Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Intellitouch i10-3D + expansion #246

Closed
ckurlinski opened this issue Nov 15, 2020 · 57 comments
Closed

Intellitouch i10-3D + expansion #246

ckurlinski opened this issue Nov 15, 2020 · 57 comments

Comments

@ckurlinski
Copy link

I am not able to pull the expansion, ID (32) into njsp running capture replay.
If I go into the indoor controller, under Advance, get image from outdoor device, njsp crashes, then restart njsp, the expansion is there.

Here is the replay.zip

replay.zip
@tagyoureit
Copy link
Owner

Hey, so I'm not getting any errors from the replay. I do have some logic in the app for the expansion panels. But there isn't enough info in here for me to debug anything.

  1. Does this replay include when your expansion panel is setup?
  2. Can you tell me what the error is when the app crashes? It should tell us what line is causing the crash. Also, can you tell me what the last packets are that were received? Since the app is crashing it may not capture these through the replay. You can turn on all logging and capture what is shown in the console log when everything goes south.

I'm pretty sure that I built the expansion panel logic from one of your previous captures. Unfortunately, I've only seen the i10X and it isn't 100% clear how to tell the different expansion panels from each other.

@ckurlinski
Copy link
Author

ckurlinski commented Nov 17, 2020 via email •
edited
Loading

@ckurlinski
Copy link
Author

ckurlinski commented Nov 17, 2020 via email

@tagyoureit
Copy link
Owner

Also, this is a master list of all the curl put json configs that work so far for me.

Nothing was attached.

Let's break this down a bit. There is a lot to unpack here...

Every time I run a replay capture, I loose the expansion.

I really don't understand what's going on here. Running a replay capture is no different from starting the app up from scratch the first time. If you delete your ./data/*.json files does that also loose your expansion panel?

The commands I use to program it are from the swagger ui page, and for the most part they work great.
I think there is an overlap in the intellitouch board specification affecting the custom names; should start with 200, but starts with 201.
So I can’t assign customs names with curl put, but I can work around that.

Are you saying IntelliTouch starts at 201 and not 200?

Scheduling is not working either with curl

Need more details here.

If you can provide me with some more specifics that would really be helpful. I still need the exact output from the console when you say njsPC crashes. For this, if you enable all logging and also console logging it should output everything to your screen. The you can manually copy/paste the results back here.

@ckurlinski
Copy link
Author

Replay Master Branch
replay-2.zip
replay-3.zip
replay-4.zip
replay-5.zip
replay.zip

@ckurlinski
Copy link
Author

logs.zip

@ckurlinski
Copy link
Author

logs_no_expansion_master.zip

@ckurlinski
Copy link
Author

This is from main branch, no expansion, indoor / get image from outdoor
NJSP crashed

packetLog-get_out_to_in.log

@ckurlinski
Copy link
Author

Main Branch, no expansion, indoor send image to outdoor
No crashes
2020-11-17_21-47-50.zip

@ckurlinski
Copy link
Author

Next Branch, no expansion, capture with reset

2020-11-17_21-58-56.zip

@ckurlinski
Copy link
Author

Next branch, no expansion, indoor get image from outdoor, no crash!

2020-11-17_21-58-56_2.zip

@ckurlinski
Copy link
Author

Next Branch, no expansion, indoor send image to outdoor, no crash.

2020-11-17_21-58-56_3.zip

@tagyoureit
Copy link
Owner

tagyoureit commented Nov 18, 2020 •
edited
Loading

Hmm... I can't seem to find where the configuration has gone for the expansion panels. Here are the 2 (status) packets:

OLD CAPTURE WITH i10X
10,47,243,63,0,0,0,1,0,[96],64,0,0,0,64,82,32,0,82,0,0,0,0,4,0,60,103,1,5

NEW WITH i10X
13,28,32,12,0,0,0,0,0,[64],64,0,0,0,136,80,0,0,78,0,0,0,[4],4,0,163,72,1,5

NEW WITHOUT i10x
17,11,33,0,0,0,0,0,0,[64],64,0,0,0,113,81,0,0,78,0,0,0,0,4,0,163,72,1,5

The number in brackets previously told me the expansions. 96 vs 64... the 7th bit told the story. But with the new firmware it is consistently a 64. I've gone through the other packets and nothing is sticking out at me as indicating an expansion panel. I'll keep looking but a bit disappointed not to find it tonight.

