Author Topic: GW1000 New Firmware  (Read 11694 times)

0 Members and 1 Guest are viewing this topic.

Offline johnny2678

  • Member
  • *
  • Posts: 1
Re: GW1000 New Firmware
« Reply #75 on: May 24, 2020, 02:32:04 PM »

I will have more information at the start of the new week

Available in the US??🤞🤞🤞

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #76 on: May 24, 2020, 02:36:53 PM »
Everything from Ecowitt should be available worldwide. The only exception is that they won't ship the WS80 to the US as Ambient has an exclusive on that for the US market.
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline StephenR0

  • Senior Member
  • **
  • Posts: 83
Re: GW1000 New Firmware
« Reply #77 on: May 25, 2020, 06:48:27 AM »
I just had an interesting experience.  My weewx installation stopped updating with data from the GW1000.  After some troubleshooting, I rebooted the GW1000 and data started flowing again.  My WS-2902A console was receiving data throughout this time.  I've just updated the GW1000 with the new firmware.  Has anyone else had this same experience with the new firmware?

Offline olicat

  • Forecaster
  • *****
  • Posts: 1515
  • GWxx00, HPx5x1C, WN1900C, WN1980C & WS3800C
    • FOSHKplugin
Re: GW1000 New Firmware
« Reply #78 on: May 25, 2020, 07:20:22 AM »
I heard a similar problem from another user yesterday.
Sending to FOSHKplugin via custom server in Ecowitt format suddenly stopped working.
Apparently the path for the custom server was missing after he had subsequently activated the Weathercloud service via WS View.
It could also be that he accidentally noticed the lack of the path after activating the Weathercloud service - but the path was missing before.
He also did the update to v1.5.9 yesterday before the problem occurred.

Regards, Oliver

Offline mcrossley

  • Forecaster
  • *****
  • Posts: 1132
    • Wilmslow Astro
Re: GW1000 New Firmware
« Reply #79 on: May 25, 2020, 08:32:42 AM »
The new firmware is ready for new soil and water temperature sensors
it's possible to add n.8 sensors, in any kind of scenario

Mauro, these sensors all appear as user temp1-8, correct?

Can you mix and match, so long as they are on different channels and you have a maximum of 8?
So channel 1 = water, channel 2 = soil etc.
Mark

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #80 on: May 25, 2020, 08:48:19 AM »
The new firmware is ready for new soil and water temperature sensors
it's possible to add n.8 sensors, in any kind of scenario

Mauro, these sensors all appear as user temp1-8, correct?

They appear as 0x63 through 0x6A (2 bytes)

Quote

Can you mix and match, so long as they are on different channels and you have a maximum of 8?
So channel 1 = water, channel 2 = soil etc.
Yes
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #81 on: May 25, 2020, 08:53:10 AM »
There is a typo in the 1.5.9 release notes. The new sensor is not WN34. The correct part is WH34. I'm basing this on the API which I would trust as a better documentation than the app pop-up release notes.
« Last Edit: May 25, 2020, 08:55:56 AM by galfert »
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline mcrossley

  • Forecaster
  • *****
  • Posts: 1132
    • Wilmslow Astro
Re: GW1000 New Firmware
« Reply #82 on: May 25, 2020, 08:58:51 AM »
Thanks, yes I have tagged them as WH34 in Cumulus. I just needed that confirmation about them sharing the same 8 channels.
Mark

Offline olicat

  • Forecaster
  • *****
  • Posts: 1515
  • GWxx00, HPx5x1C, WN1900C, WN1980C & WS3800C
    • FOSHKplugin
Re: GW1000 New Firmware
« Reply #83 on: May 25, 2020, 09:07:29 AM »
I'm basing this on the API which I would trust as a better documentation than the app pop-up release notes.
Perhaps you can also say which keyname is used to transmit these values in Ecowitt format?
soiltemp1f ..8f?
if really temp1..temp8 - in °F or in °C?

Is wh34batt1..8 or temp1..8batt used to indicate the battery-status?

Oliver

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #84 on: May 25, 2020, 01:08:57 PM »
I have not looked into the Ecowitt format.
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline _John C

  • Contributor
  • ***
  • Posts: 123
Re: GW1000 New Firmware
« Reply #85 on: June 12, 2020, 06:17:57 AM »
V1.6.0 released.

Fix for data loss to meteobridge device.


