The SM435C is a damned impressive piece of hardware.? If you are seeing them advertised for ¡°10k$,¡± I¡¯ll take a piece of that action. ? Both it and the N9020B offer 160 MHz of acquisition bandwidth, but you can actually get the data out of the Signal Hound, streamable in realtime over 10Gb Ethernet.? ?And at first glance it looks like the N9020B has about 20 dB worse phase noise than the N9020B.? ? The Signal Hound only goes to 43 GHz, though, where the MXA does 50. ?So there¡¯s that. ? ? -- john, KE5FX ?
toggle quoted message
Show quoted text
|
Re: Agilent E4404B Spectrum Analyzer With Dead PSU Repair
On 4/22/24 15:22, Allen Hill via groups.io wrote: Are RIF caps still used in new equipment, or have manufacturers found better solutions? The common RIFA capacitor failure is a materials problem, which was solved long ago. Later RIFA capacitors do not exhibit the explosion problem. Many companies manufacture safety capacitors. -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
Re: Agilent E4404B Spectrum Analyzer With Dead PSU Repair
Are RIF caps still used in new equipment, or have manufacturers found better solutions?
Thanks KI4QCK Allen Hill
toggle quoted message
Show quoted text
On Apr 22, 2024, at 2:55?PM, Nick Stanley <rosewardxiii@...> wrote:
?Hi all, figure a good old repair attempt would be a good 1st post.
I bought a broken E4404B SA in hopes to repair it. The seller said it has a dead PSU, but doesn't know if it has any other failures.
I took the PSU out to test it. I decided to test all the MOSFETs and the 2 optocouplers, all tested good.
Then I realized since this PSU can take DC input, I can just supply 12VDC and see if the DC-DC converter secondary side work. And it DOES. All voltage rails (+28V, +15V. +5V. -5V, -15V, fan) are all present. Draws about 5-6W which seems alright.
Which narrows down the failure to the AC mains sides. I checked the mains side, the 2 bulk caps hold a total of about 310VDC, which matches with the peak of rectified mains AC.
When I force the supply on while connect to mains, I hears a small click every second. Measurements of the DC rails shows them varying greatly. It's like the PSU is trying to start up but can't.
Will update with any progress and pictures soon.
Also, as I was testing the PSU, one of the RIFA caps exploded, filling my room with smoke and cover the PCB in tars. Horrible
|
Ah, you are correct; the PXA models we have at work are N9030A and N9030B.? My mistake.
Regarding the data format, that is something interesting that I have not looked into.? It may be an opportunity to speed up the tests; I will check with our instrumentation engineers.
Thanks a bunch, Hardy!
Jim
------ Original Message ------
Date 4/22/2024 12:02:51 PM
Subject Re: [HP-Agilent-Keysight-equipment] N9020A
Hi First we have to be sure what analyzer ?model we test¡N9020A and N9020B both belong to the MXA series---not the PXA series. Regarding transfer speed we also must look at the data format in use---real-floating-integer -ascii-16 or 64 bit.That makes a great difference. Hardy ? ? I note that the newer PXAs run our automated tests that check for spurious signals considerably faster than even the older PXAs at work.? The old ones take about 1.5 times longer to do the same tests.? IIRC, the model numbers are N9020A and N9020B.? I'm working from home today so I can't check until tomorrow.? But even within a product line, improvements in performance do happen. Jim Ford Laguna Hills, California, USA "And we know that God causes all things to work together for the good of those who love God and are called according to his purpose for them." Romans 8:28 (NLT)
------ Original Message ------ Date 4/21/2024 11:44:15 PM Subject Re: [HP-Agilent-Keysight-equipment] N9020A The X-series analyzers have evolved quite a bit since their introduction, as the early versions ran XP, and the newer ones are probably on Win10, maybe 11. That gets into licensing negotiations with uSoft. Even with XP, the EXA, MXA, and PXA were faster and more capable than the PSA. ? The PSA was faster than the 8566/8568, and had a better display, as well as better dynamic range and speed. Log mag display linearity was no longer an issue as the log amps in the IF were replaced with software log functions. Dynamic range was at least 20 dB better. Amplitude accuracy improved from =/- 1.5 dB or worse to less than .5 dB, almost into power meter territory. ?The limitation of the PSA was that it could not support USB devices very well. Writing thee drivers to accommodate every possible peripheral in the wild, wild west that was USB back then was a lot of work. uS was doing all of that work, so why re-invent the wheel. ? The EXA/MXA/PXA took a lot of the hardware tech from the PSA and added the speed, memory, peripheral support, and processing horsepower of a modern PC, integral to the machine. Could it have used Linux instead of Windows? Yes, but at the time, Linux did not have the horsepower and peripheral support, among other things. ? ? Some features that ship standard with every box: ? More displayed data points, at least 64,000 last I knew were available. No, the