Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: Prev 1 [2] 3 4 Next   Go Down
  Print  
Author Topic: Length of coax from xcvr to amp affecting input swr, why ?  (Read 37523 times)
W1QJ
Member

Posts: 2985




Ignore
« Reply #15 on: January 02, 2012, 08:07:24 AM »

Amp Supply amps all have arco trimmers for C1 and C2 with their ATI-6 boards. Their coils are fixed.  Just the opposite of most other amps.  Keep in mind that most of these older amps we have been discussing  L4B, SB-220, TL-922 etc. were designed to operate around their own transmitters/transceivers like T4X,TR-4, SB 401 HW-101, TS-520, TS-820.  All these had tube finals with adjustable pi nets.  None of these radios had built in SWR meters so we really didn't know what the swr was to the transmitter.  They had enough drive regardless to push the amp pretty hard as they would not cut back in power.  The intersting thing is when the swr meter reads high, is the radio actually folding back power?  SOmetimes you may notice an swr but not a foldback in power.  If that s the case it could be the harmonic power reflection.
Logged
W8JI
Member

Posts: 9748


WWW

Ignore
« Reply #16 on: January 02, 2012, 03:54:45 PM »

Hi Lou,

Actuaklly the direction coupler sees any power from the outside as full reflected power. I have a 30 dB high power (2.5 kW) pad on my test bench. If I run 1000 watts backwards into that pad, so there is 1 watt backfeeding my IC751A while it is transmitting, the 751A SWR meter moves up to ~1.2:1 and the radio just barely starts to fold back on some bands.

If I run 2 kilowatts, so it backfeeds the 100 watt rig with just 2 watts, the ICOM noticably shuts down and shows a 1.3 SWR. 2 watts of backwards power is the trigger point for my 751A's detector.

The cathode of a tube in a higher power grounded grid amp can have well over a dozen watts of harmonic energy, and if the tuned input doesn't knock that energy all out, the exciter sees it just as if it were normal reflected power.

73 Tom

PS...
Some of the other answers fly in the face of how standing waves work. Although Dan covered this, I'll repeat it. We all should know, if we understand SWR, that line length does not change SWR if the SWR detector and line are matched (as long as the line has low loss). We also should know exciter source impedance does not affect SWR on the line.

This leaves only a few things:

1.) The interconnect cable is bad

2.) The SWR detector in the radio is misadjusted

3.) The radio is oscillating or making a spurious that changes level with line length

4.) The amplifier is putting harmonics or spurious back into the radio

None of this mattered, as Lou already said, when we had old tube rigs. But we should still know how SWR works on transmission lines. The source impedance cannot change SWR downstream.
Logged
K6AER
Member

Posts: 5745




Ignore
« Reply #17 on: January 02, 2012, 06:54:55 PM »

If you have activated the antenna tuner in a transceiver and the output impedance is say 100 ohms the cable length will be critical when coupling into the amplifier. It will depend where the VSWR nodes combine or cancel to produce different impedances along the line. A VSWR meter will only read the line to the terminating load. The radio is still seeing VSWR if it is tuned to something other than 50 ohms out from the antenna tuner and is set to an impedance other than 50 ohms.

When driving the amplifier make sure your antenna tuner is turned off.
Logged
VE7RF
Member

Posts: 212




Ignore
« Reply #18 on: January 02, 2012, 11:58:55 PM »

If you have activated the antenna tuner in a transceiver and the output impedance is say 100 ohms the cable length will be critical when coupling into the amplifier. It will depend where the VSWR nodes combine or cancel to produce different impedance's along the line. A VSWR meter will only read the line to the terminating load. The radio is still seeing VSWR if it is tuned to something other than 50 ohms out from the antenna tuner and is set to an impedance other than 50 ohms.

When driving the amplifier make sure your antenna tuner is turned off.

##  I disagree. As long as the amp has an existing tuned input.... using the built in auto tuner is not an issue.   I shut off the auto tuner.  Then tune the amp.  THEN switch the auto tuner on..then tweak the auto tuner for flat swr. Xcvr is happy.  Amp is easily driven.

