Richard,

Before I respond to Bev's response let me ask a question.  Do you prefer to
be referred to as Richard or Dean?

Now for my response.  I knew Bev would be in agreement.  I've worked with
Fischer XRFs in the AT&T Richmond PWB Factory, in the Western Electric
Engineering Research Center (later to be called the Lucent/Bell Labs ERC)
and also in the Andrew facility in Warren, NJ.  Andy and I relied on the
Fischer XDAL machine for measuring plating thickness on components and PCBs.
I also agree with Bev's comments against handheld machines.  I always wanted
to have a handheld XRF but never found one that I could rely on.  The last
Fischer XDAL unit I worked with had four different collimator apertures.  We
most often used the smallest aperture because we were trying to measure the
plating composition and thickness on very small parts or fine pitch features
on PCBs.  When we had large components or large ground plane features we
would use a larger aperture because it sort of gave us a better average
measurement.

As for XRF accuracy, every XRF manufacture usually tells you their accuracy
is 5%.  Initially I didn't know what they meant by 5% because the
manufacturers never elaborated.  Over time and after ten's of thousands of
measurements I realized that the 5% meant that the XRF measurements will be
within 5% of the samples used for calibration.  The real key to accurate xRF
measurements is to have calibraton samples close to the thickness and
composition you are trying to measure.

George Wenger

-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Bev Christian
Sent: Tuesday, June 27, 2017 5:15 PM
To: [log in to unmask]
Subject: Re: [TN] XrF systems

Dean,
I agree with George, Joyce and Andy.  When Joyce and I were at BB we had two
Fischer XDAL machines. Totally agree with Andy with regards to not reaching
0.001% accuracy.  Generally I would say +/- 5% if you are talking about ENIG
gold thickness and +/- 10% if you are talking about ENIG nickel thickness.
Can't remember results for % composition results.

Most machines can do U down to, crap! I can't remember the lower limit. But
I do know if you want to do elements like F and lower you would need a
machine that can be evacuated.  Sure adds to the price!

Now I will go on my usual rant against handheld machines.
1) Aperture is such that you would need a pile of chip caps to get a reading
rather than say one lead of 20 mil width, which is all you need for a good
machine from Fischer, Oxford or Horiba.
2) easy to drop
3) easy to lose (or have stolen)
4) In some jurisdictions the licensing is onerous.

Regards,
Bev


-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Stadem, Richard D.
Sent: Tuesday, June 27, 2017 2:49 PM
To: [log in to unmask]
Subject: [TN] XrF systems

I am considering requesting capital for an additional or replacement XrF
system, but would like to go in armed with a short list of particular models
and costs and be confident of their performance records and capabilities.
The least important factor is cost, so don't hold back from sending me a
confidential email on your choice.
The machine should be able to accurately determine the percentage of gold,
silver, tin, nickel, lead, copper, etc., down to the .001% level in order to
confirm that the plating on incoming component leads meets the component
print or requirements.
It should also be able to accurately determine the metallurgic make-up of
any given metal sample. Is that asking a lot? I don't know.
I prefer a hand-held, but if what I am looking for is only available in a
base machine package, then that will have to work.
No information you folks provide will be shared. Feel free to send me an
email at either [log in to unmask]<mailto:[log in to unmask]> or
[log in to unmask]<mailto:[log in to unmask]>.
I sincerely appreciate any information you can share with me.

Thanks
R. Dean Stadem