LCD won¡¯t show them all but they are all available via GPIB, USB, or LAN ports. ? Sweep speeds are substantially faster than the 8566/8568. ? Amplitude accuracy and frequency are much better because of the large amounts of memory, both RAM and disk, which allows calibrating at more points across the range of the box, and applying the corrections immediately. ? Measurements like channel power and adjacent channel power are available. ? There are at least 4, traces available, in color, that can use any of 4 or 5 different detectors, so you can compare traces taken simultaneously with different detectors ( Rosenfell, Peak, Neg Peak, Average, and Sample). There are more trace math tricks available than the earlier models. ? There are, I think, 10 markers available, on each trace. Any marker can be a reference marker and any other marker can be a delta marker to that. Multiple reference and delta markers can exist on the same trace. ? On board storage of setups, traces, etc. is limited only by the size of the hard disk, which of course have grown larger over time. With the USB and LAN ports, off-board storage is just like you would do on a Win PC. ? There are several other cool things the X-series can do, but I think I¡¯ve hit a few of the more useful ones. ? I have used the 8566/8568 family since they came out in the late 70¡¯s, the PSA for several years, and also the X-series since they were introduced. I still own a 66, a 68, and a 3.6 GHz EXA. I don¡¯t use the 68 much. I use the 66 for microwave needs, but the EXA for most everything else on my test bench. Yeah, I hate the slow boot up time, but it does everything else so well and so much faster, and I very seldom need to post-process any sweeps, which saves me a lot more time than the slow boot up costs. ? My EXA has the noise figure option, and it does have some quirks compared the using the 8970 system that I used to have, but it is just as capable. It doesn¡¯t have quite as many tricks as the 8970 could pull off with its Special Functions menu, but I haven¡¯t missed those. ? I have had the EXA for going on 15 years and have not had any problems with it. (OK, now I¡¯m slightly apprehensive about going into the lab and turning it on, having said that!) ? Just like PCs in general, the hardware needed to support newer versions of Windows has changed, and obsoletes the older stuff. An upgrade is pretty simple, though not hobbyist cheap. The computer part ?is a plug in module that has a standard interface, so that , last I heard, an original X-series can be upgraded to the lates Wintel version just by replacing that module. The RF hardware doesn¡¯t need to change at all. Certainly, I would expect that more recently built units will have better RF stuff as that evolves over time. ? Blame the suits for going to a Windows OS? Nope, blame the engineers who finally got the freedom to build a machine that could do everything that the evolving technology of RF hardware, signal processing, and computing horsepower made possible. ? Tom, N8ZM ? ? Hi Just an extra note The preamp are inbuilt all MXA`s One of the very nice features are a package named EDP..enhanced display package¡ª its also an ?software option and works wonders-it simply lets you make normal sweeps at your own choise¡and have a split screen with a spectrogram¡. If one buy from the seller I earlier have linked to,you simply ask Hong for the options you want..and the analyzer can use!! ? ? Hi Jeremy et all The MXA have both a preamp and electronically attenuator as option.All MXA have inbuilt 28v dc for noise sources. The phase noise and noise figure options works even better than the one used in the old PSA. Regarding vsa 89600 software¡.all MXA` used to come with ?version 12.02 and version 14 preinstalled.works great in my MXA`s -and of course one can use this software from a pc as well. There exist ?cracks for version 12.02 and version 22.21. Only two of the sellers at ebay can offer the units with-a lot- of options¡have made a link. I have collected most info about these nice systems.If interested I can share!! The only caveat are the bandwidth at 3.6 Ghz,but many of us ?still have the ?old HP/Agilent 856x/859x series analyzers going to multi Ghz.Combinedwith the 3.6 Ghz MXA it¡¯s a great combo as the vsa 89600 sofware have an inbuilt up/down converter mode.The MXA come standard with a digital bandwidth of 25 Mhz!!! Hardy ? ? ? Here's the datasheet for the N9020A? MXA analyser
It may be the case that the 28V hardware for driving an external noise source is standard equipment. It might therefore be possible to measure noise figure by adding a suitable licence code. Maybe Hardy can advise about this.
The analyser would also need a preamp to cover up to 3.6GHz or an external (homebrew or HP) preamp could be fitted. The noise figure of the internal preamp isn't that great, so I'd recommend making an external preamp, especially for use below about 1GHz.
A noise source would also be needed and this could be homebrew or something like an HP 346A/B. A homebrew noise source would be a lot cheaper but it won't be as stable or as accurate.
