This is my write-up if anyone can make any additions / improvements

Testing for RoHS/PbF solder
LTSIT (Low Temperature Soldering Iron Test)
Using a never used reserve thermocouple monitored variable temp OK
soldering iron set up for a tester.
185C/365F for SnPb and 217C/422F for production PbF seems about the
temperatures. Olive oil in a heated bath (with flammability precautions in
place) will go up to 210 deg C without smoking or boiling.
I calibrated the iron tip in an oil bath with a 250 deg C glass
thermometer, to a known 200 deg C with the heater LED slow intermittent.
Tried distinguishing known 60/40 leaded solder and known silver solder
95.5/4/.5 and had to drop the set point down to 185 C to melt leaded solder
wire on the tip and not melt silver solder wire. More than just the test
part of the tip in the oil so perhaps explains the lower set point. Swathed
the tip with tallow-looking conventional solder flux over the tip to assist
action. So distinguishing Pb/Sn and production PbF should be more
straightforward with this "test iron" . Testing on IC/small transistor
pad/joints (least heatsinking) and then by inference the same solder used in
the usual suspect
areas, ie larger pads, larger leads but PbF not up to the job.
2 more physical tests.
Grey mark on paper.
Scrape back surface of the solder and rub a specific number of times
and force, with
some copier paper over a cocktail stick. Leaded solder leaves a much darker
grey mark, nearer black, on the paper.
Indentation test
The following using a well worn automatic centre punch, so a new sharper one
for use only on soft metals is probably called for. This one, existing
spring replaced with a lighter one giving a load before trigger of 3.5Kg.
Comparing block of copper,Al, roofing lead and off the roll solder
wire,63/37 and 85.5/4/.5 silver solder laid over the lead as a soft anvil.
Steel rule with 1/100 inch markings and a x30 basic microscope
Measuring diameters of the indentations (x10 thou/mil)
Cu 1.5
Al 2.5
high tin PbF solder 3
SnPb 4
Pb 5
Had to decide whether a 2003 board had failed solder due to SnPb
or PbF and excess heat , where it was used.
Yamaha started using PbF about 2001 and Marshall Amps (UK)
maybe 2003 (before 2006 there was no requirement for Pbf
or RoHS stamps or stickers).
Could not decide by appearance,
good mirror finish, lack of pastiness but solder joints too shallow to tell
domed from conical, so probably SnPb.
Low temp soldering iron test, melted - SnPb
Paper test , decidedly black - SnPb
and indentation test much nearer Pb diameter than Al so - SnPb.
As I've not got a separate centre punch yet and swapping springs for this
test
I have to check with a lump of aluminium and a lump of lead
to calibrate each time.
Had not developed the chemical test at that stage.
Chemical test
Got some spirit of salts (32 percent hydrochloric acid) from a traditional
hardware shop. No colour reaction found heated with leaded solder but
on cooling a white deposit would suggest lead chloride and then
Potassium Iodide KI added, now seems a number of repeated heating/cooling
cycles is required.
I increased sample sizes to about 3cc of 32% HCl and about 30 cu-mm of
solder wire. About 1/3 of the wire scraped with a razor blade into
scrapings and remainder
added as round and the part round remnant . NB BEWARE when boiling the
hydrochloric acid , point
the test tube away from you, perhaps a loose plug of cotton wool in the
mouth of the tube.
Especially important on re-heatings as due to lack of nucleation points , or
whatever the reason,
it does not slightly boil first like a kettle but can erupt without warning
, so only gentle slow warming.
After boiling and cooling twice
then perhaps about 30 cu-mm of KI added and a couple of repeated boiling and
cooling. For the leaded one bright orange crystals? soon appear on forced
cooling and flecks of glistening gold colour the second time of heating
cooling after KI added . The glistening gold agrees with my chemistry book
for
lead testing with KI, don't know about the orange. The colours and crystals
disappear on boiling.
The lead free one , no colours at all in the liquid but the metal goes
black (same with leaded sample but can only be seen when boiled
and clear of precipitate). With repeated boiling then still a lot of gas
comes off the metal for
some time after removing from the heat , unlike the leaded sample. Maybe
acting on the 5 percent silver content.
The KI came from a kids chemistry set in a charity shop. One way around kids
and chemistry and product liability, absolutely no name. Nothing on the box
and nothing in the accompanying book for makers name or even country of
origin.
Looks as though regular flicking of the test tube is important.
Repeated samples and again nothing appearing in the non-lead sample. The odd
flecks that would appear when shaken were probably reduced flakes of the
original scrapings, again black colour. Looks as though 30 cu-mm is about
the right amount of KI with 30 cu-mm of solder and initially 3cc of HCl. I
can only estimate using a small spatula that is in fact a stainless steel
lobster pick, fork+spoon. The 4mm wide spoon end is ideal for this. No
orange colour this time (initially) . Also useful for cleaning any crud
from the inside of test tubes as fairly long, spoon bit first
and then fork end if really stubborn. Once a very slight yellow tinge to the
liquid appears and at boiling and through cooling , much flicking of the
tube with a finger, then instead of flecks that look like golden flakes of
skin you get a dense mass of more like cream-coloured or
sulphur dust colour soap flakes or fern/fan shaped. Reheat
and all this dissolves well before boiling point. Allow to cool untouched
and then just a small scattering of fine gold-coloured flecks when you flick
the tube. The orange colour seems to only appear after repeated heating and
cooling, the second SnPb sample is now starting to show this orange colour
after 5 or 6 heat/cooling cycles and increasing each time.
Second sample of PbF shows no colouration or precipitate, again , so
seems repeatable and differentiable. Left the PbF test tube laying around
and after about a week the solution was quite red-orange and some orange
crystals.
Tin Iodide/Stannic Iodide is red-orange and so that delayed colour in the
SnPb
samples also is probably indicative of the tin content.
Potassium dichromate after HCl treatment of SnPb solder gives a dense green
colour of
chromium chloride and a non-indicative white precipitate.
Repeated a few weeks later with some solder from a 2006 amp covered in RoHS
and PbF
stickers. No gold flecks or yellow or orange colours the first day just non
glistning flecks and bits
of pcb lacquer scraped off with the solder. Second day very light tint of
yellow becoming orange and then
orange-red over the next week with time and odd tube shaking, just
indicating presence
of tin and no lead.
Physical tests on the board, LTSIT on a 2nF, 100V cap with 2mm diam pad, 1mm
wide trace
and 0.5mm pin , so low heat capacity.
First pin liberally covered in flux, took 5 seconds of heating before
melting and immediately , 0.5sec,
greyed over on removing iron. The other pin , no flux, it would not melt.
Paper test , very light grey and indentation test on 1/4 inch socket pin
joints , so
chunky solder of about 3mm diameter gave 2.5 units compared to 2units for Al
and
5units for lead sheet


--
Diverse Devices, Southampton, England
electronic hints and repair briefs , schematics/manuals list on
http://diverse.4mg.com/index.htm

worked quite well , a sample tube , 45mm long,on a scanner bed, required
disturbing just before the beam passed under. General orange background
colouration due to the tin iodide and the "gold" scintillating flecks from
the lead iodide

http://diverse.4mg.com/lead_iodide.jpg