EMRFD Message Archive 4558

4558 2010-04-23 12:59:16 Jim Miller SWR Question Message Date From Subject If I have a multiband non-resonant dipole fed by open wire balanced line how should SWR be estimated? If I don't terminate the feedline at the other end it will present a different impedence for comparison at the dipole than if it is shorted or (more desireably) fed by a matching network to 50ohms. Should the process be to model the antenna and feedline first to define the matching impedence required, implement the matching network in the model including the presumed 50ohm source then calculate the SWR on the line? The reason for all of this is to determine the SWR on the feedline to understand its loss. tnx jim ab3cv It's my understanding that the impedance of the source doesn't affect the SWR on the line, so that need not be factored in. You can figure the SWR from the line's characteristic impedance and the complex impedance at the antenna terminals. Generally with open wire line, it's assumed that losses are going to be acceptable even with some rather extreme SWRs so calculations aren't needed. In my case I used enough ladder line to reach from my 180 foot dipole to my Johnson Matchbox and then go one the air on multiple bands. But that doesn't really answer your question. You could use modeling to figure the impedance at the antenna terminals and then use the "transmission line equation", available in various spreadsheets to find the impedance at the shack end for various lengths. And you can use the published loss figures to figure the loss per length. (I've got spreadsheets for this stuff. TL Details, if I have the name correctly, should also do it.) Most people want to do the calculations only for the purpose of making sure that the impedance at the tuner is within its range. Although less sophisticated than NEC type simulations, there are simple programs that will estimate the antenna impedance and then the impedance at the rig end of a line. On W5DXP's page is a BASIC program that does this, linked at the bottom ... http://www.w5dxp.com/notuner.htm I took the BASIC code and converted it to an EXCEL spreadsheet. The intent is to find a feedline length that will give satisfactory performance (impedance at the tuner not extreme) on various bands. I think you can find some similar calculators among the programs of the late Reg Edwards. 73- Nick, WA5BDU Hi Nick Thanks for the info. For my dipole the SWR is 13:1 assuming 390ohm windowline on 10m so the SWR loss is a little over 0.5dB for 50ft. The same antenna modeled with source at the other end of the 50ft of 390ohm windowline is 10:1 for 0.4dB. So I'm a bit confused as to which I should assume. I suspect the latter is the proper one. Neither loss is a big deal but I just like to know which is right. jim ab3cv If you want to be picky, probably neither one is right :-). But arguing about 0.1dB differences in feedline loss is hard when really 1dB is hard hard to distinguish even in well-controlled conditions. There is a school of thought around ever since the SWR meter became popular, that SWR is the end-all and be-all of antenna/feedline/transmitter matching. It is not a very productive school of thought. The ARRL handbook and antenna book charts for ladder line loss (largely replicated from year to year with no source ever given) have been questioned. Some like VK1OD have modeled ladder line using numerical/computer methods and predict a much greater loss than the traditional ARRL tables have quoted. There was a measurement offered in QST a few months ago but it actually didn't measure anything. A big reason why they didn't measure anything, is that the loss was so small it could not be measured with the instruments and techniques being used. But the handbook numbers said that the loss would be so small it couldn't be measured. So while the authors touted this as experimental verification, it is at best verification with a big asterisk (as in "we didn't prove it right but we couldn't prove it wrong"). There is some animosity between VK1OD and ARRL authors/publications on this topic. I am critical of the ARRL tables because they do not cite a reference. But I'm also critical of VK1OD's calculations because they are not verified by measruement. Really you think this could be resolved by an actual measurement but it seems to me that I haven't seen an actual measurement done. It was a weakness of ARRL publications in the past that references for values in the tables were rarely offered. I think that there was a desire (unconscious?) that the ARRL handbook be "the bible" which did not need to refer to anything else. The handbook in the past decade or so has done a real turnaround, and they do offer a very nice bibliography with references for many/most measured numbers and all sorts of other things. But I haven't seen a measurement done. I'm not arguing for "ignorance is bliss", after all this is a group based around the thought that things can be measured, but really what you need to do is match your transmitter to your line and antenna. This will probably take some kind of tuner. And get it I was confused (still am, maybe) about the difference in measurement or analysis conditions that resulted in the two different SWRs with the same antenna and same feed line on the same frequency. Now I'm wondering if it's the fact that the loss in the line causes the SWR measurement at the source end to be lower than predicted (for a lossless line?) due to attenuation of the reflected signal. As a matter of taking this to the extreme, I recall an article one on testing a UHF transmitter that suggested that a 100 foot length of typical somewhat lossy RG58 could be a decent dummy load even if open or short circuited at the far end. I'm still a little surprised that a fairly short length of ladder line on 28 MHz would result in a reduction from 13:1 to 10:1. 