The noise figure GUI will probably be similar to that used in other Agilent/Keysight analysers. It's a really nice feature to have on a modern spectrum analyser.?
|
Hi Jeremy Spot on One can not? compare the performance of Signal Hound(many versions) and? the MXA. We are dealing with a low cost surplus ?instrument with specs far beyond the? usual broad band receive/monitor devices as I think signal hound belong to- up to units cost about 10k$--just my view¡..No ?this thread are about getting a instrument with performance far beyond what we are used to ?have. Hardy ? ? On Mon, Apr 22, 2024 at 03:23 PM, John Miles wrote: Frankly if I were shopping for a new MXA-class spectrum analyzer today I'd
be inclined to look at Signal Hound first. Their newer SAs don't give up
much ground to the big boys in terms of features or performance, and most of
their gear supports Linux as well as Windows.
Which SH product were you thinking of that can be compared to an MXA? Their products appear to me to be aimed at a different type of customer.
In other words, if you want something small and low cost that can sweep quickly and process data quickly then the SH products look very powerful to me. However, to get this you have to trade against some fairly grim looking performance specs elsewhere. There's too many compromises in the hardware to allow a valid comparison with an MXA analyser unless the person doesn't actually 'need' the all round excellence of the MXA.
|
Re: 5342A: sometimes counts, sometimes not
PD: service manual is pure gold. I found this thread to pull ...
|
Agilent E4404B Spectrum Analyzer With Dead PSU Repair
Hi all, figure a good old repair attempt would be a good 1st post.
I bought a broken E4404B SA in hopes to repair it. The seller said it has a dead PSU, but doesn't know if it has any other failures.
I took the PSU out to test it. I decided to test all the MOSFETs and the 2 optocouplers, all tested good.
Then I realized since this PSU can take DC input, I can just supply 12VDC and see if the DC-DC converter secondary side work. And it DOES. All voltage rails (+28V, +15V. +5V. -5V, -15V, fan) are all present. Draws about 5-6W which seems alright.
Which narrows down the failure to the AC mains sides. I checked the mains side, the 2 bulk caps hold a total of about 310VDC, which matches with the peak of rectified mains AC.
When I force the supply on while connect to mains, I hears a small click every second. Measurements of the DC rails shows them varying greatly. It's like the PSU is trying to start up but can't.
Will update with any progress and pictures soon.
Also, as I was testing the PSU, one of the RIFA caps exploded, filling my room with smoke and cover the PCB in tars. Horrible
|
Re: 5342A: sometimes counts, sometimes not
Remarkable to mention that the input 2 works fine, with both internal and external freq references.
Internal opt 001 oven it's now adjusted to be closest possible to the gpsdo phase.
Still don't get that 75000000 with self check, I'll read the service manual about.
R.
|
Hey Are we still talking about N9020A?....come on ..dont hi-jack this thread!!! Hardy -----Oprindelig meddelelse----- Fra: [email protected]< [email protected]> P? vegne af Chuck Harris Sendt: 22. april 2024 19:27 Til: [email protected]Emne: Re: [HP-Agilent-Keysight-equipment] N9020A Hi John, I tried several other words in that place before communism. I thought perhaps fascism, as defined by Mussolini, socialism, corporatism... None of them really fit quite right. The essence I was trying to convey was one of forced redistribution of wealth, which is a prime affect of communism. The government is being asked by some consumers to force manufacturers to spend their money in ways they don't consider beneficial to their companies... using methods that aren't included in the government's enumerated powers. Feudalism didn't occur to me. It requires a monarch with power over the buyers... I don't know who that would be. For every John Deere, there are ten other companies, many of them foreign, that won't lock down repairs. For every Apple, there are 1000 different companies making android phones. I haven't run factory software in any of my phones in decades. Out here in farm country, I am seeing lots of orange, red and blue tractors these days. -Chuck Harris On Mon, 22 Apr 2024 08:27:31 -0700 "John Miles" <john@...> wrote: Everything in politics is a reaction to something else. Right To
Repair is a reaction to blatant market abuses by a couple of bad
actors -- not calling any names, but, um, Apple and John Deere.
Customers were basically given the choice to shoulder unreasonable
costs to maintain products they've purchased or to switch to inferior
products... a choice that's increasingly dictated by network effects.
If those companies were surprised when their customers wrote their
Congressmen, well, gee, bummer, I guess that's on them.
Rather than being a manifestation of "communism," right-to-repair is a
reaction to the "You will own nothing and be happy" sentiment
propagated by certain people who would not normally be identified as
communists. It's not helpful or productive to use terms like
"communism" in this context. Feudalism might be a better card to play.
