Peter
Yes and no! In fact, the MCM-2 was slightly cheaper than our flagship
CM-5 model, which was our best seller, ever. It was perhaps a bit of a
solution looking for a problem, remembering that the smallest tank had a
useful volume of only 3.38 cm3 of solution (plus a little more again in
the measuring circuit and in the measuring cell)!
I agree that the CM-11 is a low-cost machine for run-of-the-mill
applications and was a replacement for our CM-3 model.
I see you are talking about resistivity of solutions. This is a relic of
the earliest work by Hobson et al. It is not clever to use resistivity,
which is why I used conductivity, even from the first CM-1 in 1977/1978.
Contamination is proportional to the delta-conductivity, not
resistivity. Resistivity meter ranges are finite. The early OmegaMeters
(used for the MIL-P-28809A equivalence value determination) had a
resitivity meter with a range of 0-20 megohm-cm (as Hobson also used).
If you regenerated the solution for more than a few minutes, the meter
would go off scale. You could regenerate it further, say to 200
megohm-cm, put a circuit in and the resistivity would drop to, say, 21
Mohm-cm and you would conclude it had zero contamination, because the
needle didn't budge. Later models went up to 50 Mohm-cm and then higher
again. I don't know what current models will measure. If you use
conductivity, not only is it linear, but it has no dead zone.
The reason the US specs say 75% IPA is also a historical legacy from
Hobson's work. His original method was to spray the stuff onto the board
from a squeeze-wash bottle, with an integration time of 1 minute. It was
necessary to dissolve as much rosin residue as possible in that time. If
you immerse the board in solution with much longer integration times,
you do not have that constraint. The UK Ministry of Defence did a lot of
work (2 years) on the "best" solution for immersion methods and came up
with 50% as being the best compromise between solubilisation, precision
and sensitivity. This was reflected in their DEF STAN 00-10/3 (1986)
standard which disallowed 75%. The European Space Agency, in 1984, also
did an evaluation and came to the same conclusion. My own tests
concurred, although these predated the ESA and MoD tests. Chapter 23 of
my book gives a discussion on this. Of course, there is also less
evaporation and fire hazard at 50%, as well.
OK, so why 50%? Well, let's say that for a given contamination level
under given test conditions, you have an initial conductivity of a 75%
solution of 0.005 uS-cm and it rises to 0.05, your delta-cond is 0.045.
Under identical conditions with a 50% solution, you may start with 0.01,
but you would end with 0.10 and a delta cond of 0.090. Which would you
say is the easier to measure with precision? (Add some practical
tolerances to see this.). The starting conductivity is not material, it
is always the delta conductivity that counts and the reason that the
starting conductivity is higher with 50% is exactly the same reason as
you get twice the sensitivity.
Second argument: if the solubility of, say lead carbonate, for example,
in cold water is 0.0011 g/l and that in alcohol is 0.000043 g/l, which
will saturate first, 75% or 50%? So ionic metal salts, resulting from
the soldering chemical reactions, will be less likely to saturate and
thereby give false results.
I'm sorry, your reading is wrong. 75% was first and dates from
1968-1969, for the reasons I stated. 50% was not proposed until almost a
decade later (I think I was one of the first, having published my tests
about 1978). Saturation of organics has never been an issue. It was a
question of speed of dissolution and more hydrolysis that were the
issues. With immersion systems, as I said, the slower dissolution rate
of rosin (in practice, about 30% with many practical rosin fluxes,
hardly used today) is a non-issue because you just wait an extra minute
or two before the reading has stabilised.
Incidentally, the lowest conductivity that is normally achievable with
75% solution is about 0.0035 uS-cm or about 285 Mohm-cm. This can vary
slightly because of different DI resins, constructional materials, ait
interfaces and so on. This ties in with the 250 Mohm-cm you mention. It
is physically and theoretically impossible to obtain better than 400
Mohm-cm, so you can safely ignore this "advice". In any case, as I
stated, the starting value is not material.
