reef-pi :: An opensource reef tank controller based on Raspberry Pi.

[pickupman66]

Can you check you have alert enabled in your telemetry configuration? Remember you have to reload reef-pi if you change telemetry configuration, for it to take effect.

@ranjib, yes telemetry is enabled. It works also when u test. Telemetry was enabled when I set it up and it has worked for the temperature function.

 

[bishoptf]

I got a nasty flu, expect sporadic presence for next few days . I plan to continue working on the ph driver, it’s almost there. Some work on integrating the calibration logic /ui left.
Meanwhile I have now a full blown test setup . I am loving my garage

This is the controller I plan to giveaway during local coral farmers market

I’m thinking if I should give this handmade controller or a pico baseboard from theatrus. Suggestions?
I have now almost 4 complete set of reef-pi for testing , each with different specs (pi zero, pi b, pi 2 and pi 3). 2019 should be fine

On a side note ,I hooked up one of my son’s n scale locomotive with reef-pi, and it worked :/) , the circuit is exactly same as led strip , using a logic level power mosfet . I’ll post some pictures soon

Take care and get some rest the flu is not something to mess with, hope you feel better soon!

 

[Bigtrout]

@ranjib, yes telemetry is enabled. It works also when u test. Telemetry was enabled when I set it up and it has worked for the temperature function.

Check and see how many feeds you have on adafruit. You are limited to 10 on a free account, any additional get ignored.

 

[Ranjib]

Check and see how many feeds you have on adafruit. You are limited to 10 on a free account, any additional get ignored.

Yes , that’s exactly what I suspect. You have exhausted 10 feeds , and then reef-pi is trying to send metrics to a feed that does not exist

 

[pickupman66]

Yes , that’s exactly what I suspect. You have exhausted 10 feeds , and then reef-pi is trying to send metrics to a feed that does not exist

I am not feeding to adafruit. I want it to email me when it has an issue. Email alert when the pump has run too long and shut itself off

 

[rushbattle]

The PCA9685 has a limit 1526Hz to the PWM output using the onboard clock. If you feed it 50MHz on the clock pin you can double that to 3052Hz. I find both too slow for all cases especially if you have low duty cycle percents as you’ll get “nice” banding on some cameras.

At boot the PCA9685 is 200Hz (strobe city).

As for the driver let me confirm the settings value is making it all the way through.

Any ideas about how to otherwise output PWM via reef-pi at higher frequency for use with your drivers?

 

[Bigtrout]

Any ideas about how to otherwise output PWM via reef-pi at higher frequency for use with your drivers?

Im looking at lower frequency which reef pi should already do, but its not workin correctly

 

[Ranjib]

Im noting down the pwm frequency issues. If I understand correctly, its not correct for pi driver, and for pca9685 its not working at all. I'll cross check with my test builds.

 

[theatrus]

I'll have hardware in front of me tomorrow to confirm the PWM issues (PWM did not apply in Kauai ). The initial read of the code vs the Linux driver looks ok, but I haven't verified the code flow end to end.

https://www.kernel.org/doc/Documentation/pwm.txt

When a PWM channel is exported a pwmX directory will be created in the
pwmchipN directory it is associated with, where X is the number of the
channel that was exported. The following properties will then be available:

period
The total period of the PWM signal (read/write).
Value is in nanoseconds and is the sum of the active and inactive
time of the PWM.

duty_cycle
The active time of the PWM signal (read/write).
Value is in nanoseconds and must be less than the period.

Its possible something dumb like wrapping around in the BCM driver is happening vs the value we are setting.

 

[Bigtrout]

Im noting down the pwm frequency issues. If I understand correctly, its not correct for pi driver, and for pca9685 its not working at all. I'll cross check with my test builds.

@Ranjib
In a nushell, I found that in the setting box for pi pwm, if you input a number, here are resulting hz readings:
Setting 25=400hz
Setting 50=200hz
Setting 60=166hz
Setting 75=133 hz
Setting 100=100hz
Setting 150=66hz
Setting 200=50hz
This formula is the relationship:
10000/pi pwm box setting=measured hz

As for the pca9685, no matter what I put in the setting box, i get 1500hz

 

[Ranjib]

@Ranjib
In a nushell, I found that in the setting box for pi pwm, if you input a number, here are resulting hz readings:
Setting 25=400hz
Setting 50=200hz
Setting 60=166hz
Setting 75=133 hz
Setting 100=100hz
Setting 150=66hz
Setting 200=50hz
This formula is the relationship:
10000/pi pwm box setting=measured hz

As for the pca9685, no matter what I put in the setting box, i get 1500hz

Awesome,this will help a lot. Thanks for tabulating this . You have taken these readings via fluke multimeter ?

 

[Bigtrout]

Awesome,this will help a lot. Thanks for tabulating this . You have taken these readings via fluke multimeter ?

Yes those readings were taken with a fluke 179 multimeter which is supposed to do read dc frequencies from 2hz to 99.99khz with an accuracy of .1%
They were backed up by a second klein multimeter as well.

