Support 1-wire

[andig]

I'm using the 2.1 adapter with attached WEMOS. As my heating does not supply VL temps for HK1 and HK2 I'd like to add those. 1-wire sensors with precision reading seem to be the obvious choice.

In https://github.com/john30/ebusd-esp/commit/7f1af8990cabd20ebc567390f1df15a6d98084e8, analog input was added. Would it be possible to consider adding 1-wire as well, at least for testing purposes to see if it can still keep the ebus timing?

[andig]

ping @john30 any chance giving this a try? I'd be happy to attempt this under NDA if that would help in any way spread the work.

[john30]

I'm working on that. what class of 1W sensor do you want to use? Is it a DS1820?

[andig]

Is it a DS1820?

Yes, I have a bunch of cheap DS18B20 ones like these: https://www.ebay.de/itm/5-10-Pcs-DS18B20-Wasserdicht-Thermometer-Temperaturfühler-Temperatursensor-Kabel/392177950949?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649.

Precision and consistency of reading results should be more important than reading multiple times per second.

[john30]

I made some progress and was able to read a DS1820 temperature. The only thing is the complication of the number of sensors to be attached since every sensor has a long unique ID and you'd basically have to have the possiblity to configure each sensor in ebusd-esp to get like a fixed index or name or something. this makes the configuration quite complicated. so I'd rather do a scan for sensors after boot, calculate a checksum over the ID, sort the entries and that's it.

[andig]

Sounds good. At least for me the attached sensors would be stable and I can adjust to the checksum. Given the temperatures it should also be easy to figure out which sensor is which. I assume the sensors would be read on schedule and the last reading available via api? Please let me know when there's anything to test ;)

[andig]

ping @john30 would be great to get this running before the heating period starts again ;)

[john30]

fixed with 350b762 sensor port(s) have to be set up first using the web frontend or commandline. Then temp sensors will be queried whenever the web frontend sensors page is loaded or when /sensor is retrieved via HTTP

[andig]

Do I have any chance to fetch this via "Check & Update" or alternatively remote? I've tried:

~/htdocs
❯ ./espota.py -i 192.168.0.46 -p 80 -f /Users/andig/Downloads/ebus-v2_d1mini-preview.bin 
12:57:31 [ERROR]: No Answer

UPDATE Nvm- my wemos still has the 2017 firmware...

[andig]

I have the new FW running now- but how would I connect and configure the 1-wires? I.e. which pins should they attach to? Is the external pull-up required? I'd probably connect them using the base adapter anyway sind the 3.3 and GND pins are already taken by connecting to the base adapter.

[john30]

you need to configure a port for being used for sensors first. pull-up is needed as required by the corresponding hardware. I've tested it successfully with a pull-up for a Dallas temp sensor found somewhere in the web

[andig]

Update: still working on the test setup. As I need to put terminals for the sensors into the enclosure now I'll rewire everything a little less space-consuming than I did before with headers mounted to the WeMos. Parts are ordered and I hope to get it done next weekend.

[andig]

I have some problems testing this on an esp8266-201:

• esp not connecting wifi (claims ip is 255.255.255.255 over serial)
• cannot save initial pin config across reboot
• cannot find sensors (have these powered with 3.3V and 4.7k pull-up)
• saw one wdt reset on serial connecting wifi

To be sure these aren't fake I'll retry once the ordered Wemos arrives- hopefully next weekend.

[john30]

don't know if this is due to the esp model you have. it works fine with my wemos 3.0

[andig]

I‘ll retry with brand wemos. I‘ll also check if sensors work better when fed by 5v (2.5m wires attached) as it seems that the esp GPIOs are 5v tolerant. Just to let you know that I‘m working on it...

[andig]

Could you upload a preview build of the "lite" binary? Got the new D1 Mini Lites now, the old lite binary is running fine but the preview one just outputs garbage.

Also, I have some problem using the config screen in the old image- seems it is entirely missing any kind of linebreaks regardless of terminal emulation I chose:

                                                                                                  Welcome to eBUS adapter 2.0, build 20190407
                                                                                                                                             Configured as WIFI access point EBUS without password.
                                                     For configuration with web browser, connect to this WIFI and open http://192.168.4.1/
                                                                                                                                          Entering configuration mode.
                        Chip ID: 00CFB492
                                         Hostname: ebus-cfb492

                                                              Configuration (new):
                                                                                   1. WIFI SSID: EBUS
                                                                                                      2. WIFI secret: 
                                                                                                                       3. WIFI IP address: DHCP (not connected)
                  4. WIFI hostname: ebus-cfb492
                                                5. eBUS RX+TX PINs: direct RX+TX (GPIO3+1)
                                                                                           6. ebusd connection: TCP on port 9999
                                                                                                                                 7. HTTP TCP port: 80
        8. LED PINs: RX:disabled, TX:disabled
                                              9. Initial PINs: D4:H

