Peter
Answers below:
[log in to unmask] wrote:
> Brian,
>
> Very many thanks for replying to my post. Part of your answer begs another
> question - if conductivity is linear and can accurately reflect
> contamination levels down to virtually zero, why are all the Ionic
> Cleanliness Test machines designed to measure resistivity and not
> conductivity?
a) They aren't. The Contaminometer uses a unique ballistic conductivity
meter. b) Some of the American ones use resistivity as a legacy from
Hobson's initial experiments, whereby he used a commercial resistivity
meter.
>
> I don't know how to set about coverting a machine to read conductivity of a
> liquid instead of resistivity. Is there an easy way to do it, or has no-one
> tried? I wouldn't want to just use a formula to convert the resistivity
> reading into a conductivity value, since I don't trust the resistivity
> readings at low contamination levels in the first place. There would be no
> point, as the results wouldn't be so linear.
You would have to change the instruments' measuring systems, recalibrate
(not easy), and alter the algorithms in the firmware or software.
>
> Sorry, I'm going to cross-examine further - using a 50:50 water:alcohol mix
> will certainly help dissolve more of the non-water-soluble residues and
> thereby release more ionic contamination into the test solution to be
> measured.
This is not the point. The IPA is there mostly to increase the
solubility spectrum, especially of rosin residues. It also decreases the
surface tension and viscosity of the solution. It is ONLY the water that
affects the readings, so if you double the quantity of water by going
from 75% to 50% v/v IPA, you double the sensitivity. There is another
advantage of 50%: it is less likely to saturate with heavy metal salts.
That would prove that you have a dirtier board than you
> previously thought, and prove that cleanliness measurement is a relative
> thing, not an absolute. It does not, however, improve the linearity or
> correlation of the readings the machine gives against the amount of
> contamination in the test solution - the higher amount of contamination
> might just take the readings into a region of the correlation curve where
> they coincide better, but that's all.
Not so. Say, with a 75% solution and assuming a constant temperature,
you start at an initial conductivity of 0.005 uS-cm and with a given
quantity of contamination you end with 0.02 uS-cm and you meter has an
absolute precision of +/- 0.002, then your start may be anywhere between
0.003 and 0.007 and your end may be 0.018 to 0.022. Your delta cond will
therefore be between 0.011 and 0.019 or 0.015 +/- 0.004 or +/-27%. If,
under the same conditions, you use a 50% solution, for the same
contamination, the nominal values, with the same meter accuracy, will be
0.010 +/- 0.002 and 0.040 +/- 0.002, so the min/max differential will be
0.030 +/- 0.004 or +/- 13%. In practice, it will not be as bad as this
sounds a) because the instrument will probably be more accurate than
that if it is a conductivity meter and b) because, if there were an
error in one direction at the start value, it would most likely have
about the same error for the end value, in which case the dirrerential
error will become negligible.
>
> I agree with your other comments about minimising atmospheric CO2
> absorbtion and restricting test-run times to 15 mins max (...?), although I
> was finding that at least one of the machines was reading low at the low
> contamination level end of my testing - bad enough if there had been no CO2
> influences on the results, almost worse if there were such influences,
> since the departure from accurate figures would have been even worse.
The European Space Agency report published about 1980 showed that low
readings were commonplace with low values of contamination with "static"
instruments employing resistivity meters for the reasons already cited.
They have improved since then, but the basic conceptual fault does remain.
>
> I should have the courage of my convictions - it had penetrated me
> reasoning that static technology ought to be better for measuring low
> contamination levels than dynamic. However, we have a dynamic machine in
> our hands for playing with, and a good price is being asked if we want to
> buy it. The man holding the budget is keen, even if he doesn't seem to
> understand the subtle minutiae of the measurement process and its pitfalls.
> We're 99% of the way to clinching the deal already, but I'm still not
> convinced that it can measure that well without a lot of work to correlate
> the readings with actual contamination levels - i.e. so an operator can
> look up the machine reading on a chart and get a "true" contamination
> figure.
>
> I would still like to know, though, if there is work being done to produce
> machines that are 'comfortable' with measuring low contamination levels to
> replace those that are operating right on their limits (or beyond) to
> measure low levels. Saying a machine is quite capable of reading 0.2ug NaCl
> eq-cm^-2, without taking the board area into account makes such a claim
> meaningless. Many a machine will return such a cleanliness level reading if
> you have a large enough board area to extract lots of contamination from.
> The real question is "What is the smallest gross weight of contamination
> that a machine can accurately measure, before it is divided by the board
> area?" If it can measure accurately all the way to zero ug (either by
> measuring resistivity or by measuring conductivity, that's the machine for
> me! Anybody have one at a reasonable price?
