[time-nuts] HP 8568B

[Dan Rae]

On 6/28/2014 6:48 PM, Hui Zhang wrote:
> Hello all:
>      Rencently I bought a second-hand HP8568B Spectrum Analyzer, it kind of vintage but works well. Only problem is its time base lost accurate. When I use it measured my HP Z3801A output (locked to GPS), the center frequency read is 9.999912MHz, about 88Hz diffrence. So I decide to use  my Z3801A for its external reference. When I connected the cable and set the reference switch in rear paneal to EXT, the CRT displayed "EXT REF", I believed it worked. But when I use it measure a nother GPSDO (Trimble ThunderBolt locked to GPS), it still have 42Hz diffrence. I am very sure both of my two GPSDO is good, if I use my HP53132A counter to compare them they will has less 1E-10 (1mHz) diffrence. And then I tie a BNC three-way connect from my Z3801A, one way to 8568B's EXT stand input, one way to the signal input in front panel, the  diffrence is still 42Hz - it's 10.000042MHz. I was fully confused, do I need a calibration after I used external 10Mhz stand? I read the manual again but not fou
>   nd any infomation about how to do it. What do I do now? Anyone give me a suggestion? Any infomation will be appreciate.
>
>
Hui Zhang, check the supplies from the A24 board.  The oscillator +20V 
supply is supposed to turn off when "Ext" is selected.  Also the series 
transistor for the heater supply I feel is very under rated; I mean 
feeding an oven heater through a 2N2222A sounds like a good idea?  It 
could be that you have the internal Osc still working when "Ext" is 
selected.

Also the older Ovens fitted to those are not the most reliable ones ever 
made.  I reported on the opening up and repair of one in an 8568B of 
mine about six months ago but that may have been on the -hp- list, not 
here...

Good luck!

Dan

ac6ao