-- john, KE5FX
-----Original Message-----
From: [email protected]
<[email protected]> On Behalf Of Chuck Harris
Right to Repair, is a communistic, people's power sort of proposal...
-- Denne mail er blevet tjekket for vira af AVG-antivirussoftware. www.avg.com
|
Hi John,
I tried several other words in that place before communism.
I thought perhaps fascism, as defined by Mussolini, socialism,
corporatism... None of them really fit quite right.
The essence I was trying to convey was one of forced redistribution
of wealth, which is a prime affect of communism. The government
is being asked by some consumers to force manufacturers to spend
their money in ways they don't consider beneficial to their
companies... using methods that aren't included in the government's
enumerated powers.
Feudalism didn't occur to me. It requires a monarch with power
over the buyers... I don't know who that would be. For every
John Deere, there are ten other companies, many of them foreign,
that won't lock down repairs. For every Apple, there are 1000
different companies making android phones.
I haven't run factory software in any of my phones in decades.
Out here in farm country, I am seeing lots of orange, red and
blue tractors these days.
-Chuck Harris
toggle quoted message
Show quoted text
On Mon, 22 Apr 2024 08:27:31 -0700 "John Miles" <john@...> wrote: Everything in politics is a reaction to something else. Right To
Repair is a reaction to blatant market abuses by a couple of bad
actors -- not calling any names, but, um, Apple and John Deere.
Customers were basically given the choice to shoulder unreasonable
costs to maintain products they've purchased or to switch to inferior
products... a choice that's increasingly dictated by network effects.
If those companies were surprised when their customers wrote their
Congressmen, well, gee, bummer, I guess that's on them.
Rather than being a manifestation of "communism," right-to-repair is a
reaction to the "You will own nothing and be happy" sentiment
propagated by certain people who would not normally be identified as
communists. It's not helpful or productive to use terms like
"communism" in this context. Feudalism might be a better card to play.
-- john, KE5FX
-----Original Message-----
From: [email protected]
<[email protected]> On Behalf Of Chuck Harris
Right to Repair, is a communistic, people's power sort of proposal...
|
Re: 5342A: sometimes counts, sometimes not
|
On Mon, Apr 22, 2024 at 03:23 PM, John Miles wrote:
Frankly if I were shopping for a new MXA-class spectrum analyzer today I'd
be inclined to look at Signal Hound first. Their newer SAs don't give up
much ground to the big boys in terms of features or performance, and most of
their gear supports Linux as well as Windows.
Which SH product were you thinking of that can be compared to an MXA? Their products appear to me to be aimed at a different type of customer. In other words, if you want something small and low cost that can sweep quickly and process data quickly then the SH products look very powerful to me. However, to get this you have to trade against some fairly grim looking performance specs elsewhere. There's too many compromises in the hardware to allow a valid comparison with an MXA analyser unless the person doesn't actually 'need' the all round excellence of the MXA.
|
Hi First we have to be sure what analyzer ?model we test¡N9020A and N9020B both belong to the MXA series---not the PXA series. Regarding transfer speed we also must look at the data format in use---real-floating-integer -ascii-16 or 64 bit.That makes a great difference. Hardy ? ? I note that the newer PXAs run our automated tests that check for spurious signals considerably faster than even the older PXAs at work.? The old ones take about 1.5 times longer to do the same tests.? IIRC, the model numbers are N9020A and N9020B.? I'm working from home today so I can't check until tomorrow.? But even within a product line, improvements in performance do happen. Jim Ford Laguna Hills, California, USA "And we know that God causes all things to work together for the good of those who love God and are called according to his purpose for them." Romans 8:28 (NLT)
toggle quoted message
Show quoted text
------ Original Message ------Date 4/21/2024 11:44:15 PM Subject Re: [HP-Agilent-Keysight-equipment] N9020A The X-series analyzers have evolved quite a bit since their introduction, as the early versions ran XP, and the newer ones are probably on Win10, maybe 11. That gets into licensing negotiations with uSoft. Even with XP, the EXA, MXA, and PXA were faster and more capable than the PSA. ? The PSA was faster than the 8566/8568, and had a better display, as well as better dynamic range and speed. Log mag display linearity was no longer an issue as the log amps in the IF were replaced with software log functions. Dynamic range was at least 20 dB better. Amplitude accuracy improved from =/- 1.5 dB or worse to less than .5 dB, almost into power meter territory. ?The limitation of the PSA was that it could not support USB devices very well. Writing thee drivers to accommodate every possible peripheral in the wild, wild west that was USB back then was a lot of work. uS was doing all of that work, so why re-invent the wheel. ? The EXA/MXA/PXA took a lot of the hardware tech from the PSA and