Purchasing HP 432A (was: eBay HP 478A thermistors. Any good?)

[tatel]

Long story short: I want to calibrate/adjust a bunch of vintage CROs, because I think repair is a good way to get skill in electronics, but over all because I have fun from it.

So I purchased a vintage military version of HP 8640B and, of course, now i want to check output amplitude, levelling, etc. Considering now to get a power meter. It could be either a 432A or one of those that work with 848x series sensors, probably a 435A or 436A. I'm leaning towards one of those since they cover the 20dBm output range of signal generator and I think they are probably good enough to adjust a vintage o'scope

However, before pulling the trigger, I see that 478A thermistor sensors are easily found on eBay and even cheaper. Moreover, it looks 432A would be more precise and easier to calibrate/adjust, so I think that could be advantageous. But I wonder if these eBay sensors are any good. See https://www.ebay.es/itm/166500473128

Any pointer/hint, not just about the sensors but about using a signal generator to adjust an old CRO, would be greatly appreciated. Apart from that signal generator, I just have a TinySA and some cheap chinese attenuators/adaptors/dummy loads

[factory]

Listing states it's working, therefore if it isn't any good you should be able to return it, note the sensors (all types) are very easily damaged and often advertised as untested for this reason. You will probably notice that there are many more meters out there, than working sensors.

David

[G0HZU]

This is just my opinion but I think the HP432A (with the 478A thermistor sensor) is now best used for metrology purposes rather than as a general purpose power meter. The stock HP478A sensor is rated from 10MHz to about 10GHz but in my experience the input VSWR is poor below about 30MHz and poor above a few GHz. It is best suited for metrology in the range of about 50MHz to about 1GHz in this respect. I typically only use mine for checking the 1mW reference from my Anritsu power meter using DC substitution and a decent Keithley DVM.

The 478A sensor is also very fragile and the sensor quotes 30mW as the maximum input power level. This means it could be damaged with the 8640B sig gen (in theory at least). One mistake and it will be cooked and then it's game over.

The thermocouple sensor based meters that replaced the 432A are more suited to general RF benchwork. They are easier to use, have a more stable zero performance vs time and temperature and it's possible to get sensors that work down to about 100kHz. They can also withstand 20dBm input power levels without damage although I would recommend using them up to about 15dBm maximum if you want the best performance in terms of sensor linearity.

For checking old CR scopes for frequency response you ideally want to use a sig gen with low harmonic distortion. The HP 8640B should be generally good here as long as you keep the power output fairly low. I haven't used one for many years but I'd expect it to be good (in terms of low distortion) at levels up to 0dBm.

The ideal setup would involve a good sig gen and an external levelling amplifier. I made my own levelling amplifier a few years ago as described here:

https://www.eevblog.com/forum/rf-microwave/homebrew-rf-levelling-amplifier/

The benefit is accurate levelling and a well defined 50R source impedance. This is probably too good for just checking a scope for frequency response though.

One issue with using the 8640B will be tuning it across the frequency range of the scope in octave ranges. This could become quite tedious to do by hand. Modern synthesised sig gens can do this via remote control or via the front panel with a rotary control. I'd recommend something like the Agilent ESG series as these use a BFO/HET system up to about 250MHz. The harmonics are usually much lower on the BFO/HET range as long as the power is kept below about 0dBm. The IFR 341x series of sig gens have a BFO/HET range up to about 370MHz and typically produce low harmonic distortion up to this frequency as long as the output power level is kept below about 0dBm. The IFR sig gens tend to be very expensive though.

The analogue versions of the Agilent ESG sig gens sometimes appear at quite low prices. Probably not much more than a healthy 8640B sig gen especially for the 1GHz variants. The RF levelling is typically excellent with the ESG sig gens as well. The harmonic distortion does degrade at frequencies above the BFO/HET range but not by that much. An external LPF followed by a 10dB attenuator could be used on these ranges.

The really cheap way to check the 8640B for flatness would be to use a decent low barrier Schottky diode detector with its own internal 50R termination. This won't be as good as a HP power meter, but it should be good enough for the stuff you are doing.

