I've just spent a frustrating afternoon trying to work out whether I can trust my thermocouple/PID combination in my Lineastein. It all started when I compared their brew boiler temperature with the temperature in the portafilter basket with an embedded thermocouple. It appeared that there was a12-15 degree F difference between the 2. This didn't seem right as other people seemed to be reporting a 4 degree F difference. I finally pulled out the their 2 thermocouples from their thermowells and put those together with the puck PID and a separate probe fitted to a Techtronix meter into a cup of water. At a temperature of about 150F there was an 8-10 degree variation between the 4. I then tried a lower temperature of around 90F and they were within a degree or 2. I realised at this point that I had no reference probe to judge others against, so i tried a few rough reference points; boiling water, my body temperature, and ice cubes in water. The variation at "boiling" is about 6F, about 5 at body temperature, and about 12F at chilled water. A further problem is that the highs and lows don't relate to the same probes; i.e. the Omega K type is high at boiling, and low at body and chilled water temps.

The probes are all K type; 1 Omega, and 3 off-brand eBay specials.

My question is this: I want to try Schomer's precise 203.5F (I've just bought his book), but I've now got no idea of which probe I should use to know what actual temperature my boiler/brew head is running at. I have a Fluke laser temperature probe, but even it's accuracy at 100C is something like +-3C.

Do I change to RTDs? It seems like I still need some way to reference/calibrate my measurements.

A way to calibrate a thermocouple and meter is to "soak" the thermocouple in a "bucket of steam" and make appropriate corrections for your altitude and barometric pressure at the time.



I'm just curious - exactly what Tektronix meter are you using and how are you "fitting" the thermocouple signal to that meter? The OEM PID retrofit to a Linea uses a sheathed Type J thermocouple adapted to a 1/2" BSPP fitting. I know you said you are using Type K's but how are you fitting this to the brew boiler? Pics would be worth many words.

The Omega should be ok; I'm surprised that it isn't a more or less consistent offset. A few atomic % off on the TC alloys and you'll get a different mV; not just high or low, but with a different slope (I know they're not perfectly linear).

I think type T is supposed to be slightly more accurate if you want to stick with TCs.

What's wrong with RTD?

It's more likely cross talk on a bad multi-channel meter (I have an old Omega badged one that does this) than the TCs. Test the TCs individually for boiling and ice point. If we were sitting in a bar, I'd bet a round on them being within 1F, 0.5C of accuracy.

The head slap mistake is mixing up J and K types, since they are close enough at low temperatures to just seem inaccurate.

Omega.com has plenty of information on thermocouple accuracy. Generally thermocouples work or they just don't give a reading. Chemical composition of thermocouple wire has a lot to do with their accuracy. Higher purity wire that is strain free gives a better answer. You are mixing too many things in your experiment. Use one readout for all of the probes. Calibrate according to Eric S suggestions. Temperature uncertainty for the method mentioned by him is on the order of 0.05C, as long as your barometric pressure correction is good.

Sounds to me like you have different thermocouple types, or readouts that are bad. Even the worst thermocouples should be within a couple of degrees at boiling water temp.

-greg

NickA wrote:I've just spent a frustrating afternoon trying to work out whether I can trust my thermocouple/PID combination in my Lineastein. It all started when I compared their brew boiler temperature with the temperature in the portafilter basket with an embedded thermocouple. It appeared that there was a12-15 degree F difference between the 2. This didn't seem right as other people seemed to be reporting a 4 degree F difference. I finally pulled out the their 2 thermocouples from their thermowells and put those together with the puck PID and a separate probe fitted to a Techtronix meter into a cup of water. At a temperature of about 150F there was an 8-10 degree variation between the 4. I then tried a lower temperature of around 90F and they were within a degree or 2. I realised at this point that I had no reference probe to judge others against, so i tried a few rough reference points; boiling water, my body temperature, and ice cubes in water. The variation at "boiling" is about 6F, about 5 at body temperature, and about 12F at chilled water. A further problem is that the highs and lows don't relate to the same probes; i.e. the Omega K type is high at boiling, and low at body and chilled water temps.

The probes are all K type; 1 Omega, and 3 off-brand eBay specials.

My question is this: I want to try Schomer's precise 203.5F (I've just bought his book), but I've now got no idea of which probe I should use to know what actual temperature my boiler/brew head is running at. I have a Fluke laser temperature probe, but even it's accuracy at 100C is something like +-3C.

Do I change to RTDs? It seems like I still need some way to reference/calibrate my measurements.

Not trying to distract from the technical nature of the OP, but...

@NickA - Are you trying to pull Vivace Dolce, or some other blend/SO with a "known" sweet spot of 203.5F? If not, this will likely prove to be much work for poor return. Many coffees taste better quite far away from 203.5F.

I have no disagreement that temperature controlability is an important aspect of espresso extraction--only with the implication that there is "an ideal temperature for all coffee."

...and now back to your regular program!

Mike

Hi Mike, no, not Vivace coffee, but a theory (Schomer's precise 203.5) is worth investigating to see what it offers. The point I was making was that whether I agree with Schomer or not; it would be good if I could know reasonably accurately what my brew head temperature is, rather than working from relative temperatures, i.e. this pull is probably 2 degrees hotter than the last one.

I had already tried changing the input settings on the PIDs between J and K because I know that they are reasonably close and that what I am describing sounds similar to the behaviour of mixing up probe types. The meter is a Tektronix DMM870, and trhe PIDs are 3 Watlow 935s.

The 935's are good controllers. But the ones on eBay may be quite old. Old enough to require recalibration. In my experience, an error of 4F with an uncalibrated 935 wouldn't be surprising. The good news: they are fairly easy to calibrate using a millivolt source and ice point simulator.

Jim

Hi all, thanks for your input. Greg, I am taking your advice and will reduce vaiables. Jim, I am going to attempt calibration of the Watlow 935s. For me it is easier to do the RTD calibration as I have the resistors and higher voltage sources, whereas I don't have the millivolt sources needed for Thermocouple calibration. (Now I just need to make an adaptor for the RTD probe to fit the steam boiler ...)