added the speed, memory, peripheral support, and processing horsepower of a modern PC, integral to the machine. Could it have used Linux instead of Windows? Yes, but at the time, Linux did not have the horsepower and peripheral support, among other things. ? ? Some features that ship standard with every box: ? More displayed data points, at least 64,000 last I knew were available. No, the LCD won¡¯t show them all but they are all available via GPIB, USB, or LAN ports. ? Sweep speeds are substantially faster than the 8566/8568. ? Amplitude accuracy and frequency are much better because of the large amounts of memory, both RAM and disk, which allows calibrating at more points across the range of the box, and applying the corrections immediately. ? Measurements like channel power and adjacent channel power are available. ? There are at least 4, traces available, in color, that can use any of 4 or 5 different detectors, so you can compare traces taken simultaneously with different detectors ( Rosenfell, Peak, Neg Peak, Average, and Sample). There are more trace math tricks available than the earlier models. ? There are, I think, 10 markers available, on each trace. Any marker can be a reference marker and any other marker can be a delta marker to that. Multiple reference and delta markers can exist on the same trace. ? On board storage of setups, traces, etc. is limited only by the size of the hard disk, which of course have grown larger over time. With the USB and LAN ports, off-board storage is just like you would do on a Win PC. ? There are several other cool things the X-series can do, but I think I¡¯ve hit a few of the more useful ones. ? I have used the 8566/8568 family since they came out in the late 70¡¯s, the PSA for several years, and also the X-series since they were introduced. I still own a 66, a 68, and a 3.6 GHz EXA. I don¡¯t use the 68 much. I use the 66 for microwave needs, but the EXA for most everything else on my test bench. Yeah, I hate the slow boot up time, but it does everything else so well and so much faster, and I very seldom need to post-process any sweeps, which saves me a lot more time than the slow boot up costs. ? My EXA has the noise figure option, and it does have some quirks compared the using the 8970 system that I used to have, but it is just as capable. It doesn¡¯t have quite as many tricks as the 8970 could pull off with its Special Functions menu, but I haven¡¯t missed those. ? I have had the EXA for going on 15 years and have not had any problems with it. (OK, now I¡¯m slightly apprehensive about going into the lab and turning it on, having said that!) ? Just like PCs in general, the hardware needed to support newer versions of Windows has changed, and obsoletes the older stuff. An upgrade is pretty simple, though not hobbyist cheap. The computer part ?is a plug in module that has a standard interface, so that , last I heard, an original X-series can be upgraded to the lates Wintel version just by replacing that module. The RF hardware doesn¡¯t need to change at all. Certainly, I would expect that more recently built units will have better RF stuff as that evolves over time. ? Blame the suits for going to a Windows OS? Nope, blame the engineers who finally got the freedom to build a machine that could do everything that the evolving technology of RF hardware, signal processing, and computing horsepower made possible. ? Tom, N8ZM ? ? Hi Just an extra note The preamp are inbuilt all MXA`s One of the very nice features are a package named EDP..enhanced display package¡ª its also an ?software option and works wonders-it simply lets you make normal sweeps at your own choise¡and have a split screen with a spectrogram¡. If one buy from the seller I earlier have linked to,you simply ask Hong for the options you want..and the analyzer can use!! ? ? Hi Jeremy et all The MXA have both a preamp and electronically attenuator as option.All MXA have inbuilt 28v dc for noise sources. The phase noise and noise figure options works even better than the one used in the old PSA. Regarding vsa 89600 software¡.all MXA` used to come with ?version 12.02 and version 14 preinstalled.works great in my MXA`s -and of course one can use this software from a pc as well. There exist ?cracks for version 12.02 and version 22.21. Only two of the sellers at ebay can offer the units with-a lot- of options¡have made a link. I have collected most info about these nice systems.If interested I can share!! The only caveat are the bandwidth at 3.6 Ghz,but many of us ?still have the ?old HP/Agilent 856x/859x series analyzers going to multi Ghz.Combinedwith the 3.6 Ghz MXA it¡¯s a great combo as the vsa 89600 sofware have an inbuilt up/down converter mode.The MXA come standard with a digital bandwidth of 25 Mhz!!! Hardy ? ? ? Here's the datasheet for the N9020A? MXA analyser
It may be the case that the 28V hardware for driving an external noise source is standard equipment. It might therefore be possible to measure noise figure by adding a suitable licence code. Maybe Hardy can advise about this.
The analyser would also need a preamp to cover up to 3.6GHz or an external (homebrew or HP) preamp could be fitted. The noise figure of the internal preamp isn't that great, so I'd recommend making an external preamp, especially for use below about 1GHz.
