|
On Sat, May 3, 2025 at 06:59 PM, vbonhorst wrote:
The N432A and Tegam use a NBS Type IV bridge. I attach three links that reflect on power measurement.
Thanks there's some interesting info there. Are you sure the N432A is an IV bridge? Maybe there are different ways to describe these circuits but I think the N432A uses a classic self balancing Wheatstone bridge system that is similar to that used in the old HP 432A.
?
I think the HP 432A is a great choice for home use if the aim is to just calibrate a 1mW power sensor and make measurements up to maybe 1500 MHz at a similar power level. Once the 432A is used below about -10dBm the drift becomes an issue and I think the overall uncertainty will creep up. To check my 50 MHz 1mW reference here, I've been doing annual DC substitution checks using my 432A here for about 9 years and I get consistent results each time and I log all the resistance, V0, V1, Vcmp readings in a spreadsheet. I also find that making a basic DVM measurement at the recorder at the back agrees very closely with the DC substitution method. Maybe I'm just fortunate with the accuracy of the recorder output from this particular 432A?
?
?
?
|
Measuring RF power is not an easy job. Testing power sensors is much more difficult. Bolometers are very good transfer sensors and are especially stable over a long time. Errors in Instrumentation accuracy can be small, depending on the capabilities in DC metrology. The limiting factor always is the knowledge of the calibration factor of the reference power sensor. Some NMIs specify around 0.2 % as their best CMCs. A calibration at this level of accuracy is time consuming and expensive. A second influence is stability of the environment temperature. ?One never should forget, RF power measurement ??is a thermal measurement. Instruments to control thermistor mounts had been made by some companies, like HP/ Keysight and Tegam. The old HP 432A (Type II bridge) is a good low cost choice, If you do not mind DC Substitution. The N432A and Tegam use a NBS Type IV bridge. I attach three links that reflect on power measurement.
? regards, Hubert ?
|
The HPAK system is easier for me to understand (in terms of error management and uncertainty) because it uses a simple resistive Wheatstone bridge and a single differential amplifier for each thermistor in the sensor. It looks like the Tegam system uses multiple amplifiers to allow them to dispense with the two 1k bridge resistors. Presumably it relies on being able to cope with and cancel the offsets in the various amplifiers in their system as it auto nulls.
?
It would be interesting to know how the amplifier offset issue is managed and how significant it can be.?
?
There does appear to be a maintenance port at the rear of the 1830A. This isn't described in the manual and maybe this can support external DC substitution?
|
Thanks. I watched a few of them on youtube and have mixed views on what I saw.
?
?
The webinar above started off really well, but it turned into a bit of a train wreck when the application engineer actually tried to use their hardware with their PS Cal software to do a basic power measurement. To save you all 53 minutes of your time, there was some sort of setup issue and they failed to demonstrate anything and the demo was halted. The only potentially useful thing in that video is a direct contact email for the applications engineer right at the end of the video. Maybe this email is still valid?
?
I also found application note 217 from Tegam
?
?
This also has issues. It really doesn't read like a professional document. I've been reading lots of (truly excellent and informative) HPAK app notes all my career and although HPAK do sometimes make typos and mistakes in their text and diagrams, I think this app note 217 from Tegam has more than just bad diagrams that concern me.
Obviously their system diagram is a hastily drawn mess, with nodes grounded that shouldn't be grounded, but it's the wording and the poor technical description and lack of analysis that sets this apart from HPAK. It also appears that they don't use a hardware based self balancing bridge. HPAK use two 1K and one 200R resistors and the thermistor in a bridge. But Tegam appear to do it a slightly different way to HPAK? Maybe the diagram is missing some key components. There's no uncertainty analysis anywhere, just an accuracy figure in the datasheet.
?
For me, even if the Tegam unit is technically comparable to HPAK, the N432A and the old HP432A have the massive advantage that they both offer external DC substitution using a DVM. So this means you don't have to return the meter for calibration. This is great for home use.
?
?
