Preamplifier and antenna relay configuration

The following discussion about the preamplifier and antenna relay configuration took place in the Moon-Net reflector. I found the subject so interesting that I decided to put together all the messages in this page.

On 5-Oct-2003 WA7SKT wrote:
Im in the process of constructiing my antenna relay and preamp panel to mount in the box at my EME array.

I would like to know if just the 2 antenna relays are enough for adequate isolation of GASFET preamp or should I add an additional relay to preamp input to terminate it with 50 ohms during transmit.

I plan on keeping power on preamp at all times because I am told that it is better protected when powered up.

What are you doing?

On 5-Oct-2003 AF6O wrote:
Loren,

You need enough isolation to reduce the transmit power to a low enough level that it won't damage the preamp. If the two antenna relays have a high enough isolation spec then that is all you need, otherwise add the additional relay(s). I use CZX3500s which provide more than enough isolation to protect my preamp from the 1000w I run without the need for additional relays.

On 5-Oct-2003 F5VHX wrote:
interesting, Jim have you ever actually measured the isolation ??

I had seen this available but noted it said "non-grounding" on the spec. page (which I assumed meant the unused port is just left open)

but then they claim 65 db isolation at 150 MHz, and I find the two hard to reconcile, especially as they are still claiming -65 at 2 GHz.

would be interested in anyone else's actual measurments on this relay also..it looks a good price if the isolation numbers are credible.

On 5-Oct-2003 AF6O wrote:
No, I haven't measured it except to say that my 144 preamp survives 1000 w and my 432 preamp survives 1200 w. Both using the CZX3500s.

On 5-Oct-2003 WB9UWA wrote:
HI Loren,

First, I am alarmed that you say you are using two antenna relays and one of them is not already switching the input to a 50 ohm load. It is standard practice to route the output of the preamp down a separate feedline to the receiver input. The relay that switches the preamp input to 50 ohm, can be a low cost, low power relay. Of course all of my relays are low cost surplus type N relays anyway. The only exception is that the relay that switches the 50 ohm load to the input of the preamp is an even lower cost BNC relay. These can be seen on the GM4JJJ web site. Here, you can see exactly what I am doing. Much of this being the result of me asking beginner questions (and getting some very good answers too!) when I designed and built my array.

http://www.gm4jjj.co.uk/WB9UWA/details/

In the preamp box photo, you can see the BNC connector that contains the 50 ohm resistor. The type N relay is actually a 6 VAC relay run on DC with a series dropping resistor. Common surplus hamfest relays with large contacts will handle EME power levels at 2M. Just make sure you clean the contacts first and make sure they mate well. Often relays are discounted because the coils are not 12 VDC. Don't let this stop you. Power supplys and extra relays are easy stuff. Some of my 24 VDC relays run on the unregulated 22 VDC on my 13.8VDC station power supply.

Why trust the specs? It is easy to add the low cost second relay on the input and helps with peace of mind in case something fails. The preamp will be more happy with a 50 ohm load on it rather then possibly oscillating away when it is not connected to anything.

On 5-Oct-2003 KL6M  wrote:
Loren,

Good advice from Jim WB9UWA. I will amplify it slightly. There is great advantage to having an "isolation" relay that terminates the input of your preamp into 50 ohms. It provides an evaluation tool. It provides a comfort level when you can see the noise produced by the termination in comparison with cold sky, to tell you that your system is functioning properly. I use it frequently to determine if birdies are internally or externally generated.

I'm sure there are other uses also. Since I have gone to this configuration about 3 years ago I have not burned out a GaASFET. Prior to that I was going through them like popcorn. (got pretty good at surface mount soldering...hi)

I have a system connection diagram on my website that you may want to look at. It was used for a 2M expedition to my home town in SD and provides the typical approach we have been talking about:

http://www.qsl.net/kl6m/144expo.htm

On 5-Oct-2003 SM5BSZ wrote:
Hi Loren,

> I would like to know if just the 2 antenna relays are enough for adequate 
> isolation of GASFET preamp or should I add an additional relay to preamp 
> input to terminate it with 50 ohms during transmit.

The below is for 144 MHz only!