EDIT:
One thing to try is to capture anything that might start/be broadcast with the OCP boot sequence:

  1. Start njsPC. Let it finish doing it's sync thing.
  2. Turn off OCP
  3. send a PUT to njsPC (/app/logger/clearMessages) to clean out the logs
  4. Turn on OCP
  5. Let it finish it's cycle of boot up messages and syncing with the other ICP
  6. Repeat with the i10x card removed

@rstrouse
Copy link
Collaborator

rstrouse commented Nov 22, 2020 •
edited
Loading

Byte 22 is an unidentified byte on the 2 message for IntelliCenter. This byte is directly after waterSensor4. Perhaps this byte is not our indicator though. To solve that mystery, the exact process @tagyoureit outlined above needs to be completed where we have two separate log files that are clearly identified. One with the i10x installed and one without the i10x installed during the time njspc is trying to get the information from the OCP. Only then can we determine if another byte (perhaps on the action 5 message) contains the data we are looking for.

Below is the 2 message for IntelliCenter and below that is the 204 message which contains the expansion panels:
image

This is a big message from IntelliCenter but byte 13 identifies the installed expansion modules for master and byte 14-18 identify which modules are installed on the expansion panels to include valve expansions and personality cards (i10x, i5x).
image

@ckurlinski
Copy link
Author

This is (4) sets of logs.
After each test, added test(#) to each folder.
This was done off a clean next install, no mods, no edits to the data pool json.

Test 1 is base system, nothing removed
Test 2 is with out the i10X
Test 3 is without the i10x and the valve expansion board inside the i10-3D
Test 4 is with everything back up and running

All test were preformed after a stop and restart of njsp, using node dist/app.js

After all the tests, I stopped, and zipped the whole logs dir.
This is what I'm posting.

Thanks

2020-11-24-logs.zip

@rstrouse
Copy link
Collaborator

@tagyoureit -- Since there is no 204 message from Touch I looked at the differences between the files provided for action 2. Here they are for the 4 tests. It appears there is no difference between them.

//#1 [255, 0, 255][165, 1, 15, 16, 2, 29][9,  3, 243, 61, 0, 0, 0, 3, 0, 64, 64, 0, 0, 0, 99, 78, 0, 0, 77, 0, 0, 0, 0, 4, 0, 163, 72, 1, 5][4, 150]
//#2 [255, 0, 255][165, 1, 15, 16, 2, 29][9, 29, 33,   0, 0, 0, 0, 0, 0, 64, 64, 0, 0, 0, 99, 64, 0, 0, 77, 0, 0, 0, 0, 4, 0, 163, 72, 1, 5][3, 144]
//#3 [255, 0, 255][165, 1, 15, 16, 2, 29][9, 33, 241,  0, 0, 0, 0, 2, 0, 64, 64, 0, 0, 0, 99, 79, 0, 0, 76, 0, 0, 0, 0, 4, 0, 163, 72, 1, 5][4, 116]
//#4 [255, 0, 255][165, 1, 15, 16, 2, 29][9, 44, 243, 60, 0, 0, 0, 3, 0, 64, 64, 0, 0, 0, 99, 78, 0, 0, 76, 0, 0, 0, 0, 4, 0, 163, 72, 1, 5][4, 189]

There is a difference for an action 109 which is identified as iLink. Here it is between the tests. Note the value in byte(1) for #1 and #4 when the i10x is installed. I don't see anything related to the valve expansion though. @ckurlinski -- do you have an iLink installed?

