So I have HP 8902A's with Option 050 (high accuracy Tuned RF Level), along with the special kit to calibrate them to the higher accuracy (HP 11812A). It is a special edition HP 355D with Type N connectors, a pair of matched 10 dB attenuators, and a matched cable. I would like to start putting the pieces together for getting this accredited, but in getting ready to send the calibration kit (HP 11812A) out to Keysight for calibration, they only do up to Z540-1-1994. Any thoughts on where else I could send for an ISO17025 Accredited calibration? Or general thoughts on this?
Hawaii
If I remember correctly the option 050 was only done in Spokane Washington HP.
This is for traceability, I believe the attenuator they used was a laser calibrated
attenuator , now why a laser used to calibrate the attenuator I don't know , maybe
distance calculations ? But I know it was a NIST done attenuator that cost a bundle
to get done . The special option 050 verf kit is just that a verf kit . Now saying that
it doesn't do the correct calibration for accuracy . But you can use it to do a excellent
non 050 cal !! So if any of the guys out their know why it was a laser calibrated attenuator
used to get the tolerances for the option 050 8902A please explain and discuss .
Now this was being done before A2lA and such and now maybe Keysight has a new option 050
attenuator to do the 4-1 ratio's or A2LA cal . Or maybe its consider state of the art cal know .
Its been along time since I discussed the 050 option and things change in this industry .
Hawaii, if you can utilize them, the Army Primary Lab does the 30 MHz Attenuators on behalf of NIST on NIST's Piston Attenuation system - since it is the US National Std it can be argued that the accreditation doesn't matter, although their A2LA scope does/did cover 30 MHz Attenuators. Only way that I know of to get better uncertainties is to go to the UK. The APSL has done them for a lot of folks out there, countries even if I recall correctly.......
To answer Top Gun's question, the only way (back then) to get the required uncertainty was using a Piston Attenuator system that uses a laser for distance measurements. Nowadays you can use voltage dividers, I think NPL in the UK is going/ went that way already, the US isn't far behind. I think the 050 option on the 8902 was just that it included the attenuator for calibrating it, but I could be wrong, otherwise you had to send your 8902 to someone that had an attenuator good enough to cal the tuned rf level. Been a few years since I did an 8902 since the Army mothballed them all a few years ago and I haven't worked the NIST system in over a year.....
Calibrator
Yep !! Thanks for the recall on the piston attenuator and laser .
That is the correct way back then . The Option 050 was the same design
as the later 8902A (SN wise ) it was just verified to meet the specs for the 050 option .
If you had a older 8902A with the relays on the front end you had to convert
the front end module . But Yes on the 30MHZ signal , would you think the
Fluke 9640A series would have the accuracy ??? Does anyone Know the
verification to do a 9640A for attenuation ? Just a thought guys ! Instead of the
attenuator for the 050 option use the 9640A ?? I haven't checked the accuracys
or anything , So I could be way off base on this .
But yes the voltage divider would be a very accurate way since Technology and the
MFG of these is WAY better then back in the day .
But thanks for the RECALL on the memory for the Piston Attenuator, CalibratorJ
Do any of you guys remember the Microtel 1295 ?
I used to love that for attenuation measurement .
Great Box when it worked but when it had a issue
then the problem was support . I believe MACOM bought
Microtel back in the day and made a few under that brand.
They where way ahead of their time . It was a great sub for the
8902A/11793A system since you need only 1 source to go way up in
frequency .
So it sounds like my HP 11812A kit isn't necessarily worth using. We inherited a number of items from our other lab that closed down. Among them was the kit, and we have at least two or maybe three option 050 HP 8902A's. And it further sounds like those are just older ones, where on later models the upgraded technology did away with opt 050 and replaced it with a more accurate rendering of the 8902A without the option (the option became the specs for the standard model).
So I guess two questions... Do I need to bother getting the 11812A kit calibrated? And/or how/where do I get my option 050 8902A's calibrated to option 050 accuracy accredited? I am prepping to get this area accredited, so I want to get it done to best accuracy.
Keysight still offering 17025 Accredited cal
I could be speaking out of turn here. But I am pretty sure any lab could calibrate the HP 8902A Option 50 using the new Agilent Dynamic Accuracy Test Set used the test the Vector Network Analyzers.