Brian
[log in to unmask] wrote:
> Hi, Brian,
>
> I share in your lamentation over the demise of the much more accurate
> contaminometer. I have previously spoken at some length with Graham
> Naisbitt about the "current" contaminometer model, the CM11. It's main
> advantage is that it's cheaper than most comparable instruments on the
> market, but not demonstrably different or better than others (sorry,
> Graham) inasmuch as it suffers from the same limitations as all the others.
> i.e. it's designed to be compliant with IPC and MIL specs and nothing
> better.
>
> I think your much more sensitive instrument was allowed to die for a good
> "Chicken and Egg" reason- folks didn't perhaps understand the nature of the
> cleanliness testing process, and if they simply wanted a machine to tell
> them if their boards passed IPC or MIL Std specs, the "normal" machines
> would do that, so why buy a machine that was doing the job differently
> and/or, (dare I say?) more expensively? Actually, I've no idea if it was
> more expensive than other machines or not, but it might be a reason why it
> isn't still around today.
>
> Today, though, we do need machines that are more sensitive. Someone in an
> earlier post mentioned requiring one with a starting resistivity of 450
> MOhm-cm. I would welcome that, but it would require a whole different
> exrtract solution to reach that level of resistivity and react linearly to
> contamination. The old DI Water:Alcohol mix still has some life ahead of it
> if the ROSE Tester software could be recoded to set a starting cleanliness
> level at around 250 MOhm-cm rather than 150 MOhms that's the current upper
> limit for most machines at present. 250 MOhms is about the limit for this
> type of solution (75% Alcohol : 25% DI Water), but it would allow much
> smaller levels of contamination to be measured accurately than present
> set-ups are capable of.
>
> Brian, large respect to your knowledge and experience but may I dare to
> challenge you on a point? You also talked of 50:50 test solutions, and at
> one time I would have agreed with that ratio as being better than 75:25. I
> changed my mind because of two factors:
>
> 1) The lower proportion of alcohol seems to reduce the achievable
> resistivity level of the test solution, thereby reducing the sensitivity of
> the machine that's so crucial to measuring small amounts of contamination.
> 2) There's a trade off when determining the Alcohol to Water mix between
> how much alcohol is required to dissolve non-water-soluble residues to
> release trapped water-soluble residues, and how much water is required to
> wash off the water-soluble residues that have been freed by the alcohol. My
> reading on the matter indicates that 50:50 mixtures were succeeded by 75:25
> mixtures as they were found to be more effective, especially when there
> isn't much residual contamination to wash off anyway. (That's presuming the
> cleaning process is working to required levels).
>
> Thanks for joining in the skeet shoot.
>
> Peter
>
>
>
>
>
> Brian Ellis <[log in to unmask]> 26/05/2003 02:12 PM
> Sent by: TechNet <[log in to unmask]>
>
> Please respond to "TechNet E-Mail Forum."; Please respond to Brian Ellis
>
> To: [log in to unmask]
> cc: (bcc: DUNCAN Peter/Asst Prin Engr/ST Aero/ST Group)
> Subject: Re: [TN] ROSE for Flip Chip
>
>
>
>
>
>
>
>
> Peter
>
> To some extent I agree with you, but not entirely.
>
> Using the pasted area is no better than using the board area. The paste
> is deposited on metal. The danger lies on substrate and the paste
> chemistry spreads over the substrate during preheat and reflow.
>
> I agree that most current models of equipment do not meet the
> requirements of cutting-edge industry. This is why I hatched the
> Microcontaminometer in 1983, culminating in the MCM-2 in the late 1980s,
> which was able to go down to 50 ng of absolute sodium chloride with a
> 50% solution (in other words over 4 orders of magnitude more sensitive
> than your instrument). The conductivity measuring system was able to
> measure 1.5 parts per billion (ppb) of salt in the solution! Multicore,
> in their wisdom, decided to abandon this model after they took over the
> range. They did make a vague effort in the late 1990s to resuscitate the
> idea with the CM-20, I believe, but I don't think that got much beyond
> the pre-production stage (I don't know for sure, as there were a lot of
> communication difficulties). Maybe Graham Naisbitt of Concoat can tell
> you what the current situation is, as I understand they are the present
> successors to the Contaminometer legacy (I have absolutely no interest
> in it, either technical or financial).
