SWR on the E44XX Series Power Meters

[HarryBee]

Has anyone ever done the SWR on the RF output of  E44XX series power meters?
In the manual they use figure out the voltage reflection coefficients.  Then they use the 100 Ω and 200 Ω settings on the 432A power meter with the thermister mount.
Here's my problem and ultimately my question.  I am getting bad SWR readings on every E44XX series power meter I check as well as the N19XX power meters.  I get consistent and OOT readings if I use a directional bridge or even a network analyzer. 
I realize that most people don't check SWR on power meters because they work for military or they are part of labs that are calibrating for profit but I need to check SWR because the customers use random adapters between the RF out and the power sensor whenever they standardize the power sensors and also we follow mfr procedures
So, the question is, if any of you ran across this problem with these P-Meters.  The spec is 1.06:1 and I'm getting readings of 1.14:1. 
Yes the contacts are clean and no I'm not going to test pin depth (I know this can be an issue but not the case with all the same models being faulty).

[CalLabSolutions]

HerryBee,
You can't check Output SWR on a powered output using a Network Analyzer of SWR bridge like you would on an input device.  The output signal is going to interfere with your measurement.
Yes, Some signal generators have this test, but they set the attenuators, blocking the applied RF power.

Do you not have a 432A and thermistor mount?

Mike

[HarryBee]

CalLab,

    I know you can't check a powered out as you've mentioned.  The power out is switchable to on/off state.  I am assuming that the  power has to be on to get the true internal load.

    I have the 432A and thermistor mount and do that 5 page long funky formula on the E44xx series manual and get OOT readings that are consistent with the network analyzer and SWR bridge with the cal output turned off.

[CalLabSolutions]

HarryBee,
Just checking.. Don't take it as an insult or any thing like that, but over the years I have see some people do some really strange things.  How some of these people qualified for this carrier field is beyond me.     

We have the Agilent 432A method written into our Automated procedures, so far, none of our customers have contacted me having a problems.  I have only ran the test personally a couple times testing the software.
 
That said, I think the On/Off setting on the ref  output on the power meter is switched.  I have never taken one apart.  But if so, your network analyzer and vswr bridge would see it as slightly loaded open. 

Mike

[HarryBee]

No Sir.  I would not take offense.  Takes a lot to make me mad these days... . .. unless i"m drinking 

I think we just have bad units. I tried multiple standards as well.  Maybe it's the way the users are connecting things? I hate to admit that I may have to break out the pin depth checkerater 5000.

I will try tomorrow.  Thank you for your assistance regarding this matter.

[djshepp21]

No Sir.  I would not take offense.  Takes a lot to make me mad these days... . .. unless i"m drinking 

I think we just have bad units. I tried multiple standards as well.  Maybe it's the way the users are connecting things? I hate to admit that I may have to break out the pin depth checkerater 5000.

I will try tomorrow.  Thank you for your assistance regarding this matter.

I have seen this issue as well.  Seems as though checking the output while "on" defeats the whole purpose of a SWR test when you take into account that when you adjust the output to a nominal 1mW you are using that same thermistor mount for the SWR test.  (Perhaps this is the reason that the USAF does not check this parameter in their procedure.)  That being said, the 432A method is repeatable and the numbers do work out.  If anyone has an excel spreadsheet that helps calculate this stuff out and could share it (for the greater good), I'm sure it would be greatly appreciated.

[Bryan]

I share your frustration, I'm not sure if the test came about with the transition from E441xA to E441XB or at a certain serial number.  It is my half-assed opinion that if it is indeed bad it will show up when measuring the 1 mW power output anyway, I have not seen a bad one yet and am anxious to examine one.  I have a calculator that I slapped together several years ago.  The procedure is pretty vague IMO, I do know that the 432 manual says to turn it off when you are switching the mount resistance switch so I do the coarse & fine zero after setting up for 100 ohm as well.  I'll forward to Hoopty and he can decide if it gets on here, I don't have the mojo to stick it on.

