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

OP
OP
Ranjib

Ranjib

7500 Club Member
View Badges
Joined
Apr 16, 2016
Messages
9,825
Reaction score
17,041
Location
Pleasant Hill, Concord
Rating - 0%
0   0   0
Hello again

Any ideas what i can do with my camera ?
Its an original raspberry cam.

330E8E6D-E888-45D1-AAE4-D161536598A0.png
Take periodic photos and upload them in google drive.

It’s not very good solution and I have not used it in a while. It takes significant amount of computing resources and the actual controller function can be at risk. I just found the 30$ security cameras (like Wyze) to be lit better at this.

That said, it’s not cool to have bugs like this, I’ll test and fix it . Thank you for reporting this. Is this happening on 3.0 ?
 
OP
OP
Ranjib

Ranjib

7500 Club Member
View Badges
Joined
Apr 16, 2016
Messages
9,825
Reaction score
17,041
Location
Pleasant Hill, Concord
Rating - 0%
0   0   0
Found a small issue with my own today. The LM7805 regulators do work fine, but require heat-sinking at the required wattage, gets too hot without it. Not a big deal but something to plan for. Since I don't have any heat-sinks around right now for that size part, I dropped in a similar sized switcher regulator from EzSbc, in this case their PSU-2-5. They make some great little switchers in a tiny package.

Back to having fun again.
Nice. I’m trying out ubec buck converter in my new builds, they are bit pricy but renowned for their superior quality and provides decent emd protection
 

Zippyfear

Community Member
View Badges
Joined
Apr 30, 2019
Messages
86
Reaction score
152
Rating - 0%
0   0   0
Take periodic photos and upload them in google drive.

It’s not very good solution and I have not used it in a while. It takes significant amount of computing resources and the actual controller function can be at risk. I just found the 30$ security cameras (like Wyze) to be lit better at this.

That said, it’s not cool to have bugs like this, I’ll test and fix it . Thank you for reporting this. Is this happening on 3.0 ?

agreed. I just picked up these.



They have magnetic bases, and are very slim attached to the side of the tank. so they actually are just about perfect. - They were recently on Black Friday sale - 2 for $45!

Please excuse the messy Reef-Pi - it's currently undergoing reconstruction with new circuits into a new case.

reef1.jpg
 

Blaxkin

Community Member
View Badges
Joined
Oct 4, 2019
Messages
27
Reaction score
46
Rating - 0%
0   0   0
agreed. I just picked up these.



They have magnetic bases, and are very slim attached to the side of the tank. so they actually are just about perfect. - They were recently on Black Friday sale - 2 for $45!

Please excuse the messy Reef-Pi - it's currently undergoing reconstruction with new circuits into a new case.

reef1.jpg


Thanks for the tip :) now i ordered one of the yi cam :)
 

Blaxkin

Community Member
View Badges
Joined
Oct 4, 2019
Messages
27
Reaction score
46
Rating - 0%
0   0   0
Take periodic photos and upload them in google drive.

It’s not very good solution and I have not used it in a while. It takes significant amount of computing resources and the actual controller function can be at risk. I just found the 30$ security cameras (like Wyze) to be lit better at this.

That said, it’s not cool to have bugs like this, I’ll test and fix it . Thank you for reporting this. Is this happening on 3.0 ?
Yes it happens in 3.0
 

Blaxkin

Community Member
View Badges
Joined
Oct 4, 2019
Messages
27
Reaction score
46
Rating - 0%
0   0   0
And now is see after the change for an extra power supply and an 3300uf capacitor i also has peaks. Not so much but the graph is invisible after 1 peak.

I cant coding :-( but is there a simple way in code to ignore 85 and 127 values ?
93732359-252E-4488-B199-595D5C0C497E.png
 
OP
OP
Ranjib

Ranjib

7500 Club Member
View Badges
Joined
Apr 16, 2016
Messages
9,825
Reaction score
17,041
Location
Pleasant Hill, Concord
Rating - 0%
0   0   0
And now is see after the change for an extra power supply and an 3300uf capacitor i also has peaks. Not so much but the graph is invisible after 1 peak.

I cant coding :-( but is there a simple way in code to ignore 85 and 127 values ?
93732359-252E-4488-B199-595D5C0C497E.png
Nothing yet. But you can use the reef-pi db command to do some surgery against the database and remove those values (replace ID with the specific id for your temperature sensor)
Code:
sudo systemctl stop reef-pi.service
sudo reef-pi db show temperature_usage <ID> > usage.json
# edit the json file and remove the bogus values
cat usage.json | sudo reef-pi db update temperature_usage <ID>
sudo systemctl start reef-pi.service

you can get the id of temperature sensors/controllers like this:
Code:
sudo reef-pi db list temperature
 

hhaase

Active Member
View Badges
Joined
Nov 11, 2019
Messages
415
Reaction score
355
Rating - 0%
0   0   0
The next version of my hat is close to being locked-down for ordering. Still developmental at this stage, but I'm comfortable/confident enough to start adding more features now. I'll keep adding stuff as my understanding gets stronger of how things work. Next round is a big layout change, which should be more enclosure friendly. I'm not quite ready to start releasing things into github yet for these but I'll again be offering up leftovers for cost when I've tested these out.