?
|
There are several videos of Tegam equipment on YouTube? Ing. Patricio A. GrecoTaller Aeronáutico de Reparación 1B-349Organización de Mantenimiento Aeronáutico de la Defensa OMAD-001Gral. Martín Rodríguez 2159San Miguel (1663)Buenos AiresT:?+5411-4455-2557F:?+5411-4032-0072
toggle quoted message
Show quoted text
On 1 May 2025, at 10:17, Razvan Popescu via groups.io <yo8ryr@...> wrote:
? I cannot find the article about the 1830A where Tegam said they used a type IV bridge and added a 6.5 digit multimeter to build the 1830A.1830A is more used in the calibration labs together with the transfer standards they also sell combined with the PS-Cal software or SureCal one which can control the power meter and usually two VNAs for low band and high band to automatically calibrate power sensors and also output the fancy report.I never saw them independently used outside the calibration rack. Tegam sells complete racks for this purpose.When doing the calibration one must take in consideration all components used in the uncertainty formulas. That includes power meter/thermistor mount/VNA/resistive splitter/attenuators...Does anyone have any calibration lab procedure and uncertainty calculation for RF power sensors? I found different white papers that have some formulas that they used but I am interested if anyone has the exact procedure used on daily basis by cal labs.On 01/05/2025 01:19, jmr via groups.io wrote:I've not used any Tegam gear but I think it's reasonable to assume they know what they are doing. I don't think there will be anything exotic about this type of meter because the aim will be to replicate the DC substitution technique used with my old HP 432A meter.
When using the old 432A you first have to measure the resistance of the mount resistance inside the 432A and you have to include the cable in this as well. This is because this acts as a series sense resistor in an IV system. Then it is just a case of zeroing the meter and measuring the voltages at Vrf and Vcomp with the source on and off.
Obviously, having to measure the mount resistance by hand is a major pain because it means unscrewing the thermistor head from the cable each time and measuring the resistance with a decent ohmmeter and then refitting the sensor. My 432A mount resistance has never changed from about 200.2 ohms but I still measure it each time.
I would expect that Tegam will use some form of four wire sensing system that lets them measure the mount resistance and all the voltages automatically with no user input. So there will be nothing to dismantle or measure manually. So it will be a lot quicker to do the DC substitution. The newer N432A will presumably be similar.
So I think it's reasonable to assume that both Tegam and Keysight know how to make a decent >6 digit DVM and four wire IV system that operates at DC. So any subtle differences in accuracy on a datasheet will largely be irrelevant as the uncertainty associated with the 478A thermistor will dominate here.
So it really is up to you if you can live with the clunky user interface of the Tegam unit. I wouldn't want to use it as a regular power meter because of the clunky menu system. At least the N432A has a proper keypad and a decent display.
If you have spotted a used Tegam unit then I guess it could be a bargain as long as it is in good working condition. I'd be concerned about long term support though. Tegam have been associated with acquisitions over the years and were acquired themselves a few years ago. So some of their product ranges could become obsolete at some point if the parent company decide to cut out stuff that doesn't generate much income..
|
On Thu, May 1, 2025 at 02:17 PM, Razvan Popescu wrote:
I cannot find the article about the 1830A where Tegam said they used a type IV bridge and added a 6.5 digit multimeter to build the 1830A.
I'm not sure what is inside it but I assume it will use a pair of auto balancing bridges like the 432A system. I'm also assuming they use some form of 4 wire IV method to measure their own mount resistance.
It could be that the mount resistance is stable enough to not have to measure it. I'm just guessing really :)
?
I mentioned 6.5 digit DVM because Keysight describe their voltage measurement as being equivalent to a 6.5 digit DVM inside the N432A. Tegam will presumably use or claim something similar.
?
With the N432A I think you still have the option of connecting an 8.5 digit DVM externally to the rear of the instrument to do DC substitution the old school way. I suspect only the strictest calibration houses bother with the external 8.5 digit DVM. This probably represents the ultimate accuracy in terms of providing a highly accurate transfer standard.
?
