Humidity sensor on exhaust - other gas sensors ? - Page 4

#31: Post by **ira** » May 12th, 2023, 1:20 pm

If you get input into Artisan, as I understand it, a sensor is just a line on a chart and you get to define how it relates to the input. But I could be completely wrong.

May 12th, 2023, 10:36 pm

Yes, in Artisan you can import custom sensor data to plot.

In the Ikawa Pro App when you export the data to CSV you get an RPM sensor and set point for example, on top of my head, I'm traveling and don't have the files here, the CSV format has a first line with date, and the the events time ( TP, FCs, FCe, ... ) then a second line with strings: Time1, Time2, BT, ET, Event followed by custom measures, such as RPM, SetPoint or humidity. Than the data, line by line.

Regarding QWIIC, I like it, but it limits the sensors, I think anyway some custom PCB might be needed ...

About logging the thing that led me to the reversing Ikawa's BT protocol was the synchronization, I like that you don't have to play with the data after capturing. You can do multiple roasts and forget about it until you want to analyze it.

Most boards have wifi , a small http server could be used to download the data. Or an easier solution would be a service online that collected the data, from there you could export to any format : excel, artisan csv or Cropster. No app needed, just a simple website.

#33: Post by **GDM528** » May 14th, 2023, 3:14 am

Esteve wrote:Regarding QWIIC, I like it, but it limits the sensors, I think anyway some custom PCB might be needed ...

I spent several hours pouring through chip datasheets for half-dozen QWIIC humidity boards, so actually kinda grateful there weren't more

I'll be testing a Sensirion SHT31-D humidity sensor so we can compare the results. Agreed that a custom PCB would be optimized for the application, but that could present a pretty significant barrier to entry.

Question about the ESP32 controller board: could it directly BT-pair with the roaster? I'm wondering if the humidity (and other) data can be synchronized with the roaster data that way. It would only require reading data from the IKAWA. The ESP32 is capable of far more than simply polling I2C devices - it would be cool if it could collate and format all the data in one place before uploading to Artisan, IKAWA Pro, etc.

#34: Post by **GDM528** » May 14th, 2023, 3:34 am

Brewzologist wrote:My drum roaster exhaust includes both drum fumes and any heated air from the burner that comes into the back of the drum. Assuming this device has merit I wouldn't be bothered by drilling a hole for a new sensor. I don't want to use the tryer or access door because it will affect the temp inside the drum too much.

I have a fan that pulls exhaust from the drum so there is a net negative pressure. Not entirely related to what you are trying, but I have had a pressure sensor connected to my exhaust stream via a ~1.5ft length of tubing for a long time. I bent a piece of brass tubing pointing downstream and have never had it clog up.

Gas combustion fumes have a lot in common with roasting coffee fumes, so once they're mixed it's game over. Care to swag just how negative is the air pressure in the chamber?

May 14th, 2023, 10:14 am

Question about the ESP32 controller board: could it directly BT-pair with the roaster?

Yes, it could and would have plenty of processing power. I wouldn't do it though because implementation is a pain in the ass, Ikawa uses Google protobuf, which with a good IDE and easy language was quite boring to implement, on arduino might be quite worse...

Another reason is that you can only have one devices paired at a time. That's one of the reasons I implemented the profile sending. You'd have to open Ikawa app, send profile, close it, make ESP32 pair and then roast.

I spent several hours pouring through chip datasheets for half-dozen QWIIC humidity boards, so actually kinda grateful there weren't more I'll be testing a Sensirion SHT31-D humidity sensor

Looks like a good option. Skimmed fast and the response is 8s, correct ? Before trying any sensor the response time is what worried me the most, but after trying I don't think is as important.

Post the results

#36: Post by **Brewzologist** » May 14th, 2023, 2:18 pm

GDM528 wrote: Gas combustion fumes have a lot in common with roasting coffee fumes, so once they're mixed it's game over. Care to swag just how negative is the air pressure in the chamber?

It will bend the flame from a lighter gently into the tryer point. I keep the fan just high enough to remove chaff.

#37: Post by **GDM528** » May 15th, 2023, 3:53 am

Esteve wrote:Another reason is that you can only have one devices paired at a time. That's one of the reasons I implemented the profile sending. You'd have to open Ikawa app, send profile, close it, make ESP32 pair and then roast.

Ooops, and duh, forgot about that single host per device thing, so the phone wins.

Esteve wrote:Looks like a good option. Skimmed fast and the response is 8s, correct ? Before trying any sensor the response time is what worried me the most, but after trying I don't think is as important.

Post the results

The SHT31 has an accelerated response time feature that can cut the response time to 4s. It also holds accuracy up to 80C, which means the cooling tube can be shortened, further reducing the lag time. The SHT45 looks like an even better version, but the QWIIC boards don't use the version with the protective membrane yet. Looking forward to posting results in a couple weeks, and really curious if the CO2/VOC sensor adds any useful additional data about the roast.

#38: Post by **GDM528** » May 15th, 2023, 4:02 am

Brewzologist wrote:It will bend the flame from a lighter gently into the tryer point. I keep the fan just high enough to remove chaff.

So that seems like a fairly mild negative pressure, given the tryer port is (I'm guessing) 20-25mm diameter. There's a wide selection of air pumps that can deal with that, and for less than $10.

#39: Post by **ira** » May 15th, 2023, 1:41 pm

Making and mounting something like that on a PCB is a rather simple task now a days. I do that all the time, so if there's a sensor you want to use and all that's in the way is lack of PCB, that should not be an issue.

May 28th, 2023, 2:47 pm

Just an update on the humidity sensor project :

3D printed a case, have not had time to do many roasts but the firsts tests seem to give a much more stable reading.
Added some roast information to the visualisation tool.

Next steps:

After building the DIY color meter and GDM528 comments on QWIIC, I really like how easy sensors connect. I'll be adding qwiic connectors to my current setup and integrate all electronics in a box (or two).
Substitute the plastic "coil" with a metal one.

GDM528, have you been able to test the QWIIC sensor ? Curious about your results.

Results from a decaf roast, was curious to see if it had a similar behaviour, FC was hard to hear but humidity "signature" is very clear.

New setup.