List your test equipment "scores" here!

[alm]

I've had close to a dozen HP 8484A sensors from eBay over the last 5 years pass over my workbench and have had pretty good luck.  I can't remember receiving any from eBay's "used" category being bad.  I've also bought a few listed under "parts/repair" as unknown that also checked out OK.

The 8484A is a diode sensor. The thermocouple sensors like 8481A and 8482A may very well be different. Despite their higher maximum power levels, they appear to die more frequently. Probably because they are more sensitive to overloads. And there appears to be almost no chance of repairing them given the lack of spare parts and special skills available if the thermocouple is blown.

[rastro]

I've had close to a dozen HP 8484A sensors from eBay over the last 5 years pass over my workbench and have had pretty good luck.  I can't remember receiving any from eBay's "used" category being bad.  I've also bought a few listed under "parts/repair" as unknown that also checked out OK.

The 8484A is a diode sensor. The thermocouple sensors like 8481A and 8482A may very well be different. Despite their higher maximum power levels, they appear to die more frequently. Probably because they are more sensitive to overloads. And there appears to be almost no chance of repairing them given the lack of spare parts and special skills available if the thermocouple is blown.

That's a good point.  I think the thermocouple is also more sensitive to shock.

I would agree that the 8484A are pretty sturdy.  Even though they have a max of -20dBm; I believe they can tolerate +30dBm for short periods.  I probably wouldn't deliberately chance it. 

You bring up a good point regarding the themocouple sensor being a bigger risk.  I can't speak for the 8481/2A's however I have had at least three 8481D's from eBay without any trouble - the "D" version might be diode sensors also. 

rastro

[Neomys Sapiens]

After fixing the cracked plastic clip that holds the sensor cartage wires to the PCB it zeroed and calibrated fine.  Yes you would need to have it re-characterized if you move or mess with the sensor cartage. 


You may also lean toward the 437 since you can store the calibration coefficients in a frequency lookup table - some added convenience.

What is the assumed meaning of 'cartage'?

[orin]

I've had close to a dozen HP 8484A sensors from eBay over the last 5 years pass over my workbench and have had pretty good luck.  I can't remember receiving any from eBay's "used" category being bad.  I've also bought a few listed under "parts/repair" as unknown that also checked out OK.

The 8484A is a diode sensor. The thermocouple sensors like 8481A and 8482A may very well be different. Despite their higher maximum power levels, they appear to die more frequently. Probably because they are more sensitive to overloads. And there appears to be almost no chance of repairing them given the lack of spare parts and special skills available if the thermocouple is blown.

There is a company on ebay that will fix 8481As or sell you a fixed one for around $450.

Expect any 'untested' 8481A or 8482A to be broken.  Even those that zero may be compromised.  I've returned two 8481As - the first because it failed on SWR though it would zero and measure power, the second because although it would zero, it didn't actually measure power.  I did however manage to get a working 8482A.  I have an 8484A as well, but it tends to drift.

[cncjerry]

I thought the 'D' power sensor models were the diode type?  for instance 8481d, 8482d, etc.

Thanks for all the input.  This is all complicated but I'm starting to understand what I need.

[cdev]

If someone says "for parts or best offer" and then you ask them if it powers up and they tell you it does, but then it doesn't power up when you get it, instead it blows a fuse.. who is right? What to do?

[HighVoltage]

If someone says "for parts or best offer" and then you ask them if it powers up and they tell you it does but they cant test it further but then it doesn't power up when you get it, instead it blows a fuse.. who is right? What to do?

Then you make a claim with the seller and tell him it is not as described and make arrangements for a lower price (partial refund) or return it for a full refund. Most sellers will agree to a partial refund in my experience.

[nctnico]

If someone says "for parts or best offer" and then you ask them if it powers up and they tell you it does but they cant test it further but then it doesn't power up when you get it, instead it blows a fuse.. who is right? What to do?

Then you make a claim with the seller and tell him it is not as described and make arrangements for a lower price (partial refund) or return it for a full refund. Most sellers will agree to a partial refund in my experience.

Then you are lucky. 'For parts' means what it means. Unless the description says 'powers up' ofcourse. However 'powers up' doesn't always have to be a good thing. If it powers up then there is usually a more complicated error. A power supply is easy to fix.

