[time-nuts] Ebay FE-5680A Rb: Connections and Features
GandalfG8 at aol.com
GandalfG8 at aol.com
Sun Dec 11 19:29:56 EST 2011
In a message dated 11/12/2011 23:01:14 GMT Standard Time, smither at c-c-i.com
Can anyone tell me if they ALL have the MAX level shifter/driver chip
> fitted, even though it may not be connected to the connector ? I don't
> want full programmability, just 10 MHz frequency correction. Perhaps I
> have to take pot luck when I buy ? What are my chances of getting one
> the 'hidden' RS232 and 1PPS output ?
For what it is worth, this unit:
appears identical to the one I purchased four weeks ago. Same seller,
off PCB, same "extras" thrown in, same price.
The FEI part number you are looking for is 217400-30352-1, it is often
clearly visible in the auction photos and can be seen in at least one of the
photos in the above link.
I'm sure all will do what you ask but the biggest problem I found was
actually getting to talk to the unit, it doesn't give you any feedback if you
get it wrong, just sits there looking at you.
Data connections are as per the manual, pin 8 of the 9 pin connector on the
module goes to pin 3 of the 9 pin RS232 port at the PC, and pin 9 at the
module goes to pin 2 at the PC. Add the ground connection and that's it.
You will need some software to send the HEX 32 bit signed integer values
plus checksum to the unit.
Programs such as Hyperterminal aren't much help here but I found the RS232
Hex Com Tool from Virtual Integrated Design to be ideal.
A demo version can be downloaded from _http://www.viddata.com/_
The demo version times out after 5 minutes but can be restarted immediately
so that isn't really a problem.
Settings are 9600 Baud, No Parity, 8 Data Bits, 1 Stop Bit, No Flow
There's plenty of utilities online to convert decimal to HEX, I
The checksum is just the XOR of the four 8 bit words and I used Windows
calculator in programming mode for this, not sure if this is relevant to all
calculator versions but it is in Win7.
My unit came with an offset of zero, and I'm guessing most others will too
as they probably weren't ever adjusted in use.
I followed Jose Camara's advice for initial offset testing, and used
offset values of +4095 and -4096 to calibrate the frequency adjustment range
with the zero offset as a confidence check.
The commands to send these offsets, without storing them, plus the
immediate effects they had for me, are as follows......
+4095 -- 2E 09 00 27 00 00 0F FF F0 -- 10.000,000,023,00 MHz approx
0000 -- 2E 09 00 27 00 00 00 00 00 -- 9.999,999,995,00 MHz approx
-4096 -- 2E 09 00 27 FF FF F0 00 F0 -- 9.999,999,967,00 MHz approx
Gaps in the command string are for clarity, and not needed when sending,
and all marked as approx because I was just looking for ballpark figures and
didn't wait too long for them to settle.
>From these results it was easy enough to calculate an initial adjustment
with corresponding offset,that would get me close to 10MHz, followed by a
further correction and offset that got me to 10.000,000,000,1 MHz, give or
take all sorts of variables, including the phase of the moon on the next
Which I decided wasn't too bad for a first try, and close enough anyway to
use the adjust and store command for the same value:-)
More information about the time-nuts