A noise source would also be needed and this could be homebrew or something like an HP 346A/B. A homebrew noise source would be a lot cheaper but it won't be as stable or as accurate.
The noise figure GUI will probably be similar to that used in other Agilent/Keysight analysers. It's a really nice feature to have on a modern spectrum analyser.?
|
I note that the newer PXAs run our automated tests that check for spurious signals considerably faster than even the older PXAs at work.? The old ones take about 1.5 times longer to do the same tests.? IIRC, the model numbers are N9020A and N9020B.? I'm working from home today so I can't check until tomorrow.? But even within a product line, improvements in performance do happen.
HTH
Jim Ford Laguna Hills, California, USA
"And we know that God causes all things to work together for the good of those who love God and are called according to his purpose for them." Romans 8:28 (NLT)
------ Original Message ------
Date 4/21/2024 11:44:15 PM
Subject Re: [HP-Agilent-Keysight-equipment] N9020A
The X-series analyzers have evolved quite a bit since their introduction, as the early versions ran XP, and the newer ones are probably on Win10, maybe 11. That gets into licensing negotiations with uSoft. Even with XP, the EXA, MXA, and PXA were faster and more capable than the PSA. ? The PSA was faster than the 8566/8568, and had a better display, as well as better dynamic range and speed. Log mag display linearity was no longer an issue as the log amps in the IF were replaced with software log functions. Dynamic range was at least 20 dB better. Amplitude accuracy improved from =/- 1.5 dB or worse to less than .5 dB, almost into power meter territory. ?The limitation of the PSA was that it could not support USB devices very well. Writing thee drivers to accommodate every possible peripheral in the wild, wild west that was USB back then was a lot of work. uS was doing all of that work, so why re-invent the wheel. ? The EXA/MXA/PXA took a lot of the hardware tech from the PSA and added the speed, memory, peripheral support, and processing horsepower of a modern PC, integral to the machine. Could it have used Linux instead of Windows? Yes, but at the time, Linux did not have the horsepower and peripheral support, among other things. ? ? Some features that ship standard with every box: ? More displayed data points, at least 64,000 last I knew were available. No, the LCD won¡¯t show them all but they are all available via GPIB, USB, or LAN ports. ? Sweep speeds are substantially faster than the 8566/8568. ? Amplitude accuracy and frequency are much better because of the large amounts of memory, both RAM and disk, which allows calibrating at more points across the range of the box, and applying the corrections immediately. ? Measurements like channel power and adjacent channel power are available. ? There are at least 4, traces available, in color, that can use any of 4 or 5 different detectors, so you can compare traces taken simultaneously with different detectors ( Rosenfell, Peak, Neg Peak, Average, and Sample). There are more trace math tricks available than the earlier models. ? There are, I think, 10 markers available, on each trace. Any marker can be a reference marker and any other marker can be a delta marker to that. Multiple reference and delta markers can exist on the same trace. ? On board storage of setups, traces, etc. is limited only by the size of the hard disk, which of course have grown larger over time. With the USB and LAN ports, off-board storage is just like you would do on a Win PC. ? There are several other cool things the X-series can do, but I think I¡¯ve hit a few of the more useful ones. ? I have used the 8566/8568 family since they came out in the late 70¡¯s, the PSA for several years, and also the X-series since they were introduced. I still own a 66, a 68, and a 3.6 GHz EXA. I don¡¯t use the 68 much. I use the 66 for microwave needs, but the EXA for most everything else on my test bench. Yeah, I hate the slow boot up time, but it does everything else so well and so much faster, and I very seldom need to post-process any sweeps, which saves me a lot more time than the slow boot up costs. ? My EXA has the noise figure option, and it does have some quirks compared the using the 8970 system that I used to have, but it is just as capable. It doesn¡¯t have quite as many tricks as the 8970 could pull off with its Special Functions menu, but I haven¡¯t missed those. ? I have had the EXA for going on 15 years and have not had any problems with it. (OK, now I¡¯m slightly apprehensive about going into the lab and turning it on, having said that!) ? Just like PCs in general, the hardware needed to support newer versions of Windows has changed, and obsoletes the older stuff. An upgrade is pretty simple, though not hobbyist cheap. The computer part ?is a plug in module that has a standard interface, so that , last I heard, an original X-series can be upgraded to the lates Wintel version just by replacing that module. The RF hardware doesn¡¯t need to change at all. Certainly, I would expect that more recently built units will have better RF stuff as that evolves over time. ? Blame the suits for going to a Windows OS? Nope, blame the engineers who finally got the freedom to build a machine that could do everything that the evolving technology of RF hardware, signal processing, and computing horsepower made possible. ? Tom, N8ZM ? ? Hi Just an extra note The