开云体育

Mixing 'n' matching S parameter test sets


 

Hi all,

I'd like to check out this old VNA I have by hooking it up to a different S parameter test set than it was intended for. I'm just wondering if I could cause any damage by so doing. I don't see any reason why it should, but you can't be too careful with this kind of thing! The test set is an 85046A and the VNA is a 8754A. The plugs and sockets all line up with each other, so that's something - but far from conclusive!
So.... safe to try?


 

Those are completely different generation VNAs. The 8754A is strickly a manual VNA in that if you want calibration of cables and fixtures you will need to have the storage normalizer box or you can use grease pencils on an overlay on the screen! The 85046A was made for the 8753 series of VNAs and has switching of signal direction built in, as well as four port capability. Your 8754A doesn't have the drivers to do that. Just because it has the type N connectors for the RF ports, it doesn't have the smarts to drive the switches. When 8754A was developed, it was impossible to predict what the future microprocessor-based VNA would be. The 85044A or B transmission/reflection test set will work with it as it has no switching and has only an attenuator with a knob. They can be found relatively cheap. I had a couple of them, sold one and now I use the other with my Rigol RSA-3000 series spectrum analyzer/tracking generator as a SNA (scalar network analyzer) to 3 GHz.


 

The 8748A is, I believe, the S parameter test set that will work the the 8754a.

You _should_ be able to get these cheap since they won't do diddly with more modern VNAs. But try convincing an ebay seller of that!

Likewise, there are 2 normalizers that can work the the 8754a and they also _should_ be cheap since all this stuff is incredibly obsolete.

As a side note, there's an article in the HP Journal on hooking up an 8754A to a computer and various bits to automate measurements and get a bit closer to a modern VNA.

Full discloser: I have an 8754a with all the trimmings. Other than the miniVNA that's available, it's about the cheapest route to a full scale VNA for occasional use. The grease pencil bit still gives me a chuckle.

If I did a lot of RF work I'd get something newer.

Paul

On Sat, Feb 19, 2022 at 01:07:16PM -0800, John Lyles wrote:
Those are completely different generation VNAs. The 8754A is strickly a manual VNA in that if you want calibration of cables and fixtures you will need to have the storage normalizer box or you can use grease pencils on an overlay on the screen! The 85046A was made for the 8753 series of VNAs and has switching of signal direction built in, as well as four port capability. Your 8754A doesn't have the drivers to do that. Just because it has the type N connectors for the RF ports, it doesn't have the smarts to drive the switches. When 8754A was developed, it was impossible to predict what the future microprocessor-based VNA would be. The 85044A or B transmission/reflection test set will work with it as it has no switching and has only an attenuator with a knob. They can be found relatively cheap. I had a couple of them, sold one and now I use the other with my Rigol RSA-3000 series spectrum analyzer/tracking generator as a SNA (scalar network analyzer) to 3 GHz.
--
Paul Amaranth, GCIH | Manchester MI, USA
Aurora Group of Michigan, LLC | Security, Systems & Software
paul@... | Unix/Linux - We don't do windows


 

开云体育

John Lyles…

?

“The 85046A was made for the 8753 series of VNAs and has switching of signal direction built in, as well as four port capability.”

?

Please explain the 4-port capability. I was not aware it had that and would like to know more.

?

Thanks.

?

From: [email protected] <[email protected]> On Behalf Of John Lyles
Sent: Saturday, February 19, 2022 4:07 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Mixing 'n' matching S parameter test sets

?

Those are completely different generation VNAs. The 8754A is strickly a manual VNA in that if you want calibration of cables and fixtures you will need to have the storage normalizer box or you can use grease pencils on an overlay on the screen! The 85046A was made for the 8753 series of VNAs and has switching of signal direction built in, as well as four port capability. Your 8754A doesn't have the drivers to do that. Just because it has the type N connectors for the RF ports, it doesn't have the smarts to drive the switches. When 8754A was developed, it was impossible to predict what the future microprocessor-based VNA would be. The 85044A or B transmission/reflection test set will work with it as it has no switching and has only an attenuator with a knob. They can be found relatively cheap. I had a couple of them, sold one and now I use the other with my Rigol RSA-3000 series spectrum analyzer/tracking generator as a SNA (scalar network analyzer) to 3 GHz.


 

The antiquity of the 8754A is part of the appeal, though! It's closer to the fundamentals of network analysis and there are no abstraction levels between the operator and the machine as is the case with more modern kit. I should mention I'm heavily into vintage test equipment which has a certain appeal for my personally, though it wouldn't suit everyone of course.
I bought the 8754A and the storage normaliser for it many years ago but since I never had the S parameter test set for it, I never did much with it. More recently I acquired an 8753 series VNA together with the matching S parameter test set and it just occurred to me that I could fully check out the old VNA by hooking it up to the newer test set if that was possible without causing damage to either component. So there's a bit of background for you. Unfortunately from what's been posted to this thread it appears they just won't work together, however. :(


 

?I have to go against the the nay-sayers.
There is no reason why you cannot use the 85046A with the 8754A. The only issue is controlling it. Fortunatly the interface is really simple. You need 4 bits of data and a strobe (trigger) plus a 22V unregulated supply. Three of the bits are for the attenuator (10, 20, 40dB) and one for the RF relay / switch. So you need as a minimum 4 toggle switches and a push button switch plus DC power supply. If you want to get fancy a 3 bit Binary / BCD switch for the attenuator will allow direct selection.? You just set the switches and push the button.
A quick look at the manual and schematic should make it clear.

Robert G8RPI.


