Temperature Conditioning for ARM Impact Testing

(and a few other practical tips)

Minus 40º (actually the same number in ºF or ºC) is specified as the conditioning temperature prior to drop dart testing, according to the ARM protocol (“Low Temperature Impact Test”, Version 4.0, July 2003).  This relatively low temperature has been chosen because it provides a better differentiation between different polyethylene grades and different processing conditions.  At a higher conditioning temperature, differences might be harder to spot.

If you go to a company that supplies commercial catering equipment to hotels and restaurants, they should be able to offer a range of freezers that go down to -40º.  A typical domestic freezer operates at a higher temperature (approx. 0ºF).  Expect a practical operating range of 35-50ºF; this level of accuracy is sufficient to perform ARM testing.

The freezer I use in my lab (see Fig1) has an internal volume of approx. 5 cu ft and is plenty big enough for me.  I do lots of impact testing, so it should also be sufficient for most people.  If I bought one similar today, I would expect to pay the approx. $500-700.  Freezers for biomedical labs are a bit more expensive ($700-1000).  A proper “laboratory” freezer will be significantly smaller, may operate on a different mechanical basis and will carry a much higher price tag.  It will also hold to much tighter temperature tolerances, but this is not necessary for our purposes.

Typically, I stack my impact specimens for a single ARM impact test together in a large plastic bag, with the top specimen marked with an identifying number (see Fig 2).  Some people separate each specimen in a “toast rack” type device, but I haven’t found that to make any discernible difference to test consistency. I condition samples for 48 hours before I test them.

Just prior to starting an impact test, I take the stack of 24 specimens out of its bag and keep the stack upright in the freezer.  Then, when I’m ready to start, I open the freezer lid, remove one specimen and impact test it within a 30 second period.  With a bit of practice, you become fairly slick at this.  Samples that fail go straight into a recycling bin, but I retain samples that pass.  This is because I frequently carry out associated tests later on (eg Part Density, Yellowness Index and Melt Index) and I need some retained specimens.  If you only choose to do impact tests, this is probably unnecessary.

Once the specimens are cold, it can be difficult to immediately spot which is the mold side (ie outside of part) and which is the air side (ie inside of part).  To do the test correctly, the mold side should always be placed uppermost in the impact specimen holder prior to impact.  If you fail to conform to this, you will get extremely inconsistent data.  In order to be sure, I usually identify the mold side of each specimen with a permanent marker pen; nothing fancy is required, just some sort of squiggle, but this really helps when you’re in the midst of testing.

For ARM impact testing, my standard mold set-up produces 24 specimens, which is almost always sufficient to carry out the test as specified.  The ARM protocol stipulates that 10 data points are required for the least frequent event, which may be pass or may be fail.  This ensures that, if the initial drop height chosen is some way from the eventual Mean Failure Height, there are still sufficient data points to derive a statistically significant result.

When I’m trying to spot small differences between grades or processing conditions, I will usually measure the thickness of each specimen, prior to temperature conditioning.  My weapon of choice is a K-Metron from 493K, because the sensor can be connected directly to a spreadsheet via Bluetooth.  This means that you can get a quick tabulation and can automatically calculate the mean thickness and the standard deviation.  However, callipers and a clip board will do the same job, albeit at a slower pace.  Over time, I’ve tweaked my molding set-up to achieve mimimal thickness variations specimen-to-specimen, so I don’t always need to do this.  It depends what I’m trying to achieve.

If you want to adapt the ARM test for non-polyethylene materials, this low temperature conditioning may not be appropriate.  Many of the other materials available commercially (eg nylon, poplypropylene and acetal) have relatively poor impact below 32ºF.  In such cases, I usually condition at a controlled Room Temperature (23ºC / 73ºF) instead.  I’ve built a simple “hot room” for this: it’s a large box made from 4 inch thick foil-backed insulation board, with a tubular heater controlled by a PID.

Dr. Nick Henwood serves as the Technical Director for the Association of Rotational Molders. He has 30+ years experience in rotomolding, specializing in the fields of materials development and process control. He operates as a consultant, researcher and educator through his own company, Rotomotive Limited, based in UK.