preamp are inbuilt all MXA`s One of the very nice features are a package named EDP..enhanced display package¡ª its also an ?software option and works wonders-it simply lets you make normal sweeps at your own choise¡and have a split screen with a spectrogram¡. If one buy from the seller I earlier have linked to,you simply ask Hong for the options you want..and the analyzer can use!! ? ? Hi Jeremy et all The MXA have both a preamp and electronically attenuator as option.All MXA have inbuilt 28v dc for noise sources. The phase noise and noise figure options works even better than the one used in the old PSA. Regarding vsa 89600 software¡.all MXA` used to come with ?version 12.02 and version 14 preinstalled.works great in my MXA`s -and of course one can use this software from a pc as well. There exist ?cracks for version 12.02 and version 22.21. Only two of the sellers at ebay can offer the units with-a lot- of options¡have made a link. I have collected most info about these nice systems.If interested I can share!! The only caveat are the bandwidth at 3.6 Ghz,but many of us ?still have the ?old HP/Agilent 856x/859x series analyzers going to multi Ghz.Combinedwith the 3.6 Ghz MXA it¡¯s a great combo as the vsa 89600 sofware have an inbuilt up/down converter mode.The MXA come standard with a digital bandwidth of 25 Mhz!!! Hardy ? ? ? Here's the datasheet for the N9020A? MXA analyser
It may be the case that the 28V hardware for driving an external noise source is standard equipment. It might therefore be possible to measure noise figure by adding a suitable licence code. Maybe Hardy can advise about this.
The analyser would also need a preamp to cover up to 3.6GHz or an external (homebrew or HP) preamp could be fitted. The noise figure of the internal preamp isn't that great, so I'd recommend making an external preamp, especially for use below about 1GHz.
A noise source would also be needed and this could be homebrew or something like an HP 346A/B. A homebrew noise source would be a lot cheaper but it won't be as stable or as accurate.
The noise figure GUI will probably be similar to that used in other Agilent/Keysight analysers. It's a really nice feature to have on a modern spectrum analyser.?
|
Everything in politics is a reaction to something else. Right To Repair is
a reaction to blatant market abuses by a couple of bad actors -- not calling
any names, but, um, Apple and John Deere. Customers were basically given
the choice to shoulder unreasonable costs to maintain products they've
purchased or to switch to inferior products... a choice that's increasingly
dictated by network effects.
If those companies were surprised when their customers wrote their
Congressmen, well, gee, bummer, I guess that's on them.
Rather than being a manifestation of "communism," right-to-repair is a
reaction to the "You will own nothing and be happy" sentiment propagated by
certain people who would not normally be identified as communists. It's not
helpful or productive to use terms like "communism" in this context.
Feudalism might be a better card to play.
-- john, KE5FX
toggle quoted message
Show quoted text
|
Just off the top of my head, it can be done, but perhaps not as easily as you'd think.
You need a central nexus, either FPGA, discrete circuitry, or microprocessor, that takes commands and makes hardware work.? It would have to be all the hardware through this central node.? If this node then has a computer interface, you can run everything remotely from the computer, or default to another, larger internal processor for standalone.
This is economically unfeasible for a manufacturer, though.
The design limitations are the complexity of the control node (and method of interface), the speed of that interface in getting data out, and of course, the tasks it needs to do.
You could distribute the machine's intelligence with a "smart RF section", a "smart Display", and so on, depending on what you want.
All very modular, but not trivial to do.? At best, you have an add-on control module for each section.? That might almost be a standardized design.
The things you control must lend themselves to this purpose, as well.
Harvey
toggle quoted message
Show quoted text
On 4/22/2024 5:54 AM, Martin via groups.io wrote: HI Hardy,
I did not mean trying this on an existing instrument.
I was wondering how one could design the architecture of a modern instrument to be more universal and less dependant on a single computer. I imagine (dream of) enabling the HW core of the instrument to be used to full extent with an external computer if the internal has failed or cannot be used anymore.
cheers
Martin
|
A good approach with respect to being "less dependent on a single computer"
is to use a headless instrument that works with any PC you plug it into.
Almost every box in a modern R&D workspace needs to talk to a PC, but there
are few good reasons to demand that the T&M manufacturers build a separate
PC and display into each of them. (Portability in a field-friendly
instrument being one of those good reasons.)
Frankly if I were shopping for a new MXA-class spectrum analyzer today I'd
be inclined to look at Signal Hound first. Their newer SAs don't give up
much ground to the big boys in terms of features or performance, and most of
their gear supports Linux as well as Windows.
Keysight has made some exploratory moves into the headless space with the
Streamline products, but SH and Tek got the jump on them in the
spectrum-analyzer department.
