| 15 Oct 2020 |
 Tammo ter Hark | We have started a FAQ / Q&A which will be open for all people interested in MeetJeStad. It is open now at: http://forum.meetjestad.net/ It is our intention to open this up soon and to add a link on the website (https://www.meetjestad.net)
Can you think of more questions (and preferably with answers...) that we can place there? Please sign up and add questions and answers! | 12:12:48 |
 Matthijs Kooijman | ommat: Wouldn't something like "questions.meetjestad.net" be more appropriate? I can imagine we'd want a more traditional forum as well later, and then it would be annoying to have forum. already taken? | 12:50:34 |
 Diana | can be used for both, it actually is meant to be a forum, but we have already put in some questions that we have recently had to answer a lot, so we can tell people: look at the forum first. | 14:15:37 |
 Harmen | Would be great to hear your suggestions on how to add structure to the site: which categories, topics, code of conduct etcetera... | 14:26:29 |
| Matthijs Kooijman | It looks very question-centered, not so much suitable for announcements or general discussion (but I guess it can be used for that if we want) | 14:41:08 |
 Paul_B | Agree with Matthijs, I see that there are catagories.
So we could make a sort of forum using them and for discussions keep questions unanswered? | 17:16:56 |
| Harmen | Maybe questions is the wrong word, I guess we could change the vocabulary into topics and contributions in stead of questions and answers... | 17:25:42 |
| 18 Oct 2020 |
| Tammo ter Hark | Dear all: next wednesday it is the 21st of October. We would like to invite all people that are and have been active during the recent years to attend the meeting and to celebrate with us the third place in the Trouw Duurzame top 100 (Sustainable top 100 of one of the main Dutch news papers). Please find details in: http://www.meetjestad.net
Ha, komende woensdag is het weer de 21e en dat is de dag in de maand dat we altijd een MeetjeStad bijeenkomst hebben. Kijk in de agenda op: http://www.meetjestad.net en vier die avond met ons mee dat we de derde plaats hebben in de Trouw Duurzame top 100 van 2020. | 17:46:10 |
| 20 Oct 2020 |
| Paul_B | I've run into an issue with adding the soil moisture (I2C) sensors to the current platform, sometimes the sketch hangs. The sketch uses the build-in Wire library and it likely hangs in the Wire.endTransmission() function. Anyone familliar with this issue / knows a work-around? | 08:27:55 |
| Matthijs Kooijman | Paul_B: Yes. The cause is probably noise on the I2C bus (I2C is not really intended to run over longer runs of cable, it was originally designed to be between chips on the same PCB). This noise likely causes the I2c hardware to mess up and lockup, which is not detected by the Arduino code. | 08:30:40 |
| Matthijs Kooijman | However, recently a timeout feature was added to the Wire library, which should at least prevent locking up in these circumstances. You need the latest AVR core for this, see https://github.com/arduino/Arduino/issues/10803 for the WIP documentation for this new feature. | 08:32:09 |
| Matthijs Kooijman | Ideally, you'd also fix the underlying cause (e.g. if it is indeed noise, by using shielded cable) | 08:32:47 |
 pilla | In reply to @ommat:matrix.org
Dear all: next wednesday it is the 21st of October. We would like to invite all people that are and have been active during the recent years to attend the meeting and to celebrate with us the third place in the Trouw Duurzame top 100 (Sustainable top 100 of one of the main Dutch news papers). Please find details in: http://www.meetjestad.net
Ha, komende woensdag is het weer de 21e en dat is de dag in de maand dat we altijd een MeetjeStad bijeenkomst hebben. Kijk in de agenda op: http://www.meetjestad.net en vier die avond met ons mee dat we de derde plaats hebben in de Trouw Duurzame top 100 van 2020. I want to attend tomorrow. But the webpage won't recognize me, and there is a "page not found" when I try to recover my account. What shall I do? | 09:42:33 |
 Kees van O | Maybe even beter to use shielded with RS432 serial protocol. This is more intended to bridge distances. I2C is more system internal device to device. I've figure how the code seize compares. | 09:43:06 |
| Matthijs Kooijman | pieter.sijbrandij: I don't think you have an account on meetjestad.net yet. But you do not need to log in to the website to attend tomorrow, you can just join by clicking the link in the news item (https://telefoon.planb.coop/meetjestad) | 09:46:12 |
| Matthijs Kooijman | kvoeveren: Yeah, using a differential protocol is even more robust (you mean RS482 I suppose though, RS432 does not seem to exist :-p) | 09:47:12 |
| Kees van O | ??? Scherp maar ik bedoel RS485. Add on modules ca. 0,40 Eu p/s. Iedere zijde 1 werkt prima ook over afstanden. | 10:06:20 |
| Kees van O | 
Download image.png | 10:06:33 |
| Matthijs Kooijman | Maar niet zomaar toepasbaar als je bestaande sensor alleen I2c doet, dan moet er weer een extra microcontroller of ander IC tussen. | 10:42:15 |
| Paul_B | In reply to @matthijs:matrix.org
Paul_B: Yes. The cause is probably noise on the I2C bus (I2C is not really intended to run over longer runs of cable, it was originally designed to be between chips on the same PCB). This noise likely causes the I2c hardware to mess up and lockup, which is not detected by the Arduino code. Thanks! I'll look into the new AVR core for the short term fix. An alternative protocol/shielded cable would be great to include in a new prototype iteration. Or switch to a stand-alone node perhaps, although we might run into LORA range issues then. | 10:58:15 |
| Paul_B | Is the latest Core 1.8.3 or is this a (beta/other) version we need? | 11:13:25 |
| Matthijs Kooijman | Should be the latest stable release | 11:15:23 |
| Matthijs Kooijman | But also check the github link, since the timeout is not enabled by default | 11:15:39 |
 Peter dM | In reply to @matthijs:matrix.org
However, recently a timeout feature was added to the Wire library, which should at least prevent locking up in these circumstances. You need the latest AVR core for this, see https://github.com/arduino/Arduino/issues/10803 for the WIP documentation for this new feature. @Paul_B, at the moment the resistors loading the I2C (R2, R3) at the sensor side are 3k3. Have you considered lowering their value ? | 11:17:30 |
| Matthijs Kooijman | Strong pullups can indeed help, yeah (though be careful to not overload the I2c devices, smaller pullups means more current to pull the bus low) | 11:21:57 |
| Peter dM | In reply to @matthijs:matrix.org
However, recently a timeout feature was added to the Wire library, which should at least prevent locking up in these circumstances. You need the latest AVR core for this, see https://github.com/arduino/Arduino/issues/10803 for the WIP documentation for this new feature. ... and there is SoftWire, which I use to connect 6 identical sensors to the same Arduino. | 11:56:21 |
| Peter dM | We have seen that Si7021 humidity measurements would reach up to 119, to then continue from -6 and therefrom upwards.
There is a fix now for that, which adds 125 to some of the measured values: | 11:59:11 |
| Peter dM | 
Download mjs_fhum_0100_02w.png | 12:00:04 |
| Peter dM | The thin blue line are the original measurements, the thick blue line are the corrected measurements.
The vague bluegrey line are the KNMI measurements.
https://meetjestad.net/static/graphs/fixhumi shows the original and corrected measurements
for 7 standard sensors and 6 experimental testsensors, for several time-spans.
It looks to me that sensors 100 and 410 and testsensors 1, 4, 5 and 6 are working in an acceptable way with this correction.
Without this correction, these sensors would probably be written off as unusable. | 12:15:32 |
| 22 Oct 2020 |
|  jose_martin left the room. | 19:50:51 |
