Thermometry for Heat Gun / Dog Bowl Roasting - Page 3

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
JimG (original poster)

#21: Post by JimG (original poster) »

another_jim wrote:Would drilling two holes, and mounting each TC wire separately with a nut and some conductive silver compound work?.
That got me thinking about another possible alternative. Let's assume that we could locate some PTFE or other non-conductive threaded parts that can tolerate ~500F.

Drill one hole through the bottom of the dog bowl. From the top, pass a pan head PTFE machine screw through the hole.

Now, on the bottom side of the bowl, make a small fork out of the stripped leads of the t/c wire. Position the bare leads on either side of the PTFE screw shank.

Slide a PTFE flat washer, then a nut (could be metal) over the end of the screw. Tighten the nut, trapping the bare t/c leads tightly between the flat washer and the bottom of the bowl.

If everything worked according to plan, this would give 2 junctions with the bowl around 1/8" apart, making a circuit just about like what you would get from spot welding. Using PTFE (or similar non-conductor) prevents the formation of unintended junctions. It also minimizes the added mass near the junction which might otherwise cause undesirable temperature lags.

A few possible flaws with this approach would be:
  • Melting/burning of the PTFE hardware
  • Poor electrical contact between bowl and t/c leads --> high resistance
  • Oxidation/corrosion at the two junctions (long term)
  • Physical abrasion of the head of the screw that ends up being inside the bowl
Might be worth a try?


User avatar

#22: Post by espressme »

I believe that that would not work. The two wire types must be in tight proximity ( read twisted or spot welded.)
You can buy a bolt ended thermocouple.
A short pan-head or carriage bolt though the bowl from the top and drilled to take the TC would work. The bolt hole would locate and allow the TC to be positioned up into the bowl and spinning beans.
Just my take on the problem.
richard penney LMWDP #090,

JimG (original poster)

#23: Post by JimG (original poster) »

espressme wrote:I believe that that would not work.
Da*n, I was afraid someone would notice that :D

Actually, rather than the beans themselves, this system is designed to measure the temperature of the steel pan. The idea is to gather kind of an average bean temp across a large surface. Makes it less susceptible to localized effects.

For the purpose of measuring the temperature of the bowl, a 3-piece thermocouple actually does work (the steel in the bowl is the third metal). Some of the older publications on thermocouple applications describe something very similar for spot checking surface temps of enclosed vessels.


User avatar

#24: Post by espressme »

Thank you! I sit corrected! :D
richard penney LMWDP #090,

User avatar

#25: Post by Martin »

to provide a path to reducing the amount of time/experience needed to arrive at a point where consistent & acceptable results can be achieved. . . . in my current state of mind I'd trade the warm fuzzy for some solid data . . . . still searching for something that gets me Hottop (or better) quality at DB prices with HG-level control.
Faster and cheaper with science?

HG/DB roasting uses "solid data" that comes from many sources. Key, of course, are the immediate sensory inputs of color, sound, smell, tactile/heat, time, etc. There's also a wealth of funded knowledge about roasting reactions generally----matters such as drying, relative and absolute time spent in "stages," and so forth.

HG/DB practices are already much more data-rich than "warm and fuzzy" implies. Incorporating bean temp data and maybe profiles into your existing data is interesting and worthwhile, but not likely to be a shortcut. Seems that even a modded popper (cheap) is more promising for learning about temps in particular.


#26: Post by frustrated_uk »

Here's how I roll...

The flour sifter conducts heat nicely, eliminates scorching and creates a bed of hot air beneath the bean pile.

Attaching a thermocouple is as simple as twisting the wire into the mesh. Secure and accurate. Simples.

User avatar
Team HB

#27: Post by another_jim »

Beautiful; thanks.
Jim Schulman