Jim  VE7RF
Logged
VK4TUX
Member

Posts: 37




Ignore
« Reply #19 on: January 03, 2012, 01:09:26 AM »

##  I disagree. As long as the amp has an existing tuned input.... using the built in auto tuner is not an issue.   I shut off the auto tuner.  Then tune the amp.  THEN switch the auto tuner on..then tweak the auto tuner for flat swr. Xcvr is happy.  Amp is easily driven.

Jim  VE7RF
[/quote]


Jim is probably not interested in responding to any of my comments, however for the rest here I had a situation where I made up a new 6ft coax length (properly soldered plug body > braid & 1000v megger tested, with 20 amp conductivity test also; I am an electrical tradesman) and tried it out, working well, apart from 15m K3 ind swr which was unusually very high. Amp is a sb-220.

I grabbed my Yaesu tuner and placed it inline with its own extra 3' patch lead. I thought; "Ok I will just use the tuner on 15m and put it in bypass for the other bands".
I did that because I use the amp 15m position for 17m, which had a good input swr as per above setup, but previously both 15 & 17 were ok, and neither perfect as you would expect on the one band input.

Well when I connected all that up I had it in bypass and went through the bands and they were all 1:1 swr on the K3.

15m had gone from very bad to very good, with the tuner not being utilised and just in bypass .

Then I think ok 6 & 3, lets try that, and did with the same good result.

I went to each band and tested tx into the cantenna load and all good 1:1 on the K3 every band once amplifier was tuned correctly.

The amplifier input network impedance match effects the linearity of the result signal if not good, and I wondered whether using a tuner cleans up a signal effected for that reason or the signal stays bad despite good power transfer?

I have always avoided using a tuner in the input line if possible and found a way here.
The input coax length appears to become part of the amp input network impedance match,
despite not agreeing with reported theory.

I can only put it down to the method the transceiver swr detection circuit uses to get it's result. I think good found lengths may vary for different amps and setups, I cannot see the problem with experimenting.

Adrian ... vk4tux

Logged
VE7RF
Member

Posts: 212




Ignore
« Reply #20 on: January 03, 2012, 03:03:44 AM »


## Perhaps the answer is to replace the silver mica's used for C1..with arco's ?

Jim  VE7RF

Jim, I guess you still have both the 3' & 6' patch coax pieces. In your massive amp resources kit can you find a 259 double female joiner to connect them together to make a 9' length?
If you could try that I would be interested to see the result if that's not too much trouble?

Adrian ... vk4tux
[/quote]

##  I'm gonna try that..if I can find a good one.  And no, the coax is not bad.   The swr is dead flat when amp in bypass mode too.   The external bird/cd/as meter's all jive with the yaesu meter.  It's the yaesu meter that's the one that will determine if the po of the xcvr folds back, not the external meters.

##This will all change..again, when the new upper shelf is finished on desk #2.  I tried this b4 and it worked good.  I string the L4B's, nose to tail, all in series..and tune em up on different bands.  They will all easily handle 2 kw on bypass. A rotary switch ensures that only one amp's key line is on at any one time. IE: all the amps are switched to operate mode.   Key line from xcvr goes into the rotary switch box...and multiple outputs, one to each L4B.  I can also tune more than one amp up on the same band as well.

##  I find this hard to believe that nobody else has this problem with rising/varying swr with different lengths!  For folks who have adjustable arco's for each band,I'll bet even they have the same issue with different lengths.  Except they can re-tweak the input swr down flat.  How many of these newer amps use arco's ??

Later.. Jim  VE7RF
Logged
VK4TUX
Member

Posts: 37




Ignore
« Reply #21 on: January 03, 2012, 05:48:16 AM »


## Perhaps the answer is to replace the silver mica's used for C1..with arco's ?

Jim  VE7RF

Jim, I guess you still have both the 3' & 6' patch coax pieces. In your massive amp resources kit can you find a 259 double female joiner to connect them together to make a 9' length?
If you could try that I would be interested to see the result if that's not too much trouble?

