My reef-pi build freshwater style!

[Bigtrout]

Have my calibration packets, now just have to get some time to calibrate my ph probe.

I have noticed something weird though, the ph board is a few inches away from my mosfet pwm modules and I notice some erratic readings during light transistioning. Someone else mentioned this as well. The probe is away from any interference just reading its reference solution, but i get a night day rhythm in time with my lighting cycle.
The ph probe and my lighting cables are shielded. The boards are not. I think it may be worthwhile to mount my ph board in a metallic box to shield it. Ph is read in millivolts, so it is very easily interfered with. Maybe i can mount it in an altoids box

 

[Ranjib]

Just wait for the proper solution it's not worth it being off.

I dont think its that big of a deal . You can delete the existing sensor, create a new one in disabled mode, apply calibration and then enable it. You'll need the earlier one to get the calibration value anyway

 

[Ranjib]

Have my calibration packets, now just have to get some time to calibrate my ph probe.

I have noticed something weird though, the ph board is a few inches away from my mosfet pwm modules and I notice some erratic readings during light transistioning. Someone else mentioned this as well. The probe is away from any interference just reading its reference solution, but i get a night day rhythm in time with my lighting cycle.
The ph probe and my lighting cables are shielded. The boards are not. I think it may be worthwhile to mount my ph board in a metallic box to shield it. Ph is read in millivolts, so it is very easily interfered with. Maybe i can mount it in an altoids box

makes so much sense. I'll see if I can reproduce this

 

[Bigtrout]

Calibrating now! This is like playing junior chemist in high school...distilled water, calibration packets, measuring out 250ml of water to make the solutions...floating the solutions in the tank to stabilize temp. Then letting ph probe stabilize for 20 mins to get a nice steady reading. Carefully rinsing the probe before the next solution. My wife thinks im nuts.

Calibration packets are 4.01, 6.86 and 9.18.

I want to see if the slope is consistent over the whole range, so getting raw numbers on all 3 and then seeing what it looks like.

Results forthcoming.

 

[Bigtrout]

Exciting!

So results so far in raw counts before applying to reef pi.

ph 4.01 =20260
ph 6.86=195
ph 9.18= -16470

Looking at slopes the 6.86 number should have been 13. But using millivolts to counts across the span with some rough calculations, even though the 6.86 number is off by 182 counts, thats only off by 1.5 millivolts. Im impressed! Thats a margin of error of roughly .03 ph units!

So when calibrating I dont think it matters much if I use the 4.01 and 6.86 numbers, or the 4.01 and 9.18 numbers.

Fun stuff, time to apply a calibration and first test is my tap water, then into the aquarium!!!

 

[Ranjib]

I love how you break down these details, so informative. Thank you for detailing this.

 

[Bigtrout]

Woohoo...success! Ph probe calibrated!
Now reading 9.18 in the 9.18 solution.

My tap water is 6.70 out of the tap which correlates well with my test kit.

Now into the tank!!!

And the tank is a nice 6.6 ph, much better than trying to decipher yellowish green from a ph chart.

 

[Ranjib]

Woohoo...success! Ph probe calibrated!
Now reading 9.18 in the 9.18 solution.

My tap water is 6.70 out of the tap which correlates well with my test kit.

Now into the tank!!!

And the tank is a nice 6.6 ph, much better than trying to decipher yellowish green from a ph chart.

woot

 

[Bigtrout]

Will be interesting to see what ph swings i get when the light ramps up and the plants start using co2...with all the rain around here, my tap water is very soft, may have to think about some crushed shells in a filter or some baking soda to raise KH a bit and keep the ph steadier.
Normally the tap water is 6.9ph and 5degrees KH which is great for tetras and angels, but lately the ph has been 6.7 and the KH 1 degree.

I should see a cycle correlating with my lighting period where PH goes up when plants are using co2 and goes down when its dark and they are not.

 

[Bigtrout]

UPDATE...

Well ph probe worked all night with good readings. Then at 6am...holy moly, my lights come on and ph readings shoot up from 6.5 to 11. Something is interfering big time!

Quick troubleshooting points to my white channels. Cables of lights and ph probe are not run near each other. So interference has to be inside my case. The ph board is located near my mosfet pwm modules. I need to track this down and figure it out. When I calibrated all my numbers seemed to be very good and ph results consistent with test kits.

This interference is from either:
Ph board being too close to mosfet modules. When I calibrated all white channels were off, but blues were on at 80 percent. If I turn on blues and turn whites off reading goes back to an expected reading. Reading doesnt change no matter where blue lights are set, but white channels cause a large effect.
The white channels do have 3.2 amps running thru the modules vs .4 amps on the blue channels.

