Odd error with Agilent 1158A probes on Keysight MSOX scope

[LapTop006]

I recently picked up a bulk lot of "known-bad" Agilent probes simply for some mechanical spare parts should I ever need them, since the price was too good to let slip by.

They arrived today and while testing them out just to be sure, one of the 1158A's actually passes signal and cal's just fine, however I get an odd error "Invalid characters were removed from the string." when I plug either of the two 1158A's into the scope (A fully updated MSOX6004A).

I can't find any reference to this error online, and since I get signal I'm guessing it's cosmetic and safe enough to ignore, but would be nice to know for sure.

While I'm at it, I think I read somewhere about the InfiniMaxx II probes being refused as incompatible is known and expected, not that I have a need for them, anyone know for sure?

[LapTop006]

While I'm at it, I think I read somewhere about the InfiniMaxx II probes being refused as incompatible is known and expected, not that I have a need for them, anyone know for sure?

Dug the reference up, the "Infiniium Oscilloscope Probes and Accessories Data Sheet says S series or 90k series.

Guess these will be my candidates to pull the board from and use as a shell to make auto-detecting SMA adapters out of (essentially making my own version of Tek's TCA-SMA since I can't find any official Keysight version).

[KrudyZ]

Guess these will be my candidates to pull the board from and use as a shell to make auto-detecting SMA adapters out of (essentially making my own version of Tek's TCA-SMA since I can't find any official Keysight version).

What does this buy you over a simple BNC to SMA adapter, since your scope already has 50 Ohm inputs?

[LapTop006]

Guess these will be my candidates to pull the board from and use as a shell to make auto-detecting SMA adapters out of (essentially making my own version of Tek's TCA-SMA since I can't find any official Keysight version).

What does this buy you over a simple BNC to SMA adapter, since your scope already has 50 Ohm inputs?

It gets me a few things, auto-detection, so less chance to screw up a measurement, second, reusing the connector should mean the BNC end is a better match at higher frequencies, plus easier to get on and off, many BNCs are extremely tight on those connectors, but the AutoProbes all seem to be just right.

[Tony_G]

I recently picked up a bulk lot of "known-bad" Agilent probes simply for some mechanical spare parts should I ever need them, since the price was too good to let slip by.

They arrived today and while testing them out just to be sure, one of the 1158A's actually passes signal and cal's just fine, however I get an odd error "Invalid characters were removed from the string." when I plug either of the two 1158A's into the scope (A fully updated MSOX6004A).

I can't find any reference to this error online, and since I get signal I'm guessing it's cosmetic and safe enough to ignore, but would be nice to know for sure.

While I'm at it, I think I read somewhere about the InfiniMaxx II probes being refused as incompatible is known and expected, not that I have a need for them, anyone know for sure?

It might be corruption in the probe identifier memory. It'd be interesting to see what is inside one and if you can read it.

TonyG

[LapTop006]

It might be corruption in the probe identifier memory. It'd be interesting to see what is inside one and if you can read it.

It if was on just the one I'd agree with you, but on both makes me think it's either a bug dealing with these probes, or whoever had them last reprogrammed them.

I should just be able to dump the I2C EEPROM on them, although clipping on to the pogo pins might be a bit of a pain, if I had contact details for the folk who make these things I'd ask them for one of the connector boards on their own.

[TheSteve]

You will have to open one up and see if it has a standalone eeprom you can read/write. Some of the probes have the data stored in a microcontroller. I would say something happened that corrupted them.
My 1158A gives no errors when used with my MSOX3024T.

[LapTop006]

You will have to open one up and see if it has a standalone eeprom you can read/write. Some of the probes have the data stored in a microcontroller. I would say something happened that corrupted them.
My 1158A gives no errors when used with my MSOX3024T.

Tah, certainly points to the probes having been messed with. Just to check you're running the current (late 2017) Keysight firmware for the X-series?

I popped one open at the scope end, it's a 24U02 EEPROM, so I should be able to dump it easily enough.

I know I must have 50 things here that can do such a job, but no standalone "i2c tool" that's actually easy to use. I should have a bus pirate somewhere which probably comes close enough.

[TheSteve]

Did you ever read the eeprom in your 1158A?

I just got a second 1158A probe and was seeing the same message. It looks like the serial # is shifted one byte in the eeprom. However it doesn't look like corruption as I believe the checksum is still correct. The scope reads the probe serial number as having an extra byte in length, that it what it is complaining about. I don't think there is an issue using the probe though. Either way I removed the shift and believe I corrected the checksum byte (the 3rd byte is the checksum). The 3000A series doesn't appear to care about the checksum, the older scopes might though. That is something I hope to test on the weekend.

When reading/writing the 24C02 eeprom set the base address to A2.

edit:

I have now tested the patched 1158A with both my DSOX3000t and a DSO80000A series scope. Both are happy with it, no errors and the DSO80000A reads the serial # just fine.

[nfmax]

I have just acquired what appears to be a NOS 1157A probe - it has the full set of accessories, unbent & un-soldered, with the full complement of tips, and the manual was still in its plastic film shrink wrap. I get exactly the same error when I connect it to my MSOX-3104T, which has the latest firmware. The probe seems to work OK, as far as I can tell with only a 1GHz scope, after calibration. However, the calibration information doesn't seem to be retained: if I disconnect the probe and reconnect it (to the same channel) the error re-appears and the scope requests a probe calibration. This fits with the idea that there is a problem reading the serial number (in my case this is a US4100 series number). I can't find any way in the scope menu system to display probe type & serial number information, but this may be a side effect of the same bug. It seems unlikely this probe's memory has been tampered with. The vendor checked it using a 54825A.

@TheSteve - does your 3000T retain the probe calibration now you have tweaked the serial number in the EEPROM?

The scope is still under warranty so I may raise an issue(!)

[nfmax]

I raised an incident with Keysight UK (ID 00508836) and they have checked with the same model probe, without errors. They suggest my probe is faulty 

However this doesn't seem right. Reading up-thread, it looks like some probes of this series give errors with the X3000T, and some don't.  There is no case where one of the error-inducing probes gives an error with an older series oscilloscope. At least one probe has been 'fixed' by (?) re-writing the serial number in the 'correct' place in the EEPROM and adjusting the checksum.

It seems to me the probable cause is that the code in the X3000 series is being more fussy about reading the serial number, rejecting some and accepting others. This ought to be fixable with a firmware update.

Edit to add:

I've passed on a reference to this thread to the Keysight contact centre handling the incident