Adrian ... vk4tux

##  I'm gonna try that..if I can find a good one.  And no, the coax is not bad.   The swr is dead flat when amp in bypass mode too.   The external bird/cd/as meter's all jive with the yaesu meter.  It's the yaesu meter* that's the one that will determine if the po of the xcvr folds back, not the external meters.

Yes I mentioned exactly that * in an earlier post here. Putting an out of rig swr meter in the line to amp is a waste of time.
I never mentioned bad coax, must have been another poster, yes swr great here too with amp on stdby when using 6', same deal.

At the end of the day as long as the feedpoint impedance (an amp or antenna input) is a good match, and the transceiver swr detection is happy then all is good. It's a shame Rich got banned here, he is very knowledgable on the subject.

##This will all change..again, when the new upper shelf is finished on desk #2.  I tried this b4 and it worked good.  I string the L4B's, nose to tail, all in series..and tune em up on different bands.  They will all easily handle 2 kw on bypass. A rotary switch ensures that only one amp's key line is on at any one time. IE: all the amps are switched to operate mode.   Key line from xcvr goes into the rotary switch box...and multiple outputs, one to each L4B.  I can also tune more than one amp up on the same band as well.

##  I find this hard to believe that nobody else has this problem with rising/varying swr with different lengths!  

Try google and you will see abundant comment on the subject. Apart from that many use antenna tuners to deal with it, but the issue with that is you may not detect a change in the amp input , relay failure, tube filament or supply failure that would show up as bad rig swr all of a sudden esp with auto-tuners.

On the sb-220 band positions 10m is used for 12m & 10m
                                           15m is used for 17m & 15m
                                           20m is used for 30m & 20m
40=40 and 80 =80 and these two can be optimised for the label band, however the others are a compromise which brings up the situation in the first place. However the right coax length can be found that keeps things happy on all bands incl warc.

Tom is right about dedicated hi-Q separate inputs for every band you use, but with a sb-220 things are different unless you want to leave the unlabelled bands out and optimise and raise Q on perfectly tuned input networks for what is on the dial.

Adrian ... vk4tux

« Last Edit: January 03, 2012, 06:00:16 AM by VK4TUX » Logged
VK4TUX
Member

Posts: 37




Ignore
« Reply #22 on: January 03, 2012, 06:11:27 AM »


##  I find this hard to believe that nobody else has this problem with rising/varying swr with different lengths!  

Try google and you will see abundant comment on the subject.

An example here;

"
>From time to time, people have mentioned how Collins used to recommend a
certain length of coax between the KWM2 and 30L1 amplifier.  As the
following anecdote may illustrate, this isn't just ancient history.  I hope
this story helps someone else troubleshoot a similar problem.

Recently, I redid the T/R relay system in my SB-220 to add QSK, and was
perplexed to discover, after the modification, that the amplifier suddenly
showed a high input SWR on 40m (only), as well as abnormally high gain,
grid current and plate current on that band only.  Visions of an HF
oscillation, or some other abnormality, had me digging for answers. 

Thanks to Tom, W8JI, who asked if I had changed anything else at the same
time.  Eureka!  I had "improved" the jumper from exciter to amplifier by
shortening it about 8 feet.  Just now, I put the old one back in, and
everything is back to normal.  I don't pretend to understand WHY this
happens, or why it happens only on one band, but there you have it.



73, Pete Smith N4ZR
"


Adrian ... vk4tux
Logged
G3RZP
Member

Posts: 1323




Ignore
« Reply #23 on: January 03, 2012, 07:03:45 AM »

I have been told by some users that the 'magic length' of cable for the 30L1 provides an impedance which keeps it more stable on 15 and 10. I cannot say that I'm impressed by the design: a 40 microhenry plate choke gets a lot of current through it on 80m, and the lack of neutralisation begs for problems at the top end of the frequency range. Add to that the low values of grid bypassing capacitors - presumably to give negative feedback - and there are troubles stored up. The mechanical mounting of the PSU equalising/bleeder resistors is a real bodge job.