I never used more than a single relay and I have never blown a GASFET. I have blown the relay a few times when doing silly things like transmitting when half the antenna had fallen to the ground after a storm. The very high SWR then caused too much voltage for the N-connectors with arcing as a consequence but the GAS-FET was always well protected by the shorting to ground provided by the shorting RK500 (CX520D) relay from SSB Electronic.

> I plan on keeping power on preamp at all times because I am told 
> that it is better protected when powered up.

I have seen this statement many times. I believe it is false!

I have done some tests and found the opposite (yes, really destroying some GASFETS). It may depend on how your input cicuit is arranged though.

I always feed DC to the preamp in parallel with the coil of the antenna relay so I have transmit mode as the default position when the rig is switched off. The idea is to have some protection for lightening discharges.

The relay isolates by 80 dB in a 50 ohm system so it gives more than adequate isolation unless one makes a mistake with the cable length between the realy and the GASFET.

In transmit position for the relay, the output impedance of the relay is extremely low, inductive and practically loss-less. If you accidentally match your GASFET to this impedance, a lot of power may be transferred. More than enough to destroy the transistor.

With power applied to the transistor, you may force the input impedance to something near 50 ohms if the amplifier uses feedback, without power, the input impedance will be reactive and nearly loss-less. Inductive or capacitive, high or low depending on circuit details. When power is sent into the preamp, the gate-to-source diode will rectify and change the impedance at the gate from very high to very low. The input impedance will change dramatically because of that.

A simple input with a tapped coil will go from a very high impedance to a very low impedance when the transistor rectifies. With a cable length of 5 cm between the relay and the preamp there is no risk at all to provide anything like a power match and the transistor is very well protected by a single relay.

Another think is that it may be very helpful to be able to switch in a 50 ohm dummy load rather than the antenna for trouble shooting.

On 5-Oct-2003 K2TXB wrote:
Hi Loren. I see that you got lots of replies to your preamp question, and most of them were accurate but I did not see one detail that I think is important mentioned.

Wire your relays so they are powered up when receiving! Remove power from the relay coils to transmit. This gives you some important safeguards. First, if you should lose power to your relays or the wires get cut or intermittent connection, etc., your relays will go into transmit position and the expensive preamp and receiver input will be protected. Wired the other way, if you are the air, or key up when the relays are not working it is often disaster.

Second, when your station is powered down your antennas are not connected to the preamp and thus the preamp is protected from nearby lightning strikes.

Jim's comment about using a separate feedline for the receiver is right on the money. You don't use a relay on the output side of the preamp, but just run it down to the shack via a separate run of lower cost coax - RG8 style is fine. Inside the shack you modify your transciever to allow a separate receive input for 2 meters, or whatever band you are using. If you are using a transverter it is easy, but I have modified my TS-2000, my FT-847, and my old FT-726R for this and it was not that difficult in every case. In fact the old FT-726R was the hardest to do.

If you happen to have a TS-2000, look at my web page where I show how to make this mod (http://mywebpages.comcast.net/russk2t/). The FT-847 mod is done in a very similar way but it is much easier to get at the internal connections on the 847.

You will never regret setting your station up this way. It is easy enough to put the transciever back to 'stock' condition if you want to sell it. As someone mentioned, switching a 50 ohm load to the input of the preamp during transmit allows you to evaluate your station. If you add a switch to switch the antenna relays while still in receive mode, you can then see and measure the noise from the 50 ohm load. When you switch back to the antenna, if the noise gets a lot stronger then you know you have a problem and you can measure how bad the problem is. When pointed at cold sky and if you don't have any serious side lobes that are picking up ground noise, you should be able to see a reduction in noise when you switch to the antenna.

Hope these comments help, Loren.

On 5-Oct-2003 PA0JMV wrote:
Hi Loren,

In my station I have always used just ONE coaxrelay (isolation 60dB or better) to switch between TX-RX and *never* blown any active device since 1988.. (and I can assure you, I've been running PWRRR :-)). (No 50 ohms R for testing, the box outside is too small for another relay ) I would strongly recommend to activate the coaxrelay during reception only (fault resistant) and switch OFF the preamp-DC during TX sequences. Please use separate coax runs for tx and rx lines, that will save you a LOT of troubles!!! During RX, I feed 12V DC via the RX-coaxline to the preamp and the outdoor relay, and a second DC-type of contact(+ cable) on my coaxrelay inhibits any RF-power to the transverter when the relay and preamp are in receive mode. This way I don't need anything like a sequencer, just a simple combination of relaycontacts is the most effective way to protect a preamp ! Looking forward to working you and others off the moon :-)

On 5-Oct-2003 G3ZIG wrote:
Exactly what I do here. As Mike say, you can also switch the pre amp between the antenna and 50 ohm load to check sky noise.