I doubt the Tegam unit supports an external 8.5 digit DVM and this probably separates the two meters in terms of ultimate performance in terms of providing a highly accurate transfer standard. I doubt the Tegam unit can compete here even this is only relevant for critical users at calibration houses. Outside a calibration house this advanced feature of the N432A isn't really relevant.?
|
I cannot find the article about the 1830A where Tegam said they used a type IV bridge and added a 6.5 digit multimeter to build the 1830A.
1830A is more used in the calibration labs together with the transfer standards they also sell combined with the PS-Cal software or SureCal one which can control the power meter and usually two VNAs for low band and high band to automatically calibrate power sensors and also output the fancy report.
I never saw them independently used outside the calibration rack. Tegam sells complete racks for this purpose.
When doing the calibration one must take in consideration all components used in the uncertainty formulas. That includes power meter/thermistor mount/VNA/resistive splitter/attenuators...
Does anyone have any calibration lab procedure and uncertainty calculation for RF power sensors? I found different white papers that have some formulas that they used but I am interested if anyone has the exact procedure used on daily basis by cal labs.
toggle quoted message
Show quoted text
On 01/05/2025 01:19, jmr via groups.io wrote: I've not used any Tegam gear but I think it's reasonable to assume they know what they are doing. I don't think there will be anything exotic about this type of meter because the aim will be to replicate the DC substitution technique used with my old HP 432A meter.
When using the old 432A you first have to measure the resistance of the mount resistance inside the 432A and you have to include the cable in this as well. This is because this acts as a series sense resistor in an IV system. Then it is just a case of zeroing the meter and measuring the voltages at Vrf and Vcomp with the source on and off.
Obviously, having to measure the mount resistance by hand is a major pain because it means unscrewing the thermistor head from the cable each time and measuring the resistance with a decent ohmmeter and then refitting the sensor. My 432A mount resistance has never changed from about 200.2 ohms but I still measure it each time.
I would expect that Tegam will use some form of four wire sensing system that lets them measure the mount resistance and all the voltages automatically with no user input. So there will be nothing to dismantle or measure manually. So it will be a lot quicker to do the DC substitution. The newer N432A will presumably be similar.
So I think it's reasonable to assume that both Tegam and Keysight know how to make a decent >6 digit DVM and four wire IV system that operates at DC. So any subtle differences in accuracy on a datasheet will largely be irrelevant as the uncertainty associated with the 478A thermistor will dominate here.
So it really is up to you if you can live with the clunky user interface of the Tegam unit. I wouldn't want to use it as a regular power meter because of the clunky menu system. At least the N432A has a proper keypad and a decent display.
If you have spotted a used Tegam unit then I guess it could be a bargain as long as it is in good working condition. I'd be concerned about long term support though. Tegam have been associated with acquisitions over the years and were acquired themselves a few years ago. So some of their product ranges could become obsolete at some point if the parent company decide to cut out stuff that doesn't generate much income..
|
I've not used any Tegam gear but I think it's reasonable to assume they know what they are doing. I don't think there will be anything exotic about this type of meter because the aim will be to replicate the DC substitution technique used with my old HP 432A meter.
?
When using the old 432A you first have to measure the resistance of the mount resistance inside the 432A and you have to include the cable in this as well. This is because this acts as a series sense resistor in an IV system. Then it is just a case of zeroing the meter and measuring the voltages at Vrf and Vcomp with the source on and off.
?
Obviously, having to measure the mount resistance by hand is a major pain because it means unscrewing the thermistor head from the cable each time and measuring the resistance with a decent ohmmeter and then refitting the sensor. My 432A mount resistance has never changed from about 200.2 ohms but I still measure it each time.
?
I would expect that Tegam will use some form of four wire sensing system that lets them measure the mount resistance and all the voltages automatically with no user input. So there will be nothing to dismantle or measure manually. So it will be a lot quicker to do the DC substitution. The newer N432A will presumably be similar.
?
So I think it's reasonable to assume that both Tegam and Keysight know how to make a decent >6 digit DVM and four wire IV system that operates at DC. So any subtle differences in accuracy on a datasheet will largely be irrelevant as the uncertainty associated with the 478A thermistor will dominate here.
?