[Tony_G]

Thought I might just add some data to what Jeremy was saying - Here are some images that show the cal section for one of my 478A's:




In comparison, here is the VSWR table for one of my 8481A sensors:



You can see that the passing spec for the 478A at higher frequencies is larger than the 8400 series sensor.

Hope this is interesting to you.

TonyG

[radar_macgyver]

Anecdotes aren't evidence, but I got my hands on a number of thermistors from a lab that was shutting down. Every single one was bad - there were about 15 sensors, with both type-N and WR-28 inputs.

By contrast, I also received from the same lab a couple of 8482s, including one W8482 (WR-10!) and they're all in good shape. The 435s, especially the B suffix, are much nicer instruments to use than the 432s, and both can be found with HPIB ports that work with the calibration processes and external leveling of some of the HP synthesizers of the 8673 vintage (though I haven't done this myself, I've seen it mentioned in the user manuals).

[alm]

My anecdote is that a couple of years ago I bought 2x HP 478A from eBay sold as working, and both appeared to work and null fine, though I haven't tested them over frequency. My impression is that later these devices were treated a bit more gently because they were used for special measurements, while the thermistor and diode sensors were used for every day measurements.

[tatel]

Well, thank you all for your answers. Much appreciated.

Are there any questions convenient to ask to the seller before buying a 478A sensor? I mean, from threads similar to this, I learned to look for the possibility of returns and for the availability of sensor calibration chart. Those 478As from eBay can be returned, but should I ask something about them? Having to return it would be annoying.

Another thing would be stability/calibration of the power meter itself. Again, from other threads, I see that 432A could not need anything done to it for years and years. The one I'm looking at seems to have been last calibrated in 2011. Not sure about 435A and 436A candidates. Not having to be worried about power meter would be lovely. But perhaps I shouldn't be worried about it with 435/436, either? Any opinions about that?

432 and 435 doesn't have HP-IB, at least that I could see. However I see it's recommended to connect recorder output to some DMM/DVM. I have HP3456, which does have HP-IB. So I guess I could use it not just for measuring, but also for logging to some PC. Is that right?

This is just my opinion but I think the HP432A (with the 478A thermistor sensor) is now best used for metrology purposes rather than as a general purpose power meter. The stock HP478A sensor is rated from 10MHz to about 10GHz but in my experience the input VSWR is poor below about 30MHz and poor above a few GHz. It is best suited for metrology in the range of about 50MHz to about 1GHz in this respect. I typically only use mine for checking the 1mW reference from my Anritsu power meter using DC substitution and a decent Keithley DVM.

I would be using it in the 20-300 MHz range. 500 MHz at most since that's what my AN/USM-323 (military version of 8640B) could output. So 432A would be OK for the frequencies I'm going to be on, I guess.

The 478A sensor is also very fragile and the sensor quotes 30mW as the maximum input power level. This means it could be damaged with the 8640B sig gen (in theory at least). One mistake and it will be cooked and then it's game over.

Yeah, that worries me. As a total noob, I don't know what would be best. On one hand, as I understand it, using about 5Vpp would be the better thing to adjust these CROs. OTOH, power meter is to be used to check sig gen, not the CROs. Perhaps some good attenuator could be used. Any hints about that?

The thermocouple sensor based meters that replaced the 432A are more suited to general RF benchwork. They are easier to use, have a more stable zero performance vs time and temperature and it's possible to get sensors that work down to about 100kHz. They can also withstand 20dBm input power levels without damage although I would recommend using them up to about 15dBm maximum if you want the best performance in terms of sensor linearity.

Understood. I think 15dBm could be enough to adjust the CROs. Perhaps one could go for a sensor able to go higher power, to have good linearity at 19 dBm -which is the limit of AN/USM-323. Any ideas?

For checking old CR scopes for frequency response you ideally want to use a sig gen with low harmonic distortion. The HP 8640B should be generally good here as long as you keep the power output fairly low. I haven't used one for many years but I'd expect it to be good (in terms of low distortion) at levels up to 0dBm.

