Clifton Laboratories 7236 Clifton Road  Clifton VA 20124 tel: (703) 830 0368 fax: (703) 830 0711

E-mail: Jack.Smith@cliftonlaboratories.com

My wife uses a Taylor digital oven thermometer (model 1470) to measure the interior temperature of roasts and the like. The first one got a bit too close to a hot stove burner and the case melted, so I replaced it with a new one earlier this year.

Recently, she remarked that the new thermometer was reading differently than the old one, so today I compared the old and new oven thermometers along with two glass "laboratory" thermometers, two analog dial thermometers and my Fluke 189 Digital Voltmeter and 80PK-25 Type K thermocouple combination.

If you don't have one, by the way, I recommend the Taylor 1470. As my measurement data shows, both samples I tested agree well with my Type K thermocouple.
 

I compared the following instruments:

• Two glass tube "laboratory" thermometers, with scales from -10°C to +110°C. I bought these at a cheap price and the  accuracy matches the cheap price.
• Two Taylor 1470 Digital Oven  thermometers, one about two years old with a partly melted enclosure and one about six months old.
• A 2" diameter Taylor analog dial thermometer my wife uses in the kitchen.
• A 1" diameter Weston analog dial thermometer I use in the shop.
• Fluke 189 Digital Multimeter with an 80PK-25 Type K thermocouple probe, and Fluke Type K thermocouple adapter I use in the shop.

The only instrument of these seven with a stated accuracy that I can put my hands on is the Fluke Type K thermocouple, which is rated at ±1.1°C over the test range.

The figure below shows how the six thermometers compare with my Fluke thermocouple. I've added lines to show the thermocouple rated accuracy, which corresponds to ±2°F.

The two extreme test temperatures I used are melting ice water and boiling water. The thermocouple was less than 1°F different from theoretical temperatures of both extremes. I know that boiling water should be around 211.7°F for my elevation and today's barometric pressure and the thermocouple read 211.3°F.

At the lower end, the ice water mixture read 31.0 °F, a bit lower than the normal 32°F. However, the ice in the mixture is fresh from the freezer and is around 0°F. I did not have the patience to do the normal ice/salt and water bath calibration, and in cooking, the normal temperature range the oven thermometer is used with is in the 140 - 180°F range.

• The two glass tube thermometers don't agree with each other, or with the thermocouple. I recall paying about $1 each for these, so I suppose their accuracy is proportional to the price.
• The old and new Taylor oven thermometers track rather closely with each other. The minimum temperature these indicate is 32°F and below that the display says 'min" so I don't have a data point for the new thermometer at 32°F. At all other temperatures, the two agree with each other within a degree or so. Both read on the low side of the thermocouple data, however.
• The large (2" diameter) Taylor dial thermometer is quite close to the thermocouple data over almost the entire temperature range.
• The small Weston dial thermometer is also quite close to the thermocouple data.

Some guidance can be found in looking at a typical internal temperature range for differing degrees of "doneness" for a beef roast:

The interval is 20°F which suggests that an accuracy of about 10 to 20% of the interval is reasonable, i.e., an error of between 2 and 4°F should be satisfactory. If we assume the thermocouple is the true temperature, then both oven thermometers easily meet the 4°F error standard, and almost meet the 2°F objective. 

If the true temperature is somewhere around the average of the ensemble, then the two Taylor oven thermometers are even better.