2. Wiring from white channels to either the pi or to lights running too close to ph board.

Fix would be either shield the wiring or the ph board, or relocate the ph board. Seeing how sensitive it is I will probably look at locating in a metal box that is grounded to the board power supply. Then recalibrate to be sure interference did not play any part in the first calibration.

If this fixes the issue, may be handy to mention this in build guides seeing how sensitive this board and probe is.

 

[Bigtrout]

Source of interference found. Its the power wiring on the mosfet pwm modules. Some quick tinfoil as shielding and temporarily relocating the ph board has sort of solved the problem for now. I used shielded cable for the pwm wiring to the lights but did not hook up the shields. I will probably still put the ph module in a shielded enclosure to stop any interference, and will redo the lighing cables using the shields.

 

[Michael Lane]

Source of interference found. Its the power wiring on the mosfet pwm modules. Some quick tinfoil as shielding and temporarily relocating the ph board has sort of solved the problem for now. I used shielded cable for the pwm wiring to the lights but did not hook up the shields. I will probably still put the ph module in a shielded enclosure to stop any interference, and will redo the lighing cables using the shields.

Nice troubleshooting! Tracking down interference has always been very frustrating to me. I'd love to see some pictures next time you open it up.

 

[Ranjib]

I’m paying attention to all of your findings, and I’ll make sure to include these in the official guide

 

[Bigtrout]

I’m paying attention to all of your findings, and I’ll make sure to include these in the official guide

Now that the holiday is over, I will probably get an altoids box for the ph board. Because the board is taking millivolt readings and converting to digital counts, and is isolated which is good for eliminating galvanic interference from the aquarium, that section of the board is very prone to induced voltages from other circuits. In my case the induction from the pwm of my lights which are several amps is interfering at certain duty cycles.
So move the mosfet modules or shield the ph board.
In my opinion shielding the board from any and all induced voltages seems like the best course of action for a truly accurate ph reading and an altoids box seems like an easy solution.

Will let you know my findings

 

[Ranjib]

Now that the holiday is over, I will probably get an altoids box for the ph board. Because the board is taking millivolt readings and converting to digital counts, and is isolated which is good for eliminating galvanic interference from the aquarium, that section of the board is very prone to induced voltages from other circuits. In my case the induction from the pwm of my lights which are several amps is interfering at certain duty cycles.
So move the mosfet modules or shield the ph board.
In my opinion shielding the board from any and all induced voltages seems like the best course of action for a truly accurate ph reading and an altoids box seems like an easy solution.

Will let you know my findings

Ack. I feel good about the fact that I built my ph controllers as independent unit

 

[Bigtrout]

Ok so today I had time to mess around with the ph controller which was giving false readings with my lights above 30 percent and good readings with them lower than 30 percent.

My first thought...interference from the mosfet modules inside my all on one box. So I did this:

I mounted the ph board inside an altoids box to shield any interference. Result was it didnt make a difference. I landed the shields on my lighting pwm cables. No difference. I grounded the altoids box to earth ground. No difference. I was getting frustrated about what was interfering with the ph probe.

So its something in my lights. I put my ph probe into 6.86 calibration solution as far away from the tank as the 2 meter cord would reach. Bad readings with lights on over 30, good reading with lights under 30.
So now i delved into the lights...and I find that when light 1 is on, im getting good readings, but when light 2 is on, whoa...bad readings. So its something about light 2 thats wrong. Each light has its own power supply like a laptop brick and 2 mosfet modules in my reef pi box to control the lights by pwm.

So I switch the cables going into reef pi, and the problem follows. Hmm is it the fet modules? So i rewire them to switch them between lights. The problem does not follow. So the moafet modules are ruled out. Now its down to power supplies. I switch them and the problem follows the power supply.
I disconnect the suspect supply and switch the other one between lights...no matter how I set the lights and whichever light I use, PH is good! I Do the same with the other supply and ph readings bad again. I have a bad power supply!
So this bad supply must be making some hellacious EMI. So bad that it was affecting the ph board somehow even though it is now shielded. Cant explain exactly whats wrong with it and how its causing the interference but its doing something bad.

A replacement is 18 dollars, but the supply is rated at 4 amps at 15 volts, and the light draws 3.9 amps with blues and whites all on at 100 percent. Seems like this supply is being pushed to MAX. Why buy another to chance the same thing happening. Why not power both lights with one power supply?

Just so happens good ole Meanwell makes a 15 volt 10 amp power supply. Both lights on at full power draw 7.8 amps, so 10 amps gives me a nice cushion so its not maxxing out the supply. I wanted to upgrade this anyways because the power bricks that came with the light start dropping voltage when the lights are full on.
So a meanwell LRS 150-15 is on the way. I may have to mount this into a box of its own so the lighting mosfet modules may be moved there so all the lighting is self contained with only signal wires coming over from my reef pi controller.

Bad power supplies can do some crazy things!