I'm putting a couple screenshots from AN/USM-323 manual as attachments, with the specs for harmonics, output level accuracy, flatness and VSWR. How does it look?

The ideal setup would involve a good sig gen and an external levelling amplifier. I made my own levelling amplifier a few years ago as described here:

https://www.eevblog.com/forum/rf-microwave/homebrew-rf-levelling-amplifier/

Yeah, I was reading that thread two days ago. Very interesting. As I understand, it implies to have, among other things, a power meter, a DVM at recorder output, and a sig gen able to go very low in frequency. I guess power meter 435 would be adequate for that. HP3456A would have to be at recorder output. Not sure about sig gen, thoug. It looks the lowest frequency would be 500 kHz.

The benefit is accurate levelling and a well defined 50R source impedance. This is probably too good for just checking a scope for frequency response though.

8640B was recommended to me precisely for levelling and low harmonics. Not really sure it wast the best recommendation, and I got the military version. Since I already got it, I'll have to try and get the best from it

One issue with using the 8640B will be tuning it across the frequency range of the scope in octave ranges. This could become quite tedious to do by hand. Modern synthesised sig gens can do this via remote control or via the front panel with a rotary control. I'd recommend something like the Agilent ESG series as these use a BFO/HET system up to about 250MHz. The harmonics are usually much lower on the BFO/HET range as long as the power is kept below about 0dBm. The IFR 341x series of sig gens have a BFO/HET range up to about 370MHz and typically produce low harmonic distortion up to this frequency as long as the output power level is kept below about 0dBm. The IFR sig gens tend to be very expensive though.

I want this able to go up to 300 MHz at least. I got for peanuts a quite beaten and partially scavenged (knobs) Tektronix 485. I'm hoping I could resurrect it as a kind of Franketronix. Anyway, even if I fail to resurrect it, I'm pretty sure I will get quite a bit of fun, and learn a couple things BTW.

My upgraded Rigol AWG isn't good over 60 MHz so I hope to use this thing with my most used 100 Mhz DSO. I'm also interested in long range RC. The probability for me getting more of an oscilloscope than a 500 MHz one is near zero, I think. So AN/USM-323 seemed to fit the bill pretty well, and it was relatively cheap. I don't think tuning across frequency ranges will be so tedious to me, but I'll have to live with it anyway.

The really cheap way to check the 8640B for flatness would be to use a decent low barrier Schottky diode detector with its own internal 50R termination. This won't be as good as a HP power meter, but it should be good enough for the stuff you are doing.

Well, I still have at least some days before the time to pull the trigger. I'm leaning to 435A since it would the cheapest power meter I could get now. Even on a budget, I don't mind getting quality tools if possible.

Again, thank you all

[alm]

By contrast, I also received from the same lab a couple of 8482s, including one W8482 (WR-10!) and they're all in good shape. The 435s, especially the B suffix, are much nicer instruments to use than the 432s, and both can be found with HPIB ports that work with the calibration processes and external leveling of some of the HP synthesizers of the 8673 vintage (though I haven't done this myself, I've seen it mentioned in the user manuals).

Are you saying the HP 435 can be found with HP-IB ports? Are you sure you are not confused with the HP 436A or 437B? As far as I know the 435 is a purely analog device.

Are there any questions convenient to ask to the seller before buying a 478A sensor? I mean, from threads similar to this, I learned to look for the possibility of returns and for the availability of sensor calibration chart. Those 478As from eBay can be returned, but should I ask something about them? Having to return it would be annoying.

I'm certainly no expert, but other than hooking them up to a 432 and seeing if they null and measure the reference of a 435/436/437/438, probably return loss, and maybe DC resistance?

Another thing would be stability/calibration of the power meter itself. Again, from other threads, I see that 432A could not need anything done to it for years and years. The one I'm looking at seems to have been last calibrated in 2011. Not sure about 435A and 436A candidates. Not having to be worried about power meter would be lovely. But perhaps I shouldn't be worried about it with 435/436, either? Any opinions about that?

HP 432A adjustment can be done with purely DC voltages, if I remember correctly. 435 and later models basically need a 432 or other RF reference to adjust.