So it really is up to you if you can live with the clunky user interface of the Tegam unit. I wouldn't want to use it as a regular power meter because of the clunky menu system. At least the N432A has a proper keypad and a decent display.?
?
If you have spotted a used Tegam unit then I guess it could be a bargain as long as it is in good working condition. I'd be concerned about long term support though. Tegam have been associated with acquisitions over the years and were acquired themselves a few years ago. So some of their product ranges could become obsolete at some point if the parent company decide to cut out stuff that doesn't generate much income..
?
?
|
On 30/04/2025 22:52, Paul Bicknell via groups.io wrote: Hi Razvan
Hi Paul OK it is frustrating that it seams so difficult to find users of the Tegam 1830 A
Over the next few weeks I would like to get onto the Tegam forum so you might possibly send the link you have to me The link is not existing anymore. It was available on the old Tegam website but it seems that Advanced Energy, the new owner of Tegam, migrated them to the new website. For me the link was not working anyway, probably the forum was already closed when I tried it?! Regarding calibration have you down loaded the user's manual ?? As from that it seems that the user can calibrate the Tegam 1830 A using standard Dc lab standards Advanced Energy provides the manuals and other docs for the 1830A except the firmware. If you manage to get the firmware please let me know. Manual link: Other Technical docs: Procedure to calibrate it is simple if you have the 3 items that they mention in the manual. To initiate the calibration you need to connect to the power meter via ethernet port and there is a menu for calibration and verification in the webUI. Seems to be straight forward. I think I will also have to try and contact a few Calibration houses in the UK
If they are 17025 accredited labs you can check online their accreditation to see what standards they use for the "RF Power Measurements". I checked last year a few cal labs in the EU and none of them had any Tegam stuff. They had Keysight and R&S power meters and sensors and probably they will do a comparison with their calibrated power sensor head in a setup similar to the one in Jeremy's photo. They normally use a software to do that fast and also compute all data in a report and print a new label for you or update the cal table in the EEPROM via software. Check out Cal Lab Solutions, they have a software called PS-Cal that supports the Tegam hardware. Regards Paul
Regards, Razvan
|
Hi Razvan
OK it is frustrating that it seams so difficult to find users of the Tegam 1830 A
Over the next few weeks I would like to get onto the Tegam forum so you might possibly send the link you have to me
Regarding calibration have you down loaded the user's manual ?? As from that it seems that the user can calibrate the Tegam 1830 A using standard Dc lab standards
I think I will also have to try and contact a few Calibration houses in the UK
Regards Paul
toggle quoted message
Show quoted text
-----Original Message----- From: [email protected] [mailto: [email protected]] On Behalf Of Razvan Popescu via groups.io Sent: 29 April 2025 22:34 To: [email protected]Subject: Re: [HP-Agilent-Keysight-equipment] 1 mw - 0dbm lab calibration On 29/04/2025 23:06, Paul Bicknell via groups.io wrote: I have been talking to people today about the Tegam 1830 A and as yet
I cannot find anyone in the UK that has actually sold one or used one I was also looking for someone that used any Tegam products so I can ask more questions about the procedures for the calibration of power sensors and uncertainty calculations but I couldn't find anyone. I wrote to 2 persons that advertised calibration software and services for Tegam but no reply.
I assume those that are using the Tegam 1830 A are talking to each
other either through standards or possibly a user group for Tegam
instruments
Tegam had a forum (community) but for me the link never worked. I couldn't find any other Tegam related forum or group. You will need to make some friends at some Metrology or Calibration Lab but first check their 17025 accreditation, they should specify what hardware they use for RF Power Measurements.