[CalibratorJ]

I think one of our techs has done the SWR test on an N191X power meter...... using Agilent's TME software. I have not had the displeasure of doing it yet (I do recall the tech telling me how much of a PITA it was to do), but I think our power meters are coming up for cal later this month. It might be worth a gander to wander over to their website and look at how they perform the check, I believe they use an E5071C and a power splitter or two, amongst other things. The help files for the TME procedures will show you the connections, but not the math involved. If I have a chance tomorrow, I will try to look at the actual procedure......

[CalibratorJ]

I have seen this issue as well.  Seems as though checking the output while "on" defeats the whole purpose of a SWR test when you take into account that when you adjust the output to a nominal 1mW you are using that same thermistor mount for the SWR test.  (Perhaps this is the reason that the USAF does not check this parameter in their procedure.)  That being said, the 432A method is repeatable and the numbers do work out.  If anyone has an excel spreadsheet that helps calculate this stuff out and could share it (for the greater good), I'm sure it would be greatly appreciated.

The Army omits the SWR check as well, probably since if it is out there is nothing you can do about it and/or you adjust the power ref out for 1 mW using a thermistor with very low SWR anyway.

In my opinion, unless you are dealing with some seriously tight uncertainties (or a REALLY crappy power sensor), the small amount of power bouncing back from the sensor @ 50 MHz, hitting the power reference and bouncing back is negligible anyway, considering that most power sensors are mfr spec'd for several percent and everything is referenced back to 1 mW @ 50 MHz anyway, which is traceable from your standard thermistor mounts.

If I was at work, I might take the time to do the math, but since I am at home..... sorry, can't be bothered to look up all of those specs and do all of the calculations (plus I don't have all of my excel cheatsheets at home).

[CalLabSolutions]

No Sir.  I would not take offense.  Takes a lot to make me mad these days... . .. unless i"m drinking 

To bad.. I like having a beer or two with my fellow calibration friends. 

[HarryBee]

Ha, CalLab maybe we can have one some day.  Calibrator J, SWR is an important measurement which can be affected if the user uses an adapter or cable at the end of the RF out. Even though it may seem like common sense to some to do this you can't always count on the user to do this as well.
I am aware of many people not doing this in commercial sector because they've never been accustomed to it in the service.  There are reasons the military doesn't do it.
The check with the 432 and mount is a PITA, but basically all you are doing is checking the 1 mW out at 200 and 100 ohm and doing funkified match that eventually spits out a SWR reading.
You are correct in the setup for the N19xx series.  I don't have the agilent software and would like to know how I can do this manually. 

[HarryBee]

Great!!  I figured out what was wrong.  The typical values for the voltage reflection coefficients listed in the description of the test for the thermister bridge set to 100Ω with an RF impedance of 25Ω is ~0.33. This is a typo in the agilent manual I think as Appendix A lists ~0.033 for these "typical values".

What a relief.  Now the next question is "how do I do uncertainty calculations on a reading that uses typical values" (lol, nah luckily I don't have to do that for this unit).

[CalDude]

Typically speaking, 95% of the equipment is good 85% of the time, so why are you bothering to do the swr test anyways? Let alone that uncertainty analysis stuff...

The formula used for SWR is in the Agilent manual, and use the ISO GUM to figure out uncertainties. I'm sure Lost In Place will help you with that if you ask.   

[HarryBee]

You go to hell!! You go to hell and you die!!

The formula was wrong.  There is a typo in the procedure right off Agilent's website.
And leave Lost in Place alone or you're gonna find ya self on the wrong side of a plunger!

[CalDude]

I'm not sayin, i'm just sayin...Lost keeps telling me he did this stuff when he owned his own lab.

So tell us what happened, and why did you determine that it was important you shave the sheep AFTER (and not before) you were done with it in the barn?