I power my relays on receive. Just make sure there is enough contact pressure in the unpowered position.I have had problems with burning of the contacts on transmit

On 5-Oct-2003 WA7SKT wrote:
Hello all,

Thanks so much for the many replies to my post...I have archived every one to refer to again.

>From all the responses I have these common configurations.

1: Run separate transmit and receive line between station and array.

2: Use isolation relay with 50 ohm terminator for preamp.

3: Of course have antenna relays powered during receive..de-energize during transmit.

4: Some power down preamp during transmit..some dont.

One thing though is I didnt realize so many had transmit and receive separated on transceiver or some using separate transmitter and receiver.

I dont have mine split and wont for a while so if I run separate transmit and receive lines I may just have to use an interlocked relay to switch the 2 lines back to transceiver. At least only exciter power will be at this point.

On 5-Oct-2003 VK3AUU wrote:
One thing that nobody mentioned, was to make sure that you use reasonably low loss coax between the preamp and the receiver in the shack otherwise you will degrade your performance on receive. Just any old coax will definitely not do.

On 6-Oct-2003 UT5ER wrote:
With the interest I read discussion about RF switching. I regret, but the accessible Russian relays (REW-14; REW-15) have isolation between ports less than 30 on 144MHz (I have measured with the the shaking frequency generator SRI, I am not sure in term correction). Declared -60DB have struck me. Prompt please the reference where it is possible in more detail to learn a name and to get acquainted with performance attributes such relays.

On 6-Oct-2003 PE1ECM wrote:
Apparently I have missed some of the discussion on the Russian REW14 relays. I cannot find any Moon-net REW 14 related mail.

Since I have bought some of those relays for my EME station ( in progress ) I am " deadly " interested in possible lack of isolation.

The relays are supposed to isolate 2.5 kW each on 144 Mhz ( Alex, UR4LL, told me that they could be used on 432 with > 2 kW ). Obviously, 30 dB Tx/Rx isolation on 144 Mhz would destroy my preamp(s) and/or maybe my transceiver.

That is why I ask for respondents who have used these REW 14 relays on 144 Mhz and report about their experiences. Help me save my Rx front ends... :-)

BTW : I cannot find any specification on this REW 14 relay even not on ND2X's page. Who has ?

Thanks for cooperation.

On 6-Oct-2003 YO4FRJ wrote:
Well Peter, from my past research experience I will advice you always to double check all the ex-USSR mil components specs and data. If your coax relays deal was based on Alex's ( UR4LL ) advice, I will suggest you to believe him. However, you can ask his support with additional measurements, specs, etc., or a double check on to this isolation parameter issue. In respect with Arkady's measurement procedure we can ask him more detailes and might help if we can bring Alex on the same Moon-Net table of talks.

On 6-Oct-2003 WB9UWA wrote:
Hi Arkady,

Did you measure the isolation with a 50 ohm generator and a 50 ohm load? If either is not 50 ohm then the measured isolation is likely to be different.

If you are using a single relay ahead of your preamp on 2m, you must be very careful that you get the expected isolation. The preamp is usually not a 50 ohm input so isolation can be much less then expected. You may have to plan for a specific cable length between the relay and the input of the preamp to maximize the isolation.

