HP 8664A faults

[plazma]

I got a HP 8664A for free. It gives error messages "Hardware failure 5", "Calibration error 1" and "Calibration error 2. Does anyone have experience in debugging and fixing these?

[Tony_G]

Hardware Failure 5 is in the ALC. The result code of 8,043,311 is a calibration error code (hence the Cal Error messages) - The manual suggests running the ILD (Instrument Level Diagnostic) test function to get a full error result.

• INSTR PRESET
• SPECIAL
• 3,2,0
• SPEACIAL
• If the display says "320 Test" then press ENTER (ON)
• Otherwise rotate the RPG Knob until "320 Test" is displayed and then press ENTER(ON)

This is in section 2 of the service manual on the Keysight web.

[plazma]

Thanks, I'll try this.

[plazma]

I had to open the top lid and enable "Service Special Function" DIP switch. After that the "320 Test Instrument" worked. The result code was -10,124,311.

I have to disconnect the HF DRVR - J6 connector and connect the cable to the power meter for further debugging (Test 2A).
To be continued tomorrow.....

[markus]

Did you manage to repair your 8664A?

[plazma]

Did you manage to repair your 8664A?

I did not go further. The unit is waiting next to my desk.

[analogRF]

you have an error related to REFERENCE section (A9 + A20). The HW error #5 could be the result of that, so you should try to fix the REF first.
The ILD test stops at the first error which in your case is in the reference board. you can run individual tests for all other sections separately but it is better to start by fixing the reference section.
do you have the 10MHz output on the back of the unit? if not, your timebase osc might have problem. you can try to feed a 10MHz to the reference input at the back
which of the red LEDs on the I/O board remain on? In normal operation they must all be off.

there are three outputs from the reference (A9) at J1, J4, J5. you can easily disconnect them and examine on (50ohm input) oscilloscope.

[tkamiya]

Have you opened it up and inspected?

The first thing I would do is to open it up and make sure all cards and connectors are seated well.  I usually remove and re-seat everything to break oxidation on connector surfaces.  While card is out, look closely, especially capacitors.  Look for cold solder joint. 

If your unit has an optional oven crystal oscillator, (High stability time base option)  there has to be a coax jumper on the back.  Make sure it's there. 

I've recently repaired my HP8644A.  An odd sequence of event and failures can cause calibration error when CMOS backup battery is dead.  It should be on A16 board. 

Negative prefix on error code means it was NOT able to POSITIVELY identify that code is accurate.... 

[plazma]

Thanks for the hints. I'll continue the repair when I have a less busy time at work.

[analogRF]

If your unit has an optional oven crystal oscillator, (High stability time base option)  there has to be a coax jumper on the back.  Make sure it's there. 

8664A with OCXO option does not need coaxial jumper at the back. the OCXO is connected to A9 inside the unit

EDIT: 10MHz output in the back actually comes from A9 board (REFERENCE). If you don't have anything on the 10MHz output you will have error(s) related to hardware #8 which is the REF module. You can temporarily feed the 10MHz input in the back and see if you will get anything in the RF output.

[lugaw]

For anyone who is having this error code the fix is to use an external 10Mhz source or repair the OCXO which have a habit of turning off the OCXO because some components on the module degrades because of time.

If the HP OCXO is still outputing 10Mhz you can just jumper it directly to J2.  The A20 module have extra circuitry that detects if there's another timebase present and suppress the OCXO.  This circuit is not used by the 8664A.  This is what I did to my 8664A and it's still working.