//#1 [255, 0, 255][165, 1, 15, 16, 109, 10][1, 15, 0, 0, 0, 0, 0, 0, 0, 0][1, 76]
//#2 [255, 0, 255][165, 1, 15, 16, 109, 10][1,  0, 0, 0, 0, 0, 0, 0, 0, 0][1, 61]
//#3 [255, 0, 255][165, 1, 15, 16, 109, 10][1,  0, 0, 0, 0, 0, 0, 0, 0, 0][1, 61]
//#4 [255, 0, 255][165, 1, 15, 16, 109, 10][1, 15, 0, 0, 0, 0, 0, 0, 0, 0][1, 76]

@rstrouse
Copy link
Collaborator

Wait a minute is byte(7) part of the features? If it isn't it's right there in action 2 byte(7).

@ckurlinski
Copy link
Author

No ilink installed
I will run the same tests tomorrow with the Indoor unit and the Spa Controller disconnected.

@rstrouse
Copy link
Collaborator

Please do me a favor for all of the tests make sure everything is turned off. The extra messages are noise that make it harder to sift through.

@tagyoureit
Copy link
Owner

Hmm, I was pretty sure that 109-111 was iLink. But if it's config info that could be different. But I've never seen these packets on any other system then @ckurlinski .

@ckurlinski
Copy link
Author

Could 109, 110, 111 be the new Expansion addresses?
Anyway, I will do the test without the scheduled items running.
Thanks

@ckurlinski
Copy link
Author

Dumb Question, could 109 be NJSP?
It is showing up on the logs with nothing attached and in service mode

@rstrouse
Copy link
Collaborator

No, this is a broadcast message that is on the RS485 bus. The destination is 15 which tells me it is a panel 16 sending out information to whom it may concern. Bear in mind, not all messages follow this as a rule but most do. For IntelliTouch I believe all messages come from source 33 when njspc originates them.

I don't know how the expansion panels work in IntelliTouch so are you placing the master in service mode or can you place the expansion in service mode?

@ckurlinski
Copy link
Author

The system is in service mode
Process for this test; rm -rf data, mv logs ~/i10-3D_{test_type}, mkdir data mkdir logs, then start it up.
Doing following tests: i10-3D_only, i10-3D_Valve_expansion, i10-3D_INDOOR_UNIT, i10-3D_i10x

@ckurlinski
Copy link
Author

Here's an interesting set of logs
Wiped / reset , i10-3D and i10x only, nothing else.

i10-3D_AFTER_UNIT_WIPE_i10x_2nd.zip
i10-3D_AFTER_UNIT_WIPE_i10x.zip

@ckurlinski
Copy link
Author

And after a clean njsp next install
i10-3D_AFTER_UNIT_WIPE_i10x_CLEAN_NJSP_NEXT_INSTALL.zip

@rstrouse
Copy link
Collaborator

So you placed the panel in service mode. That will probably mess with our results. Still I see 109s with a 2 now in the first byte but this could be related to the fact that the master is disabling all incoming command related messages.

Is the process you are using to enable the expansion panel described here on page 26?
https://www.pentair.com/content/dam/extranet/pentair-pool/residential/manuals/automation/intellitouch/intellitouch-pool-control-system-users-guide-english.pdf

If it is I want a capture of you doing this process. I don't want you to set service mode, have anything turned or wipe the unit, just enable the i10x using the described process. Please also do not issue any commands from njspc or press any commands on any of your controllers. All of this just becomes noise that obscures the messages that we are interested in.

Please have njspc running then turn on the capture without reset before you start the process. Then when you are finished, wait about a minute, then turn off the capture and upload those files.

@ckurlinski
Copy link
Author

Just did it, forgot to turn on packet capture
trying again

@ckurlinski
Copy link
Author

No Pumps, Valve Expansion Boards, no Chlorinators, no INDOOR Unit attached
Just the i10-3D and the i10x

i10-3D_AFTER_UNIT_WIPE_i10x_CLEAN_NJSP_NEXT_WIPED_5_FUCKING_TIMES.zip

@ckurlinski
Copy link
Author

With pumps and some config to get some things running

i10-3D_AFTER_UNIT_WIPE_i10x_CLEAN_NJSP_NEXT_WIPED_ADDED_PUMPS_BASIC_CONFIG.zip

@rstrouse
Copy link
Collaborator

Ok, just to be sure. The i10-3D_AFTER_UNIT_WIPE_i10x_CLEAN_NJSP_NEXT_WIPED_5_FUCKING_TIMES.zip was the process from page 26 of the manual doing a capture without reset or was this something else?