On 6-Oct-2003 UT5ER wrote:
Hello, Peter. In general all of Russian manufacture is made with safety factor! The relay not exception of this rule. Regret, but I can not quickly find the passport of the relay. But if it is necessary for you I will do later. Coil voltage -24 Volt (reliably works at 19 V), current - 0,2?. To get a rise attenuation between ports apply two relays in series. Similar system using Alex, UA4AQL and as I remember - Victor, UA4API (in the near future 4 years - QRT) For this connection cable is not necessary. Certainly it is necessary to have a socket "daddy" for one relay. Into "mum" of the relay is inserted the Cu wire with appropriate diameter (hardly) and long; opposite wire end soldered to to the central contact of a socket "daddy". On rear part of a socket, intended for cable input, there is a groove, it corresponds to a groove on the relay. Finish, we have two relays, connected by ports in series, on 150 MHz VSWR less than 1,1(by ut5er). In the last letter I mentioned attenuation -30DB. The measuring device is intended for measurement of amplitude-frequency respons. between two-port, an error also is possible. But similar results told me RA3LE

On 6-Oct-2003 UT5ER wrote:
> Hi Arkady, 
> Did you measure the isolation with a 50 ohm generator and a 50 ohm load? 
> If either is not 50 ohm then the measured isolation is likely to be different.

 Certainly, but for the relay 75 Ohm.

> If you are using a single relay ahead of your preamp on 2m, you must be 
> very careful that you get the expected isolation.

 I am compelled to apply the second tiny relay, not coaxial. I do not know the appropriate term in the English language. It is a glass pipe abt 3mm diameter and inside 3 gilt contacts for switching. A pipe is filled with inert gas. It is a bad choice, such relays (it is strange, but fact) at passing-through small current of high frequency (and audio too, in a telephony) it is necessary as well voltage for break-down (puncture) of an oxidation layer. I collided at work repeatedly with this phenomenon, but I can not explain, gold is not oxidable.

> The preamp is usually not
> a 50 ohm input so isolation can be much less then expected. You may have to plan for a
> specific cable length between the relay and the input of 
> the preamp to maximize the isolation. 

It stands to reason. Quantity of MGF-1302 which I have burnt out for 7 years comes nearer to 8 pieces(plus Rus. GaAsFETs). At me 2 cobles, but one ant.connecor in FT-847. I had to refuse much for replacement 1302. Also problem in the FT-847 can not direct operate REW (100ma max current).

On 6-Oct-2003 WB9UWA wrote:
Hi Loren,

I would still be concerned about preamp burnout even if only exciter power level is switched at the output of the preamp. All sorts of things can go wrong. For one, you could TX using a TX button on your rig. If this bypasses the sequencer it may roast the preamp. Since you are on JT44 and your PTT is interfaced to the computer, what happens when your PC boots up? My rig goes into the TX mode, twice during boot up. Also many rigs send a breif TX signal during power up. It is usually easy to break into the rig for a direct connection. This may involve getting a proper miniature connector. Some of the people on this reflector have probably made this direct connection on your model rig. I use the Yaesu FT736r and have a relay inside the rig that switches between the normal input and the feedline from the output of the mast mounted preamp. This way, I can still easily use the rig on the terrestrial antenna. I also have the IC final in my rig controlled by the sequencer so it is impossible to transmit a carrier before the sequencer gives permission. I have not lost a preamp since implementing this system in spite of making many errors when up at 3AM.

I also use the second relay in front of the preamp to switch in a 50 ohm dummy load during TX or non-operation. I consider the possible, say .05db of additional loss a good trade-off for the convenience of being able to compare sky noise against the noise from my dummy load. My SWR is 1.0/1 on both my EME array and dummy load, so the noise probably compares fairly realistically. I feed power to the preamp (currently an MGF1302) all the time via RG59u coax. Since the preamp always sees a 50 ohm input, it never oscillates (perhaps only a few milliseconds during relay transition?). Relay coils are separately controlled and diodes are used to suppress voltage spikes. MOV's and Zener diodes are used on the voltage supply line to the preamp. 7805s alone are not to be trusted since they can pass voltage spikes caused by switching relay coils or near by lightening.

The second relay may also save the preamp in case something happens to the primary relay. In practice, at 2M, the additional loss of the relay will be made insignificant by other noise sources. A possible exception might be when averaging several transmissions to copy the signal. In that case a very slight reduction in nf *can* make a difference in copy. Leif is of course correct when he says that using a single relay is optimal to station performance. Not all of us are comfortable with the procedure for ensuring sufficient isolation. This can involve choosing the correct cable length.

