Add RFID support and upgrade OCPP library #325

matth-x · 2022-03-08T13:20:49Z

The RFID integration supports two modes:

Local authorization: The ID tags are stored locally on the ESP32 (managed via the dashboard) and the charging session is only authorized if the user presents an RFID card with a matching ID tag (based on the work of Oliver Norin)
OCPP authorization: The EVSE triggers the standard OCPP charging process with RFID authorization

The RfidTask contains a custom driver for the NXP PN532 to fit best into the OpenEVSE architecture.

The changes from the OCPP library include:

Improved charging process with input from users of my library
Added option to redirect all debug output to DBUG, added log levels
Complete OCPP deinitialization
OCPP TriggerMessage support improved, added ClearCache message type and more configuration keys
Code refactoring (fix compiler warnings, field type fixes ect.)

Changes of the OCPP integration:

Full BootNotification support
Integrate more EVSE error codes
LCD shows more details
URL validation fix (Error in MongooseOcppSocketClient::isValidUrl #319)

Further changes are:

Upgrade ArduinoJson to v6.19.1 (the newest ArduinoOcpp version requires it)

Voltage is used by divert mode, as well as sent along to the OpenEVSE module for power calculations. Setting this to zero could result in a lot of issues. The divert code would see divide by zero happen, and the energy montoring code on the OpenEVSE module would result in the recorded energy usage being incorrect. While this shouldn't happen reguarly, I found myself in this situation by sending a null/zero-length message to the relevant topic that the ESP32 was listening on. Because there is no error checking in the float parsing code, a null value parsed to zero, which resulted in the voltage value being set to zero. An alternative here would be to reset the voltage to `DEFAULT_VOLTAGE` if a zero or invalid value was received. As written, this will simply do nothing and ignore the invalid/zero value.

A value between 60 and 300 volts, while being a generous range, should prevent issues where the value read from MQTT is read as a `0` or erroneously specified in millivolts.

…to rfid_and_timer

Jeremypoulter/issue5

Voltage is used by divert mode, as well as sent along to the OpenEVSE module for power calculations. Setting this to zero could result in a lot of issues. The divert code would see divide by zero happen, and the energy montoring code on the OpenEVSE module would result in the recorded energy usage being incorrect. While this shouldn't happen reguarly, I found myself in this situation by sending a null/zero-length message to the relevant topic that the ESP32 was listening on. Because there is no error checking in the float parsing code, a null value parsed to zero, which resulted in the voltage value being set to zero. An alternative here would be to reset the voltage to `DEFAULT_VOLTAGE` if a zero or invalid value was received. As written, this will simply do nothing and ignore the invalid/zero value.

A value between 60 and 300 volts, while being a generous range, should prevent issues where the value read from MQTT is read as a `0` or erroneously specified in millivolts.

chris1howell · 2022-03-09T20:35:14Z

I have tested the RFID support (local authorization) with great results. I was able to scan and save 25 tags and start charge sessions with tags.

I used an Adafruit RFID reader https://www.adafruit.com/product/364. A custom OpenEVSE RFID reader will be tested and avaliable soon.

jeremypoulter

A few comment, overall looks good, however would be good to do lots of smaller PRs, one for each issue, rater than a single large PR, a lot easer to review.

platformio.ini

jeremypoulter · 2022-03-08T22:45:36Z

src/web_server.cpp

+  rfid.waitForTag(60);
+}
+
+void handlePollRFID(MongooseHttpServerRequest *request) {


Can we not poll, better to add the data to the status endpoint and also send them via event_send.

Alternatively an additional WebSocket endpoint just for RFID events would be better than polling.

Okay, I will rework that

src/rfid.cpp

jeremypoulter · 2022-03-08T23:00:19Z

src/rfid.cpp

+}
+
+void RfidTask::setup(){
+    if (config_rfid_enabled()) {


Will this require a reboot after RFID is enabled?

Actually, the lifetime of the connection to the PN532 is fully handled in loop() so this is kind of redundant. I better remove it.

jeremypoulter · 2022-03-08T23:06:44Z

src/rfid.cpp

+        // Send to MQTT broker
+        DynamicJsonDocument data{JSON_OBJECT_SIZE(1) + uid.length() + 1};
+        data["rfid"] = uid;
+        mqtt_publish(data);


Should use event_send to also send over the websocket

Is this sufficient for that purpose?