@tagyoureit - Once his system stabilizes, a 109 is sent on periodically and address 17 responds with an ack. 17 is a likely candidate for an expansion panel. 16 is the originating address and 17 responds.
image

@ckurlinski
Copy link
Author

Yeah but 109-11 was showing up on a fresh wiped system, with nothing but RPI w/NJSP connected to it....

i10-3D_AFTER_UNIT_WIPE.zip

@rstrouse
Copy link
Collaborator

Yeah but that was with the comm port connected to the expansion load center... correct?

@ckurlinski
Copy link
Author

no, just the i10-3D

@rstrouse
Copy link
Collaborator

So without the load center wired in at all? What else was connected to RS485 at the time? The OCP is clearly communicating with something over RS485.

@ckurlinski
Copy link
Author

Only (1) ONE thing was connected to the i10-3D
I will do it again
Including the reset with live capture

@rstrouse
Copy link
Collaborator

Please do not do a capture with reset. That causes njspc to interrogate your controller which sends hundreds of requests and responses. Do a capture without reset

@ckurlinski
Copy link
Author

I meant resetting the i10-3D while doing a capture, else NJSP does this virtual controller thing

@rstrouse
Copy link
Collaborator

Ah yes.

@rstrouse
Copy link
Collaborator

After weeding through about 1,600 messages I cam across this sequence in your i10-3D_AFTER_UNIT_WIPE.zip it seems to indicate the i10-3D asking for control panel asking for response on 109. Notice there is no ACK and byte(24) of the 2 message is 0.
image

Then later in the sequence it tries to ping 18 and 19 to see if there is anything there directly. The timing between 776 and 777 however was almost a minute so this tells me that the panel was taken off line then brought back online in that time. The state and config calls indicate that you refreshed -webClient after you restarted or it detected it and refreshed on its own.

In that time byte(24) switched to 4. Unfortunately, I have no idea what is located on byte(24) and the act of taking the panel offline and bringing it back up may simply cause it to delay.
image

@ckurlinski
Copy link
Author

Wiped like my dainty ass....

i10-3D_AFTER_UNIT_WIPE.zip

@ckurlinski
Copy link
Author

sorry wrong log

@ckurlinski
Copy link
Author

People were talking to me, how dare they

i10-3D_ONLY_RESET_WIPE.zip

@rstrouse
Copy link
Collaborator

Ok so here is what I know. On startup the OCP asks on address 109 who is out there. If no expansion panel exists then it gets no response. However, when an expansion panel is installed it responds with an ACK on the address of the expansion panel. This occurs during restart until it establishes that the control panel exists or it doesn't. If the expansion panel exists it continues to ping it on 109. The payload on that ping appears to be [1,x,x,...] where byte(0) indicates the expansion number and at least byte(1) indicates the relays that should be active. So when we were seeing [255, 0, 255][165, 1, 15, 16, 109, 10][1, 15, 0, 0, 0, 0, 0, 0, 0, 0][1, 76] it was telling EXP1 to activate relays 1-4.

During startup, it goes looking for any controller on action 110 which we do not process These addresses are 19, 18, and 17 working backwards to 17. If it already has confirmation for 17 on the 109 message then it doesn't ask for it on 110. That is the process that the OCP goes through to detect whether there is an expansion panel on the system. I suspect this is true for all IntelliTouch boards at reboot of the OCP.

Now what we need to know from the OCP is whether it has detected one or not. I suppose we could look for ACKs on 109 to determine if the expansion exists but there has to be a more direct route. For example, ACK 109 from address 17 = EXP1, ACK 109 from address 18 = EXP2, ACK 109 from address 19 = EXP3.

