[time-nuts] FE-5680A Mechanical Question

EWKehren at aol.com EWKehren at aol.com
Fri Jan 13 13:49:20 EST 2012

I think the dither issue should be explored separately. Its finding should  
be included in the Controller. During development I understand the need for 
on  the board programming, but unlike OCXO's which there are many of, I 
envision  this controller dedicated to the FEI 5680A using 10 MHz and a 1 pps 
input. With  no flexability the total cost can be kept below $ 15. If cost is 
not an issue  many things can be done.
Bert Kehren
In a message dated 1/13/2012 1:34:51 P.M. Eastern Standard Time,  
albertson.chris at gmail.com writes:

On Fri,  Jan 13, 2012 at 8:37 AM, Jim Lux <jimlux at earthlink.net> wrote:
>  On 1/13/12 8:15 AM, Chris Albertson wrote:
>> On Fri, Jan  13, 2012 at 3:19 AM,<EWKehren at aol.com>  wrote:
>>> What we know is that you can set  the Rb in 7 E-13. Dithering would 
>>> even closer setting,  the question is what rate will the Rb accept with
>>>  out
>>> loosing lock or deterioration of the performance. Some one  should 
>>> that.  I am still waiting to se some  aging. Taking the 10 MHz output 
>>> than
>>> use  an  analog loop and something like a Morion OCXO and you have  the
>>> best of
>>> all.
>>> Digitally  controlling the Rb will cut the cost of the control loop in
>>>  half.
>>>  For $10 in parts and a PC board for less than $10  using Shera like
>>> controller  can be realized. What is  needed is some one able to do the
>>> PIC.
>>> If some  one is interested and able, please contact me off list. A low
>>>  cost
>>> GPS or a 1 pps output of a Tbolt be perfect  source.
>>> Bert Kehren
>> Don't  use a PIC for the prototype.  A desktop PC could work as well
>>  and everyone here already has one.   Connect the FE5680 to the  PC's
>> serial port and send commands to adjust it.  The PC also  needs to be
>> able to read a voltage.   Many already have audio  input with 24-bit
>> ADC chips.
> But those  audio inputs are almost always AC coupled.
> Bringing up a  question: Does anyone know of a cheap (<$20ish) USB voltage
> sensor  (16 bits or better, ideally)..  I can see one of those Atmel USB
>  capable micros (like the one on the Arduino Uno) hooked to a dual slope  
> successive approximation ADC.
> There seem to be an  amazing number of times that I want something like 
>  The  DATAQ $29 widget is only 10 bits, unfortunately.  A USB interface  
> would work nicely, but I haven't found one that's in the under $50  price

Many computer audio interfaces are AC coupled but not all of  them.
But even with AC coupling what you can do is use the VCO portion of  a
CD4046.  This will convert voltage to frequency in the audio  range.
It's a cheap work around.   But really not because  frequency in nearly
imune to noise and can be sent over long  cables.

What I really want is a lower price GBIP interface.    I just bought an
HP5328A on eBay that has the option 21 DVM.  This  meter can measure
volts to about 5 digits and send the data out the GPIB  but getting
that into a computer is the hard part.

OK so I check on  eBay.   Most are $300 but If you can find a computer
with an old  ISA slot then there are working GBIB cards for about $20.
The trick is  finding an old PC.  These are AT class or early Pentium
type  computers.   Most of these are now in landfills some place.
But  now I will be looking.

I do software all day, every day.  From  experience, I can say it is
MUCH more productive to develop code on a  larger desktop computer.
Better tools are available.  Then move it  down to the target computer
after you have it debugged.   The  process is helped if the target
computer is like the desktop  computer.

There is a lot we don't know about the FE5680, like how fast  can you
move the phase of the PPS?  Does the FE5680 maintain lock when  you
step the DDS?   how fast can you step the DDS.    All this will take
experiments.  best to do those on the big desktop  machine where we
have tools to log data to disk, make plots and so  on.

Again, if anyone makes PCBs PLEASE include a way to program the uP  on
the card without need of extra hardware.   The firmware will  get
upgraded and not everyone has a programmer.  There must be a way  for
end users upgrade the firmware.

Chris  Albertson
Redondo Beach,  California

time-nuts  mailing list -- time-nuts at febo.com
To unsubscribe, go to  
and follow the  instructions there.

More information about the time-nuts mailing list