You would have to understand how the standard works. Using the older one you would be comparing the linearity of a power sensor's square law regions linearity accuracy to the linearity of the 8902A.
Mike.
Hawaii
The option 050 is just enhanced verified accuracy . AKA a special calibration
to verify the specs to a higher accuracy . Still the 11812A verifys the specs
but not to the proper specs . All option 050 will have the new front end .
The later Sn# units will also have the new front end , but Keysight considered
the special cal option 050 as a option. So that being said If its not a option 050
then you cant get the 050 cal . They check the SN# vs the options .
But still I would say use your 050 8902A as a standard 8902A and then you will be good .
Just put in your records that the 050 was a special cal for accuracy of attenuation .
Wow a lot of words for saying its not worth getting the 050 verified and don't get the
11812A calibrated . But still Mike might have something to check out on the dynamic
accuracy test set . But still having one of those is very rare .
I personally would not get the 11812A calibrated just for the 8902A 050 verification .
I hope that doesn't confuse you more , It did me .
I'm looking at getting accredited in this area. We inherited a bunch of units from other lab in another part of the country that closed. This included a few 8902A's all with Opt 050 and a very nice looking 11812A kit. I was looking at the manual that came with the kit, which does go into how to cal the 050. It included doing it at 30 MHz, using the C.V. numbers on the 11812A, then entering them in as offsets between each of the Tuned RF ranges, etc. So this kind of begs the question as to what improved hardware is there in the 050 systems. seems like those has the LO loop (hard plumbing with two Type N connectors on the back - unless that is for another option).
I was also thinking of sending my 11812A to someone else with good accredited attenuation uncertainties to use to cal the 8902A-050's. It looks like the tolerance is a bit better with the 050.
The spec for the kit itself is (at 30 MHz only):
+/-0.003 DB + 0.003 DB/10 DB STEP (0 - 50 DB Settings).
Then the -050 spec of the 8902A spec (when cal'd with the 11812A) is:
+/-0.010 DB/10 DB STEP (0 - 100 DBM)
+/-0.050 DB/10 DB STEP (-100.01 to -120 DBM).
I don't know. Its such a nice kit in a beautiful wooden case.
The 8902A with the rear type n connector for the LO are a option 030 and probably
corresponding filters . On the option 050 you start getting into the argument of verification
testing vs calibration . Yes the 11812A is used to verify the 050 but calibration to those specs
might be in question if your trying to get accreditation for that . Notice they state check relative accuracy .That the question is it good enough ??? I just find it a trap and would not want to go there.
But we are always questioning in our field . Maybe a call to Keysight is in line for you to question
the option 050 and the 11812A . If its good enough . We don't want you spend money on something
that is not right .
I absolutely appreciate the continuing conversations on this. And I most definitely don't want anything that isn't correct. I was reading in the service manual and saw some explanations of the special codes to both read and write the range to range offsets (38.x and 39.x series) for the tuned RF. It looks kind of like they are accidentally giving you part of the adjustment procedure. It looks like special codes 39.1 through 39.4 or so (as I recall) are to write in the correction constants for the range to range accuracies. The rather troubling one is that in the procedure in the 11812A manual it tells you to use 39.9 to clear all codes before doing verification, but it doesn't explain entering the new codes in. My thinking is that possibly those 39.1 through 39.4 codes are the ones to do the Option 050 adjustment procedure (which sounds about like a normal procedure). So what exactly is the hardware difference in 050? Is it just factory trimmed a little tighter for the range to range DB error, and/or a little better filtering for more stable readings?
Hawaii
Yes the option 050 is a guarantee rating of specs to a better tolerance .
Aka this combination of assembly's will meet this spec.
No real trimming of parts , its just select assemblys . And
a normal 8902A will meet specs of a 050 about 90% of the time .
The older units would require the full new front end to upgrade
it away from Coaxial switches and a new amplifier . Yep it was just
one of those things . But the unit had to have a option 050 on it
or HP/Keysight would not certify / cal it to the 050 specs .
I hope all this helps and you can get yourself dialed in on this .
The more info the better .
"It looks like special codes 39.1 through 39.4 or so (as I recall) are to write in the correction constants for the range to range accuracies. The rather troubling one is that in the procedure in the 11812A manual it tells you to use 39.9 to clear all codes before doing verification, but it doesn't explain entering the new codes in."