Of course, the low values of grid bypass cap do give negative feedback in a perfect world with no stray inductance or capacitance and thus zero size tubes to have zero lead length.
Logged
W8JI
Member

Posts: 9748


WWW

Ignore
« Reply #24 on: January 03, 2012, 07:09:24 AM »

Jim,

The problem you have is not unusual, and manifested itself in the 1980's when solid state rigs appeared. You aren't the first one to notice it. Ameritron ended up with a Drake TR5 from Dentron that had some weird big finals subbed in for the OEM parts, and no active SWR foldback. I suspect that was to make it  work with the poor or no tuned inputs. The Q and arrangement of ATI6 boards, when they used them, was all over the place. On ten meters, some just had one trimmer and a wire jumper to replace the coil!

No one needed to pay attention to back-fed harmonics when exciters had no SWR detection and foldback.

Of course the TR5 had other issues, it would oscillate out-of-band with some loads, making me wonder how anyone used it to adjust amplifiers.

73 Tom
Logged
W8JI
Member

Posts: 9748


WWW

Ignore
« Reply #25 on: January 03, 2012, 07:17:16 AM »

I have been told by some users that the 'magic length' of cable for the 30L1 provides an impedance which keeps it more stable on 15 and 10. I cannot say that I'm impressed by the design: a 40 microhenry plate choke gets a lot of current through it on 80m, and the lack of neutralisation begs for problems at the top end of the frequency range.

The 30L1 makes a great oscillator on 15 and 10 meters. Try running some stability tests as you mesh the load cap. Collins fought for years because they insisted on wapping feedback around two stages in transmitters, and those floating grids. The floating grids were a good idea in the 30S1, but that amp had no grid current and had a tetrode with a screen to shield the input from output.

Quote
Add to that the low values of grid bypassing capacitors - presumably to give negative feedback - and there are troubles stored up. The mechanical mounting of the PSU equalising/bleeder resistors is a real bodge job.

Of course, the low values of grid bypass cap do give negative feedback in a perfect world with no stray inductance or capacitance and thus zero size tubes to have zero lead length.

Not even then. The problem is grid current.  Calculate the phase and level of feedback as the grid cycles through its current range.

The grid caps are the lower part of the divider, and the grid-cathode time-varying and frequency-varying impedance is the upper half of the divider. It has to be the silliest circuit in the world for a mult-band amp with grid current.

Logged
G3UUR
Member

Posts: 119




Ignore
« Reply #26 on: January 03, 2012, 07:45:31 AM »

Some of the other answers fly in the face of how standing waves work. Although Dan covered this, I'll repeat it. We all should know, if we understand SWR, that line length does not change SWR if the SWR detector and line are matched (as long as the line has low loss). We also should know exciter source impedance does not affect SWR on the line.

None of this mattered, as Lou already said, when we had old tube rigs. But we should still know how SWR works on transmission lines. The source impedance cannot change SWR downstream.

There's no new physics here, just some more you don't fully understand. A pair of GG 3-500Z tubes present a load to the driver that is a non-linear function of the drive voltage, and as such it is quite different from the constant impedance loads used in the elementary transmission line theory you're considering, Tom.

The source impedance does affect the distortion caused to the drive signal voltage and the reflected harmonic energy seen on the transmission line. What is needed right at the cathode/filament is a voltage source with very little source impedance to drive the linear without too much distortion being caused by non-linear loading. That's very dependent on the phase delays of all the networks and cables between the driver and tube cathode/filament.

Bruene VSWR sensors are fairly flat with frequency, whereas pick-up line types are more sensitive to harmonics unless some form of compensation is used, but they'll all indicate harmonic reflections pretty well unless the input matching network attenuates them well.