[rastro]

After fixing the cracked plastic clip that holds the sensor cartage wires to the PCB it zeroed and calibrated fine.  Yes you would need to have it re-characterized if you move or mess with the sensor cartage. 


You may also lean toward the 437 since you can store the calibration coefficients in a frequency lookup table - some added convenience.

What is the assumed meaning of 'cartage'?

Sorry for the spelling error.  I meant to say 'cartridge'.  The front end after the RF connector that holds the actual sensor/diode in the case of the HP 8484A. 

rastro

[rastro]

I thought the 'D' power sensor models were the diode type?  for instance 8481d, 8482d, etc.

Thanks for all the input.  This is all complicated but I'm starting to understand what I need.

The 8484A is diode sensor; so their is at least one 'A' model that is an exception to a "D" naming convention.  It's my understanding that thermocouple sensors don't have sensitivity to reliably measure below -30dBm.  So if the lowest range on the datasheet is significantly below -30dBm (e.g. -70dBm) it's probably a diode sensor.   However, to confuse the issue more, there are some HP sensors that use both both diode and thermocouple to get a much wider range.  I believe this is only on newer models (E-series) that have the correction coefficients stored internally in each sensor.  I don't think these sensors are backward compatible with the HP 436/7/8 power meters.

As a side note the 8484A and 8481D seem almost identical except that the 8481D is smaller.  Maybe some one knows what the differences are?  Again I've had pretty good luck on both models but as they say "your mileage may vary"....

rastro

[cncjerry]

thanks again.  still debating.  Now I'm thinking I want a 438 instead of a 437.  I found sensors in the 150 range, cable in the $50 range and I can get a 438 for around 150 or less.  So for 350 I can get one channel configured and then work on the second at a later date.  The power sensors look like the are very band and range specific.  If doesn't look like many have more than about 50db of range.  So it looks like you have to have a stack of attenuators.

The problem I have is after picking up the 8350 for a song, I find it hard to pay double for a power meter, or more.  I pulled the plug-in out of the 8350 today and it is absolutely spotless.  Nice and clean - almost polished - rails.  I doubt it was used much.  The liquidator that sold it really doesn't have much equipment so it must have been a fluke.

I posted another thread asking if I can somehow mix it to use as a TG for my 8566b second band from 2.0 to 5.8Ghz.  I can use it with max hold, of course.

Thanks for all the info on power sensors.  Still looking.

Jerry

[orin]

thanks again.  still debating.  Now I'm thinking I want a 438 instead of a 437.  I found sensors in the 150 range, cable in the $50 range and I can get a 438 for around 150 or less.  So for 350 I can get one channel configured and then work on the second at a later date.  The power sensors look like the are very band and range specific.  If doesn't look like many have more than about 50db of range.  So it looks like you have to have a stack of attenuators.

Mini-Circuits will sell you a stack of attenuators... I have a K1-UNAT+ - good to 6GHz.  I tend to keep the 30dB attenuator on the 8484A.

[cncjerry]

And Orin, nice to hear from you, hope all is well up north.  I'm still taking good care of your Geller standard and will repay the favor someday I hope.  Playing with any 3456s lately?  I still have one in the garage a patient gentleman sent me to fix about 18 months ago and though I said "at my convenience" I think this is pushing it and I should get a move on.  They are fun to fix on a rainy day.

Jerry

[orin]

And Orin, nice to hear from you, hope all is well up north.  I'm still taking good care of your Geller standard and will repay the favor someday I hope.  Playing with any 3456s lately?  I still have one in the garage a patient gentleman sent me to fix about 18 months ago and though I said "at my convenience" I think this is pushing it and I should get a move on.  They are fun to fix on a rainy day.

Jerry

Hi Jerry,

All is well up here - at least after getting a new furnace on Monday.  Glad that the Geller is still alive and well and has a good home.  I haven't been doing anything with the 3455 or 3456s.  I should get back to that nanoprocessor FPGA project.

Orin.

