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

[pickupman66]

Hey I assume you back up and things are back to a working condition, congrats.

Yes all turned back over to the pi for control. I haven't looked at logs but seems good this far.

If the Archon had the speed of the reef pi it would have been way better.

 

[b4tn]

Reboots can be Linux kernel panics, or power issues.

Do you port forward just 443 or put the device in a DMZ (is port 22 also exposed to the internet?)

I assume you changed the default raspberry pi password for the pi account?

That made me laugh, I even add nologin to the pi account.

Lol. I do Infosec also. More plans and policy these days so I’m rusty on the technical aspect. But we had a surprise pen test years ago. The network team left the default password on a riverbed appliance they where doing a demo on. It was pretty ugly lol

 

[daveCB]

Does anyone know what this HTML error refers to? I get it several times per day. Seems harmless but just would like to know.

2/12/2019, 9:32:59 PM UI ERROR {"error":"strconv.Atoi: parsing \"150\\n\": invalid syntax"} | HTTP 404

 

[rushbattle]

Wow , just noticed we have 500 pages worth of controller posts
Keep it up folks . And thank you for making thing community awesome

I was going to comment on that too! Thanks Ranjib et al for all of the hard work!

 

[bigcheese]

Does anyone know what this HTML error refers to? I get it several times per day. Seems harmless but just would like to know.

2/12/2019, 9:32:59 PM UI ERROR {"error":"strconv.Atoi: parsing \"150\\n\": invalid syntax"} | HTTP 404

Something is trying to cast a string ("150") to an integer, and it seems to be failing because of the trailing newline character (\n). Just a guess it's probably on the UI side rather than the underlying logic... check around any user input boxes on the page when you see the error and look for any stray whitespace after a 150.

 

[geekengineer]

10,000 posts too!

That means I never be able to catch up.

 

[Ranjib]

Does anyone know what this HTML error refers to? I get it several times per day. Seems harmless but just would like to know.

2/12/2019, 9:32:59 PM UI ERROR {"error":"strconv.Atoi: parsing \"150\\n\": invalid syntax"} | HTTP 404

Where you have set 150? PWM frequency or something else? You have a new line character in that field. Delete it an type 150 and hit save/update.

 

[dmolavi]

Hardware question ... the last connector I need to get for my case is a BNC connector for the pH probe. Why do the probes use BNC? Is it for positive-action on the connection itself (IOW - a 'twist lock' so that it doesn't work it's way loose), or do they actually require the shielding provided by the cable? I ask because I'd rather not fuss with crimping BNC connectors if I don't have to (I have the crimper already from other projects, but...blah).

I'm hoping to be able to get a "thin" BNC cable (like this one: https://www.amazon.com/CablesOnline-1ft-RG316-Cable-RF-BC101/dp/B078SNS7C6/) and just 'nip the tip' and solder to the panel mount connector. Anyone see this as a problem, or should I try to find a panel-mount (F) with the cable going to a male end, if such a beast exists?

 

[dmolavi]

Hang on...found these:
https://www.arrow.com/en/products/031-4803-rfx/amphenol-rf
https://www.arrow.com/en/products/73412-0032/molex

Which is the correct one? The only difference that I can see is the impedance. @theatrus , any idea? or does it not matter for this application?

 

[Bigtrout]

Hang on...found these:
https://www.arrow.com/en/products/031-4803-rfx/amphenol-r

Which is the correct one? The only difference that I can see is the impedance. @theatrus , any idea? or does it not matter for this application?

Did some digging, on the other reef site I found a thread where there was no consensus and they didnt think it mattered. The neptune rep was quoted as saying:
"The impedance of the cable doesn't matter. Don't extend the cable too much and use proper BNC connectors as the signal is very high impedance and extremely small."

Someone else mentioned bnc connectors from radio shack worked fine.

Have not been able to find specs on what type ph probes use.

However, the atlas scientific extension cables for their ph probes are 75 ohm cables.

I think it wont matter much for our application, impedance matching is more important for a high speed signal transmission. I think the shielding is the important thing because we are reading very small voltages which is prone to interference from other equipment or cables in close proximity.

 

[dmolavi]

Did some digging myselfm...the connector on the pico board is a 50 ohm Amphenol

 

[Bigtrout]

Did some digging myselfm...the connector on the pico board is a 50 ohm Amphenol

Read above i edited my post. But we are not transmitting a highly variable signal so impedance probably doesnt matter, its more having the cable shielded that does matter.

 

[Ranjib]

Hardware question ... the last connector I need to get for my case is a BNC connector for the pH probe. Why do the probes use BNC? Is it for positive-action on the connection itself (IOW - a 'twist lock' so that it doesn't work it's way loose), or do they actually require the shielding provided by the cable? I ask because I'd rather not fuss with crimping BNC connectors if I don't have to (I have the crimper already from other projects, but...blah).