It may be a bit aggressive of me to do this big a change so soon, given how little I've tested, but I'm already doing a board order this weekend, and I can piggy-back this revision in and save a lot on shipping.

1: Adding spots for dual Atlanta Scientific probe modules, including an isolation circuit on-board.
2: Making the 4-pin headers both compatible for temp probe circuits.
3: Aligning the Pi-Zero so it can mount on top, and all connectors other than power face the same direction.
4: Moving all the resistors to surface mount, as this is already getting crowded quickly, and I have plans still.

Reefpi v0_2.JPG
 
OP
OP
Ranjib

Ranjib

7500 Club Member
View Badges
Joined
Apr 16, 2016
Messages
9,825
Reaction score
17,041
Location
Pleasant Hill, Concord
Rating - 0%
0   0   0
The next version of my hat is close to being locked-down for ordering. Still developmental at this stage, but I'm comfortable/confident enough to start adding more features now. I'll keep adding stuff as my understanding gets stronger of how things work. Next round is a big layout change, which should be more enclosure friendly. I'm not quite ready to start releasing things into github yet for these but I'll again be offering up leftovers for cost when I've tested these out.

It may be a bit aggressive of me to do this big a change so soon, given how little I've tested, but I'm already doing a board order this weekend, and I can piggy-back this revision in and save a lot on shipping.

1: Adding spots for dual Atlanta Scientific probe modules, including an isolation circuit on-board.
2: Making the 4-pin headers both compatible for temp probe circuits.
3: Aligning the Pi-Zero so it can mount on top, and all connectors other than power face the same direction.
4: Moving all the resistors to surface mount, as this is already getting crowded quickly, and I have plans still.

Reefpi v0_2.JPG
i love seeing these iterative improvements. Curious why atlas scientific board over reef-pi ph board. Purdon me if yuu have already answered.
 

hhaase

Active Member
View Badges
Joined
Nov 11, 2019
Messages
415
Reaction score
355
Rating - 0%
0   0   0
doesn’t the reef-pi ph board use an atlas module? I was using the guide linked on page 1 of this thread. Or was there a different version I didn’t see? I’ve only got this in cad so far, easy to change still.

I’m still learning the system and I’m not familiar at all with the coding side, so I’m trying to stick with known-supported hardware.
 
OP
OP
Ranjib

Ranjib

7500 Club Member
View Badges
Joined
Apr 16, 2016
Messages
9,825
Reaction score
17,041
Location
Pleasant Hill, Concord
Rating - 0%
0   0   0
doesn’t the reef-pi ph board use an atlas module? I was using the guide linked on page 1 of this thread. Or was there a different version I didn’t see? I’ve only got this in cad so far, easy to change still.

I’m still learning the system and I’m not familiar at all with the coding side, so I’m trying to stick with known-supported hardware.
Till reef-pi 2.0 atlas scientific ezo board was the only way to get ph readings. Post 2.0, @Roberto_b designed and graciously contributed a opensource ph board https://github.com/reef-pi/ph-board for reef-pi project. In US there are couple of folks who assemble it. @Michael Lane has it in his tindie store. 3.0 has built in support for this and I intend to update the adafruit guides with that board instead. EZO board support will still be there, but reef-pi ph board will be the official recommendation / present in standard documentations. Its lot cheaper and opensource, I think this is the right thing to do, as long as supply chain is present
 

hhaase

Active Member
View Badges
Joined
Nov 11, 2019
Messages
415
Reaction score
355
Rating - 0%
0   0   0
Ahhh, now I see it, and yes I think I'll restructure around that schematic in its place. Very clean schematic, easy to read. The good part is I'm already using an isolation circuit based around the same ADM3260 chip, so a lot of the work is already in place. And I don't think it will affect my footprint at all for the PCB.

My reasoning for two probe modules was one Ph and one ORP. So I'll just plan it out that way from the start.
 
Last edited:

Des Westcott

Well-Known Member
View Badges
Joined
May 29, 2018
Messages
645
Reaction score
1,034
Location
Durban - South Africa
Rating - 0%
0   0   0
Quick question about something I've noticed.