Not really, that is used for filling in the status when polled or reporting to EmonCMS, event_send will send the value over MQTT and the WebSocket so any clients get updated immediately rather than up to 30 seconds later.

src/rfid.cpp

src/evse_man.h

matth-x · 2022-03-10T20:55:40Z

Thank you for the input, @jeremypoulter ! I will take a deeper look into everything and will continue this conversation when I have further questions.

matth-x · 2022-03-24T22:04:17Z

Now the PN532 driver and the RfidTask are separate modules. And I changed the RFID section in the dashboard so it doesn't require polling from the UI anymore.

jeremypoulter · 2022-03-24T22:48:37Z

src/pn532.cpp

+
+    if (!config_rfid_enabled()) {
+        status = DeviceStatus::NOT_ACTIVE;
+        return SCAN_FREQ;


No need to poll if RFID is not enabled, can just return MicroTask.Infinate and manually wake the task when the config is changed.

Okay. Only for the PN532 driver or should I also do it so in the RfidTask? And what would be the best way to detect when config_rfid_enabled() turns on again? (another case as example is sufficient)

jeremypoulter · 2022-03-24T22:53:50Z

src/pn532.cpp

+
+    listen = true;
+    //"flush" ACK
+    delay(30);


Shouldn't use delay should set some state and drop out of loop with a return of 30 and perform the action next time round, as it happens read is the first thing that is done in the loop so might be ok

Okay, will change it

jeremypoulter · 2022-03-24T23:03:31Z

Also can you merge in the changes from master?

jeremypoulter · 2022-03-24T23:26:59Z

The other thing that needs doing is adding a build option to enable the RFID code as it needs to be disabled on the ESP32 Gateway board. In the default config the I2C conflicts with the Ethernet and despite the large expansion connector there are not suitable alternative pins for I2C so it is best to remove.

matth-x · 2022-03-25T14:22:40Z

Okay, will merge from master.

Do you have an opinion about to what extend the build option should skip the RFID code? My ideas are:

Replace the PN532 driver with an empty stub
~~Handle this inside the RfidTask class~~
~~Remove RFID modules and all references to it with precompiler statements~~

Edit: Option 1 seems to have the the least potential for errors. A possible improvement would be to surround everything in pn532.* with #ifdef ENABLE_PN532.

matth-x · 2022-03-26T17:10:15Z

I have implemented your feedback, except to stop polling when config_rfid_enabled() is false. As ocpp.cpp has the same behavior and I should rework both, can we maybe postpone this to the next PR?

jeremypoulter · 2022-03-26T18:28:57Z

Thanks will check it over. As a preference I would prefer the code is compiled out rather than Dummy/Null drivers. Will make for a smaller binary.

You can leave the infinite return for another PR if you would prefer. There is a function in app_config that notifies of changes to the config. Not an ideal way of doing it, but is ok.

jeremypoulter · 2022-05-16T22:25:06Z

@matth-x I have done some basic testing and it looks mostly ok but there are still a few issues. I need this merged to unblock some other bits, so I am going to merge as with the latest changes it is mostly ok.

Please raise tickets for the following and fix as separate PRs:

With RFID (build option) enabled the GUI options should still be hidden if no RFID module detected
Add build option to completely remove the RFID code
Stop polling the OCPP/RFID tasks then they are disabled
Sometimes it looks like the card gets stuck in a reading state and won't read any additional tags, It is fairly easy for me to tell this as the dev set up is really noisy from an RF POV and I can hear the activity on my speakers and there is a very distinct sound when scanning vs reading a card

retrocoder-78 · 2022-06-21T21:55:12Z

I have tested the RFID support (local authorization) with great results. I was able to scan and save 25 tags and start charge sessions with tags.

I used an Adafruit RFID reader https://www.adafruit.com/product/364. A custom OpenEVSE RFID reader will be tested and avaliable soon.

Any news on the OpenEVSE reader hardware @chris1howell? Thanks!

Oliver Norin and others added 30 commits February 11, 2021 15:21

Added RFID support and pause timer

6124a1f

settings for RFID and sleep timer in gui

829353d

Moved PN532 library to platform.ini

0585bac

Restrict voltage via MQTT to a known sane range

c334b5c

A value between 60 and 300 volts, while being a generous range, should prevent issues where the value read from MQTT is read as a `0` or erroneously specified in millivolts.

