As far as full queue error, it looks like common errors. When things go bad on the hub, one of the error that may come up is queue full. It did happen with my sensor to veechoo in 2018. I believe the error is also report recently in the community forum with other sensors or even applications. It is not specifc to this dth. I think this error is aftermath of something bad had happened. It is not the true root cause of the issue.
The concern that I have is actually more to why your Zigbee become offline. I did search this forum. I would not link the result here. There are a few of them. I also am not finding anything specific to a hub or firmware version. Yes... there are some zigbee issue is reported in the community recently.
I think the zigbee offline could be the start of your issue. If it happens again, I would be interested to know the root cause of it.
I read not all of the zigbee offline threads. Some of them happen similar to you interm of timing. Waking up in the morning and finding out that zigbee stopped working.
There are some members stated like you that rebooting the hub get their zigbee back to work.
I wonder that something that happened overnight may cause this. This is speculation on my part because I do not have extensive data. It is just my observation from reading a few posting here.
@parish99 & @douglaspitman I had request to make BME680 version. Thank you for bring this up.
I certainly am looking to find a way to get it integrated. There is a major issue that I need to overcome today. The BME680 cost $10 to $13 for my kind of size. We can bring the sensor to $5 ish if we are making 1000+ sensors. If we can get such demand here, this should be do able.
There are some technical issue. However, they are minor issues. They are something that I can handle.
It is just basically come down to the cost of the BME680. The only solution at the moment is to purchase it at 1000+ quantity to make it viable.
I would be willing to pay the difference between the lower and higher quantity purchase. In other words, I think I am seeing the whole device would be maybe $5 more if made in smaller quantities, like make 200 instead of 1000? That $5 would be worth having this sooner rather than wait for 1000 other people to commit to purchasing.
It will be hard to ask the manufacturer to switch a part for just small batch in the middle of assembly. It could translate to much higher total price of the sensor. This custom request is a way for contract manufacturer to charge premium fee. It is just much easier to negotiate and deal with contract manufacturer on a simple plain uniform parts assembly.
I really like to have BME680 for all of us. It is make sense to have the additional sensor.
I will try what I can do to get the BME680 on the Environment Sensor. BME680 and BME280 is pin compatible. In the next batch, I will ask the manufacture not to populate the BME280 on some of the boards. I will solder the BME680 once they got here. I hope this is easy enough for them to follow. I cannot promise you when this will happen. But, I will try.
Meanwhile, If anyone else would like one with BEM680, please let me know here or PM. If I do get a sense that we have a demand close to 1K unit, we can just jump right to make the Environment Sensor with BME680. So far the "want" count is around 10 ish, we have 990 to go.
There is no way I could do that, not even with the gear I saw when I looked into the topic. I find it amazing that you can do this at home.
Any chance we could upgrade our existing modules? Just throwing this out there and it may not be practical. But it would provide another path to reaching 1000 total.
I'm wondering how you might be able to reach a broader audience with your question of demand as I suspect there are some previous purchasers who won't read this here. Maybe a note on Ebay? Do they even allow it?
The cheapest I’ve found for the BME680 sensor is a board by CMJU - you can pick up on aliexpress pretty cheaply bu can also get on amazon: (ignore the price in the preview below its £12)
The BME680 sensor to make the most of it and get IAQ requires a closed source library from Bosch. I initially wrote my own code, but then switched to ESPHome and now about to tinker with ESPurna.
Displays are these(I received a dodgy batch as you can see from the streaks on the screen, but replacements are on route):
Case is 3d printed
I’ve also started to play around with e-ink displays and managed to get a small POC up and running to display the readings - idea being to have these as battery powered informational displays .... it wakes up every 30 mins and pulls the latest data before going back to sleep ..
I just noticed that there is "Queue full (refresh)". There is a method in the DTH named refresh. I never seen this called every 30 seconds. I manually call refresh every so often by hand. I have never seen this issue.
If I can speculate for now and assuming that refresh is called every 30 seconds, perhaps this is the trigger of the issue. A refresh will issue several zigbee.readAttribute commands. There are a lot of attributes there in the DTH. Each of them will have one attribute read. In ideal world, the refresh ever 30 seconds goes through perfectly. But, once there is a hick up on the Zigbee mesh, maybe, there is some internal queue got filled up.
I would be curious whether "refresh method" is called every 30 seconds. I would like to find out who call refresh at this interval. We should reduce the call interval. This may flood the Zigbee mesh.
If there is a trigger that call refresh at certain interval, I suggest that you try to disable it to see if the issue can be resolved.
Again I'm the idiot. I've been doing a refresh every 10 seconds from node-red because I use the illumination from that sensor to decide if it's light enough to not need the lights turned on when motion is detected. That flow has gone through many changes and the refresh isn't even needed anymore.
Not long before the first incident I added a dozen water leak sensors to the mesh so I think the added traffic stressed the mesh to where the refresh became a problem.
I've eliminated the refresh so it probably won't happen again.
@daniel.john.edge thanks for sharing the information. It is a cool project you got.
BME680 has closed source library that perform additional signal processing to give indoor air quality measurement as well as estimated of a couple gas value. Without using the library, we can only get raw reading of the "gas sensor". With the Zigbee platform that I am working on, this closed source library is not available.
The raw value of the gas sensor may be a reasonable estimation to start with. It may not have all the bell and whistle of the detail data. However, there could be a threshold where it would indicate a bad/good air quality. Bosch called it a gas sensor. Typically, a gas sensor is a resistive sensor. They way it work is to heat up an element. Then, measure the resistance of this heating element. The theory is that the gas surrounding the heating element will change the resistance while it is heated. BOSCH manage to shrink the sensor to a very small size. I also trust that BOSCH with the German spirit has also figure out the precision.
This is probably a good time to bring up that there are other solutions as well. It may not be small, precise and sexy as the Bosch solution. But, it is cheap. There is a variety of MQX sensors that would plug right into the Environment Sensor Analog input. Amazon has a lot of assortment of these. Here is an example.
I play around with one of them. It is a MQ9. I have a DTH for it as an example to use the Environment Sensor Analog input.
If there is enough interest to go this route, I can revive the topic and document it better. In term of cost, this is probably the most cost effective way to detect gas in your home.
Just a warning, the MQX solution is not sexy. There will be wire to connect. It also detect sensitive to smaller set of gas per MQX type. You would use MQ9 for Carbon Monoxide, Coal Gas, Liquefied Gas. You have to use MQ2 for Combustible Gas, Smoke. In comparison, I do not know how many type of GAS the Bosch sensor is sensitive to. In term of basic fundamental how it work, they should be similar. Bosch may just have perfect it.
In the meantime, the BME680 is still very high on my list to be worked on with Environment Sensor. In short term, I will try to get some Environment Sensor with BME280 unpopulated. I will try to get BME680 on it. I may also have some solution in the medium term if you really need to have the more accurate BME680 gas reading using dual MCU.
In long term, I am working on much newer platform for my Zigbee sensor based on NRF52840. The Bosch closed source library should be use able in this platform. I am really excited with NRF52840. You may have seen my Bluetooth gateway with this platform. I have a few thing line up with this platform in Zigbee and Bluetooth at the same time running on one NRF52840.
I also want to express my thanks to everyone here. Your support and purchases allow me to fund my hobby in making sensor for the community. In this holiday season and the spirit of Thanks Giving, I thank full for this community in supporting my work and hobby.