A preamp designed for best noise figure will rarely have a 50 ohm input impedance, so a single relay in front of the preamp may not give the expected isolation. Any miscalculation here could blow the preamp. Using two low cost relays ahead of the preamp will save money and give peace of mind compared to using a single expensive relay. Please note that these comments are directed at 2m use. At 432 Mhz, the overall noise temperatures are much lower so losses must be driven down to 0db if possible.

On 6-Oct-2003 AL7EB wrote:
Loren,

Just another reply. I'm not sure I will offer anything new to what has already been said, but I just made these changes last year in my 2m-eme station: 1) Converted my antenna relays to default to the Tx path when not powered. Not only a safety issue but allows me to still operate if somehting goes bad on the tower (albeit with much reduced Rx sensitivity). 2) Replaced old Two Transco-Y coax relays and one Dow-Key type with three new CN-600NL for switching my H/V pol system. The Transcos were damaged and suffered high insertion loss. The CN-600NL is rated at 0.09 dB insertion loss, 48-dB isolation at 1 kW at 150 MHz. I eliminated using a secondary Rx relay for now. 3) Converted my FT-847 for separated Rx antenna input. This one action will save me so many preamp repairs! 4) Routed both mic-PTT and cw-keying thru a DEM sequencer. It controls preamp power, antenna switching, PA activation, and FT-847 PTT. It has switching on the input for either computer control, mic, or manual (for normal cw). This is CWKEY-5, PSK-31 and WSJT/JT-44 compatible.

Yet to do: 5) Upgrade my Rx preamp output feedline from two RG-8/x to one LMR-400. 6) Install MiniCircuits Phase shifters after my H/V preamps for continuous Rx polarity adjustment; 7) Upgrade my azimuth rotator to Ham-IV from Ham-III (its on its way via UPS from Norm's Rotor service). Ham-III did not survive 65+ mph winds! 8) Upgrade power lines to preamps to shielded cable. 9) Upgrade power lines on elevation actuator to shielded lines (noise suppression). 10) Install video camera for optical moon tracking.

...of these the separation of antenna line to the Rx in the FT-847 probably is the most significant for survival of my preamps! I only operate at 150w so may encounter other issues when I get medium power or QRO!

On 6-Oct-2003 LZ2US wrote:
Hi Peter, REW14 can handle 2,5 kw at 144 mhz and 2 kw at 432 mhz for sure. Its isolation is another "question". As I remember the isolation at 144 mhz (properly loaded with 50 ohm) is around or above 45 dB. I can do these measurements again and tell you the results.Will try to do these tests agan. I saw a talks about CX520 or CZX......... relays. These relays are much different compare to russian REW14 (15,16) relays. For sure CX and CZX relays have much better isolation.

On 7-Oct-2003 F5VHX wrote:
> My SWR is 1.0/1 on both my EME array 
>and dummy load, so the noise probably compares fairly realistically.

is it extremists week on here or something ?
Jim, this is not measurably possible, and practically, absolutely not even approachable on an antenna.

On 7-Oct-2003 WB9UWA wrote:
Hi Graham,

What can I say? EME is an extremist hobby! I use two SWR bridges and an MFJ259b to check SWR. They are in good agreement with each other and they all like 50 ohm loads very much at 144 Mhz. I use a T match on the antenna. The balun is carefully cut to resonance. The driven element is cut to length to be close on SWR and the T bars are set to a 50 ohm point. The length of the T bars are used to fine tune the resonance of the driven element. They are telescopic and are soldered in place when 1.0/1 SWR is achieved. They are brass. This is done when the yagis are in the final mounted location on the roof. Type N connectors are used throughout the system.

The 50 ohm dummy load is very easy at 144 Mhz. The resistor leads were simply kept very short.

A 1.0/1 SWR at 144 Mhz is a simple matter of making sure everything is exactly 50 ohm and there is no reactance. I have not yet met an antenna that I could not tame to a 1.0/1 SWR provided sufficient adjustments were available.

Did I miss something?

On 7-Oct-2003 SM5BSZ wrote:
Hi Graham, Jim and all,

> What can I say? 
> EME is an extremist hobby! 
> I use two SWR bridges and an MFJ259b to check SWR. 
> They are in good agreement with each other and they all 
> like 50 ohm loads very much at 144 Mhz. snip

The 1.0 to 1 SWR means SWR < 1.05 to me. That translates to 50 ohms +/- 4% or so.