Do you remember if there is a screen on the ICP that displays the installed panels? If there is we may be able to spy on the ICP asking whether they exist.

@ckurlinski
Copy link
Author

Zoom?

@rstrouse
Copy link
Collaborator

First do you have dashPanel set up? We will need that for the MessageManager tools.

@ckurlinski
Copy link
Author

AllI have to does start it

@rstrouse
Copy link
Collaborator

Do you know if there is anywhere on the ICP that displays the installed expansion panels?

@ckurlinski
Copy link
Author

It does, when the expansion panels are on the system, the ICP adds additional screens , Main Display, Display 1, Display 2, Features
If the expansion is missing, then Display 2 isn't there

@rstrouse
Copy link
Collaborator

Hmmm... is this automatic or does it require you to do the get image from outdoor to have the Display 2 show up?

@ckurlinski
Copy link
Author

It's Pentair-a-matic, the ICP is a POS, I think it intercepts the commands from NJSP.
I tried to create a Feature, shows up on the ICP, but not the OCP, so now I have to send the image to the OCP

@ckurlinski
Copy link
Author

Wait, it took (5) minutes to update the OCP, the feature is there now.

@rstrouse
Copy link
Collaborator

I ditched IntelliTouch. I am so glad I did. IntelliCenter is not without its warts but at least there are a couple of spots with clear skin. All told, when you add up the amount of time I spent shaking a chicken bone in the air to get IntelliTouch programmed and working and the time I spent troubleshooting equipment I could have built my own.

Fwiw, the system for when it was built did plenty more than any other. However, it is very long in the tooth. You can bet the OCP had your feature but it needed to send 40 circuits & features out to everyone involved. Then when the ICP go it it needed to ask for every one of the 99 schedules so what you have is a few hundred tiny messages being shot back and forth between everyone involved.

So the question stands. If the ICP currently does not have Display 2 tab (no i10x installed). Does the act of installing it on the OCP automatically make the Display 2 tab appear or do you have to get the image from the outdoor panel?

@ckurlinski
Copy link
Author

I have the i10x up, there is something going on with the system, it's a mess, I have factory reset the whole system literally (10) times today.
I have a meeting in the morning to pass judgement on this thing.....
Your opinion, after seeing my logs and config, dump or stay?

@rstrouse
Copy link
Collaborator

My opinion is to dump it. You are at it's limit when it is working properly. My experience over the years with IntelliTouch includes having to replace the board twice. The third time I decided it was time to move on because I thought the personality board was fried but thinking back it was probably the power supply. There were times when the ICP would lock up and the wireless controller would cease to operate as well. It always seemed to happen when I didn't have time to troubleshoot it or it was raining. Heck to replace the wireless controller battery is $400 every 2 years. A new wireless controller for IntelliCenter is less than that.

So I went searching for alternatives. Jandy & Hayward... nope (I'm not even sure Jandy has a system that would accommodate your needs, certainly not iAquaLink). I came across this project and thought that maybe I would build my own but IntelliCenter seemed to be able to do what I wanted so I went with it. I have to say I am actually pretty happy with IntelliCenter at the moment. Sure one fw ago it was tweaking my body temp setpoints but it has been pretty solid on 1.047.

Another bonus it that the control panels are just a better interface. They are point and jab and don't require tapping out morse code to get to the menu you need. Even the wife can operate it without destroying the programming. If I had a nickel for every time she deleted the solar heater when she wanted to shut it off... or somebody house sitting had trouble changing the setpoint on the pool or spa.

Heck I can't tell you how many times I found the Gas Heater heating the pool because both bodies were set when they only wanted the spa heater. IntelliCenter solves these problems by assigning equipment to the body explicitly.

@tagyoureit
Copy link
Owner

Thanks guys, for continuing to work on this... Good catch on the 109. Strange that I've never seen that before from any other system.

@ckurlinski
Copy link
Author

Upgrading the system to IntelliCenter.
The issues might be bad hardware, and not related to NJSP.

Thanks

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@tagyoureit @ckurlinski @rstrouse