 [ You are not allowed to view attachments ]
« Last Edit: June 12, 2020, 09:48:48 AM by galfert »
Ecowitt GW1000CA x2,   HP2551CA,   WH65 Outdoor 7in1 2 arm array,  WH32B Indoor Temp/Hum/Baro, WH45 Co2/PM2.5/PM10,  WH41 PM2.5,   WH57 Lightning Detector,   WH31 Temp/Hum x8,  WN34L Liquid Temp x2,  WN34S Soil Temp x1, WH51 Soil Moisture x8,   WH55 Water Leak x4,   Meteobridge (TL-WR902AC v3),   Da9L PWT x2, Home Assistant Integration

Offline mcrossley

  • Forecaster
  • *****
  • Posts: 1132
    • Wilmslow Astro
Re: GW1000 New Firmware
« Reply #86 on: June 17, 2020, 08:17:50 AM »
The version 1.6.0 seems to have removed the battery status information for WH34 devices that got added in 1.5.9.
Mark

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #87 on: June 17, 2020, 10:34:26 AM »
The version 1.6.0 seems to have removed the battery status information for WH34 devices that got added in 1.5.9.

Are you referring to 0x6B battery for all sensors 8 bytes length?


Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline mcrossley

  • Forecaster
  • *****
  • Posts: 1132
    • Wilmslow Astro
Re: GW1000 New Firmware
« Reply #88 on: June 17, 2020, 12:43:52 PM »
Are you referring to 0x6B battery for all sensors 8 bytes length?
Yes
Mark

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #89 on: June 17, 2020, 12:59:46 PM »
The 0x6B was removed. But in place for that each sensor payload changed from 3 bytes to 4 bytes.

It used to be 3 bytes (for each 0x63 - 0x6A):
1 byte for ID, followed by 2 bytes for data (temperature).

This is why before you had to look at 0x6B separately to see battery status.

It is now 4 bytes (for each 0x63 - 0x6A):
1 byte for ID, followed by 2 bytes for data (temperature), followed by 1 byte for battery status.

It is simpler now because each sensor has the battery status right after the data bytes.

So battery status was not removed, rather you just will find it in a different and better place.
« Last Edit: June 17, 2020, 01:07:40 PM by galfert »
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline mcrossley

  • Forecaster
  • *****
  • Posts: 1132
    • Wilmslow Astro
Re: GW1000 New Firmware
« Reply #90 on: June 17, 2020, 04:29:59 PM »
Great!  #-o  I'll have to make a new release of Cumulus MX now to try and fix this  ](*,)
The 1.6.0 change will have broken Cumulus MX's support for these devices.

I'll have to detect the firmware version and if 1.5.9 do a different decode than if it is 1.6.0
Regressing API's is never a good idea :(

What are the possible values for the battery byte? (awaiting new API document).
« Last Edit: June 17, 2020, 04:45:33 PM by mcrossley »
Mark

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #91 on: June 17, 2020, 11:17:38 PM »
Mark,
I don't think you need to worry about testing for 1.5.9 or prior and having to do different things based on firmware version. Because nobody has this sensor yet. Just make Cumulus MX work with 1.6.0 or greater for this sensor. Besides with the GW1000 the firmware only goes up, you can't go back down. Anyone with a GW1000 will be also using the WS View app and it always give you a pop-up notification that a new firmware is available and it is a one click install that takes less than 10 seconds. I can't see anyone not upgrading their firmware. If it doesn't work and they get no battery status so be it. If it bothers them then they will troubleshoot and upgrade as that is always the first logical troubleshooting step. Another note is that this new sensor was just introduced with 1.5.9, so anything prior would not support the sensor anyway. Version 1.5.9 was out a very short time before 1.6.0 came out, barely 3 weeks prior. You shouldn't need to target anyone running 1.5.9 as I can't think that there will be too many people doing that. People are either actively upgrading firmware or they aren't and so then they will likely be on a very old firmware or something current. Also anyone setting up a new station will be prompted to upgrade.

EDIT: Striking this part out because it is wrong for the new WH34. See my next post in this thread.
For battery status, I'm not positive (documentation is vague), but I believe 1 is low and 0 is normal for the battery (unsigned short), same as all the other sensors.

Also don't kill yourself rushing this update...nobody has this sensor yet.