144 MHz is nearly DC so it should not be a problem;)

To get my own idea about this I made a test of all the dummy loads I have accumulated over the years. The network analyzer showed SWR < 1.03 on all except one of the 13 attenuators i have. the thirtienth was 1.10 but with a rather steep slope for 1.2 at 200 MHz so I have thrown it away. This one did not have a proper BNC connetcor, I would call it "computer grade", the connection to ground was not 100% reliable.

I made no calibration whatsoever, it is obvious that the precision of "telecom" dummy loads as well as HP analyzers are within 2% or better. Tuning an antenna for a null at a particular frequency is trivial. Having it stay at null with changes in temperature or when turning or particularly elevating is another thing.

The SWR bridge is a null detector. Presumably there is a diode there that may lead to grossly incorrect readings. To try your swr bridge, try this test:

1) Find a BNC dummy load that shows SWR = 1.01 or better on your meter.

2) Add a BNC T-connector. SWR should go up to about 1.025. This is a poor measurement because the capacitance you add will be added to the reactive part of the dummy load, but you should clearly see the degradation, particularly if you tweak the dummy well below 1.01.

3) Add a second BNC T-connector for the signal to pass through both connectors. The SWR should go up to about 1.05.

Each BNC T-connector will add 1 ohm capacitive at 144 MHz.

4) Use two T-connectors, but this time put the dummy directly on the T that is connected to the SWR bridge and put the other T at the other side, leaving both females open on the second T. In this case SWR should rise to about 1.20, the added impedance is 8 ohms.

To decide how sensitive your preamp is to impedance errors, use a good cable, make several lengths so you can insert a T-connector or two between the preamp and your dummy load. If adding 1 ohm or perhaps 8 ohms makes a small difference at three different points that are separated by wl/8, and if you can see the impedance change well on your bridge, then you are safe. Otherwise, be very careful with conclusions based on noise level readings when switching the input.

Best is to have a directional coupler at the RX input so you can inject a signal towards the antenna and check SWR live. Then you can also inject a signal towards the RX and look at S/N with some computer software. With a constant S you will get a very accurate value for N.

On 7-Oct-2003 Dominique wrote:
I made no calibration whatsoever, it is obvious that the precision of "telecom" dummy loads as well as HP analyzers are within 2% or better.

Accuracy of the network analyser cannot be better than the load you are using to calibrate it.... When you measure SWR , you don't do it , but you compare the load you are measuring to the one you used to calibrate .

On 8-Oct-2003 F5VHX wrote:
Leif, Jim, et al...

>The 1.0 to 1 SWR means SWR < 1.05 to me.

Leif, if I transfer to you 10,000 euros a week, can you send me back 10,000 at your exchange rate please ?? :-)

>To get my own idea about this I made a test of all the dummy 
>loads I have accumulated over the years. The network analyzer 
>showed SWR < 1.03 on all except one of the 13 attenuators 
>i have. the thirtienth was 1.10

While I can be very sure that you know, it seems worthwhile to revisit this...

1.0:1 swr is an infinite (and totally unachievable or measurable) return loss

1.03:1 is 36.6 dB return loss, a VERY different thing, I'm totally certain that all 'amateur' equipment is incapable of measuring even that and I remain absolutely certain that no amateur radio 144 antenna will have a return loss of 36 dB either, although there is no doubt that test equipment used may tell you so...

To measure a 36 dB return loss properly you would need a network analyser and even then a rather decent one.

Probably the most common/best used meter in amateur circles is the Bird. They are very careful not to quote directly any directivity spec., from memory they mention in passing (buried somewhere in a long text) words to the effect of "never worse than 25 dB". Measured it is closer to 20 than 25 (we could argue for some weeks the measuring method/accuracy)

To measure a return loss of even 20 dB accurately you need directivity approaching 30 dB.

So please let's not have anymore infinity return loss claims !

I mentioned here some months ago the excellent FREE software available from Agilent (Appcad) , included in this little package is a good reflection calculator which can be used as a simple swr/return loss converter.

Leif, I like your 'T' tests, I do the same/similar here with pre amps for stability testing.....I also add a hand held helical antenna (rubber duck) or even just a piece of wire, to one side of a 'T' and then you can use a hand running up and down it to generate a range of mismatches, adding a load to one side of the 'T' and doing it again gives you a second range...also a shorted plug on the 'T' is a good test to do.

