| Products > Test Equipment |
| Agilent E835xA PNA VNA Notes |
| (1/2) > >> |
| joeqsmith:
Some time ago I purchased an old Agilent PDN to replace my 1970s HP8754A. While I was familiar with the the 8753, I was wanting something newer. While it's not a bad VNA and certainly out performs my old HP8754A in every way, the lower frequency is limited is 300kHz making it useless for measuring PDNs. The UI is alright but some parts of how it drives are cumbersome. On the plus side, interfacing it with LabView is no problem over Ethernet. They do offer a TDR option for it but the price far exceeds the value of the VNA. Of course with LabView, adding TDR (or pretty much anything else) is not a problem. Hopefully we will continue to see the prices on some of this older equipment come down for the hobbyist. Attached is a document talking about the design goals for the PDN and comparing it with the 8753 series. These are a few of my notes: The PNA uses the 33321-60056 70dB attenuators. Administrators Username Password: Admin/agilent for newer systems Older units use: blank or tsunami Agilent has some notes about modifying the OS for the PNA!!! The disk controller does not handle inverting the IDE jumpers. The original Fujitsu MHV2040AH requires no jumpers to become a master. If a replacement drive does require a jumper, it must be installed before the cable. Snip the pins and jumper!!! The first partition is NTFS, the second FAT32. They use Ghost to image the drive. After using Drive Image to image the drive, use MBRWORK to copy the first track of the original drive. Then save this to the SSD. There appears to be a level shift at each bands switch point. Insert a thru, set the center to the switch point with a 2MHz span. Measure difference of the two sides. Band Level shift error 10MHz 0.070 748 0.007 1500 0.020 3000 0.150 4500 0.175 Calibrate the unit, replace the thru with an attenuator and measure the level shift 10MHz Thru 0.00dB 10dB 0.1dB 20dB 0.12dB 30dB 0.11dB Measuring S21 and using the spectrum analyzer to measure Port 1, it appears flat. Injecting a known good signal into Port 2, then moving the signal +/- around the band's switch point, we see the level shift. This is a design limitation and appears to the a non-linear behavior in the receivers. Looking at S12, we see the same problem. See Caesar's comments attached. I have also attached the FAQ from Keysight. *** Added some other possible notes that I found in the KeySight forums. *** When I was looking, I had came across a matrix of various VNAs. I had thought about buying that PicoVNA but after downloading at trying the software, I decided to go with a used system. |
| joeqsmith:
The attenuators on these will fail with time. There is a self test you can run to check them (and other components). I watched this video of a young man attempting to rebuild one. Do not use this method. The rail will slide and the dog bone holes will align over the pins. There should be no force required. |
| joeqsmith:
I've mentioned that I had attempted to make an up converter for the HP8754A. Test1 is showing the very first time I attempted to use the 8754A to collect data up to 6GHz. Sadly, I was far to ignorant at the time to pull it off with my home made calibration standards. By the time I bought the PNA, I knew enough that cal standards would be a problem. Used standards were priced far to high for my comfort zone. Knowing how easily they are damaged and the cost of sending them in before service made this a poor choice. I borrowed a set of metrology grade 3.5 and N standards that I used to characterize my home made shorts and opens. For the loads, I just sorted parts for the best ones. The PNA is a 4-receiver system and supports more complex methods. Shown was my attempt and making some TRL standards as well. I attempted to made a T-Checker but that was a total bust. If you are looking for some low cost standards, you may want to check out this company: https://appliedeminnovations.com/index.php/product/vna-calibration-kit/ |
| joeqsmith:
I suspect many of you working on these old relics have similar transfer cables you use to setup these old drives. Also shown is the PNA testing w/ an SSD. Note the jumper I previously mentioned. Shown with the final drive mounted into the unit. Note the lack of plastic. |
| nexus:
Hi Joe, Excellent post! I am in need of your expertise as far as restoring OSs for the E835XA series analyzers. I have a E8357A which is running Windows 2000 on the early 266 MHz CPU. While the analyzer was bootable and working, I made a full forensic image of the disk using macrium reflect (including the original MBR). However, after accidently deleting an option on the unit, I tried restoring the disk image I made to the original drive. This made it completely unbootable into either NTFS partition or the "Recovery" FAT32 partition. I have tried everything I can think of to get it booting. I've made the four windows 2000 bootable floppies and run "chkdsk", "fixboot", "fixmbr" on both OS partitions. I've tried marking the partitions as active using "fdisk" as well from a 98SE bootable floppy. Nothing I do can get it loading the OS. When I try to boot to the NTFS partition, the machine just sits at a blinking cursor right after BIOS, with no messages. When I try to boot to the FAT32 recovery partition, it shows a "Missing operating system" error with a blinking cursor again. Any ideas on what can be done to make it bootable? Do you have a copy of the first track of the original drive that you extracted using MBRWORK? Maybe writing that to the imaged drive will make it bootable at least. Also, if you have any versions of the PNA FW you could share, that would be extremely helpful. My analyzer has a very old version, A.01.40. I think the latest version I can use with this OS is around A04.2X. Appreciate any help you could provide. |
| Navigation |
| Message Index |
| Next page |