PS - if you are looking at API documentation 1.5.5, I have found a typo. It erroneously mentions that USR8 is 0x69 and that would be wrong because that is for USR7. The correct identifier for USR8 is 0x6A. I've notified Ecowitt. I'm sure the next version will correct this. It is a minor typo that you would catch if you noticed that they doubled up on 0x69 instead of going to the next hex number. Interestingly the typo doesn't exist in 1.5.3 when this sensor was introduced.

« Last Edit: June 18, 2020, 01:15:40 AM by galfert »
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline lemuba

  • Member
  • *
  • Posts: 48
    • Fitzbek-Wetter.de
Re: GW1000 New Firmware
« Reply #92 on: June 18, 2020, 12:46:59 AM »
For battery status, I'm not positive (documentation is vague), but I believe 1 is low and 0 is normal for the battery (unsigned short), same as all the other sensors.


This is valid for the 7in1 Sensor, but e.g. the WH57 Lightning Sensor counts the Battery status down from 5 to 1 (or zero?) I could believe that all newer sensors work like this. My WH57 is currently at „4“...

« Last Edit: June 18, 2020, 12:49:27 AM by lemuba »

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #93 on: June 18, 2020, 01:00:53 AM »
For battery status, I'm not positive (documentation is vague), but I believe 1 is low and 0 is normal for the battery (unsigned short), same as all the other sensors.


This is valid for the 7in1 Sensor, but e.g. the WH57 Lightning Sensor counts the Battery status down from 5 to 1 (or zero?) I could believe that all newer sensors work like this. My WH57 is currently at „4“...
Okay yes you are correct. There are some sensors that have a power level and other sensors that don't and only show 0 good or 1 low. In this case the WH57 is one with levels. Power level 0~5, when the power level <=1, it displays low voltage. But the WH34 is different and has even more complexity in how it works. See below.

It may seem backwards for the ones that only have a good or low indicator. But it is basically reporting a condition of low battery. In computers 1 means true. The check is for low battery condition. I suppose they could have made check for good and then made that be the 1 for true, but that isn't what happens in the software. In the software the low power indicator is either on or off. Therefore it would only need to be turned on if low and that is the true statement that you are looking for. In the case of power levels then they are always reporting and then the logical choice is that the higher number means more power.

EDIT: The following has been corrected for prior errors from initial post

These sensors have just the indication as 1 low power indicator and 0 being good (no levels):
WH65
WH40
WH25
WH26
WH31
WH51

These are the sensors with power level 0-5 with <=1 being low:
WH41
WH57
WH55

Special case:
WH68 Voltage=val*0.02V When <=1.2V, display low voltage
WS80 Voltage =val*0.02V When <=1.2V, display low voltage
WH34 Voltage=val*0.02V When <=1.2V, display low voltage

« Last Edit: July 28, 2020, 01:41:22 PM by galfert »
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline mcrossley

  • Forecaster
  • *****
  • Posts: 1132
    • Wilmslow Astro
Re: GW1000 New Firmware
« Reply #94 on: June 18, 2020, 04:37:44 AM »
I will make v1.6.0 or later a requirement to support WH34 type sensors.

The problem is you do not just lose battery status. Because of the way Ecowitt have implemented their data structure as a binary blob, you need to keep track of the offset of every bit of data. Them adding in a extra byte means you are unable to decode anything from that point onwards successfully.

It would have been much better in this day and age to implement in a self referencing format such as JSON.


These sensors have just 1 low power indicator and 0 being good (no levels):
WH65
WH40
WH25
WH26
WH31
WH51
WH41

These are the sensors with power level 0-5 with <=1 being low:
WH57
WH55

Special case:
WH68 Voltage=val*0.02V When <=1.2V, display low voltage
WH34 Voltage=val*0.02V When <=1.2V, display low voltage


Isn't the WH41 a 0-5 level device?
And WH57 and WH80 both volts * 0.02?
« Last Edit: June 18, 2020, 05:11:37 AM by mcrossley »
Mark

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #95 on: June 18, 2020, 09:56:35 AM »
I will make v1.6.0 or later a requirement to support WH34 type sensors.

Good idea

Quote

The problem is you do not just lose battery status. Because of the way Ecowitt have implemented their data structure as a binary blob, you need to keep track of the offset of every bit of data. Them adding in a extra byte means you are unable to decode anything from that point onwards successfully.

I see your point from a programing perspective.

Quote

It would have been much better in this day and age to implement in a self referencing format such as JSON.