[cncjerry]

OK, I've had a run of good luck.  I picked up two 3586 Selective Level Meters for $50 per.  These are basically a radio that was used for testing channelized voice on T1 and other lines.  I can't remember what the voice technology was called but it was still around in 2001 when I was at cisco.  Very accurate for testing filters, etc. up to 30Mhz with some limitations above 20Mhz.  There are a number of boxes that look like this including the 3336 Synthesized Level generator, 3325 Synthesize/ Function Generator, 5325 Universal Counter, and I think one or two others.  I have one of each or so, ok, two of each or so.  I've had a 3586B that I've used for the Frequency Measuring Tests but the more desirable of the 3586's are the 'C' model as it has a true 50ohm input.  The two I bought today for their Option 004 ovenized 10811 oscillators are an 'A' and 'C' with the 'C' being in the better shape.  It has some hanger rash on the cases but not real bad.  You can see the front panel is in pretty good shape.  The issue with this line of products from HP were the white line across the front under the LED display always discolored.  But it can probably be fixed with the peroxide and sunlight trick. 

The 'A' unit didn't fully power on but the fans started and it went into standby, but no lights, etc.  Probably a power issue.  The 'C' model powered up with an error.  I wasn't going to fix it but when I opened it up to gut it, it was so absolutely gorgeous on the inside I thought I would take a crack at fixing it.  Turned out to be the fractional N VCO so I swapped it with one from the unwanted 'A' unit and off it goes.  Works perfectly.  I tweaked the oscillator to my Cesium beam and ran if for a couple of hours and I'm happy with it.  I'll put it on top of my 'B' unit and decide which to use in the future.

So take a look at how these are built.  The HP 8566B and 8568B spectrum analyzers were built the same way with bolt-down panels holding the cards sliding into a frame.  These units weight a ton.  The 8566B I know exceeds 140lbs with the two units.  This thing has to be over 60lbs.  It's a pain to get the cards out because of all the bolts but it is solid.  After removing the bolts you pull on the silver spindles to remove the cards

Anyway, that's all for now.

Jerry

[Berni]

Ah yes this is classic HP construction of RF gear back then.

I had two repairs recently involving power issues on such era instruments. One was my HP 8566B that just needed a new fuse, no idea why it went(Looked fine in the glass but didn't conduct) and another was a HP 4191A (RF Impedance analyzer) that also had fans running but shown nothing on the front panel. Turned out it was a bent pin in a connector for the transformers 5V winding.

It was more work moving these heavy boatanchors onto the bench than actually fixing the trivial problem.

[cncjerry]

On your 8566B, though I am against wholesale cap replacement, I had a leaky cap on one of the supplies so they do fail.  The supply was like the -10V or 10V, can't remember exactly, but the cap was a Sprague Powerlytic, 13000uf at 25V. The next time you have the RF unit it on its side you might want to check the ESR of some of the caps.  Mine just game me unlocks and other silly errors, fuse stayed ok.

Jerry

[Berni]

On your 8566B, though I am against wholesale cap replacement, I had a leaky cap on one of the supplies so they do fail.  The supply was like the -10V or 10V, can't remember exactly, but the cap was a Sprague Powerlytic, 13000uf at 25V. The next time you have the RF unit it on its side you might want to check the ESR of some of the caps.  Mine just game me unlocks and other silly errors, fuse stayed ok.

Jerry

I was mostly blaming the mains for that fuse. As the local area developed more houses ware added to the already long line to the transformer substation, additionally heat pumps became a popular way of heating homes here and that put a heavy load on the grid here. As a result i would get fairly low voltage in the mains of around 210 to 220V, this caused the HP 8566B brown out and reboot on a few occasions as it was set for 240V. To fix that i changed the mains selector to 220V and it was fine. In the mean time dad had a solar array installed and things got even worse, to the point where the solar inverter was throwing errors. So i set up a DMM to log the mains voltage over a few days and was shocked to find out it would at times dip to 180V while other times rise to 245V. I shown the graph to the power company and they hooked up there own voltage logger to indeed find the voltage is way out of spec. After some back and forth they finally ran a dedicated new line from the transformer station for this area. And that made the voltage rock solid.

So yeah id assume my unit didn't like being in 220V mode with those sort of swings. Who knows how big of a swing it saw since i was only logging for 3 days and saw overvoltage.