I was told that the Tegam 1830 A 1has been around for about 10 years
and is ?6600
The first dated firmware release for 1930A was in 2011. There were a few other firmware releases before but I couldn't find any release date for those, so I would assume that 1830A is at least from 2011 on the market. Comments please Paul
Regards, Razvan
|
On 29/04/2025 23:06, Paul Bicknell via groups.io wrote: I have been talking to people today about the Tegam 1830 A? and as yet I cannot find anyone in the UK that has actually sold one or used one I was also looking for someone that used any Tegam products so I can ask more questions about the procedures for the calibration of power sensors and uncertainty calculations but I couldn't find anyone. I wrote to 2 persons that advertised calibration software and services for Tegam but no reply. I assume those that are using the Tegam 1830 A? are talking to each other either through standards or possibly a user group for Tegam instruments Tegam had a forum (community) but for me the link never worked. I couldn't find any other Tegam related forum or group. You will need to make some friends at some Metrology or Calibration Lab but first check their 17025 accreditation, they should specify what hardware they use for RF Power Measurements. I was told that the Tegam 1830 A? 1has been around for about 10 years and is ?6600
The first dated firmware release for 1930A was in 2011. There were a few other firmware releases before but I couldn't find any release date for those, so I would assume that 1830A is at least from 2011 on the market. Comments please? Paul
Regards, Razvan
|
I have been talking to people today about the Tegam 1830 A? and as yet I cannot find anyone in the UK that has actually sold one or used one ? I assume those that are using the Tegam 1830 A? are talking to each other either through standards or possibly a user group for Tegam instruments I was told that the Tegam 1830 A? 1has been around for about 10 years and is ?6600 ? Comments please? Paul ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of jmr via groups.io
Sent: 29 April 2025 19:24
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] 1 mw - 0dbm lab calibration ? On Mon, Apr 28, 2025 at 07:27 PM, John Griessen wrote: How about this idea: us the 432A for its DC cal ability, and transfer it to a Boonton 4231A that has GPIB?
In that case, I'd only use the 432A occasionally? Or are you suggesting there's no need for the 432A?
I think many people effectively do just that by being able to check and adjust the 1mW reference on a modern power meter using the 432A and the DC substitution method. However, my setup has that subtle blip in the efficiency at 50 MHz because my 478A sensor has a subtle blip in the VSWR at 50 MHz. So I have to know the efficiency at 50 MHz and how the efficiency changes just above 50 MHz. Having the H7x option would help a lot here but this would restrict the best performance to less than 1 GHz.? I typically only use the 432A about once a year to do a check on the 1mW output on my Anritsu power meter. It's not much use (to me) for anything else although it can be used to level the 83752A sweeper in the range of about 70 MHz to 2GHz. Above this the efficiency of the 478A tails off a bit. One other niggle with the original 432A meter is that the rotary cal factor switch on the front panel only allows 1% step sizes. This is OK most of the time but not so good if you want to connect a DVM to the recorder output at the rear and you want to dial out the efficiency as accurately as possible. None of this was really an issue for the majority of users 50 years ago, but it is a niggle today. The N432A meter and the Tegam meter will be able to offer much finer adjustment for the efficiency at each frequency point.
|
On Sun, Apr 27, 2025 at 11:37 PM, Lynn C wrote:
I'm interested in the specifics on the types of "hobbyist?grade" references jmr mentioned.? There are voltage references in varying grades of accuracy available to a hobby user (from a new 1.5V D battery to IC references and published board designs using these IC references).? Current references for the?hobbyist are a little harder to come by.
?
What are possibilities for RF power references for the hobbyist?? Some power meters have a 0dBm output at 50MHz.? But this reference depends on some previous adjustment for output power, and perhaps frequency.
?
Specific examples of such hobbyist-grade references and where to obtain such references would be helpful.
?
Lynn C.
KA7FPH
?
?
For a fairly advanced reference, I think it would be possible to replicate the 50 MHz oscillator in a typical HP power meter from the 1970s or 1980s. This would be a weekend project rather than something made in under an hour. This would give a very stable output vs time and temperature but it would obviously need to be calibrated on day 1.? So if a calibrator isn't available then this probably isn't a viable option.
?
Another option for lower frequencies is to use something like a 74HC4060 to make a crystal oscillator and take one of the /2 flip-flop outputs and add a series resistor and blocking cap and and some form of L or Pi filter to filter it to a sinewave. Then fit something like a 6dB attenuator to get about 0dBm out. This wouldn't need anything other than a DC voltmeter to check and adjust the 5V supply to calibrate the output. This ought to be able to make something within +/- 0.2dB at a few MHz. With adjustment, it ought to be much better than this if there is something available to calibrate it. This ought to be quite stable vs time. Not as good as an HP reference but still quite good.