73- Nick, WA5BDU Hi Nick First of all I apologize for the confusion caused by my use of 10m when I mean 10Mhz or 30m. The difference in loss for me isn't so much a problem of actual (or predicted) loss of power rather that I just am not understanding which representation is the accurate one or that perhaps both are but I'm lacking an understanding of why. Thanks Jim Miller AB3CV Nick Kennedy wrote: > I was confused (still am, maybe) about the difference in measurement or > analysis conditions that resulted in the two different SWRs with the same > antenna and same feed line on the same frequency. > > Now I'm wondering if it's the fact that the loss in the line causes the SWR > measurement at the source end to be lower than predicted (for a lossless > line?) due to attenuation of the reflected signal. As a matter of taking > this to the extreme, I recall an article one on testing a UHF transmitter > that suggested that a 100 foot length of typical somewhat lossy RG58 could > be a decent dummy load even if open or short circuited at the far end. > > I'm still a little surprised that a fairly short length of ladder line on 28 > MHz would result in a reduction from 13:1 to 10:1. > No surprise at all. Read http://www.qsl.net/4z4tl/pub/Feeder_tunes.pdf to find out why. -corey - AE5KM Here's an online calculator that should help: http://www.microwaves101.com/encyclopedia/calvswr.cfm The SWR at the antenna terminals is 13. Plug that into the calculator and note that the equivalent return loss comes out to 1.34 dB. At the source end of the transmission line, the 0.4 dB line loss will increase this return loss to 1.74 dB. When you plug that number back into the calculator, the SWR comes out to 10.02. This is a good illustrati If the idea is to get an efficient match, then all you need to do is add an impedance bridge and adjust your balanced line tuner to null. If on the other hand, you investigations are more demanding, then i would recommend a VNA. - Farhan On 4/24/10, Nick Kennedy wrote: > It's my understanding that the impedance of the source doesn't affect the > SWR on the line, so that need not be factored in. You can figure the SWR > from the line's characteristic impedance and the complex impedance at the > antenna terminals. > > Generally with open wire line, it's assumed that losses are going to be > acceptable even with some rather extreme SWRs so calculations aren't needed. > In my case I used enough ladder line to reach from my 180 foot dipole to my > Johnson Matchbox and then go one the air on multiple bands. > > But that doesn't really answer your question. You could use modeling to > figure the impedance at the antenna terminals and then use the "transmission > line equation", available in various spreadsheets to find the impedance at > the shack end for various lengths. And you can use the published loss > figures to figure the loss per length. (I've got spreadsheets for this > stuff. TL Details, if I have the name correctly, should also do it.) > > Most people want to do the calculations only for the purpose of making sure > that the impedance at the tuner is within its range. > > Although less sophisticated than NEC type simulations, there are simple > programs that will estimate the antenna impedance and then the impedance at > the rig end of a line. On W5DXP's page is a BASIC program that does this, > linked at the bottom ... > > http://www.w5dxp.com/notuner.htm > > I took the BASIC code and converted it to > an EXCEL spreadsheet. The intent is to find a feedline length that will > give satisfactory performance (impedance at the tuner not extreme) on > various bands. I think you can find some similar calculators among the > programs of the late Reg Edwards. > > 73- > > Nick, WA5BDU > > > I "rediscovered" TLW and loaded it on my computer. It had gotten lost in my last HD kerfuffle. My version 2.02 is from the ARRL Antenna HB 20th edition. It is a really nice accompaniment to EZNEC since it does all the transmission line calculations as well as drawing several tuners complete with loss calculations. I reran my doublet on EZNEC without the feedline then plugged the resulting antenna impedence numbers into TLW. For my case of 50ft of "450" window line driving a 88ft doublet at 50ft on 30meters it gives me: matched line loss of 0.041dB plus an SWR loss of 0.188dB for a total line loss of .229dB and for a low pass L tuner it shows an additional 0.23db of loss with the assumed values of Q. So the overall loss is 0.46dB or about 10%. This doublet has been my only antenna since becoming licensed and like many is a byproduct of the site restrictions. It has worked "good enough" for me so far on 80-20m. But I recently discovered that the pattern on 80m was pretty much a cloudwarmer and have started using the feedline as a 50ft vertical with the 88ft doublet acting as a top hat. This greatly improved the performance on 80m for my needs and even gives me a loadable antenna on 160m as well. I have 8 sixty foot radials as a ground plane. I'm planning to add a 5 band hexbeam to cover 20-10 and now the horizontal 88 will only be used for 40 and 30 so I've been more closely investigating its performance on those bands. 73 jim ab3cv Tim wrote: > >The ARRL handbook and antenna book charts for ladder line loss (largely >replicated from year to year with no source ever given) have been >questioned. Some like VK1OD have modeled ladder line using >numerical/computer methods and predict a much greater loss than the >traditional ARRL tables have quoted. A greater less in some circumstances, but less loss in others... see below. >There was a measurement offered in QST a few months ago but it actually >didn't measure anything. A big reason why they didn't measure anything, >is that the loss was so small it could not be measured with the >instruments and techniques being used. But the handbook numbers said >that the loss would be so small it couldn't be measured. So while the >authors touted this as experimental verification, it is at best >verification with a big asterisk (as in "we didn't prove it right but >we couldn't prove it wrong"). > >There is some animosity between VK1OD and ARRL authors/publications on >this topic. I am critical of the ARRL tables because they do not cite a >reference. But I'm also critical of VK1OD's calculations because they >are not verified by measruement. > >Really you think this could be resolved by an actual measurement but it >seems to me that I haven't seen an actual measurement done. > The key point is that most transmission lines suffer greater losses at locations where there is increased current than they do at locations where there is increased voltage. "VSWR" on its own doesn't tell us which of those two cases applies; we also need to know the actual length of the line and the actual value of the terminating impedance. In other words, a chart of "increase in loss" versus "VSWR" has got to be assuming something that may not always be true. The differences will tend to average out over a very long line that has several maxima and minima of both voltage and current; but that will not be true for a short line. The size of the error is not the main issue here. What's important is the accuracy of our understanding about transmission lines. -- 73 from Ian GM3SEK http://www.ifwtech.co.uk/g3sek