HP 438A power meter with HP8481A power sensor 'input overlaod' without

[mk_112358]

Hi All,

I got an HP438A power meter and HP 8481A power sensor from eBay.
When I connect the power sensor to the power meter, the power meter displays 'input overload', even without any RF power on the sensor.
I tried zeroing it with the reference output, but it fails.
I'm not sure the problem is with the sensor or the meter. I don't have another power meter or sensor to check which is the faulty one.

I greatly appreciate any help/comments on this. Thanks!

[Hamelec]

you bought it in "fully working condition" ?
hopefully payed with paypal?
open case and return to seller!

otherwise:
- crosscheck with a 2nd powermeter/powersensor/cable..
- check powersensor s11 with VNA..
...

[Tony_G]

I suspect that Hamelec is correct here - unless you paid in the US$300-400 range for the 8481A then I'd put money on the sensor being blown.

There really isn't any good way to test the unit without another known working sensor or an HP 11683A Range Calibrator.

I'd be inclined to return it for a refund if it was supposed to be working.

TonyG

[alm]

I thought you could test the 438A by feeding it some DC voltages? The service manual tells you how, I believe. The 11683A is just a convenient way to generate those voltages.

But my guess would indeed the sensor being broken. In the absence of a network analyzer, would a DC Ohms measurement of the sensor's input yield anything useful?

[Stray Electron]

  I've tested a LOT of those used power sensors and they were ALL bad! I hope that you bought it "used and working" and paid for it with PayPal, if not you just wasted your money.

  On Ebay,  "untested" translates as "we tested it and we know that it's bad but we hope that you're foolish enough to buy it without a guarantee that it works".

  And NO, those sensors can NOT be repaired.

[Tony_G]

ALM - The way the 8400 series sensors work means that you can't just feed it some DC values - You need to simulate the sensor with the appropriate sampled AC signal etc. The sensor and the meter join together to create one "Hybrid OpAmp".

You could probably make one using the 11683A schematic and a dead 8481A though.

EDIT: I thought I might check to see if there was a way to measure the resistance as a quick check and the manual does have a procedure for when you don't have a known good power meter/sensor

Troubleshooting - Eliminating the Power Meter and Sensor Cable
Where a “known good” power meter and/or sensor cable is unavailable, another means must be used to isolate the fault to the Power Sensor.

This is done by ensuring the power meter is providing the correct 220Hz drive signal. Check the following levels of the square wave with an oscilloscope.
• At the black/white wire: -0.05 ± 0.05 Vdc (top of square wave).
• At the brown/white wire: -9Vdc (bottom of square wave).

If the levels are incorrect, then the power meter or sensor cable is at fault. Refer to the power meter service manual for troubleshooting information.
If the levels are correct then the Power Sensor is at fault. Continue by troubleshooting the A1 Bulkhead Assembly.

I'm not sure if this would rule out the corner cases but it seems to be a good starting point to at least identify if the sensor is good.

EDIT #2: As Stray Electron says if they aren't guaranteeing them as tested and you're not paying in the US$300-400 range then they are almost certainly dead. It's easy to blow the thermocouple in the units.

You actually might have better luck with an 8481D as these are diode-based and while they only range up to -20dBm they're fairly safe up to something like 20/30dBm (from memory, never taken one up there so no empirical evidence).

TonyG

[bdunham7]

  And NO, those sensors can NOT be repaired.

Well they can, and they even sell the parts for them--or at least they did.  I have a working 438A and cable and I've repaired some of those sensors for a surplus eBay seller.  The ones that I manage to fix get sold as working and the ones I don't probably end up as 'not tested'.  I don't think I've see one come in already working--they don't show up on the surplus market unless someone blows it up or drops it.

I don't recall whether I've seen the overload message, but that could be the meter or the board in the sensor.  Without a known good meter, I'm not sure you can test the sensor properly, although you can check it's 50R input resistance.  The test procedure Tony_G posted will confirm a bad meter, but IMO won't verify a good one.

Here's the guts of one.  The gold cylinder contains the 50R terminator/thermocouple hybrid and those little gold wires attach to that IC.  That cylinder or 'pellet' is or was sold as a spare part, but isn't cheap. 

[mk_112358]

Thank you for the valuable comments.

I checked the power meter with a working sensor and found out the power meter is fine. So the culprit was the power sensor.

I followed the operating and service manual for HP 8480 series (https://xdevs.com/doc/HP_Agilent_Keysight/HP%208481A,%208482A,%208483A%20Operation.pdf) to troubleshoot the power sensor unit.

It turned out the 'input overload' error was due to poor contact of gold wires with the board.

Thanks!

[alm]

Having a known good sensor is really helpful in troubleshooting indeed. Congrats on the repair!

[Tony_G]

Congrats as well - So do you think it was shock damage that impacted the wires? Normally these are burnt out, I've never seen one where the damage was just having wires detached so I'm really interested?

Did you grab any photos/images?

TonyG

[mk_112358]

Hi Tony_G

I think it is probably the shock damage during shipping. The gold wires were not in contact with the circuit board. The seller had listed it as a working condition on eBay.
Unfortunately, I did not take any pictures inside.
-MK

[bdunham7]

The gold wires were not in contact with the circuit board.

Was the clamp still in place?  Was it loose?

[gsmarte]

Any repair component sources/links/sellers that you can remember? I have a couple of bad sensors that I'd like to tackle. Thanks.

[bdunham7]

Any repair component sources/links/sellers that you can remember? I have a couple of bad sensors that I'd like to tackle. Thanks.

No, I don't think there are any economically reasonable ways to get parts for these.  My fixing has consisted of the 'make 3 working units from 8 broken ones' sort of parts swapping.  My current inventory is a small tray of junk.

[gsmarte]

Thank you. I had done some searching but couldn't find any components anywhere. Will have to accumulate more bad sensors and perhaps cannibalize them. thanks again for the reply.

[Arhigos]

  I've tested a LOT of those used power sensors and they were ALL bad! I hope that you bought it "used and working" and paid for it with PayPal, if not you just wasted your money.

  On Ebay,  "untested" translates as "we tested it and we know that it's bad but we hope that you're foolish enough to buy it without a guarantee that it works".

  And NO, those sensors can NOT be repaired.

Do you know if there is any way to check 8481D power sensor without power meter? The thing is i have a sensor in unknown condition acquired at local hamfest and now when building an amplifier I need a good power meter.
There is tons of cheap 437 and 438 meters on ebay in sub-$100 range but I don't want to buy a meter just to figure out that I have a fried sensor

[Stray Electron]

I suspect that Hamelec is correct here - unless you paid in the US$300-400 range for the 8481A then I'd put money on the sensor being blown.

  Same here. I used to test a lot of those used HP power sensors from E-bay, Lab-X, ham fests and the like and I never found one that wasn't bad!

   Those things are easy to test and considering what a working one is worth, NOBODY is going to sell an "untested" one.   If it was categorized as "new" or "used" on Ebay then you're entitled to get your money back.  End of story. Notify the seller and then file a complaint with E-bay and don't dick around.