You need to be careful though because most amateur pre amps have poor input/output isolation and can be sensitive to output changes as well, for this reason I do the same 'T' testing at output with various loads at input.

On 8-Oct-2003 SM5BSZ wrote:
Hi Dom,

> I made no calibration whatsoever, it is obvious that the precision 
> of "telecom" dummy loads as well as HP analyzers are within 2% or  better. 

> Accuracy of the network analyser cannot be better than the load you are 
> using to calibrate it.... When you measure SWR , you don't do it 
> , but you compare the load you are measuring to the one you used to calibrate .

Of course, what I was trying to say: I did not calibrate the instrument to show SWR=0 on any of the dummy loads, I used the factory calibration that I can not change.

My conclusion is that the instrument factory calibration is pretty good as well as the dummy loads that come from several different manufacturers and are of types BNC, SMA and N.

On 8-Oct-2003 WB9UWA wrote:
Hi Graham,

I think all of us realize that there are limitations to the accuracy of any measurement. If I use the 4% accuracy limitation that Leif suggest, then the error would be something on the order of .2db ( 1.00 / .96 ).

This is certainly acceptable and when sky noise can be as low as 170 degrees K on 2 meters, then the expected 2db difference should be easily seen. After a period of time, the readings can be a very good relative indicator of noise performance. The very nature of noise makes readings more precise then a .2db resolution difficult, so there is no reason not to compare the noise output of a dummy load to the array noise in a well matched system. Would you recommend against such practice just because it can't be done quite perfectly (but practically perfectly)?

I see no reason that a perfect 1.0/1 SWR is not aproachable on an antenna even if it is not possible to measure it with perfect precision. Limitations in test equiptment does not make a 1.0/1 SWR impossible. It simply makes it difficult to know exactly when 1.0/1 SWR has been achieved.

I think any reasonable person would consider SWR in steps of .1. That is to say 1.0/1, 1.1/1, 1.2/1 etc. My antennas are set much closer to 1.0/1 rather then 1.1/1. With this understanding, I can legitimately say in front of technical people that the SWR is 1.0/1. Notice that I never stated 1.00/1 SWR. This would imply a whole new level of accuracy. The test Leif suggest using "T" connectors are certainly interesting, but I don't think I would let the results, good or bad prevent me from making noise measurements based on my dummy load. I will try the test with the "T" connectors so I can know the level of accuracy of my MFJ259b. Using 20/20 hindsight, it would be more accurate for me to say that the SWR of my array and dummy load is below 1.1/1.

I have checked my array SWR at several key elevation settings. Since my array does not illuminate any metal in the near feild, the SWR does not change (within the practical step of .1). The sidelobes are very low (a measured fact), so any metal near the side and rear of the array (perhaps 25 feet) does not effect SWR. There are some guy wires 25 feet to the North. Fortunately the moon is never North of me.

So far, nobody has argued with me when I claim that my array gain is 18.5dbd. It can be argued that my gain is different depending on whether I am using vertical or horizontal polarization. The differences would be primarily because of spacing differences, construction tolorances, and support frame effects. The gain could be as high as 18.6dbd or as low as 18.4dbd. YO6.5 tells me 18.55dbd for both polarities. My point being is that arguing about my array gain would be nit-picking or perhaps showing off. Arguing that my SWR might be 1.05/1 instead of 1.0/1 is, in my opinion similar.

At the end of the day, my earlier statement that my noise compares fairly realistically between my dummy load and array is still basically true. I recommend that anyone wishing to do the same, should acheive a 1.0/1 SWR on both the dummy load and array to the best of your ability to measure and not be discouraged by limitations in instrumentation. Being aware of limitations is sufficient. Knowing the extent of the limitations? Even better.

On 9-Oct-2003 F5VHX wrote:
before folks start tearing their systems to pieces to find their 'problem'...

>This is certainly acceptable and when sky noise can be as low 
>as 170 degrees K on 2 meters, then the expected 2db difference 
>should be easily seen.

this is only true if Tant=Tqsky=170k which it cannot be.

painting with broad, but what I consider to be realistic, brush strokes......

at 144, Tant of a typical 4 bay array pointing at quiet sky is more likely in the 220 to 270 k range.

