CT Clamp/ADS1115 - Githubissues

ToViNi commented 12 months ago

The problem

Setting up 2 ADS1115 with 4 ct clamps sensors each, and using continuous_mode: on gives one sensor with wrong value.

With continuous_mode on, one random sensor will show a value on a sensor with no current.

Example. 4A on sensor A2 gives a readout on sensor A1, even when there is no current in the other tree sensors.

A1: 1.1 A (Wrong) A2: 4.0 A A3: 0.0 A A4: 0.0 A

Restart the ESP, and the (low) value moves to another sensor.

This shows that there is something leftover from last sensor reading somewhere.

Which version of ESPHome has the issue?

2023.8.0

What type of installation are you using?

Home Assistant Add-on

Which version of Home Assistant has the issue?

2023.9.2

What platform are you using?

ESP32

Board

dev

Component causing the issue

ADS1115

Example YAML snippet

ads1115:
  - address: 0x48
    id: ads1115a
    continuous_mode: on

sensor:
  - platform: ads1115
    ads1115_id: ads1115a 
    multiplexer: A0_GND
    gain: 4.096
    id: a1
    accuracy_decimals: 3
    icon: "mdi:gauge"
    state_class: "measurement"    

  - platform: ct_clamp
    sensor: a1
    id: ca1
    name: Current A1
    accuracy_decimals: 3
    update_interval: 2s
    sample_duration: 0.2s

..... and more sensors like the one

Anything in the logs that might be useful for us?

No response

Additional information

No response

ToViNi commented 12 months ago

Issue not present with longer update intervals, and sometimes disappear with same values.

ToViNi commented 12 months ago

Debug gave me the answer. To long sample duration or to short update intervals gives very unstable sample count. One sensor gets 60 samples, another only 7 samples. Testing shows that with 2 ADS1115 = 8 channels, update intervals must be minimum 5s for 0.25s sample duration. This gave me logging like this line: Raw AC Value: 0.000A after 179 different samples (716 SPS) and the values are good. (I2S speed 800kHz)

ssieb commented 12 months ago

I think the reason is that there is no scheduling to make sure that only one input is getting sampled at a time. So what happens is that ct clamps on multiple channels can try to sample at the same time and it doesn't work. For now, I suggest disabling the update interval and use an interval: to run updates for each channel in sequence with an appropriate delay between each one.

ToViNi commented 12 months ago

I think the reason is that there is no scheduling to make sure that only one input is getting sampled at a time. .......

You made my day! Excellent idea. The number of samples was very random. Now, with interval/sequenced updates, it is VERY consistent:

ssieb commented 11 months ago

Leave this issue open for scheduling support in the voltage sampler.

ssieb commented 11 months ago

Can you try this with your original config? Also have one of the channels without a ct_clamp to test if regular updates work too.

external_components:
  - source: github://ssieb/esphome@vsampler
    components: [ voltage_sampler, ct_clamp, ads1115 ]
    refresh: 1min

ToViNi commented 11 months ago

Your code works great. Original components:

01:52:59 [D] [ct_clamp:041]
'Current A1' - Raw AC Value: 0.000A after 14 different samples (46 SPS) 01:52:59 [D] [ct_clamp:041]
'Current C1' - Raw AC Value: 0.001A after 25 different samples (83 SPS) 01:53:03 [D] [ct_clamp:041]
'Current C3' - Raw AC Value: 0.000A after 213 different samples (710 SPS) 01:53:03 [D] [ct_clamp:041]
'Current C4' - Raw AC Value: 0.001A after 223 different samples (743 SPS) 01:53:03 [D] [ct_clamp:041]
'Current A4' - Raw AC Value: 0.000A after 215 different samples (716 SPS) 01:53:04 [D] [ct_clamp:041]
'Current A3' - Raw AC Value: 0.000A after 79 different samples (263 SPS) 01:53:04 [D] [ct_clamp:041]
'Current A2' - Raw AC Value: 0.000A after 59 different samples (196 SPS) 01:53:04 [D] [ct_clamp:041]
'Current C2' - Raw AC Value: 0.001A after 112 different samples (373 SPS) 01:53:04 [D] [ct_clamp:041]

