[time-nuts] FE 5680A Warning

Peter Bell bell.peter at gmail.com
Sun Nov 13 14:24:20 UTC 2011


Yes, I'm quite sure this is how it works - if you scope the snap diode
drive you can quite clearly see both the 60MHz and the 5.3MHz from the
DDS chip mixed together.  My assumption is that the DDS output is
subtractively mixed with the 114th harmonic of the 60Mhz inside the
cavity:  60*114 = 6.840Ghz - ~5.3Mhz = ~6.8347 GHz clearly puts you in
exactly the right ballpark.

This would also explain the very fine tuning steps - a 1Hz change on
the DDS output would only need a ~1.45mHz change of the unit output
frequency (at 10MHz) to bring the loop back into lock.

I still have the unit with the bad Rb cell in bits, so I will have a
poke around it and see if I can figure out how the drive on the
C-Field coil is derived.

Regards,

Pete




On Sun, Nov 13, 2011 at 7:40 PM,  <EWKehren at aol.com> wrote:
> Pete
> Thank you for your information. I agree with your findings and as I pointed
>  out previously it support the math that the AD 9832 is used for tuning.
> That  allows them to set the Rb at the optimum operating point based on their
> gas  mixture. What  I like to know is how they control the C field. Again I
> think they picked an optimum point and only vary it slightly for
> temperature  compensation. I would like to use the C field to fine tune using a
> modified  Shera controller. Latest voltage measurements on the C field across the
> coil  are 2.5110 and 2.5238 down from 2.5134 and 2.5256.
> Bert Kehren
>
>
> In a message dated 11/13/2011 4:38:09 A.M. Eastern Standard Time,
> bell.peter at gmail.com writes:
>
> I can  let you know what I've found out about these units - I got given
> a couple  of dead ones and did a bit of poking around trying to get
> them  working.  I didn't get that far, since one of the units was
> easily  fixed (the MAX882 3.3V regulator had failed) and the other
> seems to have a  faulty Rb absorption cell (it works if you swap the
> cell from the other  unit into it).
>
> The first is, as you noticed, that they have 3 voltage  regulators -
> but apparently couldn't spare one to generate the +5V line -  they seem
> to run (1) The Rb lamp / heater (the one nearest the connector)  (2)
> the cell heater (the middle one) and (3) the analog electronics  @8.8V
> (the last one).
>
> The quickest way to verify this is to apply  +15V / GND to pins 1 and 2
> of the d-type and then check the voltage on the  VCC pin (20) of the
> 74ACT240 buffer chip next to the connector - there is  no 5V there.
> There is also no 5V on pin 5 of the MAX882 - and hence no 3.3V  line.
>
> There is a connection between these points and pin 4 of the  d-type -
> and applying an external +5V to this pin will bring up the 3.3V  line
> and allow the unit to lock.  When locked, pin 3 on the d-type  goes
> low, and will sink enough current to operate an LED.
>
> Pin 5 is  also connected to GND, and was (on the original PCB) also
> connected to the  ground plane along with pin 2.  It doesn't seem to be
> actually  necessary, as the unit will operate with a single ground
> connection, but  would it would seem to be good practice to use this
> pin, too.
>
> It's  been alleged that pin 6 is a 1PPS output - but I can't see
> anything on it  at all except a constant logic 1 - I tried changing the
> 'ACT240 in case the  buffer was damaged, but this had no effect, and
> the pin is still a constant  1.  I suppose it's possible that there is
> an extremely narrow -ve  going pulse on it, but if there is it's too
> narrow for the trigger circuit  on my (100MHz) scope to see it.
>
> Pin 7 outputs a 10MHz sine wave - about  800mV p-p unloaded.  I haven't
> put it into an analyzer, but it looks  pretty good.
>
> Pin 8 and 9 are connected to the MAX3232 on the board, and  from them
> onto the serial port on the DS80C323 CPU.  Pin 8 is RXD, pin  9 is TXD
> (from the unit's point of view). This serial port accepts the  commands
> described in the FEI document that's floating around the net  for
> trimming the oscillator.  They appear to operate as  described.
>
> Internally, the main VXCO runs at 60MHz - it's buffered and  fed into
> the Xilinx XC9572XL PLD on the back of the board.  This chip  appears
> to contain mostly frequency dividers - it generates the 10MHz  output,
> 20MHz for the MCLK to the DDS chip (although with a slightly  nasty
> 40:60 duty cycle) a 416.6666Hz drive for the fselect pin on the  DDS
> chip, a 833.33333Hz drive for the lock in amp.  There are also a  bunch
> of signals that go to the test connector, and are presumably used  for
> factory programming.
>
> The output of the DDS is about 5.3MHz with  FM on it at a 416.666Hz
> rate - this is mixed in with the 60MHz output from  the VCXO to give
> the drive signal to the snap diode in the physics  package.
>
> There are 3 pots on the board - the one that's normally  accessible via
> a hole in the case seems to be intended as a C-field adjust,  but does
> nothing on the units I have - the other one near it appears to  trim
> something in the photocell amplifier circuit.  The one on the  other
> side of the board (near the VCXO) sets the bias on the snap  diode
>
> That's all I've got, since I then worked out that I didn't have  an
> electronics problem - hope it helps.
>
> Regards,
>
> Pete  Bell
>
>
> On Sun, Nov 13, 2011 at 1:46 AM, Don Latham  <djl at montana.com> wrote:
>> I just bought a couple of 5680A's that  have the same FEI p/n. They are
>> pop-riveted to a heavy double-sided  board., obviously cut off from a
>> longer board. I've removed one of  these from it's board/heatsink. There
>> are two pots on the board, only  one of them can be trimmed with the case
>> on. The sinewave forming  filter and maybe countdown chips are not there,
>> and never were. There  are three identical voltage regulators on the
>> heatsink spine.
>>  I'll be doing some more reverse engineering on these units; they were  so
>> cheap I had to have a couple! If only FEI was more helpful! hey  must
>> have moved on from these, and could put one manual out  there?
>> Oh Well!
>> Don
>>
>>  EWKehren at aol.com
>>> The FE 5680A's advertised as NEW are not NEW!  A  friend  that  made me
>>> aware of the sale  did notice clear signs of use, so I did a  closer
>>>  inspection
>>> under a Microscope and found clear indications that the  unit  has been
>>> bolted down in the past. Does work fine and I  am looking at ways to go
>>> directly
>>> to the C field to  disciplin the Rb. Does any one have info that would
>>>  help
>>> to do so.
>>> Bert Kehren  Miami
>>>  _______________________________________________
>>> time-nuts mailing  list -- time-nuts at febo.com
>>> To unsubscribe, go to
>>>  https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>> and follow  the instructions there.
>>>
>>
>>
>> --
>>  "Neither the voice of authority nor the weight of reason and argument
>>  are as significant as experiment, for thence comes quiet to the mind."
>>  R. Bacon
>> "If you don't know what it is, don't poke it."
>> Ghost  in the Shell
>>
>>
>> Dr. Don Latham AJ7LL
>> Six Mile  Systems LLP
>> 17850 Six Mile Road
>> POB 134
>> Huson, MT,  59846
>> VOX 406-626-4304
>> www.lightningforensics.com
>>  www.sixmilesystems.com
>>
>>
>>
>>  _______________________________________________
>> time-nuts mailing list  -- time-nuts at febo.com
>> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the  instructions  there.
>>
>
> _______________________________________________
> time-nuts  mailing list -- time-nuts at febo.com
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the  instructions there.
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>



More information about the time-nuts mailing list