Update readme.md

2547b50

Update wired-ethernet.md

334553b

Added RFID support and pause timer

b9edf7f

Moved PN532 library to platform.ini

68f4e6c

RFID MicroTask

b06d727

Merge branch 'rfid_and_timer' of github.com:notima/ESP32_WiFi_V3.x in…

fd54b0c

…to rfid_and_timer

removed sleep timer

9be12bc

wake up PN532 every 30 seconds

1b278ab

server endpoints for adding rfid tags

61cb664

timer utilizing scheduler

a0b669c

Wire library initiated in main

c8db15d

Merge pull request #2 from OpenEVSE/jeremypoulter/issue5

35ccdb0

Jeremypoulter/issue5

Merge branch 'jeremypoulter/issue5' into rfid_and_timer

a97215c

renamed lastState to state

cd3ed56

Added RFID info to lcd

e03e9f6

Using OpenEVSE api instead of rapiSender

597e9a0

Restrict voltage via MQTT to a known sane range

d5963f9

A value between 60 and 300 volts, while being a generous range, should prevent issues where the value read from MQTT is read as a `0` or erroneously specified in millivolts.

Update readme.md

2e529d1

Added RFID support and pause timer

991c5ce

Moved PN532 library to platform.ini

cb83f99

RFID MicroTask

d5bf8c9

Added RFID support and pause timer

d92e04d

Moved PN532 library to platform.ini

3465f59

removed sleep timer

24bdd82

jeremypoulter reviewed Mar 9, 2022

View reviewed changes

jeremypoulter added a commit that referenced this pull request Mar 17, 2022

Merge of #325 as that is needed for the latest ArduinoOcpp library

70a1812

This was referenced Mar 20, 2022

Rfid and timer #167

Closed

Platform IO ESP32 core 4.4 #207

Merged

matth-x added 2 commits March 23, 2022 19:49

change dependencies to pio packages

ea25345

separate PN532, no UI polling when adding ID tag

7c24239

jeremypoulter reviewed Mar 24, 2022

View reviewed changes

matth-x added 6 commits March 26, 2022 17:21

add build option to enable PN532

69d34e6

send rfid uid immediately as event

430a10b

remove delays

351ee0e

add RFID failure indication in UI

64d5d49

update gui

b04ffd2

Merge remote-tracking branch 'upstream/master'

bb8f921

matth-x added 4 commits May 9, 2022 13:24

RFID invisible on GUI when disabled

856d970

commit auto-generated GUI files

11ac3dd

Merge remote-tracking branch 'upstream/master'

3483e75

fix visibility property

8da38a4

jeremypoulter approved these changes May 16, 2022

View reviewed changes

jeremypoulter merged commit 923bba4 into OpenEVSE:master May 16, 2022

matth-x mentioned this pull request Dec 31, 2022

OCPP upgrade #506

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add RFID support and upgrade OCPP library #325

Add RFID support and upgrade OCPP library #325

matth-x commented Mar 8, 2022

chris1howell commented Mar 9, 2022

jeremypoulter left a comment

jeremypoulter Mar 8, 2022

matth-x Mar 10, 2022

jeremypoulter Mar 8, 2022

matth-x Mar 10, 2022

jeremypoulter Mar 8, 2022

matth-x Mar 24, 2022

jeremypoulter Mar 24, 2022

matth-x commented Mar 10, 2022

matth-x commented Mar 24, 2022

jeremypoulter Mar 24, 2022

matth-x Mar 25, 2022

jeremypoulter Mar 24, 2022

matth-x Mar 25, 2022

jeremypoulter commented Mar 24, 2022

jeremypoulter commented Mar 24, 2022

matth-x commented Mar 25, 2022 •

edited

Loading

matth-x commented Mar 26, 2022

jeremypoulter commented Mar 26, 2022

jeremypoulter commented May 16, 2022

retrocoder-78 commented Jun 21, 2022

Add RFID support and upgrade OCPP library #325

Add RFID support and upgrade OCPP library #325

Conversation

matth-x commented Mar 8, 2022

chris1howell commented Mar 9, 2022

jeremypoulter left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

matth-x commented Mar 10, 2022

matth-x commented Mar 24, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jeremypoulter commented Mar 24, 2022

jeremypoulter commented Mar 24, 2022

matth-x commented Mar 25, 2022 • edited Loading

matth-x commented Mar 26, 2022

jeremypoulter commented Mar 26, 2022

jeremypoulter commented May 16, 2022

retrocoder-78 commented Jun 21, 2022

matth-x commented Mar 25, 2022 •

edited

Loading