>
> I agree that to calibrate and drive these ultra-sensitive instruments is
> not great fun, unless you are a masochist, but they were able to produce
> the desired results accurately, with care and, above all, knowledge. I
> had several customers for them within the great names of the Swiss watch
> industry (including Swatch!), the hybrid circuit industry and component
> manufacturers and they were all able to use them without problems, even
> if it required a day's training.
>
> I see, in my book pp 316-320, I was suggesting that 0.5 ug/cm2 eq. NaCl
> was necessary, even then (written 1983-5, published 1986, for hybrid
> and SM circuits). So, what you are saying has a 20 year history, already!
>
> Hope this helps.
>
> Brian
>
>
>
> [log in to unmask] wrote:
>
>>Ah, Glen!
>>
>>I've been battling this one for far too long now myself, and one of the
>>questions I posed was just the one you are posing here - is it imore
>>realistic to calculate the solder paste area, since this is where all the
>>flux contamination wil be concentrated, or go with the spec requirements
>>which gives a cleanliness requirement calculated from 2x entire board
>
> area?
>
>>I have my own opinion which I'll explain in a moment, but let's look at
>>history first. The cleanliness figures in the IPC and MIL specs are now
>>over 30 years old and date from the days when components were all PTH and
>>wave soldered. With wave soldering, the entire board area is fluxed, and
>>with the technology of those days, trace weights, widths and spacings
>
> were
>
>>much greater than they are today. Hence the cleanliness level required
>
> was
>
>>not as stringent as it needs to be today. The cleanliness level, taken
>
> over
>
>>the entire board area made sense then, but has not been updated since to
>>account for the much finer traces, closer spacing and SMT technology of
>>today. SMT technology, of course, largely uses Convection Reflow
>
> soldering
>
>>techniques in which solder paste (flux) is deposited only in localised
>>areas.
>>
>>Doug Pauls from Rockwell Colins will probably give you a very
>
> knowledgeable
>
>>answer to your question, but my opinion is this:
>>
>>a) if you want your boards to meet IPC or MIL spec, it can be easily done
>>using the current ROSE test equipment, which also has changed very
>
> littlle
>
>>in principle in 30 years. Take whatever reading you get from the machine
>
> at
>
>>face value, divide it by the total board area and you have the
>
> appropriate
>
>>figure against which to judge if you've passed or failed. However ...
>>b) Industry concensus seems to have it that a much tighter cleanliness
>>figure is now required than in days of yore - 0.2ug NaCl eq cm^-2 instead
>>of 1.56ug is the figure that's been bandied about, based, I believe, on
>
> SIR
>
>>and Ion Chromatography results for modern boards. It's a figure I prefer
>
> to
>
>>believe than the IPC spec figure.
>>c) Using the solder pasted area as the divisor for the cleanliness result
>>makes the test much harder to pass, but will give you (to my mind) a more
>>realistic idea of how clean the fluxed areas are. I think, too, that the
>>real truth about overall board cleanliness will lie somewhere in between
>>the two areas, though where I cannot say for sure but probably nearer to
>>the pasted area than the overall board area.
>>d) as boards get smaller and clealiness level requirements become
>
> tighter,
>
>>we run into problems with the ROSE testers, which are still happily
>>claiming compliance with the IPC and MIL Specs and tend to assume a board
>>area that may be much larger than you normally deal with. The problems
>
> are
>
>>those of working at the extreme limits of the machines' capabilities.
