Hi there!

Paul Dennis wrote the following about Source Linearity:

> I have a source from a SIRA series instrument that is performing very badly
> in terms of linearity. There is a shift of as much as 8 per mille in the
> 46/44 ratio during a standard run with the gas depleting from the cold
> finger of a dual inlet. With retuning of the source I can reduce this to ca.
> 5 per mille but no better. Does anyone have any hints regarding tuning,
> filament change etc. which may help solve this problem or will I need to
> change the complete source assembly.

Well, I've never seen a SIRA source go /that/ non-linear.  And we've
got 3 of the babies too :-)  Here's some general tips on inlet blocks,
which may be relevant, and if not, it's Friday night, and I'm a typin'
fool.....

You didn't say what was happening to the 45/44 during the run, but
I'll take a pop at it....  My guess is a leak on one of the inlet
valves, probably SF, SI, or SP (or RF, RI, RP).  Have you tried doing
a run on each side of the inlet, without the changeover?  Just type
<CO> as the sample name, and the changeover won't toggle during the
run.  Try this on the reference side, and on the sample side on
different runs and see what happens.  If one is OK, and the other
drifting, try doing the same thing but with equal pressures across the
valves.  i.e., if the sample data is drifting, try flooding the back
manifold (the bit with LV and HV on it.  SP, SM, SF open, SV, SI
shut), and closing SP just prior to starting a <CO> run with the
changeover on the sample side.  If the data improves, then SP is your
man, otherwise try the same thing across SI. There /are/ better ways
of leakchecking the inlet valves, but your problem sounds pretty
extreme, so the above test should work :-)

The other possibilty is a leak on the changeover valve.  I've just had
the opposite problem on our SIRA 10, where the actuator for RS crudded
up, and wasn't opening fully, resulting in big pulses of ref gas
whenever changing from sam to ref.  Leak checking the changeover is
not too tricky:

Fill the sample side with CO2.  Have the reference side pumping on HV
(RP, RF, RM open).

Switch the changeover valve to the reference side.  There should be no
visible beam.  If there is, SC is leaking.

(Careful here:)  Close SS, wait about 20s, and open RS.  There should
be no visible beam.  If there is, SS is leaking.

(Still careful:  DON'T toggle the changeover after doing the above)
Close RP, and pump out the sample side completely including SM.

After pumping for a minute, you can toggle the changeover.  Put it
onto the reference side, and fill the ref with CO2.  Do the above
tests again but back to front, 'cause you're checking RC and RS now.
[BE CAREFUL when you have both RC and RS shut.  Opening RC will trip
the source, and opening RS will probably do the same.  Make sure you
have pumped up to RM for at least a minute before opening either
valve]

Well, I'll leave it there...  Feel free to contact me/the list with
results or developments :)

Totty  <8^)