Our electricity supplier here in SA has just started implementing "load shedding" where there is supply problems and everyone has scheduled power outages. My area typically has 2 hours a day, occasionally 2 sessions of 2 hours. This will typically last a week and then they will be back on track. It doesn't worry me and my tank. I'm equipped to handle it.

BUT i noticed something interesting with the ATO usage graphs below. My main Tank ATO usage in the top centre has data saved from as far back as the 26th November. The AWC Out and AWC In at the bottom left and centre lose all their data every time the Pi loses power. Not sure if this is some sort of bug or not.


2019-12-06 (2).png
 

Blaxkin

Community Member
View Badges
Joined
Oct 4, 2019
Messages
27
Reaction score
46
Rating - 0%
0   0   0
Sorry for a temperature question again :rolleyes:

its not possible for me to write here if its 85 ignore or skip this value ?

func (c *Controller) Read(tc TC) (float64, error) {
log.Println("Reading temperature from device:", tc.Sensor)
if c.devMode {
log.Println("Temperature controller is running in dev mode, skipping sensor reading.")
if tc.Fahrenheit {
return telemetry.TwoDecimal(78.0 + (3 * rand.Float64())), nil
} else {
return telemetry.TwoDecimal(24.4 + (1.5 * rand.Float64())), nil
}
}
fi, err := os.Open(filepath.Join("/sys/bus/w1/devices", tc.Sensor, "w1_slave"))
if err != nil {
return -1, err
 

buddy.

Active Member
View Badges
Joined
May 16, 2018
Messages
291
Reaction score
617
Location
Australia
Rating - 0%
0   0   0
@Blaxkin I used to get those 85 deg errors a while back but i changed out a power supply which fixed it.
Since updating to 3.0 im getting these errors, which I never had in previous versons, but in negative 2035 and 2046... I wonder if they are related.

temp.JPG
 

Blaxkin

Community Member
View Badges
Joined
Oct 4, 2019
Messages
27
Reaction score
46
Rating - 0%
0   0   0
i found some infos in the sensor documentation.

"Under parasite power, if tRSTL > 960μs, a power-on reset can occur."

Maybe this is my Problem ? I use all Sensors in parasite.

@buddy. ok negative i don't have an the power supply i changed

i think a good way is to ignore the error values and to fix the graph maybe from 15 to 30 or what else
 
Last edited:
OP
OP
Ranjib

Ranjib

7500 Club Member
View Badges
Joined
Apr 16, 2016
Messages
9,825
Reaction score
17,041
Location
Pleasant Hill, Concord
Rating - 0%
0   0   0
@Blaxkin I used to get those 85 deg errors a while back but i changed out a power supply which fixed it.
Since updating to 3.0 im getting these errors, which I never had in previous versons, but in negative 2035 and 2046... I wonder if they are related.

temp.JPG
No clue :0-(
 
OP
OP
Ranjib

Ranjib

7500 Club Member
View Badges
Joined
Apr 16, 2016
Messages
9,825
Reaction score
17,041
Location
Pleasant Hill, Concord
Rating - 0%
0   0   0
i found some infos in the sensor documentation.

"Under parasite power, if tRSTL > 960μs, a power-on reset can occur."

Maybe this is my Problem ? I use all Sensors in parasite.

@buddy. ok negative i don't have an the power supply i changed

i think a good way is to ignore the error values and to fix the graph maybe from 15 to 30 or what else
Thats a good find. Let us know if changing that fixes your problem
 
OP
OP
Ranjib

Ranjib

7500 Club Member
View Badges
Joined
Apr 16, 2016
Messages
9,825
Reaction score
17,041
Location
Pleasant Hill, Concord
Rating - 0%
0   0   0
Sorry for a temperature question again :rolleyes:

its not possible for me to write here if its 85 ignore or skip this value ?

func (c *Controller) Read(tc TC) (float64, error) {
log.Println("Reading temperature from device:", tc.Sensor)
if c.devMode {
log.Println("Temperature controller is running in dev mode, skipping sensor reading.")
if tc.Fahrenheit {
return telemetry.TwoDecimal(78.0 + (3 * rand.Float64())), nil
} else {
return telemetry.TwoDecimal(24.4 + (1.5 * rand.Float64())), nil
}
}
fi, err := os.Open(filepath.Join("/sys/bus/w1/devices", tc.Sensor, "w1_slave"))
if err != nil {
return -1, err

yes, thats absolutely possible. But its very specific to your issue. It will not work if the controller used in a different range and it may mask real issues. A good way of fixing this will be to apply moving average instead, which will work irrespective of what temperature range the controller is operating on. This will take a bit more time and effort, but thats the right thing to do.
 
Back
Top