I'm hoping to be able to get a "thin" BNC cable (like this one: https://www.amazon.com/CablesOnline-1ft-RG316-Cable-RF-BC101/dp/B078SNS7C6/) and just 'nip the tip' and solder to the panel mount connector. Anyone see this as a problem, or should I try to find a panel-mount (F) with the cable going to a male end, if such a beast exists?

I don’t know the exact reasoning, but all ph probes I’m aware of (atlas scientific, American marine, milawukee) uses bnc, hence we went with it

 

[daveCB]

Where you have set 150? PWM frequency or something else? You have a new line character in that field. Delete it an type 150 and hit save/update.

Nothing that I can locate on active modules is set to 150. Is it possible that its picking this up for a hidden module that I am not using.

 

[bigcheese]

Read above i edited my post. But we are not transmitting a highly variable signal so impedance probably doesnt matter, its more having the cable shielded that does matter.

Typically, 75-ohm and 50-ohm BNC connectors are made to mate with different sized cable. The center pins and outer shield crimp rings are slightly different sizes. You can get dissimilar ones to mate in most circumstances, but there will be signal distortion.

Since we're individually calibrating the sensors on each build, it *shouldn't* make a difference to you.. what won't work is if Reef-pi gets a hard-coded reference which expects an absolute mapping (say, Milwaukee pH sensor is 2V at pH 7.0), that 75/50-ohm mismatch may cause yours to be off (like 2.2V, and read as pH 7.4).
Stick with all 75-ohm if possible, AFAIK, 50-ohm BNC is really only used nowadays for older CCTV systems.

I was looking at the Vernier sensors, and if they weren't so overpriced, I'd be an advocate. They use RJ-45 with twisted-pair Cat5, which is just a way easier form factor to deal with (and I'm betting most of us here have lots of experience terminating those).

 

[theatrus]

Read above i edited my post. But we are not transmitting a highly variable signal so impedance probably doesnt matter, its more having the cable shielded that does matter.

The RF impedance doesn’t matter, so feel free to use 50 or 75 ohm here. The shield is important, as well as leakage. You could just use some bare wire over smaller distances to be honest.

The source DC impedance of pH probes is very very high, so avoid big globs of flux and other debris which can cause offset voltages to leak onto the center conductor. In the base board design the shield is actively driven.

 

[dmolavi]

Typically, 75-ohm and 50-ohm BNC connectors are made to mate with different sized cable. The center pins and outer shield crimp rings are slightly different sizes. You can get dissimilar ones to mate in most circumstances, but there will be signal distortion.

Since we're individually calibrating the sensors on each build, it *shouldn't* make a difference to you.. what won't work is if Reef-pi gets a hard-coded reference which expects an absolute mapping (say, Milwaukee pH sensor is 2V at pH 7.0), that 75/50-ohm mismatch may cause yours to be off (like 2.2V, and read as pH 7.4).
Stick with all 75-ohm if possible, AFAIK, 50-ohm BNC is really only used nowadays for older CCTV systems.

I was looking at the Vernier sensors, and if they weren't so overpriced, I'd be an advocate. They use RJ-45 with twisted-pair Cat5, which is just a way easier form factor to deal with (and I'm betting most of us here have lots of experience terminating those).

The Pico board uses a 50ohm connector, I've found a panel-mount Amphenol 50ohm connector and cable, as well. Those will be my components.

 

[dmolavi]

https://www.arrow.com/en/products/031-220n-rfx/amphenol-rf and https://www.arrow.com/en/products/115101-01-m0.25/amphenol-rf are what I'm going to use.

 

[geekengineer]

Read above i edited my post. But we are not transmitting a highly variable signal so impedance probably doesnt matter, its more having the cable shielded that does matter.

I would agree with Bigtrout's advice also.

 

[geekengineer]

I going to have a hundred questions to ask and not sure what place is best to ask them, but for starters.. (coming from limited knowledge about this whole project)

1) Is there a particular reason why nobody seems to be using the PI-zero with the reef-pi pico board? Maybe I just missed that there are people using it. All I saw so far is people using ras-pi 3 B+.

2) I have my own ras-pi hat design based off the Tentacle T3 design you can get at Atlas. I Just have some additional customization I added to it like 1-wire headers and a power strip relay header. Will Atlas modules still be supported in the future?

3) I have a 4 pump doser design that works off the I2C bus via a micro Arduino clone board. I made it so works similar to the way the Atlas pumps work. It's a bit different in that it doesn't have built in calibration. You just tell it how many mS for a particular pump to run. How hard would this be to interface something like this to the reef-pi API? Also would calibration be possible?

4) I like to be able to add an atlas dissolved oxygen sensor because of some issues I had last time with my tank. How hard would that be to add to the reef-pi API?