Letter from the Publisher: ± 0.1°F – The Debate

Heat Treat Today publishes eight print magazines a year and included in each is a letter from the publisher, Doug Glenn. This letter first appeared in Heat Treat Today's February 2023 Air & Atmosphere Furnace Systems  print edition.

When dealing with temperatures in excess of 1000°F, one would think that a ±0.1°F variation would not be a big deal. Apparently, not!

As of the most recent AMS2750 standard, 1/10th of a degree Fahrenheit matters — and if your process recorders are not recording temperatures down to 1/10th of a degree, you are out of compliance.

This is a big deal and a real hardship for many in the Heat Treat Today audience.

At the most recent Nadcap meeting held in Pittsburgh this last October, I had the chance to discuss this most recent stringent requirement with some of the people who were responsible for putting it in the standard. Even after talking to them, I’m not sure I fully understand why it is we went in this direction, and I’m not alone.

The Background

Here’s a very short explanation of how we got here. Both Revision D and E of AMS2750 required compliance temperatures to be ±2°F or ±1.1°C (“or ±0.2%” was added in Revision E). That pesky “.1” in ±1.1°C appears to be the source of this most current “situation.” The folks using °C were recording temperatures down to 1/10th of a degree, while the folks using °F — which was not a small number of people — were
not. So, the standards committee needed to make a decision on what to do about this discrepancy. The options were to round up or down or to the nearest integer for both °F and °C people OR require EVERYONE to record their temperatures down to 1/10th of a degree. After surveying end-users, the committee decided that end-users wanted to be required to record the 1/10th of a degree rather than round it up or down to the nearest integer. Thus, the new AMS2750 standard requires accuracy to 1/10th of a degree.

Thoughts

1. Even as I type it, it doesn’t make sense. Why would end-users want to record temperatures down to 1/10th of a degree? If you’re at 1750°F, a full 1°F amounts to only 0.05% of your total temperature. It is inconceivable that 1% makes that much of a difference in nearly 100% of all standard heat treat processes. In those very few processes where temperature tolerances ARE required to be that tight, SAE’s AMEC committee could have come up with a separate standard.
2. Most temperature recorders and reporting devices don’t currently allow for the display of anything to the right of the decimal, especially above temperatures at or above 1000°F. That’s because no instrumentation company in the history of heat treating ever anticipated that end-users would want to know, much less be required to record, anything to the right of the decimal.
3. Even if recorders and other instruments were capable of displaying 1/10th of a degree readings, most temperature sensing devices are  nowhere near that accurate. Special case T/Cs can do it in certain situations, but by and large, thermocouples are calibrated to ±2°F or higher. How much sense does it make to worry about recording 1/10th of a degree accuracy from a thermocouple (and wire) that is rated at ±2°F or ±5°F.
4. Let’s pretend for a minute that our thermocouples could accurately and consistently record temperatures down to 1/10th of a degree. The question that really needs to be asked is: Just because we CAN do it, does that mean we SHOULD do it? As stated earlier, for that vast majority of heat treatment processes a full degree of temperature variance won’t typically make a difference.

As some of the people I’ve talked to about this situation have readily admitted, well-intended quality committees such SAE’s AMEC committee, who have inadvertently started this little kerfuffle, are not perfect. This would be a case in point. The men and women who make up the heat treat industry’s quality systems are excellent people: highly detailed and well-motivated. But, as all of us are, they are prone to over-do the things they’re good at. In this case, that’s deciding to take it down to 1/10th of a degree when rounding to the next closest integer probably would have done the trick.

Postscript: I’m open to your responses to this column, positive or negative. And, assuming there is no foul language or threats of physical violence (!), we would be glad to publish your comments. Please let us know what you think: htt@heattreattoday.com

Find heat treating products and services when you search on Heat Treat Buyers Guide.com