 

[Ranjib]

Ok so today I had time to mess around with the ph controller which was giving false readings with my lights above 30 percent and good readings with them lower than 30 percent.

My first thought...interference from the mosfet modules inside my all on one box. So I did this:

I mounted the ph board inside an altoids box to shield any interference. Result was it didnt make a difference. I landed the shields on my lighting pwm cables. No difference. I grounded the altoids box to earth ground. No difference. I was getting frustrated about what was interfering with the ph probe.

So its something in my lights. I put my ph probe into 6.86 calibration solution as far away from the tank as the 2 meter cord would reach. Bad readings with lights on over 30, good reading with lights under 30.
So now i delved into the lights...and I find that when light 1 is on, im getting good readings, but when light 2 is on, whoa...bad readings. So its something about light 2 thats wrong. Each light has its own power supply like a laptop brick and 2 mosfet modules in my reef pi box to control the lights by pwm.

So I switch the cables going into reef pi, and the problem follows. Hmm is it the fet modules? So i rewire them to switch them between lights. The problem does not follow. So the moafet modules are ruled out. Now its down to power supplies. I switch them and the problem follows the power supply.
I disconnect the suspect supply and switch the other one between lights...no matter how I set the lights and whichever light I use, PH is good! I Do the same with the other supply and ph readings bad again. I have a bad power supply!
So this bad supply must be making some hellacious EMI. So bad that it was affecting the ph board somehow even though it is now shielded. Cant explain exactly whats wrong with it and how its causing the interference but its doing something bad.

A replacement is 18 dollars, but the supply is rated at 4 amps at 15 volts, and the light draws 3.9 amps with blues and whites all on at 100 percent. Seems like this supply is being pushed to MAX. Why buy another to chance the same thing happening. Why not power both lights with one power supply?

Just so happens good ole Meanwell makes a 15 volt 10 amp power supply. Both lights on at full power draw 7.8 amps, so 10 amps gives me a nice cushion so its not maxxing out the supply. I wanted to upgrade this anyways because the power bricks that came with the light start dropping voltage when the lights are full on.
So a meanwell LRS 150-15 is on the way. I may have to mount this into a box of its own so the lighting mosfet modules may be moved there so all the lighting is self contained with only signal wires coming over from my reef pi controller.

Bad power supplies can do some crazy things!

I hear you and i dealt with my share of gremlins due to complication arising from power supply. So much so, I would say most of my multi-module build errors are due to power supply and the variance across them. I have found some decent UL listed, regulated, switching supply in amazon (chizon, 12V 3A) that im sticking with for the time being, while I my long term idea is to use power monitoring to detect and deal with this.
Somewhere in internet i read designing a good PSU is an art. There was a time when this was such a pain for me that i read at length around meanwell psu (multi output, 12V, 5V, 3V), usb C based apple power adapters, and the new GAN based super efficient, super small power adapters.
I am going to keep investing software, hardware effort on this and keep you all posted. Let me know if you have any thoughts.
Relevant to this issue, i think having a built in under voltage detection will be good thing to have, pi kernel already provide this information (vcgencmd get_throttled)

 

[Bigtrout]

Update:
To double check myself I repeated all my testing and this is reproducible, when power supply X is running either of my 2 lights, the result is crazy ph readings. When power supply Y runs either light, ph readings return to normal.

I dont know how power supply X is throwing the emi or whether its on the DC end or feeding harmonics back thru the ac line and affecting it, or both, but both supplies are only rated for the lights on at max. They are 15V 4Amp supplies and when both channels are on they drop from 15.1 volts down to 14 to 14.2 volts, telling me they are overloaded at 4amps and strange things happen when you overload a power supply.

Time for a Meanwell sized to power both lights!!

 

[Bigtrout]

Amazing!!! My new power supply came in. I decided that I didnt want to rewire everything so got the same configuration as the beamswork lights only I got a 15volt 8 amp which is double the rsting of the lights.

Bought only one as a test to see gow clean the power is. Immediately cleaned up my ph graph. Woohoo!!! And some tests with the multimeter show it as 15.3 volts not under load and 15.25 under full load from my light!

Look at the sharp drop to 7.4 when i unplugged the suspect supply. I tested at a few differing intensities and still 7.4...wow im glad I chased this down. Now to order a second one for the other light. Its not causing interference but it gets too warm for my tastes. The 8 amp one feels room temperature.

Next up...recalibration to make sure the bad supply didnt skew the calibration.

 

[Bigtrout]

Didnt do a full recalibration, as 10 minute dips in 4.01, 6.86 and 9.18 cal solutions gave readings of 4.05, 6.89 and 9.20, showing my first calibration was good because it was late night and the offending power supply was only powering a few moonlights at the time(it got noisy when white circuit was above 20 percent)

Finally...a working...interferemce free PH probe!