-
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.
-
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.
-
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/
-
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.
-
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. -
I checked to see if I had saved the original images I made but no luck.
When I work on old systems like this, the first thing I will do is make a replicate of the drive. Normally to a newer mechanical drive. Once I have the new drive booting, I pack away the original for just in case and do all the remaining work on the new drive. I suspect with this VNA, I may have just used the PC to mirror it and did not actually create an image.
According to Caesar, the newer software (XP) requires the newer computer. The other problem you may find is that the newer firmware my require that the system is sent in for alignment. Once he retired, I'm not even sure if that's an option. The new firmware will look for calibration files that were not present with the older firmware. To create these files, the system must be aligned.
It seems someone had offered a harddrive for one that was setup but I am not sure if that help you. You could try contacting Keysight as see if they would service it. -
I see. I've imaged lots of newer machines without issues making bootable drives from the image alone. I have never worked on Windows 2000 before, but XP being heavily based on it, I did not anticipate this issue. I should definitely make at least 2 physical, bootable drives before modifying the original hard disk. You live and you learn.
Yup, I'm eyeing some 500 MHz cpu boards that I can swap in. Only thing is, they are quite expensive (~2K). Not sure if I want to drop that much money on a 20 year old VNA. I would rather just get the old 266 MHz CPU booting again if possible, as the issue is purely OS related.
Would you happen to have a copy of that first line on your hard disk that you made with MBRWORK? Perhaps the MBR was affected in the image. I'm sure it is probably running a custom/nonstandard MBR due to the custom nature of the overall unit.
If anyone is a Windows 2000 (or NT OS) expert and knows what other tricks I could try to get my unit booting, that would be great.
-
I see. I've imaged lots of newer machines without issues making bootable drives from the image alone. I have never worked on Windows 2000 before, but XP being heavily based on it, I did not anticipate this issue. I should definitely make at least 2 physical, bootable drives before modifying the original hard disk. You live and you learn.
Yup, I'm eyeing some 500 MHz cpu boards that I can swap in. Only thing is, they are quite expensive (~2K). Not sure if I want to drop that much money on a 20 year old VNA. I would rather just get the old 266 MHz CPU booting again if possible, as the issue is purely OS related.
Would you happen to have a copy of that first line on your hard disk that you made with MBRWORK? Perhaps the MBR was affected in the image. I'm sure it is probably running a custom/nonstandard MBR due to the custom nature of the overall unit.
If anyone is a Windows 2000 (or NT OS) expert and knows what other tricks I could try to get my unit booting, that would be great.
I am currently in the process of making an image of a 233MHZ CPU / Windows 2000 system. Did your system stand up in the end? -
I am currently in the process of making an image of a 233MHZ CPU / Windows 2000 system. Did your system stand up in the end?
Sorry for the late reply! I did NOT get windows 2000 to boot no matter how hard I tried! I made all sorts of recover floppies, tried to install windows 2000 fresh, etc. No luck.
I ended up upgrading the CPU to the 500 MHz pentium III version and was able to get a bootable image of windows XP running! If you have a disc image of the windows 2000 version, that would still be useful for possible future use! I have the original CPU board set aside.