Your code:

01:55:54 [D] [ct_clamp:049]
'Current A2' - Raw AC Value: 0.000A after 222 different samples (740 SPS) 01:55:54 [D] [ct_clamp:049]
'Current C4' - Raw AC Value: 0.001A after 206 different samples (686 SPS) 01:55:55 [D] [ct_clamp:049]
'Current A4' - Raw AC Value: 0.001A after 224 different samples (746 SPS) 01:55:55 [D] [ct_clamp:049]
'Current C1' - Raw AC Value: 0.001A after 205 different samples (683 SPS) 01:55:56 [D] [ct_clamp:049]
'Current A3' - Raw AC Value: 0.000A after 218 different samples (726 SPS) 01:55:56 [D] [ct_clamp:049]
'Current C2' - Raw AC Value: 0.000A after 213 different samples (710 SPS) 01:55:56 [D] [ct_clamp:049]
'Current C3' - Raw AC Value: 0.001A after 229 different samples (763 SPS) 01:55:56 [D] [ct_clamp:049]
'Current A1' - Raw AC Value: 0.001A after 221 different samples (736 SPS)

Tried one and one channel. All works great and do not affect any other channel. (at least not in this test). Values used: Interval 5s , 300mS samples.

github-actions[bot] commented 7 months ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

Denicolomanuel commented 7 months ago

Hello, I used your code and it works great. i only have some noise time to time. but when i look at the esp log, i have some warning/informations :

[13:22:07][D][ct_clamp:048]: 'Courant P1' - Raw AC Value: 0.003A after 458 different samples (458 SPS)
[13:22:07][D][sensor:093]: 'Courant P1': Sending state 0.08755 A with 2 decimals of accuracy
***[13:22:08][W][component:214]: Component ct_clamp.sensor took a long time for an operation (0.67 s).***
***[13:22:09][W][component:215]: Components should block for at most 20-30ms.***
[13:22:20][D][ct_clamp:048]: 'Courant P2' - Raw AC Value: 0.003A after 472 different samples (472 SPS)
[13:22:20][D][sensor:093]: 'Courant P2': Sending state 0.10274 A with 2 decimals of accuracy
***[13:22:21][W][component:214]: Component ct_clamp.sensor took a long time for an operation (0.61 s).***
***[13:22:22][W][component:215]: Components should block for at most 20-30ms.***
[13:22:23][D][ct_clamp:048]: 'Courant P3' - Raw AC Value: 0.003A after 478 different samples (478 SPS)
[13:22:24][D][sensor:093]: 'Courant P3': Sending state 0.09837 A with 2 decimals of accuracy
***[13:22:25][W][component:214]: Component ct_clamp.sensor took a long time for an operation (0.60 s).***
***[13:22:26][W][component:215]: Components should block for at most 20-30ms.***
[13:22:27][D][ct_clamp:048]: 'Courant P1' - Raw AC Value: 0.003A after 471 different samples (471 SPS)
[13:22:27][D][sensor:093]: 'Courant P1': Sending state 0.07020 A with 2 decimals of accuracy
***[13:22:28][W][component:214]: Component ct_clamp.sensor took a long time for an operation (0.67 s).***
***[13:22:29][W][component:215]: Components should block for at most 20-30ms.***
[13:22:29][D][ads1115:233]: 'adc_sensor3': Got Voltage=1.572625V
[13:22:30][D][sensor:093]: 'adc_sensor3': Sending state 1.57263 V with 3 decimals of accuracy
***[13:22:31][W][component:214]: Component ads1115.sensor took a long time for an operation (1.09 s).***
***[13:22:32][W][component:215]: Components should block for at most 20-30ms.***

i don't know if it's a problem or not and how to solve it. any idea?

here is my code:

esphome:
  name: tableau-electrique-general
  friendly_name: Tableau Electrique général

esp8266:
  board: nodemcu

external_components:
  - source: github://ssieb/esphome@vsampler
    components: [ voltage_sampler, ct_clamp, ads1115 ]
    refresh: 1min