The L4B input matching has, as Tom said, an output Q of 2.1. It also has an input Q of around 1.5 and overall the circulating current is 2.3 times the output current at a load of 80 ohms. This is apparently what the network was designed to match, but the actual impedance is half this at the peak of the combined grid and cathode current and is infinite over part of the cycle when the tubes are cut off. Obviously, a circulating current of 2.3 times at the average impedance is not sufficient for any length of connecting line and a higher Q is required for solid-state drivers to smooth out the load variations. A Q of 10 would definitely be enough, but many handbooks say more than 5 should be used - this lower figure is close to the Q (6.28) that provides energy storage in the network equal to that passing through.  

The real question is why just 3ft of coax between the transceiver and the linear works so well and other longer lengths don't. Perhaps, if you'd read Warren Bruene's article more thoroughly, Tom, and understood it, rather than dismissing it because he disagrees with your views on the HFTPA conjugate match question, you'd have been able to model it and tell us!

Without sufficient Q and a good flywheel effect to smooth out the non-linear load at the cathode/filament, the low-pass filter and the connecting cable, along with the input matching network phase delays will determine the source impedance seen right at the tube and affect the extra distortion created there. Rigs withe different low-pass filter designs will require different "magic lengths" of patch lead for the same type of amplifier and the same band, and different magic lengths will be required on different bands. You might find a length that works on most bands, but it might take quite some effort to find it. The simplest solution is to just re-design the input networks, Jim. Trying to understand exactly what's going on may take quite a lengthy study and not be worth to time required.  

Happy New Year de Dave, G3UUR.
  
Logged
AF6LJ
Member

Posts: 2




Ignore
« Reply #27 on: January 03, 2012, 12:01:04 PM »

Hi Lou,

Actuaklly the direction coupler sees any power from the outside as full reflected power. I have a 30 dB high power (2.5 kW) pad on my test bench. If I run 1000 watts backwards into that pad, so there is 1 watt backfeeding my IC751A while it is transmitting, the 751A SWR meter moves up to ~1.2:1 and the radio just barely starts to fold back on some bands.

If I run 2 kilowatts, so it backfeeds the 100 watt rig with just 2 watts, the ICOM noticably shuts down and shows a 1.3 SWR. 2 watts of backwards power is the trigger point for my 751A's detector.

The cathode of a tube in a higher power grounded grid amp can have well over a dozen watts of harmonic energy, and if the tuned input doesn't knock that energy all out, the exciter sees it just as if it were normal reflected power.

73 Tom

PS...
Some of the other answers fly in the face of how standing waves work. Although Dan covered this, I'll repeat it. We all should know, if we understand SWR, that line length does not change SWR if the SWR detector and line are matched (as long as the line has low loss). We also should know exciter source impedance does not affect SWR on the line.

This leaves only a few things:

1.) The interconnect cable is bad

2.) The SWR detector in the radio is misadjusted

3.) The radio is oscillating or making a spurious that changes level with line length

4.) The amplifier is putting harmonics or spurious back into the radio

None of this mattered, as Lou already said, when we had old tube rigs. But we should still know how SWR works on transmission lines. The source impedance cannot change SWR downstream.

5. the amplifier input is not presenting a 50 ohm impedance to the radio cable and bridge.

Logged
G3UUR
Member

Posts: 119




Ignore
« Reply #28 on: January 03, 2012, 12:22:03 PM »

5. the amplifier input is not presenting a 50 ohm impedance to the radio cable and bridge.

Not only is it something other than 50 ohms, it's not constant over the RF cycle either!
Logged
AF6LJ
Member

Posts: 2




Ignore
« Reply #29 on: January 03, 2012, 03:36:44 PM »

5. the amplifier input is not presenting a 50 ohm impedance to the radio cable and bridge.

Not only is it something other than 50 ohms, it's not constant over the RF cycle either!

That is very true.
The older radios were better suited to handle this situation.
Logged
Pages: Prev 1 [2] 3 4 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!