If quiet site folks are seeing 0.5dB to 1 dB Y factor, quiet sky to a load, on a 144 * 4 bay array, they are doing very well.

(but you cannot easily quantify or isolate the component of gain change or other factors in the switching/measuring process)

I use this test myself as more or less a 'go' or 'no go' test. With an unknown error factor it is ok if that is near constant, if 'this time' is the same/close to 'last time' and close to 'everytime'... then it is a good system confidence check.

local environment factors play a large part here though , just a little power line noise throws the whole thing out when playing with such small levels/measurements, even the array & ground around it being wet or dry has influence, life is tough for system checks at 144, on the higher bands it is easier (but not easy)

those not interested in the 'discussion' can tune away at this point....if you didn't already :-)

Jim, I'm not looking for an argument, a point scoring match, or even a debate, but I will not sit by and watch profound and blatently incorrect statements pass by in a forum where skill sets and understanding cover such a broad range. I am happy to have discussion but if that is required please:-

a) quote/read properly

what I said was...

"Russ, true enough and I do the same here, but be aware that pre amp gain changes, because of input load changes , limit the accuracy of this method. "

and not what this statement of yours implies

>Would you recommend against such practice just because it can't 
>be done quite perfectly (but practically perfectly)?

I stand by my statement/advice.....

b) don't try to baffle the readers !

> With this understanding, I can 
>legitimately say in front of technical people that the SWR is 1.0/1. 
>Notice that I never stated 1.00/1 SWR.

I did not challenge the fact that it might compare "fairly" realistically, merely that 1.0/1 is not achievable or measurable

(if by 1.0 using engineering numerical conventions you mean 0.95 to 1.05, then you still can't measure/claim that either, 1.05 is still > 32db return loss and I remain very confident that you can't measure and don't have that either, more importantly OTHERS can't, so don't send them on a wild goose chase, and further, better to select/use everyday conventions and language lest the readers get confused)

are you still sure at this point that it was me "showing off" or perhaps you ??

I'm not nit picking, attempting to display knowledge (for indeed mine is limited) or "showing off" (covert insults don't bother me BTW so fill your boots).

Who knows what folks might do as a result of 'advice' that appears here ?? tear their arrays to bits in search of the holy grail of infinite return loss, or change all their coax because they cannot measure 2db Y from cold sky to load ?

I'm trying to add a touch of realism to wild statements and claims often presented in a confusing and muddled manner even to some of us that have at least a basic grasp of what your point is. Actually I don't do it as often as I feel like doing it lest I become a list 'bore', or be accused/pigeon holed of the sins mentioned above.

When common sense and/or physics tell's you your measurement doesn't compute, its generally the measurement that's broke not the physics.

On 21-Sep-2006 F6CER wrote:
On 2 meters ,before the big storm of dec.1999 ! , I have always used a single high isolation relay energized on receive with the LNA , and separate coaxial lines to the shack ,I never burned a GasFet even at 1500+ watts ; As a bonus , when the station is shut-off , the antenna is not connected to the LNA this offers protection against statics . One important point is to feed the 24V (in my case) via the receiving coaxial cable and not via a separate cable to avoid any kind of loop in case of nearby lightning .


(Note: I'm not showing any E-Mail address here in order to avoid them from being collected by SpamBots. You can possibly find the E-Mail addresses of the above OM at QRZ.COM.)

This discussion is still not closed. If you have any opinions or additional related information you would like to be published here, just send me an E-Mail

Read other Moon-Net discussions

 
Share: Twitter   Facebook   Buzz  
Home page Site map Radio Sherlock search engine Join the Mailing List ! 


Select Language English EspaƱol

0 user comments


Add your public comment to this page. (Important: Only respectful comments directly related to the content of this page will be accepted here. Inquires and complains must be sent to the webmaster.)

Name / Callsign
E-mail (Will not appear online)
Comment
To prevent automated Bots form spamming, please enter the text you see in the image below in the appropriate input box. Your comment will only be submitted if the strings match. Please ensure that your browser supports and accepts cookies, or your comment cannot be verified correctly.
»