By definition, a conductivity meter is linear down to 0 uS-cm +/- its
tolerance. Provided it is sensitive enough, then there should be no
problem. The ballistic meters we used in my days with the
Contaminometers were able to measure reasonably accurately down to 0.001
uS-cm for the CM series and 0.0005 uS-cm for the MCM series. With a 75%
solution, the best starting conductivity we could reasonably attain in
practice was about 0.0042 and for a 50% solution about 0.007 uS-cm. This
was the equilibrium point of deionising in the resins and leaching ions
from the air and constructional materials. We used the ballistic system
because, with DC systems, polarisation became a problem and reduced the
readings, while with AC systems, even at low frequencies, the capacitive
current became much greater than the resistive current (several orders
of magnitude) and phase detection became inaccurate with that
difference. Don't forget that the dielectric constant of the solution is
about 50!
I am confident, with more modern technology (remember I stopped
designing the systems when Multicore took the CM range over in 1991), I
could conceive instruments today that would meet your requirements
fully. However, I'm no longer in the business. My starting point would
not be an IPA solution. I would still use the ballistic conductivity
meter, though. Unfortunately, I think the technology of ICT or ROSE has
not progressed at all over the last decade. The main reason for this is
the competition and the pressure on prices, which have typically halved.
Something has to give to achieve this and this is at the level of the
instruments and what they are able to do. One cannot make a Rolls-Royce
out of the components of a Ford Escort.
I hope these explanations will help you.
Brian
>
> Thanks again.
>
> Peter
>
>
>
> Brian Ellis <[log in to unmask]> 15/04/2003 03:20 PM
>
> To: "TechNet E-Mail Forum." <[log in to unmask]>, DUNCAN Peter/Asst Prin Engr/ST
> Aero/ST Group@ST Domain
> cc:
> Subject: Re: [TN] ICT machines and high cleanliness requirements
>
>
>
>
>
>
>
>
> Peter
>
> As the original designer of the Contaminometer, but no longer involved
> with it, in any way, I guess that I'm as qualified as any to respond to
> you.
>
> You are partially right, but not entirely. First of all, it must be said
> that the 1.56 ug/cm2 eq. NaCl or, better, ug.cm^-2 eq. NaCl(which is the
> correct way of expressing the units cf. ISO 2000) dates from an
> interpretation of Hobson's original tests, way back in 1969. Please
> don't tell me that modern electronics are the same as they were 34 years
> ago. This figure is as antiquated as the dinosaur and has no meaning
> today. The voltage gradients on our substrates have increased by at
> least an order of magnitude, actually to the limit of acceptable with
> HDIS, even for low voltage operation. I therefore concur that 0.2 ug/cm2
> eq. NaCl would seem reasonable for modern electronics with conductor
> spacings of less than 0.2 mm, **provided** that it is measured
> reasonably accurately and for hi-rel end-products.
>
> So, how can this be done? First of all, you need to maximise the
> sensitivity. The first and obvious way is to use the UK Mod standard
> solution of 50% v/v IPA and not the US MIL standard of 75%, which is a
> relic of Hobson's work which had an integration time of 1 min, because
> it was sprayed on from a wash bottle, and not immersed with longer
> integration times. The 50% solution gives twice the sensitivity of the
> 75% solution.
>
> Secondly, contamination in these equipments is directly and linearly
> proportional to the delta conductivity, so is inversely and non-linearly
> proportional to the delta resistivity. Unfortunately, for fine
> measurements, because resistivity meters have a range, you are working
> at the cramped end of the scale. It is therefore more accurate to use
> conductivity with its linear scale.
>
> Then the board size enters into play. If you are stuck with a small
> board and a large tank, you simply measure on a plurality of boards. For
> example, if you have a tank size of, e.g., 250 x 350 mm, you can put in
> 8 boards 100 x 100 mm at a time, to obtain maximum sensitivity. Of
> course, your answer will be an average of the eight, but this is not a
> disadvantage.
>
> Then there is the reduction of errors. These can come from many sources.
> A bad one comes from CO2 absorption. This can be reduced by presenting a
> minimal surface area of solution at the air interface (rectangular,
> rather than funnel-shaped, tanks) and further reduced by software
> compensation of the absorption, as well as keeping the lid on during the
> test, with a minimal air volume over the solution. Another point is to
> keep the test down to under, say, 15 minutes, so that what you measure
> is the dissolved contamination and not leachates from the substrate or
> the instrument constructional materials. For the same reason, you should
> use solution at low temperatures (20-25 deg C). At higher temperatures,
> the heterogeneous structure of the polymeric surfaces tend to open and
> provoke more leaching from the substrate.
>
> I won't enter into the so-called "static" and "dynamic" methods. Our old
> CM-5 and MCM-2 models could use either. Each had advantages and
> disadvantages but, overall, the "static" method is the better,
> especially for low contamination levels. (I put the words between
> inverted commas, "", because they are very misleading misnomers.)
>
> If you observe all these points, it is possible to have a reasonable
> accuracy at 0.2 ug/cm2 eq. NaCl.
>
> For the anecdote, I used to manufacture the Microcontaminometer MCM-1
> and MCM-2, which had a sensitivity that it could measure down to better
> than 0.05 ug eq. NaCl absolute. This would have measured better than
> 0.01 ug/cm2 eq. NaCl on a 100 x 100 mm board (it had an interchangeable
> tank for a 4" hybrid substrate). I believe Multicore stopped the MCM-2
> in 1993. It was partially replaced by the CM-20, but I don't know what
> happened to that. I can only answer for the CM-1 to CM-5 and MCM-1 to
> MCM-2 series, which were the ones I designed.