While i was in there troubleshooting the PSU i had poked around to check all supplies being correct and clean of noise, it looked fine, but id did not measure any caps. I have seen stories of the round silver polarized capacitors failing, but there are so many of those spread across pretty much all the boards that replacing those would take forever. Im always a bit intimidated by the complexity of this thing so i sort of hope i don't have to mess with it too much (Tho a lot of it is also due to the fact i hate moving this heavy bastard)

I don't use it all that often but so far it seamed reasonably reliable for how old it is. If i really needed a spectrum analyzer on a regular basis i would have probably went for something a bit more modern and convenient. Heck my 2nd Gigatronics 8GHz synthesizer still has a stuck solenoid in its mechanical attenuator that i still haven't got around to fixing.

[Harfner]

A few months ago, while going back to my office, I stopped dead in my tracks in front of a pallet cage full of T&M gear heading for disposal. Obviously they had to clear out an ancient electronics lab.
As I was limited in time (travel next day which still needed preparation) and am limited in space (rented flat, which I share with a female which inexplicably has other priorities) I had to severely limit my choices. So I did not take the 2 oscilloscopes (one 4-channel !) and did not even look up the specs or even remember the model, as that memory probably would have been too painful. I also did not rescue the 4 Fluke 8020A, as my TEA has brought several multimeters in my hands and I do not need another one, however unconvincing that may sound.
What I managed to rescue: HP3478A, HP8116A, Racal Dana 1991, Keithley 2410. You can see the stack before I moved it to the attic.
The HP3478A works as far as I can test, I had to get new button caps and had to replace the ram battery.
The HP8116A seems to work, the NiCd battery just started to leak.
The Racal Dana 1991 had the usual problem, nearly all switches dead, I had to replace them
The Keithly 2410 will definitely need much work. The power section does not work at all, perhaps the custom transformer is dead, it looks overheated.

[Specmaster]

Not a bad score that, especially the 3478A.

[cncjerry]

I was mostly blaming the mains for that fuse. As the local area developed more houses ware added to the already long line to the transformer substation, additionally heat pumps became a popular way of heating homes here and that put a heavy load on the grid here. As a result i would get fairly low voltage in the mains of around 210 to 220V, this caused the HP 8566B brown out and reboot on a few occasions as it was set for 240V. To fix that i changed the mains selector to 220V and it was fine. In the mean time dad had a solar array installed and things got even worse, to the point where the solar inverter was throwing errors. So i set up a DMM to log the mains voltage over a few days and was shocked to find out it would at times dip to 180V while other times rise to 245V. I shown the graph to the power company and they hooked up there own voltage logger to indeed find the voltage is way out of spec. After some back and forth they finally ran a dedicated new line from the transformer station for this area. And that made the voltage rock solid.

Same issue.  New neighborhood, aluminum wire, poor connections, and the legs were out of balance.  One leg around 70 and the other 110.  When the neighbors would flip on their oven, my lights would dim. They ultimately pulled all new wire in my neighborhood which fixed it all.

Jerry

[cncjerry]

By the way, many of the cylindrical silver caps are not electrolytic but tantalum, expensive little devices.  I picked up a couple of thousand of various values from a guy who repaired Tek and HP scopes.  Let me know if you need any by PM.

Jerry

[Smokey]

cool.

[Epatsellis]

Loads of fun, especially for low band dxing. As an aside, the LO has exceptionally good phase noise performance, and I use mine as an RF signal generator below 32Mhz, with a step attenuator it really is useful on the bench.

Should you decide that the OCXO in you spare is surplus to your needs let me know, I've been looking for one half heartedly for a while.

[Berni]

By the way, many of the cylindrical silver caps are not electrolytic but tantalum, expensive little devices.  I picked up a couple of thousand of various values from a guy who repaired Tek and HP scopes.  Let me know if you need any by PM.

Jerry

I just remember them being polarized capacitors. While i was in there i randomly measured a few in circuit and they seamed fine. Some have said they are unreliable and start to fail horribly with age. But there are so many of them in there that i don't really want to recap the whole thing unless they do start to go bad.

Do you need to replace these with genuine parts? Wouldn't a modern electrolytic do just fine?