432 and 435 doesn't have HP-IB, at least that I could see. However I see it's recommended to connect recorder output to some DMM/DVM. I have HP3456, which does have HP-IB. So I guess I could use it not just for measuring, but also for logging to some PC. Is that right?

Yes, you could definitely hook up a DMM and record that way. It's actually recommended in the 432A manual for increased accuracy. It obviously wouldn't allow you to remotely control the range or anything else.

[radar_macgyver]

Are you saying the HP 435 can be found with HP-IB ports? Are you sure you are not confused with the HP 436A or 437B? As far as I know the 435 is a purely analog device.

Oops, yes, 436A was the one I was thinking of.

[jfet]

I have worked in RF (satellite and microwave) most of my life and the 478/432 has been the work horse since the 1970s. 

The thermistors are sensitive to shock and burn out.  If you drop one, you can kill it.  If you overload it, it can burn out extremely fast.

I have two sets of meters and thermistors.  I keep a 10 db attenuator in the thermistor head and you learn to measure on the highest range to make sure your not going to burn it out.  Then remove the attenuator when you are sure.  That attenuator is much cheaper to replace, and its cheap insurance.

I knew a tech who kept bubble wrap around his heads and he slid them up when using them and then covered the head when he was storing them.  He used rubber bands.

The worse problem, is the cable between the meter and the thermistor head.  HP used epoxy on the back of the connectors and its almost impossible to repair them.  Go search for a replacement cable and you will find they are rare and very expensive.

[cncjerry]

This is for sale locally on Craigs List.  Though I have two 438a, I think that is the number, dual channel power meters, plus a module in my MMS, I've not run into the thermistor mount as he calls it, on this model. That would be the 478a mentioned in this thread, right?  All my models take the 8xxxx series sensors, of which I have several,  I understand the 478a to be more accurate as it measures true heat power,  I guess.  But for academic reasons, do you need a special mount for the 478a?  Does the 478a attach to most of the power meters in the line (think those that take the 8xxx sensors) starting with the analog meters as pictured? I understand all power sensors on ebay to be very risky.  All my sensors came tested locally.  This meter, though priced at 150, seems like a really nice copy and the seller seems more knowledgeable than the potential buyers.

Thanks, this is an interesting thread.

Jerry 

[alm]

This is for sale locally on Craigs List.  Though I have two 438a, I think that is the number, dual channel power meters, plus a module in my MMS, I've not run into the thermistor mount as he calls it, on this model. That would be the 478a mentioned in this thread, right?  All my models take the 8xxxx series sensors, of which I have several,  I understand the 478a to be more accurate as it measures true heat power,  I guess.  But for academic reasons, do you need a special mount for the 478a?  Does the 478a attach to most of the power meters in the line (think those that take the 8xxx sensors) starting with the analog meters as pictured?
Jerry 

There were too different lines of power meters with incompatible cables and sensors. One was the thermistor type with the 431 and 432 series power meters and 478, 8478 and 486 series sensors. And the second, later series starts with the 435 and goes own from there, with 848X and later series sensors that were either thermocouple or diode based. There were also some older monolithic power meters. The 848X and later sensors have the advantage that they are more robust, drift less with temperature, have a wider dynamic range and respond faster than the thermocouple sensors, but the thermocouple sensors have the unique advantage that they can be calibrated against a DC voltage, while the 848X sensors need an accurate RF reference (kind of a chicken-and-egg problem). For example traditionally (a special version of) the 478A was used to tune the 50 MHz reference of 435 and later series power meters. So the 848X sensors are better in most aspects, but the 478A has advantages for some specific applications. But they are not compatible with each other. See HP AN 64-1A: Fundamentals of RF and Microwave Power Measurements about the different types of sensors.

[tatel]

Well, I got 432A power meter. Now I'll have to look for good 478A thermistor and attenuator.

I see 355D is the recommended one. However at this moment I'm only interested in work at 10-19dBm range so I guess I could go for a simpler, fixed value one. Unless I'm missing something, of course. Any suggestions?

Edit: also got the sensor linked in OP

Edit2: looking at 8491A, available in 3, 6 and 20 dB attenuation. I guess I should get at least 6dB to be safe?

