? So it is length dependent?
As far as I know, the parameter that changes output is length.
I can only say that in BMEWS, on several occasions I had to use RG 59 for places that would have normally used RG 58, on a common amplifier that used RG 58 in other like locations except for shorter runs, RG 59 where the line length was very long, per schematics with a special note there regarding that substitution. From that requirement, and the lowered output of the amplifiers when the usal cable was used, it certainly had an effect.
Later in my career, I had used equations provided by a cable manufacturer to determine the maximum power transfer (referred to by them as "effective impedance") of their coax based on line length and frequency, and installed mixed RG58/59 systems in accordance with them. (I saw them still in place last month)
see ohms law above, in spite of magazine articles that says otherwise
NO !!! THERE IS NO IMPEDANCE VALUE _ANYWHERE_ IN THE DEFINITION -
ONLY -_ONLY_ IN THE MEASUREMENT PROTOCOL IS THER ANY IMPEDANCE VALUE LISTED.
Measurement to get a value, yes - in the defintion - no.
STANDARD impedance, as in testing standard, and PROPER, as in desired by the design
No confusion in that whatsoever.
shift, as in broaden.
1) You are not privy to all designs using splitters.
2) If I want to do so using a 5K ohm splitter on an 8 ohm amp to get more bandwidth, then it is proper -
"As measured herein, insertion loss is represented as the ratio of input voltage required to obtain constant output voltage, with and without the component, in the specified 50 ohm system".
As _MEASURED_ herein... in the specified 50 ohm system.
what I said was look at the first 90 degrees of a sine wave and look at the rise time of any wave (as defined as the point between 10% and 90% of peak). That is what I said, and that is what I meant.
No, that is NOT what I meant. I mean EXACTLY what i wrote . Read it.
OK, since this is becoming difficult for you - the waveform creates an impedance which the amplifier sees.
do try to stay with it and stop wandering off - ONLY the rise time is being adressed here
characteristics
So in your world amplifers can recognize a signal of any size? Signals below the threshold of recognition? 100 billion light year away starlight in a TV set? You really are not versed in amps, are you?
In particular it is invalid
totally wrong - "symbols rate" has NOTHING to do with recognizing a single pulse
it is not aboput lower - it was about faster. And after what you just said - you DO know what a rise time is, don't you? It specific definition?
Not the rate of pulses, the rise time of an individual pulse... this explaining the basics of how an amp recognizes ONE pulse is next to hopeless with you because you keep wansering off.
We are talking about the rise time of ONE pulse only. Not some broadband series of pulses.
You are never going to get the basics - you keep running off to pulse rates.
I am quite aware of what is BMEWS and where it is and exactly what it has/had. You, who have never even worked there and probably haven't even seen an antenna\, are suddenly a BMEWS expert?
I didn't say there was.
I said the other site was at Clear, and we called the other site Point Barrow.
I also asked if you were near Clear AFB.
(our system's third site was at Fylingdale moors, tracker only, and we called that Fylingdale Moors.)
I could give a shit less if the site was at Point Barrow or Piss Ant Burrow. Never been there, had no need to go there, just know that "Point Barrow" is what we in Thule called the other fixed site.
The
Ok - so now I wonder why we called it "Point Barrow".
It is total nonsense. A sine wave is *not* "identical to the rise time of a pulse" if the pulse is a '("square wave")'.
You therefore meant nothing, because what you said does not make sense.
The waveform does not create an impedance. The amplifier does not see impedance. You are not making sense. (The amplifier is what creates an impedance, and the waveform is what will see it. Not the twist...)
The above addresses "ONLY the rise time". If the input circuit is designed to be sufficiently broadband, the rise time of the signals seen in actual operation will be longer than the minimum rise time acceptable to the circuit.
Again, you are not making sense.
I'm obviously far more versed in digital amplifiers than you, and realize that there are other forms of "digital" signals than just binary pulse trains. Regardless, the input to the tuner that was being discussed when you originally made that sort of statement, is *analog*.
You do realize that the pulse, in the systems that you are discussing, *is* nothing but a "symbol", and that the pulse rate is the symbols rate.
The difference is that your terminology is very restrictive, while what I'm saying is very generic. You are limiting what you say to binary digital systems, and I'm not.