-- john, KE5FX
toggle quoted message
Show quoted text
-----Original Message-----
I was wondering how one could design the architecture of a modern instrument
to be more universal and less dependant on a single computer. I imagine
(dream of) enabling the HW core of the instrument to be used to full extent
with an external computer if the internal has failed or cannot be used
anymore.
cheers
Martin
|
Re: N9020A - Right to repair
In general I agree with Chuck.
Communistic may be a bit heavy handed.
Hard to fix modern stuff for sure, but farmers, due to the John Deere Company, would argue that the inability to source third party solutions for the modern software driven $250,000+ farm tractor is a real problem, especially when it is sitting broken in the field due to a software/ firmware or electronics module problem. Then JD tells you a serviceman will show up in a month, after the planting season, and the cost is thru the roof... Others could and would source solutions and maybe non-factory repair technician services would appear, who knows, not allowed.
You want all the bells & whistles "needed" for the modern world, well, the problems come too...
Jeff Kruth
|
Right to Repair, is a communistic, people's power sort of proposal. It sounds nice, but technology may have progressed beyond a point where it is practical. It requires the manufacturer to spend money to modify its products, and documentation, and to make special parts available long into the future, so that the purchaser can substitute his own unverified level of competence for the manufacturer's. It requires the manufacturer's warranty not to be voided by customer specified repairs. In most cases it is totally impractical, as the products were developed using hugely expensive CAD/CAM systems, and the product's documentation cannot even exist outside of those systems in a usable form. Who gets the liability when the purchaser finds that he, or his chosen independent repair organization, does bad work? Who will compensate the manufacturer when bad outcomes invariably happen, and the manufacturer's product and good name get smeared in the public's eye as the result of a faulty repair? Even something as prosaic as a 1980's automobile has already found this to have happened. Ford Explorer's developers tuned the suspension on the SUV to the tires they specified too tightly. Because the car's owners could use any tires they wished on their cars, some chose tires that did not meet the characteristics of the Firestone tires Ford specified, and unsafe handling was the result. If the car owners had to come to Ford for the tires, the problem might never have been revealed. Was its Ford's fault? Sure, but it only became a problem when incorrect materials were used in the customer ordered repairs. -Chuck Harris On Mon, 22 Apr 2024 10:29:33 +0200 "Martin via groups.io" <musaeum@...> wrote: On 21. Apr 2024, at 23:13, jmr via groups.io
<jmrhzu@...> wrote:
...A fairly decent analyser will have a 14 bit digital IF, a real
time bandwidth of 10MHz to maybe 160MHz and very good
specifications for dynamic range... One part of the analyser is merely a computer doing signal analysis
in SW, displaying measurements, doing automation etc.
Not sure about the architecture of these "modern" instruments, but I
guess their specs mainly depend on the hardware parts. Talking about
"right to repair", would it be possible to pilot these parts
exclusively with some (well documented) logic or FPGA that provides
an interface like HPIB?
If the internal computer (win, linux etc.) goes bad one could still
connect to HPIB and to everything via an external computer. A bit
like when the internal display goes bad and you connect an external
monitor to a VGA port.
cheers
Martin
|
Re: HP 140B LVPS -12.6V Not Working
Hi David,
I saw some of your comments on another site (eevblog?) regarding 1854-0071.? It appears I have the 4-071 TI001 version.? If these are indeed failure prone, then possibly that's what happened in my case.
I didn't want to remove it unless I had good cause but since it tests in circuit the way it does, then I think there's good reason to at least remove it and test it out of circuit.? If it is indeed bad, then I can at least try a 2n3904.
If HP did switch to 2n3904s for that P/N, then I wonder if those are good for any/all instances?? If not, then I might want to get a supply of 2n3704s since, as you can see, that's not the only one on this board (and, maybe, on other boards in this scope).
Thanks again, Barry - N4BUQ
1854-0071 is one of the very common small general purpose transistors used all over HP TE, 1853-0020 is the other common one.
It depends on which manual you look at, I've seen 1854-0071 (marked 4-071) listed as 2N3391 and as a selected 2N3704 in different manuals, they were made by various manufacturers and pinouts vary.
I've found original parts fitted marked MPS3391 in a HP 180, the 4-071 or 1854-0071 seems to be the most often failed part I find in HP scopes & other TE.
Later they used 1854-0215 which is a 2N3904.
A few of the common one and the generic part listed, from manuals.
1854-0071 = selected from 2N3704
1853-0020 = selected from 2N3702
Those are listed in the 432A pdf.
1854-0215 = 2N3904
1853-0036 = 2N3906
And those are listed in the 3310A/B pdf.
David
|
Dave McGuire
RubenRubio