>
> Anyway, what you wish to do is possible, provided you have the right
> instrument and know how to use it correctly.
>
> Brian
>
> [log in to unmask] wrote:
>
>>Morning, All
>>
>>I was interested to see the cleanliness level of <0.2ug NaCl eq/cm^2 that
>>is being adopted by many manufacturers, actually being stated in a data
>>booklet published by Concoat Systems. Having done a bit of poking around
>>into the world of cleanliness testing, I have a question for discussion
>
> if
>
>>you're up for it:
>>
>>Most ionic cleanliness testers are still marketed as measuring
>
> cleanliness
>
>>to the level specified in the MIL and IPC specs (1.56ug NaCl eq/cm^2).
>
> Fair
>
>>enough. Having looked quite closely into such machines as the Ionograph,
>>Concoat's own CM11 Contaminometer, Omegameters and the Zero Ion, which
>>seems to be teacher's pet in the IPC books (TR-583, anyway). I find that
>>none of the machines reads at all reliably at levels as clean as 0.2ug,
>>very especially whan the boards are small. Does anyone know, and is
>
> willing
>
>>to discuss, what work is therefore being done to improve ICT machines in
>>terms of the following:
>>a) using the test solution at much higher starting resistivity (e.g.
>>
>>
>>>200MOhm-cm +), assuming that at this level of resitivity, small amounts
>>
> of
>
>>contamination will have a greater impact on the drop in resistivity when
>>added (?)
>>b) how the test solution is presented to the resistivity probe for
>>measurement of these slight amounts of contamination. This is of especial
>>concern since small amounts of contamination don't affect the resistivity
>>of the test solution very much at current values (150MOhm-cm or so),
>
> which
>
>>thus remains relatively close to its starting value. If the test solution
>>is not homgeneous and the measuring portion of the probe is not in
>
> contact
>
>>with all the test solution that passes it, solution of a higher or lower
>>resistivity will not be measured and included in the machine's reading.
>>c) the capability of the probes to accurately measure these relatively
>
> very
>
>>small differences in high resistvity values.
>>d) the capability of the filters to clean the test solution without
>
> adding
>
>>further contanination back in.
>>
>>Suppose you have a board that only measures 4cm x 4cm, and it has to be
>>0.2ug NaCl eq/cm^2 clean. This means that the max amount of
>
> salt-equivalent
>
>>contamination allowable is only 6.4ug. To measure this board alone is
>>impossible - no machine is remotely capable of accuratey measuring a
>>contamination level this low - so a batch of these little boards would
>
> have
>
>>to be tested all together and the reading taken across the combined board
>>area. After being divided by the number of boards, you are left with an
>>average reading only. This is maybe OK if you have a lot of boards to
>
> test,
>
>>but not so OK if you're repairing only one or two of them and want them
>>clean again to original requirement.
>>
>>What is on the cards for future Ionic Contamination Testing equipment?
>>
>>Peter
>>
>>[This e-mail is confidential and may also be privileged. If you are not
>
> the
>
>>intended recipient, please delete it and notify us immediately; you
>
> should
>
>>not copy or use it for any purpose, nor disclose its contents to any
>
> other
>
>>person. Thank you.]
>>
>>---------------------------------------------------
>>Technet Mail List provided as a free service by IPC using LISTSERV 1.8e
>>To unsubscribe, send a message to [log in to unmask] with following text in
>>the BODY (NOT the subject field): SIGNOFF Technet
>>To temporarily halt or (re-start) delivery of Technet send e-mail to
>
> [log in to unmask]: SET Technet NOMAIL or (MAIL)
>
>>To receive ONE mailing per day of all the posts: send e-mail to
>
> [log in to unmask]: SET Technet Digest
>
>>Search the archives of previous posts at:
>
> http://listserv.ipc.org/archives
>
>>Please visit IPC web site http://www.ipc.org/html/forum.htm for
>
> additional
>
>>information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700
>
> ext.5315
>
>>-----------------------------------------------------
>>
>>
>>
>
>
>
>
>
>
>
> [This e-mail is confidential and may also be privileged. If you are not the
> intended recipient, please delete it and notify us immediately; you should
> not copy or use it for any purpose, nor disclose its contents to any other
> person. Thank you.]
>
> ---------------------------------------------------
> Technet Mail List provided as a free service by IPC using LISTSERV 1.8e
> To unsubscribe, send a message to [log in to unmask] with following text in
> the BODY (NOT the subject field): SIGNOFF Technet
> To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
> To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
> Search the archives of previous posts at: http://listserv.ipc.org/archives
> Please visit IPC web site http://www.ipc.org/html/forum.htm for additional
> information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315
> -----------------------------------------------------
>
>
>
>
---------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/html/forum.htm for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315
-----------------------------------------------------
|