Rise time is not necessarily related to decoding/encoding digital signals. And worse yet you are again talking strictly about binary systems with pulses of voltage or current. Such systems are very common, but there are many others that are different.
And, as I noted, sometimes there simply is *no* rise time between two adjacent symbols, even in a binary encoded digital system. Consider an RS-232 circuit, with the bit pattern 101. Clearly the second bit is encoded in a pulse that has some significant rise time relative to the value of the preceding 0 bit. Now think of the bit pattern 1011111. There are 3 bits there which have a pulse that exhibits *no* rise time at all.
How can you say that rise time is the determining factor in decoding a sequence digital symbols?
Clue: rise time of the individual pulses is *not* significant.
Take a look at a diagram I drew a few years ago:
formatting link
That shows the specifications for the waveform of a T1 pulse as seen at the receiver input. Note that rise time is not specified. Pulse width, overshoot, amplitude and wavelength are all specified... because *that* is what determines the value of the pulse.
The only frequences that make any difference at all are the above mentioned range from DC to 775KHz. Give it some thought for a moment, and try to see how a 1.544 Mbps signal can be decoded using *no* frequency higher than 775KHz. Obviously rise time for pulses *cannot* be significant!
If what you are claiming were correct, a T1 circuit's bandwidth would necessarily have to be specified at some frequency greater than 775KHz, which is indeed the pulse rate frequency. In fact, it would have to be at least 3 times that frequency in order to retain sufficient information in the pulse to indicate rise time.
But... it is *not* required. Hmmm...
What, you think "ONE" pulse requires less bandwidth than a series of pulses???? You do realize that in order to have your "square wave" pulse, the bandwidth has to be *much* higher than that required to simply pass a series of pulses...
Ahem, a square wave results from odd *harmonics* of the basic pulse rate frequency...
The rise time is *directly* related to the pulse rate.
You are so confused that you have yet to understand what we are even talking about. You do realize that in a T1 half of the information bits are sent without sending any pulse at all??? No pulse, no rise time. How can you claim rise time is used to distinguish the value of a symbol???
I doubt you understood the signficance of that statement. The input to the T1 equipment is broadband, and can accept a very high frequency (compared to either the pulse rate or the rise time of the fastest pulse that will be seen). And that makes little difference simply because the cable connecting a T1 receiver to the output from a T1 transmitter is in fact a low pass filter which removes virtually all high frequency components. 775KHz is the highest frequency component required to allow a T1 to function. Lines are specified by the amount of loss at 775KHz, and no higher frequency is ever measured.
-- Floyd L. Davidson Ukpeagvik (Barrow, Alaska) snipped-for-privacy@apaflo.com
Oh, I've never worked there and never seen an antenna?
You are grossly confused. Point Barrow *is* a specific site location, located approximately 600 miles from the BMEWS site at Clear. (If it hasn't sunk in yet, yes I've been to Clear AFB, to Point Barrow DEWLINE, to Eielson, etc. etc.)
So you don't have a clue where any of this is located...
Wellll... I was wondering something along those lines myself. Nobody else that I've ever met had that same misunderstanding...
The measurement of insertion loss is an implementation of the definition, otherwise it is not insertion loss that is being measured.
You are confused...
The intended use is as a design parameter, allowing a loss budget to be completed for a circuit. The technique described does not provide a value which can be used for anything other than comparison to another device, which makes it superfluous. A real insertion loss measurement provides a value that is useful for both the design and the maintenance of equipment.
You claimed 6 dB, roughly 3 dB more than I said. But that clearly is going to cause a measurement which is *lower* than my stated 3.5 dB.
How can you say it is 6 dB if the measurement is going to be slightly greater than 3.5 dB?
You are confused...
So you are claiming that description meant you should use 50 Ohms, even if the device is 135 Ohms and so is the circuit?
Astounding.
That may or may not be 50 Ohms. It is determine by the impedance of the circuit the device is designed to be used in.
I'll grant that *you* probably don't have enough sense to avoid using a
50 Ohm device in a 600 Ohm circuit.
But you are confused!
That is more total confusion on your part.
Of course I am assuming that we are talking about transmission lines. If you mean a piece of wire, that's different. But coax, for example, is a broadband device. It's impedance does not vary over a significant frequency range. It's impedance does not vary with length at *any* frequency.