                                                                    p. Set current PINs: D4:H
                                                                                              t. Toggle current output PIN
                                                                                                                           c. Connect WIFI
                                                                                                                                           e. Dump EEPROM content
                    f. Load factory settings
                                             F. Factory reset (i.e. erase EEPROM)
                                                                                  o. OTA enabled: waiting
                                                                                                          r. Reboot (without saving)
                                                                                                                                     0. Save configuration and reboot

                       Enter your choice: 

[andig]

To be more precise regarding the line break: I don't think this is a terminal problem as the normal WDT/Reboot messages have line breaks:

 ets Jan  8 2013,rst cause:2, boot mode:(3,6)

load 0x4010f000, len 1384, room 16 
tail 8
chksum 0x2d
csum 0x2d
v951aeffa
~ld

[john30]

simply activate "Implicit CR in every LF" and "Implicit LF in every CR" in Putty config, that's it.

new preview binaries are here (zipped): ebus-v2_d1mini-preview.zip ebus-v2_d1mini_lite-preview.zip

[andig]

simply activate "Implicit CR in every LF" and "Implicit LF in every CR" in Putty config, that's it.

I'm on Mac using serial. Neither CRLF, CR nor LF settings work.

[andig]

Using the new mini lite image- same as with the release mini lite image- I'm unable to connect to wifi (can open separate issue):

   Welcome to eBUS adapter 2.0, build 20191012
                                              ebusd device string: 255.255.255.255:9999
                                                                                       WiFi signal: 100% (-44dBm)
                                                                                                                 Press return within 5 seconds to change the configuration...

I do see it in the fritzbox config and it has an IP (or had one for a moment) but it doesn't answer to ping or http. I can't connect from serial menu either: it always shows DHCP (connected: 255.255.255.255) as IP.

[andig]

simply activate "Implicit CR in every LF" and "Implicit LF in every CR" in Putty config, that's it.

I'm on Mac using serial. Neither CRLF, CR nor LF settings work.

nvm. serial has the option, too:

[andig]

Not only DHCP, also fixed IP is not working although it says "connected". There is no IP conflict on the network.:

Enter your choice: c
Connecting WIFI...
WIFI connected.

ebusd device string: 255.255.255.255:9999
WiFi signal: 100% (-47dBm)
Configuration (changed):
 1. WIFI SSID: XXXXXXXX
 2. WIFI secret: ****
 3. WIFI IP address: 192.168.0.87/24 (connected), gateway: 192.168.0.1

Using "connect" inside the menu does not fix this. IP remains at 255... in DHCP mode or at fixed IP if fixed. Device remains unaccessible.

Then... when I finally "save and rebooted" with fixed IP set, device came up. Changing back to DHCP afterwards (save & reboot) continue to work. It looks as if there is a problem with the initial DHCP network connection and maybe with the connect menu option.

[andig]

OK. Now that I got WIFI working I've moved to sensor. I have one attached locally:

Sensors:
D1: 2817B650040000: temperature=127.50
1 sensors found

Now temp is at a wrong but stable value when sensor (on breadboard) is powered from 3.3V.

When sensor has 5V connected (Wemos should be 5V tolerant), sensor is not readable. 4.7k pullup is connected.

UPDATE seems the sensor is rogue. Moved to the "production" sensors with 3m cable attached and those are working, even on 3.3V:

Sensors:
D1: 28AA36D11D1302: temperature=23.00
D1: 28B5D14D331401: temperature=22.75
2 sensors found

It seems though that the readings need be calibrated (would be nice as part of the firmware, but not necessarily needed).

To summarize: preview firmware is working. It seems to have a problem with initial wifi connection that is also present in release firmware.

Thank you so much for baking 1wire in! I'll move this to my production setup now and see how it behaves with ebus attached.

Update remaining thing that might be nice would be increased readout precision. Not sure it that would help with the offset issue but seems worth testing.

[andig]

Update remaining thing that might be nice would be increased readout precision. Not sure it that would help with the offset issue but seems worth testing.

After reading some more I guess isn't. We already have 11bits (0.065°C) and the sensors precision is apparently 0.5°C.

[andig]

@john30 I have since moved my setup from breadboard to actual boiler. Downturn: sensors are no longer detected. I've checked the wiring without ripping everything apart again and it still looks ok.

Is there any chance that the new sensors support might not work when ebus adapter is actually connected? If not I'll disassemble and retry...

[john30]

no, I've tested it. The sensors are only queried from ebusd-esp when requested

[andig]

Is there any chance that the new sensors support might not work when ebus adapter is actually connected? If not I'll disassemble and retry...

I've verified that the sensors are not detected/ not visible in the UI if RX/TX of an actual ebus adapter are attached to the WeMOS.

[john30]

it works fine on mine with sensors connected on D7. will try to verify your setup with sensors on D1

[john30]

it also works fine for me when setting D1 to be used for sensors. Anyway, I've updated the binaries in c8c2916 which also includes an updated SoftwareSerial library, so please try again with those.

[john30]

the DHCP issue might be solved now. It seems a hard set to not using an IP was potentially missing