Clearing this is not an issue. This is a special function of the 8902A that allows the user to clear mistakes made when entering cal factors from a tuned rf level calibration, or to clear previous cal factors before entering new ones. Basically, the 8902A generates the cal factors and you can read and enter them with special functions 38 & 39. Think of this as a self cal the 8902A performs similar to the power reference cal.
Remember, you are only performing a verification with the 11812A and not affecting the calibration of the 8902A for anything except that specific verification.
The 11812A is utilized because the supplied 355D is calibrated against a primary piston type standard, so if you require accreditation, make sure the lab has an accredited one. I would check with Keysight first.
And herein lies the problem.... Keysight just informed me that although they are accredited for the measurements (as I read their scope of accreditation, which covers it), they will not provide accredited readings for the 11812A. I am currently shopping around scopes on the various Accreditor websites for someone with adequate uncertainty at the required 30 MHz to do this to ISO17025.
So since it is so early in the morning, I am hearing some things about the 11812A is for verification, and that the WRITE codes may be for something akin to the 0 dBm self cal. Does this mean that with a properly accredited 11812A that I will be able to (at least theoretically) achieve the improved accuracy/uncertainty for my Opt 050's? I do understand what the 38 & 39 codes signify. Just not sure if I can optimize my 8902A-050 to meet the improved accuracies, and subsequently to be able to accredit based on those improved accuracies. I absolutely want to be fully "street legal" in how I do this, but want do my best (if possible) to calibrate to Opt 050 specs (and improved uncertainty). As I have about 3 of these with Opt 050, if it is feasible, getting the 11812A calibrated would be a cost effective way to go.
Good to hear about the 11812A and how it is verified and again
we get into the Piston attenuator and laser used for distance .
Know I don't want to burst the bubble but I still think the
11812A is just used for the verification and not really to the
proper specs of the option 050 . So with that being said
and we where alluding to the piston attenuator being used to
calibrate the option 050 , I would conclude that you look for this
type of service , because of course you will be getting a accredited
cal and it has to be correct . I know the option 050 cal has been always
a strange topic that everyone sherked around . Boy the confusion
is getting larger . Hawaii did Keysight say that they could do a accredited
8902A option 050 cal ???? OR are you SOL ??? We will get this solved for you !!
And MetVets that shed a lot of light on the subject !!!! Thanks .
If you want to prove your 8902A meets option 050 specifications, you will require a calibrated 11812A verification kit. This kit will help you prove it meets the specifications. If you want this discipline accredited, it is my understanding that the kit will need to be accredited, as well as all the other items used in the calibration of the 8902A. You will then have to do a line item uncertainty calculation and have it approved by your accreditor before it can be added to your scope.
You can choose to have as many parameters of the 8902A accredited as you want, but your accreditor will require the proof of your uncertainty calculations (at least they are supposed to require that), for your scope to include those parameters. You will also need to have a measurement assurance program in place to make sure those parameters are scored for proficiency testing during the period of your accreditation.
What I'm really getting at is this is a complicated, time consuming, and expensive venture and if you want to have all the parameters of the 8902's performance specifications accredited, please make sure the payback is good. If it is, then this is very doable, but if Keysight won't accredit the 11812A, you may have issues getting it listed on your scope. The best step is to contact your accreditor and ask for guidance.
As far as optimizing the 8902A's specs for option 050, it is my understanding that if you have performed the adjustment procedure for the 8902A (some work but not too complicated), then the unit should pass the verification described in the performance test. If it doesn't pass, then you have to start looking for errors in your calculations, the measurements provided on the 355D, or worse.....
To clarify to MetVet, I do have an HP 11812A, along with some HP 8902A's with -050.
I do understand what is involved in this accreditation. We are already accredited in many areas, and I am fully familiar with doing the uncertainty budget etc. We do have all of the other pieces in place, and we are in process of adding a number of RF areas to the scope of accreditation. This one has just been a little troublesome with Keysight not providing accredited service for this. Even in looking at their scope of accreditation, it looks like they are well able to do ISO17025 calibrations on this. But it also looks like it might be a business decision (which I find a little confusing) to no longer do what I think is a relatively simple calibration for a high end RF/microwave lab such as theirs - even though it appears they are capable. My guess is that they may review line items to maintain their fully automated procedures, etc. (including automated collection of uncertainties for), and if it is out of support/too old, it gets dropped if there isn't enough demand to pay for continuing maintenance of the above. A notable exception is the 478A-H76, which is one of the only standards available, and it would create a big hole in the metrology community if that were dropped. But the 11812A is likely not so fortunate (old, probably very low demand).