 

S parameters, as you know, can be characterized by S11, S21, S12 and S22. The two port test sets have enough directional couplers and power splitters inside to measure all four, by switching the direction of the RF feed, if directed to do all four measurements by the VNA controller. However, the 8754A was built long before that became a necessity for fast bench measurements. There are still just two ports on the test set, I am sorry if I confused the topic. I meant four S parameters being measured.


 

开云体育

8754A Operating Information page 4 - last paragraph:
S-Parameter Test Set. Transmission and reflection measurements on two port devices wh.ich require
measurement of both forward and reverse characteristics can be accomplished using the 8748A S-Parameter
Test Set. This test set contains the necessary splitter, couplers, and switches to measure forward and reverse
characteristics without disconnecting the reversing the test device, a 0 to 70 dB, 10 dB step attenuator to
control the incident signal level, and includes the necessary equal length signal cables and interface cable to
connect the test set to the 8754A. Connect the test device using the 11857A Test Port Extension Cables or
the 11608A, 11600B, or 11602B Transistor Fixtures.

Am 20.02.2022 um 17:36 schrieb John Lyles:

S parameters, as you know, can be characterized by S11, S21, S12 and S22. The two port test sets have enough directional couplers and power splitters inside to measure all four, by switching the direction of the RF feed, if directed to do all four measurements by the VNA controller. However, the 8754A was built long before that became a necessity for fast bench measurements. There are still just two ports on the test set, I am sorry if I confused the topic. I meant four S parameters being measured.


 

One could do that, but it isn't preserving the 8754A as it was originally intended to be used, it is making a hybrid VNA with the automatic switching of a 8753 etc. When we bought 8754A in the 1980s, it was the cat's meow for doing measurements up to 1 GHz. We stocked up on good directional couplers, like the Narda 3020A, and did whatever test we needed with a single path setup. I was designing high power cavity amplifiers for FM broadcast transmitters at the time, so it was great for sweeping the big circuit for resonances and gain. People designing antennas had them strapped on their range to sweep the antenna with a solid state power amplifier to boost the level. Using external directional couplers, that was very easy to do using the 8754A. I even saw one up on a tower, as heavy as they were.

Then came the Hp3577A. It had calibration built in, that could be run to normalize the setup. We bought that for the transmitter work, only ran to 200 MHz though. I moved on to another company, where we had a Hp8409A VNA, which was a Hp8410C network analyzer, test set, switches, Hp8620 sweeper, and a 9836C computer. It was two racks of equipment. Clunk clunk clunk. The Hp8510A came out and revolutionized measurements, all in a single rack on wheels. I donated our 8409 to a university. Quickly after 8510A, the Hp8753A revolutionized the world again, and they were hot items! I moved to another employer and by now I have operated every network analyzer until the latest models from Keysight, and I just got one of them at work, getting used to it. The only one i never laid hands on was the 8753D and some of the PNA models after the 8753 was out of production.

If it were me, and it isn't, I would not put a lot of work into trying to make a 8754A do things that weren't intended, when you can find a lot of amazing network analyzer capability either used or new from so many sources. People are making measurements with a hundred dollar devices now, although they are no where as good as a real bench VNA. I do recommend that you pick up a couple of Narda 3020A broadband directional couplers, or the HP equivalents, to make good use of the 8754A for insertion loss, gain, reflection, all done by hand. It will make you understand the measurements very well using that box!

Sadly, an old engineer at work bought a 8754A back in end of the 1980s, when it was about to be discontinued. They didn't have the foresight to spend on the 8753 at the time. That 8754A sat in a cabinet, hardly ever used. I ran across it about 5 years ago, with the original manual. As much as I hated to do it, I sent it to our surplus/salvage group and hopefully someone out there got a unit in perfect condition after it was auctioned off at Bentley's in Albuquerque. I strapped the manual to it with tape as our salvage/surplus operations tends to throw out all the paper and cables. Cool old box for sure. Probably fun to repair too, so much RF and analog in it.
John Lyles


 

I think I have the original Hp APP NOTE describing measuring S parameters using the 8754A. If I can find it, I shall scan it for here. The little T/R test set that they sold worked nicely, but it didn't have automatic switching as controlled by the 8754A itself. Nor could it change signal directions to make the complimentary measurements (S22, S12), which was done by hand of course. The 8748A appears to do exactly that, although I never saw one in the flesh in my career. I wonder how rare those are?


 

That's a good point, Rainer. I don't have the phase-matched cables necessary. I did try making some up a while ago, but unfortunately they were not 'flat' all the way up to 1300Mhz. It's a lot harder than one might think!


 

HP stuff in general is very serviceable as I'm sure everyone here knows. The 8754A I have has been apart quite a few times since I've owned it. It's a dream to work on and pretty much infinitely reparable into the future. That's another great attraction for me with vintage test equipment. If you buy new today you haven't a hope in hell of fixing it if it goes wrong. So it seems to me, anyway, but of course YMMV.


 

开云体育

Thanks, John. I was hoping that was what you’d meant.

?

From: [email protected] <[email protected]> On Behalf Of John Lyles
Sent: Sunday, February 20, 2022 12:36 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Mixing 'n' matching S parameter test sets

?

S parameters, as you know, can be characterized by S11, S21, S12 and S22. The two port test sets have enough directional couplers and power splitters inside to measure all four, by switching the direction of the RF feed, if directed to do all four measurements by the VNA controller. However, the 8754A was built long before that became a necessity for fast bench measurements. There are still just two ports on the test set, I am sorry if I confused the topic. I meant four S parameters being measured.