?
Otherwise, a decent/fast DAC or DDS could be used if you don't expect or demand high accuracy. There are loads of modern options available today that can outperform the old 74HC4060 for example and the output should be self calibrating (to a reasonable degree) as long as the DAC reference can be measured and adjusted.
|
Another missvalued instrument is HP434A… I connected it to an external voltmeter and workd very good as direct transfer to DC to RF …?
Ing. Patricio A. Greco
Taller Aeronáutico de Reparación 1B-349
Organización de Mantenimiento?Aeronáutico de la Defensa OMAD-001
Laboratorio de Calibración ISO 17025?AREA: RF/MW?
Gral. Martín Rodríguez 2159
San Miguel (1663)
Buenos Aires
T: +5411-4455-2557
F: +5411-4032-0072
toggle quoted message
Show quoted text
On 29 Apr 2025, at 15:23, jmr via < jmrhzu@...> wrote:
?Boonton 4231A
|
On Mon, Apr 28, 2025 at 07:27 PM, John Griessen wrote:
How about this idea: us the 432A for its DC cal ability, and transfer it to a Boonton 4231A that has GPIB?
In that case, I'd only use the 432A occasionally? Or are you suggesting there's no need for the 432A?
I think many people effectively do just that by being able to check and adjust the 1mW reference on a modern power meter using the 432A and the DC substitution method. However, my setup has that subtle blip in the efficiency at 50 MHz because my 478A sensor has a subtle blip in the VSWR at 50 MHz. So I have to know the efficiency at 50 MHz and how the efficiency changes just above 50 MHz. Having the H7x option would help a lot here but this would restrict the best performance to less than 1 GHz.?
?
I typically only use the 432A about once a year to do a check on the 1mW output on my Anritsu power meter. It's not much use (to me) for anything else although it can be used to level the 83752A sweeper in the range of about 70 MHz to 2GHz. Above this the efficiency of the 478A tails off a bit.
?
One other niggle with the original 432A meter is that the rotary cal factor switch on the front panel only allows 1% step sizes. This is OK most of the time but not so good if you want to connect a DVM to the recorder output at the rear and you want to dial out the efficiency as accurately as possible. None of this was really an issue for the majority of users 50 years ago, but it is a niggle today. The N432A meter and the Tegam meter will be able to offer much finer adjustment for the efficiency at each frequency point.
|
On 4/27/25 16:02, jmr via groups.io wrote: You only have to press a toggle switch to zero the 432A but you will soon learn that you will be pressing this toggle switch a lot on the sensitive ranges. This quickly becomes very tiresome... On 4/27/25 16:02, jmr via groups.io wrote: all in all the 432A is really only suitable (in 2025) for metrology based measurements over a limited dynamic range. How about this idea: us the 432A for its DC cal ability, and transfer it to a Boonton 4231A that has GPIB? In that case, I'd only use the 432A occasionally? Or are you suggesting there's no need for the 432A?
|
Absolutely.
I suffered the 432A through my PhD days and the repeated zeroing was a complete pain- but fine training in persistence for a “gradual” student.
Regards,
Alwyn
_____________________________________________________
Alwyn Seeds, Director
SynOptika Ltd.,
114 Beaufort Street,
London,
SW3 6BU,
England.
SynOptika Ltd., Registered in England and Wales: No. 04606737
Registered Office: 114 Beaufort Street, London, SW3 6BU, United Kingdom. _____________________________________________________
|
Both a te great. 8478B goes up 18GHz it has a better connector and a more precise thermistor mount.?
Ing. Patricio A. Greco
Taller Aeronáutico de Reparación 1B-349
Organización de Mantenimiento?Aeronáutico de la Defensa OMAD-001
Laboratorio de Calibración ISO 17025?AREA: RF/MW?
