Oscilloscope bandwidth 1 Meg Ohm input Z

[Bryan]

Looking at Tektronix DPO4054, testing using Fluke 9500B calibrator and active head.  Some Bandwidth tests say to set the calibrator & UUT 1 Meg ohm.
Procedure has a statement to use 380 MHz instead of 500 on 5mv/div scale, due to impedance mismatch between calibrator & scope.   Supposedly when the calibrator is set to 1 Meg ohm it has a Thevenin equivalent 25 ohm source Z.  ?

All I know it that the heads are tested at 50 ohms & other models typically you use a 50 ohm source with a 50 ohm feedthru to test 1 Meg ohm BW.  The specs for the head are written with "(into 50 ohms)" included and there is a note under the spec table saying Frequencies above 500 MHz are not recommended for 1M input applications.  I have some heads that will peform the test, some that will not.  All will pass performance their own performance test as published by Fluke.

This is a Tek written Tech Ref procedure, I think it is flawed, anyone familiar with the scenario?

[CalLabSolutions]

Bryan,
We have not seen that problem. 

Our automated procedure supports both the DPO and the MSO Models using the latest version of the Tek Manual.  Keeping these procedures up to date has been a bear, because Tek has made a lot of changes to the manuals over they years. Also the verification procedure resides in both the Service Manual and the Technical Reference. I would double check that you have the latest version. 

We have done it both ways.  Most of our customers are using 4 9530 heads (3.2 GHz).  So for them we use the 1 MOhm settings of the calibrator.

We also have customers that are using the 9560 heads (6 GHz) those heads do not support 1 MOhm bandwidth, so we have the user install a 50 Ohm load.
   
Some of my customers have told me that Fluke has been replacing the output connectors on the heads.  Are you calibrating your heads in house or are you sending them to Fluke?

[Duckbutta]

CalLabSolutions is right about the connectors on the heads. The ones Fluke originally installed on these units were garbage and they have since recommended replacing the older connectors with a Precision connector.

As to the other point, to the best of my recollection, the 1M ohm specification is calibrated by inference. That simply means that if it's good at 50 ohms, 1M ohm performance is assumed to be good. I typically use a 1M Output Z with a 1M Input Z and 50 ohm with 50 ohm. I have never encountered the problem you describe though I suspect you have the garbage connectors on your heads.

[Bryan]

Thanks Gentlemen, we have had connectors upgraded and they are more robust than the originals but it has been a few years, most of the time they have elbow adapters so when they hang on the scope they don't pry on the connectors, of course we removed those while working on this.  Do you know how to crack them open?  I'd rather be able to change them myself then send them out.  On the other hand it seems like they develope a problem that results in them going back every couple of years.  These are 9530 with Wavetek labels, some that predate the Fluke purchase, others that must have been leftovers. 

[CalLabSolutions]

I have one customer that has taken the time to cut some foam holders for the heads. She has one for each series of scope. It is really slick, the idea is to stabalize the head so that the weight of the head is not pulling down on the scope's input connector.  It is a budget version of the RF jacks, and it seems to work great. 

[RFCAL]

We also have not seen this problem.We run Cal Lab solutions software and it tracks the manual perfectly.If you have a Fluke Gold procedure,that's where YOUR problem is

[Bryan]

Well I sure do appreciate all of your input.  I am looking at several performance tests, It appears the initial procedure tested 100 mV/div to 380 MHz only.  These books are marked as for firmware 1.0 & above, then I have a couple that mention 2.0 & above, these test 5/2/1 mV div to 380/300/175.
And as Mr RFCal noted I believe this is a Fluke Gold program, I have not watched it run but the test report indicates 5/2/1 mV/div and from the looks of the data sheet frequencies are 500/350/200 MHz.  It appears to be the same values as the 50 ohm Z so I'll watch one run myself and see what happens.
Thanks agin

[Bryan]

Program appears to set correct frequencies despite what the test report notes.
Will check on the heads and see when connectors were last changed and move on til next time. 
I appreciate all the suggestions & comments.

[RFCAL]

We never had a problem with the bandwidth test here.  The biggest problem we had with the Gold Procedures is they are not kept up to date, and they're buggy.  (And I just did not have time to fix them.)



Cal Lab Solutions procedure are more expensive, but I just find it easier to use them because they are done right and keep up to date.   If we had a problem like you are having, they would help us find a resolution.  If it was in the procedure, they would fix it and get it back to us, usually within a day or two.

[CalLabSolutions]

Thanks RF-Cal,
It is a lot of work keeping this procedures up to date.  And I appreicate the positive feed back.