Also to note for all pwm(pi and pca9685), everything else works as its supposed to, with the exception of setting the pwm frequency correctly

Have u looked at output freq on your ocilloscope

 

[daveCB]

So here I am chasing my tail for the last two weeks with my temperature sensors. They were going off line frequently and my email was getting flooded.
I upgraded to mini XLR plugs and hard soldered the leads on the pi hat. I upgraded the sensors to rather costly sealed sensors ( posts on the thread )
So then back to the testing; Still off line about every two hours with my checking frequency set to 60 minutes.
Ok, as a test I changed the time to 30 seconds. After two days, not one off line event !
This points me to the program and not the equipment or build.
I will be interested in your comments.

 

[Ranjib]

So here I am chasing my tail for the last two weeks with my temperature sensors. They were going off line frequently and my email was getting flooded.
I upgraded to mini XLR plugs and hard soldered the leads on the pi hat. I upgraded the sensors to rather costly sealed sensors ( posts on the thread )
So then back to the testing; Still off line about every two hours with my checking frequency set to 60 minutes.
Ok, as a test I changed the time to 30 seconds. After two days, not one off line event !
This points me to the program and not the equipment or build.
I will be interested in your comments.

I cant think of anything specific in the code that will impact this. I have several temperature sensors (and some of them is controlling heater/cooler) that operates at 15 or 30 seconds interval

 

[theatrus]

So here I am chasing my tail for the last two weeks with my temperature sensors. They were going off line frequently and my email was getting flooded.
I upgraded to mini XLR plugs and hard soldered the leads on the pi hat. I upgraded the sensors to rather costly sealed sensors ( posts on the thread )
So then back to the testing; Still off line about every two hours with my checking frequency set to 60 minutes.
Ok, as a test I changed the time to 30 seconds. After two days, not one off line event !
This points me to the program and not the equipment or build.
I will be interested in your comments.

How much wire length and what size pull-up?

ReefPi should probably do retries at some level instead of reporting bogus values

 

[Ranjib]

Talking about temperature sensors.. I am doing some experiments with the heaters internal thermostat vs reef-pi controls and I am not very happy with the heater's internal control. Particularly for pico tanks this might be a problem. I am observing almost 1 degree temperature swings in 30 min period

I wanted to observe the results for several days before posting here. This is the dashboad of my 29G biocube, with a Eheim 100W jaeger heater in it. It was controlled by reef-pi till 1st Feb. Initially I played with reef-pi's internal temperature controller to check if i can dial down exact temperature (79.5, 79.6, 79.3 etc) and then with reef-pi configured to only monitor and let the heater's internal control do the on/off. The results are not very satisfactory. I'll be using the the heater's temperature graph to implement hysteresis in reef-pi (sane defaults).
Here is the dashboard of my pico tank (3.7G, with 25W aqueon heater)

Irrespective of the observations, I am excited about this data, I think we can pin down such behaviors across tank volume and heater types (wattage and brand) and hopefully push the boundary of what temperature stability means in reef tanks

 

[Ranjib]

How much wire length and what size pull-up?

ReefPi should probably do retries at some level instead of reporting bogus values

we do check CRC code/checksum. any reason for you to think that we are reporting bogus value? or any checks we should be making?
I have thoughts about using moving average and auto correcting anomolous values.. but not for current releases (3.0)

 

[Ranjib]

reporting back on the memory issue, not seeing it in all of my builds. I am running 2.1 where this should be visible. There are some cases where I can see the jump, but its not necessarily due to leak, and its not fatal. Here is a dashboard of ph monitor running for almost 3 weeks continuously now. I dont keep the dashboards open .. so thats a factor which is not tested

 

[Ranjib]

Here is another test build (the one running in the garage with a 5.5 G test tank) dashboard. This has two ato sensors, and three temperature sensors (two inside, one outside the tank)

 

[Bigtrout]

Talking about temperature sensors.. I am doing some experiments with the heaters internal thermostat vs reef-pi controls and I am not very happy with the heater's internal control. Particularly for pico tanks this might be a problem. I am observing almost 1 degree temperature swings in 30 min period

I wanted to observe the results for several days before posting here. This is the dashboad of my 29G biocube, with a Eheim 100W jaeger heater in it. It was controlled by reef-pi till 1st Feb. Initially I played with reef-pi's internal temperature controller to check if i can dial down exact temperature (79.5, 79.6, 79.3 etc) and then with reef-pi configured to only monitor and let the heater's internal control do the on/off. The results are not very satisfactory. I'll be using the the heater's temperature graph to implement hysteresis in reef-pi (sane defaults).
Here is the dashboard of my pico tank (3.7G, with 25W aqueon heater)

Irrespective of the observations, I am excited about this data, I think we can pin down such behaviors across tank volume and heater types (wattage and brand) and hopefully push the boundary of what temperature stability means in reef tanks

I agree, and this is alot more important in smaller tanks..I tested my pair of aqueon pros on their own co trol in my 75 and only see a variation of .2 to .3 degrees F over the course of a day, and that is directly related to room temperature(keep the house a few degrees cooler at night)
Because they have accurate solid state controls, reef pi is set up as my failsafe to shut them off if one runs away stuck ON.