Gral. Martín Rodríguez 2159
San Miguel (1663)
Buenos Aires
T: +5411-4455-2557
F: +5411-4032-0072
toggle quoted message
Show quoted text
On 28 Apr 2025, at 06:55, Harke Smits via < yrrah@...> wrote:
I have a couple of HP8478 sensors, no 478A. Is there any reason
why the HP478A is advised/preferred and not the HP8478? 73 de Harke
On 28/04/2025 00:02, jmr via
wrote:
In case anyone is smitten with the concept of the HP 432A and
478A, it's probably worth mentioning that there are very good
reasons why we are not all using 432A power meters and 478A
sensors instead of the more modern thermocouple and diode sensor
based meters.
?
Just spend half an hour using one on a bench where you want
to measure RF power across the range of the meter. A few things
will then become apparent and you will probably want to walk
away from it :)
?
First of all the 478A sensors are very fragile, the max level
is only 30mW. I'm not sure what the damage level is, but it
doesn't take much to damage one or age it so it can't zero
anymore.
?
The stability of the zero is quite poor on the more sensitive
ranges. I also have an older 431C meter and this is even worse
to use as the zero controls are manual rotary controls. You only
have to press a toggle switch to zero the 432A but you will soon
learn that you will be pressing this toggle switch a lot on the
sensitive ranges. This quickly becomes very tiresome...
?
The ranging is manual via a rotary switch that shifts in 5dB
steps. This is another clue about the limited range of this
system as you might expect to see 10dB range steps here.
?
Touch or hold the 478A sensor on the most sensitive range
(-20dBm FSD) and you might be forgiven that something is faulty
because the zero will drift around a lot. Enough to make it seem
like something is wrong. This is in stark contrast to the modern
thermocouple meters.
?
The big analogue dial looks lovely but it contributes to
uncertainty. There is a B version with a basic digital display
but these don't seem very common and I think it only reads out
in mW rather than dBm. I don't think I've ever used one so I
can't be certain.
I usually use my 432A with a DVM connected to the recorder
port at the back but this can be a pain to manage due to the 5dB
steps and the need to convert volts to dBm using a different
equation for each manual range.
?
There's no GPIB port at the rear so to log it you have to do
it via a GPIB enabled DVM connected to the recorder BNC at the
rear of the instrument. You have to manually tell the system
each time you change the front panel range control because the
DVM won't know you have done this. This is fairly grim if you
want to measure over more than about a 10dB range.
?
So all in all the 432A is really only suitable (in 2025) for
metrology based measurements over a limited dynamic range. My
431C meter and 478A sensor used to be used at work to calibrate
the noise marker on the company's HP 8568B spectrum analyser.
The idea was to use a precision (flat) noise source and pass it
through a 70 MHz BPF that had good return loss and that had been
calibrated at NPL for its noise bandwidth in dBHz. By measuring
the noise power with the 431C and the 478A we could check the
accuracy of the HP8568B noise marker when doing BER tests at
various signal to noise levels. The noise source has precision
0.1dB and 1dB step attenuators made by Telonic. This was all
about 35 years ago.
?
I now own both the power meter and the HP 8568B spectrum
analyser as they both became obsolete many years ago at work :)?
?
?
?
|
I have a couple of HP8478 sensors, no 478A. Is there any reason
why the HP478A is advised/preferred and not the HP8478?
73 de Harke
On 28/04/2025 00:02, jmr via groups.io
wrote:
toggle quoted message
Show quoted text
In case anyone is smitten with the concept of the HP 432A and
478A, it's probably worth mentioning that there are very good
reasons why we are not all using 432A power meters and 478A
sensors instead of the more modern thermocouple and diode sensor
based meters.
?
Just spend half an hour using one on a bench where you want
to measure RF power across the range of the meter. A few things
will then become apparent and you will probably want to walk
away from it :)
?
First of all the 478A sensors are very fragile, the max level
is only 30mW. I'm not sure what the damage level is, but it
doesn't take much to damage one or age it so it can't zero
anymore.
?
The stability of the zero is quite poor on the more sensitive
ranges. I also have an older 431C meter and this is even worse
to use as the zero controls are manual rotary controls. You only
have to press a toggle switch to zero the 432A but you will soon
learn that you will be pressing this toggle switch a lot on the
sensitive ranges. This quickly becomes very tiresome...
