Troubleshooting Temp Sensors - Sample Roast Profile

ericshum

This graph looks really weird.

Details are as follows:
Beans: Peru Ceja De Selva Estate
Batch Size: 400gm
FC @ 11:37
Pulled @ 15:04

Result: Dark Roast (before SC)

Koffee_Kosmo

Looks to be a software refresh rate issue But I am not an IT geek so I could be wrong KK

UNM

a few guesses based on the graph and what passed by on chat last night. as you checked sensor reading at room temp that is likely to be good. just confirm the TC4C is set for a K type thermo couple (it probably is) The spike is probably a single incorrect reading on the USB input. There is usually no smoothing applied.  if it happens regularly at the same point we can check, but ignore it for now. environment temp probably being logged accurately if the up and down matches he TO heater switching on and off. bean temp is  wrong - as stated in chat, beans always hit first and second crack at a known range and your reading is way lower. where exactly is the thermocouple? can you post a photo? also remember - the exact numbers are not that important as long as it is consistent from roast to roast

beanflying

I have had this occasionally using the Meters and the Roastlogger software. My thoughts on this were noise on the I.R. to USB cables but it is possible there is something in the software? I think it would benefit from doing some smoothing of the numbers over time and even some simple error checking wouldn't be hard. i.e. if temp current to temp previous is more than x degrees different then it must be an error so ignore and use average of previous and the next to keep the curve smoother (sorry for rough logic). This would take the spikes out and no reason why it number of errors shouldn't be stored in the log to look at reducing them?

UNM

My software (aBourbon) displays a smoothed value on the LCD, but  delivers raw data to the PC monitoring the roast. The pBourbon software on the PC doesn't smooth, but could easily do so. Eric - what software you using?

beanflying

It is Roastlogger UNM so the same I am running.

ericshum

The E thermocouple is by itself, but the BT was guided to the location of the beans using a pito-tube. Did some tests and suspect that the bead was pushed into the tube and not measuring the temp of beans but the temp inside that small little tube. Quite confident that it's the case. Tried pulling the bead into the tube and the temp is much lower than E. As for the spike, I suspect it's due to the contact with the pot or any metallic surface, causing the temp to spike. It happened when something was moved. Happened again during the test runs without beans when the trying to pull the thermocouple back into the tube. It probably did touch the tube surface and hence the spike. Sent an email to the supplier. We'll see what he says. ON A SLIGHTLY OFF TOPIC: - The thermocouple we bought was probably not suitable. We suspect the resin for the braided fibreglass is rated only up to some 204dC. Could find any datasheet from the supplier we bought it from but saw something similar in Omega. I'm glad I only did one roast to realize this. I guess it's going into the bin. We'll have to go get a stainless steel probe one.

ericshum

Much better with a probe thermocouple. Spiking does happen on rare occasions but less (seems to be due to some grounding issues according to the supplier). It also seem to differ from one laptop to another. Can't seem to get my head around this. Having said that, here's the profile of the Peru (attached) that I have roasted. FC ~198dC at 9:40. This time, I was trying not to burp and manually control the temperature through the TO based on reading.