[Tony_G]

I must admit that if you're playing in that around 20 dBm range then you should really look to a range of attenuators and include a 20 dB one.

It's generally safer I think to start with high attenuation and then move down rather than risk starting low and burning out the sensor.

That said, I use my 432/478 solely to set the references on 84xx compatible meters, so I defer to those who use them in broader scenarios.

TonyG

[tatel]

Well, it all has been delivered.

However, I can't plug the cable into 432a thermistor mount jack. Jack has a protuberance so the plug doesn't fit. Please see attached pictures.

I wasn't expecting this. Any hint would be greatly appreciated

[alm]

Looks to me like the cable is the problem. Here's what my cable looks like. Though it looks fixable with some grinding / filing.

[Tony_G]

My 432A has the same connector as yours:



However, my cable (connecting the 432 to the 478) has a notch in it for the connector:



Not sure where your cable came from.

EDIT: I checked my 431C and it also has the same connector as the 432 - However, I went to see what the 431B has and, while I couldn't get a good picture, this one seems to show the connector without the tab in it so maybe the cable is from a 431B device (which also used the 478):



Can't really tell though

TonyG

[tatel]

I can only assume that cable wasn't for connection to 432A. Perhaps it's for some other model (431 perhaps)

Do you guys think it would be safe to make that notch in the connector?

[Tony_G]

The 432 schematic seems to imply that it is a straight-through cable so it should be but I would recommend checking the schematics against the cable to be sure before turning anything on:

432:



478:



TonyG

[tatel]

Yeah, it seems to be for 431.

Well, it seems I found the limits of the forum. Of course, the probability of some guy being here, having both 431 and 432, is probably next to zero. I have never seen the slightest mention of cables for 431 and 432 being different. So, here it remains this piece of information. Hope it will useful for somebody in the future.

I already ordered another cable, this one no only has the notch, but also has golden pins

Best wishes

[alm]

I haven't used the 431, but I know the HP 431 series used AC excitation for the bridge instead of DC (see HP AN 64-1), so I could imagine the cable being different, though I'm sure what the difference would be.

[tatel]

Yep, evidently you are at least a couple levels above me in the food chain

Well, you know "Take the experience with a smile"

[Tony_G]

...the probability of some guy being here, having both 431 and 432, is probably next to zero.

I actually do have a 432A, 432B, and a 431C and they all have the tab connector.

What I don't have is a 431B, which is the one that doesn't seem to have the tab - I expect that the cables are both the same and they just added  the notch due to customer feedback about getting the cable seated correctly.

Anyway, a new cable will solve the problem and you can confirm if your "spare" will work by comparison now.

Glad you solved the problem.

TonyG

[tatel]

Then it looks as if only 431 in marks A and B got that no-notch-connector. Since very few people seems to have 431 and most (all?) of them have the C one, no wonders this notchy thing passed under the radar. Seller is willing to give a partial refund so no problem here. New cable looks in pretty good condition and shows the right reference number from the parts list, HP 8120-1082, 5-foot cable for HP432A.

I did catch some 432 models have a sticker warning they are prepared for different, longer cables, and cable long must match what the sticker says for it to work OK. Mentioning it here, just in case.

So, when new cable arrives, I could finally check if both power meter and thermistor work fine. I have seen some video in eBay listings where 432A needle goes full scale when moving resistance lever from 100 to 200 ohm. On mine needle barely moves to near 0. Should I be born in Finisterre, by now I for sure would be consulting some witch-wizard to know who jinxed me... Anyway that seller is willing to replace that device if needed so no problem either.

Changing thread title so it agrees more with what this is becoming, how to purchase a 432A

[G0HZU]

...the probability of some guy being here, having both 431 and 432, is probably next to zero.

I've also got a 431C power meter and two 432A power meters and two 478A sensors.  One of the 432A power meters is as new and unused and has the original HP seals and stickers on it. The 478A sensor is still factory (vacuum) sealed in its packaging and so is the sensor cable and the manuals. It's never been unpacked and used. It must have been stored from new and forgotten about.