?
The ranging is manual via a rotary switch that shifts in 5dB
steps. This is another clue about the limited range of this
system as you might expect to see 10dB range steps here.
?
Touch or hold the 478A sensor on the most sensitive range
(-20dBm FSD) and you might be forgiven that something is faulty
because the zero will drift around a lot. Enough to make it seem
like something is wrong. This is in stark contrast to the modern
thermocouple meters.
?
The big analogue dial looks lovely but it contributes to
uncertainty. There is a B version with a basic digital display
but these don't seem very common and I think it only reads out
in mW rather than dBm. I don't think I've ever used one so I
can't be certain.
I usually use my 432A with a DVM connected to the recorder
port at the back but this can be a pain to manage due to the 5dB
steps and the need to convert volts to dBm using a different
equation for each manual range.
?
There's no GPIB port at the rear so to log it you have to do
it via a GPIB enabled DVM connected to the recorder BNC at the
rear of the instrument. You have to manually tell the system
each time you change the front panel range control because the
DVM won't know you have done this. This is fairly grim if you
want to measure over more than about a 10dB range.
?
So all in all the 432A is really only suitable (in 2025) for
metrology based measurements over a limited dynamic range. My
431C meter and 478A sensor used to be used at work to calibrate
the noise marker on the company's HP 8568B spectrum analyser.
The idea was to use a precision (flat) noise source and pass it
through a 70 MHz BPF that had good return loss and that had been
calibrated at NPL for its noise bandwidth in dBHz. By measuring
the noise power with the 431C and the 478A we could check the
accuracy of the HP8568B noise marker when doing BER tests at
various signal to noise levels. The noise source has precision
0.1dB and 1dB step attenuators made by Telonic. This was all
about 35 years ago.
?
I now own both the power meter and the HP 8568B spectrum
analyser as they both became obsolete many years ago at work :)?
?
?
?
|
Thanks! Yes, it is possible to use the 432A to test attenuators but it's probably best to limit this to small value attenuators below about 6-10dB. I have used it to measure a 20dB attenuator but this doesn't inspire a lot of confidence even if I get a good result. This is because of the need to change ranges and zero the meter again. It really is a clunky process compared to doing the same thing with a modern meter.
?
With care, it is possible to make precision RF attenuators that are good through VHF by using hand selected chip resistors. This can be used to check the linearity of the meter at home if there is nothing else available to do this.
?
I've made precision 10dB and 20dB attenuators in the past to help test my old HP 8405A VVM. In each case I hand selected the resistors using a 4 wire measurement and selected the resistors that gave the best return loss whilst also giving an accurate 20dB insertion loss. It's possible to make something that would be very difficult to beat in terms of VSWR and accuracy if you opted to buy a commercial attenuator instead. I tested across about 50-100 resistors to find the ones that gave the best performance on the 4 wire ohm meter. This doesn't really take very long.
?
The end result was an attenuator that measured 49.95 ohms when terminated in a precision 50 ohm load and the attenuation measured at DC using a 6.5 digit DMM was within 0.005dB of 20dB. I still have some old plots of one of the attenuators as below. Check out the VSWR and the flatness to 100 MHz. It's asking a lot of the VNA to measure the attenuator to a low level of uncertainty but I'm quite happy that I have made a precision attenuator that would be very difficult (and very expensive) to match if I tried to buy a commercial equivalent.
?
It is possible to make some really good test gear suitable for 'home metrology' and the attenuator below is another example. It can be used to check a power meter sensor from -20dBm to 0dBm or from -15dBm to +5 dBm or from -10dBm to +10dBm for example. It can't compete with Keysight or NIST but for home use it works really well as long as the source match is very good and the load match is something like a power sensor or VNA that has very low VSWR as well. This minimises mismatch uncertainty when the attenuator is fitted inline. At frequencies below about 100 MHz, the attenuator is typically going to have lower VSWR than virtually any sensor made by Keysight or R&S etc.
?
?
?
|
Patricio A. Greco
Paul Bicknell