Your reference to loss tables is another indication of gross confusion on your part. If the characteristic impedance of a cable was a direct indicator of loss, then 75 Ohm cable would necessarily have more loss than 50 Ohm cable. And 600 Ohm cable would be even higher!
Do you actually find that pattern in the loss tables... ;-)
So? It still have the same characteristic impedance.
Do you have a table that shows impedance vs. frequency????
The loss varies by frequency, but the characteristic impedance of the cable is relatively a constant.
Astounding. You are seriously confused about transmission line theory, and given that you can't even understand what defines a coaxial cable, it comes as no surprise.
Cable *loss* increases with length. That is simply the amount of power dissipated per unit of length, and of course it will be twice as much for a cable twice as long. That does not mean the impedance of the cable is different for shorter or longer cables. It just means the losses are parallel.
Yes! Typically the lower the load impedance the higher the output.
Oh, wait... what was that you claimed earlier about maximum power transfer when the impedance is *matched*. How can you relate that to lower impedance causing more power output...
(Not that I'd want to confuse you further, but I am expecting a response to the above that would come from Dilbert's Pointy Haired Boss.)
But you are confusing the impedance of the transmission line cable with the impedance of the load. Two distinct animals, not at all the same...
But that is not an indication that the impedance of the line is different.
But you *don't* know... (apparently anything).
Actually, that is just plain silly.
Oh my. You really are confused. That is purely bullshit on your part.
I don't think there is such a thing as a schematic that suggests exchanging RG59 for RG58...
Read any good book on transmission line theory. Look up things like "stub matching", and "impedance transformation". It will explain what that actually was all about.
You are confused. What has Ohms law got to do with loss related to the characteristic impedance of a transmission line?
No shit, Sherlock! That is because a hybrid can be designed to work with various different circuit impedances; but you *must* measure it at the impedance that it is going to used at, which *must* be the impedance it is designed for (if you want to ensure proper operation).
Exactly. The definition does not specify that we can only have insertion loss at 50 Ohms. That is because we might very well be measuring a 75 Ohm device used in a 75 Ohm circuit.
Your claim that 50 Ohms is a "standard" impedance which must be used is horse pucky.
Okay, you don't. The isolation pads are used to assure voltage measurements are a true indication of the correct power level at a given point, which is done by padding to assure that the impedance is the specified impedance.
Hence if you have a device claimed to be 50 Ohms, it is not valid to connect a signal generator to it and measure the RF voltage across the input, because the voltage will vary with the impedance, which might not actually be 50 Ohms.
If a pad is place between the signal generator and the device, the pad will provide a known impedance. The power level at the device input is determined by the pad, hence a 3 dB pad means there will be half the power measured at the input to the pad. The larger the pad, the more accurate the impedance is guaranteed to be. Typically 3 dB or 6 dB pads are used with most devices. The 10 dB pad specified by the Mil-Std procedure will assure that almost any variation will still result in an accurate reading. In practice, with a
20 dB pad the reading will be accurate (I forget now, but within something like 0.3 dB) even if the output of the pad is either open or a dead short.
So now you admit that the circuit impedance *is* what counts, not some mythical "standard impedance".
I'm sure glad you got over that confusion... ;-)
But I do know that a mismatch will *not* broaden the bandwidth...
Won't work. You are confused.
???? What do you think "tested" means, if not "measured"?
That is exactly my point. It is *not* a "standard impedance", it is the impedance of the system under test that determines the impedances used for test equipment.
Hee hee, I doubt you caught the significance of that...
I see the rest of that article got too deep for you to even attempt a response.
I trimmed the excess. We can assume that points you fail to respond to are stipulated to be as stated. That of course pretty much deletes *everything* you've said so far... :-)
sometimes you get lucky with an easy fix. of course the first time i ran into a goofy situation where an audio production studio just had to have a cable installed and it created a buzz in the audio i was beating my head against the wall (figurititivly) until i hit upon this workaround.
some additional info can be seen here
formatting link
next step is to eyeball the coax cable and see if there is a grounded (earthed i guess you would say) feed through 'block' somewhere outside. if not add one. something like this
formatting link
the idea is to try and keep lightning outside your house and anything you touch (like the connector) at ground potential.
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.