# Enable logging
logger:

# Enable Home Assistant API
api:
  encryption:
    key: XXX

ota:
  password: XXX

wifi:
  ssid: !secret wifi_ssid
  password: !secret wifi_password

  # Enable fallback hotspot (captive portal) in case wifi connection fails
  ap:
    ssid: XXX
    password: XXX

captive_portal:

# configuration du I2C
i2c:
  sda: 4
  scl: 5
  scan: True
  frequency: 200000

# configuration de l ADC
ads1115:
  address: 0x48
  continuous_mode: True

# configuration des capteurs de courant et pinces
sensor:
  # sensors source
  - platform: ads1115
    multiplexer: 'A1_GND'
    gain: 4.096
    name: "adc_sensor1"
    id: "adc_sensor1"
  - platform: ct_clamp
    sensor: adc_sensor1
    name: "Courant P1"
    sample_duration : 1s
    update_interval : 20s
    filters:
      - calibrate_linear:
          # Measured value of 0 maps to 0A
          - 0 -> 0
          # Known load: 4.0A
          # Value shown in logs: 0.1333A
          - 0.3 -> 8.3 

  - platform: ads1115
    multiplexer: 'A2_GND'
    gain: 4.096
    name: "adc_sensor2"
    id: "adc_sensor2"
  - platform: ct_clamp
    sensor: adc_sensor2
    name: "Courant P2"
    sample_duration : 1s
    update_interval : 20s
    filters:
      - calibrate_linear:
          # Measured value of 0 maps to 0A
          - 0 -> 0
          # Known load: 4.0A
          # Value shown in logs: 0.1333A
          - 0.27 -> 8.2

  - platform: ads1115
    multiplexer: 'A3_GND'
    gain: 4.096
    name: "adc_sensor3"
    id: "adc_sensor3"
  - platform: ct_clamp
    sensor: adc_sensor3
    name: "Courant P3"
    sample_duration : 1s
    update_interval : 20s
    filters:
      - calibrate_linear:
          # Measured value of 0 maps to 0A
          - 0 -> 0
          # Known load: 4.0A
          # Value shown in logs: 0.1333A
          - 0.27 -> 8.2

Have a nice day.

Manu

ssieb commented 7 months ago

I don't know why it would be taking so long. I'll try to take a look.

nobthenob commented 4 months ago

Can you try this with your original config? Also have one of the channels without a ct_clamp to test if regular updates work too.
external_components:
  - source: github://ssieb/esphome@vsampler
    components: [ voltage_sampler, ct_clamp, ads1115 ]
    refresh: 1min

Hello, I experienced the same kind of "random crosstalking" with 4 ct clamps connected to one ADS1115 and was happy when I found the issue described here. I tried your code and after first tests, it seems to solve the problem, channels are now independent. Thank you very much!

Can you can help me with the following questions: Where can I find the doc on the voltage_sampler component? I am interested to understand how it works and what are limitations, in order to tweak my ct clamp parameters for fast update rates. thx, nob

ssieb commented 4 months ago

The voltage_sampler is an interface that adc devices can provide. The ct_clamp component tries to read as fast as possible or at least every 1ms.

nobthenob commented 4 months ago

Hello Samuel, I have now monitored the behaviour of my 4 channel CT Clamp sensor a while and played around with the update_interval and sample_duration values. So far all 4 channel are working independently, as they should :)

Nevertheless, I still see that the number of samples (SPS) of the 4 channels differ very much, like what ToViNi described in the first place: Typically one channel shows significantly lower SPS values, leading to small data errors. Interestingly, the channel that show this significanty lower SPS values is changing with every restart. In average I get about 200 SPS per channel with a ESP8266 D1 mini.

Should I open a new issue on that to provide code and logs? with best regards, nob

esphome / issues

CT Clamp/ADS1115 #4890

The problem

Which version of ESPHome has the issue?

What type of installation are you using?

Which version of Home Assistant has the issue?

What platform are you using?

Board

Component causing the issue

Example YAML snippet

Anything in the logs that might be useful for us?

Additional information