I've asked other questions to the engineers like adding basic networking tools (ping, traceroute, netstat) and I was told not possible. I think they are working with a basic chipset and they have to develop everything for it from scratch. A JSON format would be a tall order.

Quote

Isn't the WH41 a 0-5 level device?
And WH57 and WH80 both volts * 0.02?
According to the documentation, with many parts in Chinese that I have to use Google Translate to understand, the WH41 is not a 0-5 power level device. But I have one and I only ever see good or bad even though in WS View it does show it as having multiple levels. So either this explains why WS View shows the WH41 suddenly die or there is another typo in the documentation. I've just never experienced levels on the WH41, it is always full or low. Should be an easy test because 0 is good in the 0/1 binary option. So see if you get a 0 with fresh batteries. I fact I'll test that in a few minutes and get back to you on this.

The WH80 in the documentation which I assume to be the WS80 leaves out any reference to supporting power indication at all.
UPDATE: The WH80 does have the same battery level structure as the WH68/WS68. It is further referenced in the document where it is missing in the prior section. So yes volts * 0.02 for WH80
« Last Edit: June 18, 2020, 01:14:10 PM by galfert »
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline Mandrake

  • Forecaster
  • *****
  • Posts: 1102
Re: GW1000 New Firmware
« Reply #96 on: June 18, 2020, 11:22:02 AM »
Your observation on the WH41 behaviour also fits my observations!

As for the WS80 it certainly seems to support battery information as I get a battery gauge showing in WSView
G1ZFO

Ecowitt HP2551A + WH65 Tri-Wing (Wunderground: IGUILDFO67)
Ecowitt GW1000 (Wunderground: IGUILDFO68)
Ecowitt GW1000 (Mk2) test environment driving CumulusMX on a RPi 3b
Ecowitt GW2000 (Test)
Ecowitt WS90 Wittboy - Test
Ecowitt WH51 (x6) Soil Moisture Sensor
Ecowitt WH41 PM2.5 AQM Sensor
Ecowitt WH31 (x8) Thermo/Hygro Sensor
Ecowitt WS80 Ultrasonic Anemometer (pre-prod test)
Ecowitt WH57 Lightning Sensor -test
Ecowitt WH32-EP (SHT35) + Davis 7714 Screen
Ecowitt WH45 CO2/PM2.5/PM10 -Test
Ecowitt WN34 Soil Temp Sensor -Test
Ecowitt WN34 Water Temp Sensor -Test
Ecowitt WN35 Leaf Moisture

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #97 on: June 18, 2020, 12:30:56 PM »
Okay did some testing with the GW1000 API. Found some interesting results...

Yes it appears that the WH41 does have levels 0-5 for battery. There is an error in the documentation as it says 0 or 1 for battery level, but in another place on the documentation I did find that it does say 0~5 for WH41. So it is defined both ways...only one is right. So I tested pulling data...

WH41 battery indicators in API (my test results):
0F = sensor is paired and dead not reporting. Probably is level 0
05 = sensor is working and it shows power level is 5 (seemingly always - levels don't really work except for 5 and 0)
00 = sensor not paired - not present

For other sensors that use the 0/1 binary indicator it seems FF is shown when sensor is not pair (not present).

I had a dead WH41 that had not been used in a couple months. I just plugged it in and while charging it showed 05 battery (also full bars in WS View). But I only let it charge for like 10 minutes and removed the charger. Then I let it run for 30 minutes (not charging) and it still shows 05 battery level. Although this sensor may have technically been designed for battery levels it must be a limitation of the batteries used (rechargeable NiMH) or some other technical aspect of the device where you never actually see levels. I do believe now having seen the raw API data that it was designed to show levels though. Shame it doesn't actually work as intended.

By the way the WH43 does not show up in the documentation. Technically the WH43 is really just seen as a WH41 internally. The WH41 and WH43 can be had in any combination up to 4 total. Meaning total number of WH41 + WH43 cannot be greater than 4 devices.
« Last Edit: June 18, 2020, 12:37:17 PM by galfert »
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline Mapantz

  • Forecaster
  • *****
  • Posts: 756
    • Wareham Weather
Re: GW1000 New Firmware
« Reply #98 on: June 18, 2020, 01:23:52 PM »
What sensor is the WH41?


Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Re: GW1000 New Firmware
« Reply #99 on: June 18, 2020, 01:25:22 PM »
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

 

anything