I just looked on Keysight website, and when you search for calibration prices for 8902A-050, it recognizes the model, but does not list what services are available. When you search on 8902A, they do list up to and including ISO17025 Accredited for the standard model, but no reference to Option 050. I'm not sure why the issue. I just sent an additional email to a customer support rep at Keysight to see if they will do an accredited cal to the 8902A including accrediting the option 050 to better uncertainties. We'll see how that goes.
What I believe I will pursue is finding an accredited vendor to calibrate the special edition 355D (with N connectors, and 10 dB stick attenuators attached).
So I am guessing that if I find another lab to do an accredited cal on the 11812A, they do not measure the attenuation values of the two stick attenuators, and that they are attached to minimize SWR. And, I should get the attenuation values of the 355D in accordance with the manual (of which I have a hard copy). I'll have to take a look to see if there is an SWR spec to be tested. Basically I just need a good accredited lab that can do high accuracy attenuation measurements at 30 MHz. Any recommendations of a high accuracy attenuation calibration lab?
Check with U.S. Army Primary Standards Laboratory. They can provide the accreditation for the 11812A. Our organization has been using them since Keysight no longer provides the accreditation.
Thanks very much for the input regarding APSL. I am looking into that. I'll be looking into what ever level of uncertainty I will legitimately be able to use based on the whole context on them. APSL looks to be a very high end lab. Just need to go through all of the various uncertainty contributors, etc. Have to get accredited for Signal Generators, Power Sensors, etc. They both will be fun. If anyone has any uncertainty budgets for any of those they may be willing to share offline, send me a PM and we can discuss email address (don't want to post them here).
Quote from: Hawaii596 on 05-08-2015 -- 15:26:51
APSL looks to be a very high end lab.
They are. The APSL is actually the standard NIST uses for some measurements
And in full fairness, after a few emails, I received input from Keysight that they actually can do an ISO17025 Accredited cal on the 11812A. Price to have it done be APSL is a little higher than Keysight. Any thoughts on which would give me the better numbers (MU and readings)? I have my thoughts, but want to hear from some who have worked with both labs (and I know they both are very good labs).
You might also look into using an NMI but for this I would recommend a Programmable Step Attenuator, benefit is many frequencies and the entire attenuation range (even though with the 8902 procedure you are only looking at the first 50 dB), you can expect something like this:
S2,1 S2,1 S1,2 S1,2
(MHz)Mag(dB) u(Mag)(dB) Mag(dB) u(Mag) (dB)
0.03 -10.028 0.003 -10.028 0.003
0.03 -20.043 0.003 -20.043 0.004
0.03 -30.071 0.003 -30.071 0.003
0.03 -40.03 0.003 -40.03 0.003
0.03 -50.061 0.003 -50.06 0.003
0.03 -60.074 0.006 -60.074 0.006
0.03 -70.102 0.008 -70.102 0.008
Now that is a good question?
Who has lower uncertainties / better calibration APSL or Keysight...
I have been to both labs and they are both at the top of their class. To that is a very hard question.
Mike.
Quote from: Hawaii596 on 05-11-2015 -- 08:38:57
And in full fairness, after a few emails, I received input from Keysight that they actually can do an ISO17025 Accredited cal on the 11812A. Price to have it done be APSL is a little higher than Keysight. Any thoughts on which would give me the better numbers (MU and readings)? I have my thoughts, but want to hear from some who have worked with both labs (and I know they both are very good labs).
I might be a bit biased, but I trust the numbers out of the APSL more than I do Agilent/Keysight/HP...Danaher?
In part because of Agilent's tendency to fudge some things, for instance the calibration of an N1911 (needs .1%) with an E4419 (.6 %) and then implying that the E4419 was calibrated at NIST to a spec a little over 400% better than it's 24 hour specs. and then doing the email equivalent of mumbling and pointing the other way when you request the cal cert for the E4419...
But don't let that sway you...