>
> Sure,
>
>>you can continue to do what you do now, which is to test a number of
>
> boards
>
>>at the same time to, as you say, give the machine something to measure,
>
> but
>
>>I believe it's time that the ROSE tester manufacturers looked to updating
>>their technology.
>>I've been giving Mike Konrad at Aqueous Technologies a hard time,
>
> demanding
>
>>"official" figures for how well their Zero Ion reads at low levels of
>>contamination, because from tests I have carried out, machines tend to
>>start reading inaccurately when total contamination weights drop much
>
> below
>
>>1000ug NaCl eq. I have boards that measure just over 9" x 6" and have a
>>target cleanliness figure of 0.2ug NaCl eq cm^-2. This means that the
>>maximum permissible contamination weight must not exceed 143ug per board.
>>The ROSE test tank I have is large enough to allow two boards to be
>
> tested
>
>>at one time, which means that the machine has to be capable of reading
>>accurately 286ug NaCl eq (board area is irrelevant as it's just a divisor
>>that you programme into the machine yourself - it doesn't affect the
>>machines ability (or otherwise) to detect and measure the contamination
>>level. 286ug is well below the level at which the machines seem to read
>>accurately, or consistently. The variance at these levels is also great,
>>and if my tests are accurate, makes the measurement at these
>
> concentrations
>
>>meaningless.
>>
>>As more people move to "no-clean" technology, the only folks left using
>>ROSE testers will be the Class 3 board builders who continue to use
>>water-soluble fluxes. Maybe that market is not big enough to justify
>>re-evaluating the capabilities of the testers, in which case ways round
>
> the
>
>>capability problem have to be found.
>>
>>Glen, whether you choose to use overall board area or pasted area for
>
> your
>
>>cleanliness calculation is up to you. If you want a better idea about the
>>board cleanliness, use pasted area, but if you want the boards to pass
>
> test
>
>>for shipment, use overall area adn no-one can legitimately criticise you.
>>
>>Hope some of the other gurus will give me a debate on ROSE tester
>>capability.
>>
>>Peter.
>>
>>
>>
>>Glenn Pelkey <[log in to unmask]> 23/05/2003 09:59 PM
>>Sent by: TechNet <[log in to unmask]>
>>
>>Please respond to "TechNet E-Mail Forum."; Please respond to
>
> glenn.r.pelkey
>
>> To: [log in to unmask]
>> cc: (bcc: DUNCAN Peter/Asst Prin Engr/ST Aero/ST Group)
>> Subject: [TN] ROSE for Flip Chip
>>
>>
>>
>>
>>
>>
>>
>>
>>Hi Technetters,
>>
>>
>> My buddy and I have a disagreement on using Resistivity of
>
> Solvent
>
>>Extract (ROSE) for measuring cleanliness on a Flip Chip board. So, we're
>>throwing it to league of experts on this forum for feedback
>>
>>
>> Here's the story: The board and assembly are very simple. One
>>flip chip die at about 0.002 square inches and one board about 0.490
>
> square
>
>>inches. Flux is applied only to the flip chip die, reflowed, and DI
>
> inline
>
>>washed.
>>
>>
>> Here's the disagreement: I say use 2X the surface area of the
>
> die
>
>>for the calculation. My good friend says use 2X the board surface. Of
>>course, you can see one number is much higher than the other in ug/sq
>
> inch
>
>>NaCl equiv. My logic for using only the die area is to understand the
>>process cleanliness at that location only, but I recognize the entire
>
> board
>
>>surface could be cross contaminated in some manner. We also loaded the
>>machine with 32 of these units to get something measurable. When we only
>>put 4 in, nothing showed up, but the smallest entry for surface area is
>
> 0.1
>
>>square inches.
>>
>>
>>Thanks for your input!
>>
>>
>>Glenn
>>
>>
>>---------------------------------------------------
>>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
> -----------------------------------------------------
>
>
>
>
> [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
-----------------------------------------------------
|
