Maelstrom Reef Pi

[AbjectMaelstroM]

EDIT: I was incorrect with the color codes.

Pins are: PWM-GND-ON/OFF-12V.

Wiried up as such:

Also, I was able to get the light to turn on, however, then this happened.

ULN went up in smoke, guessing a short. Or I'm missing something. I have another hat, but I think I'm going to wait for someone else to chime in as I really don't want to burn another one. Honestly, other than a shoddy connection (I was using the breakout female/male cables) I don't know why it would burn out.

More science needed.

 

[Michael Lane]

I'm interested in figuring out the right way to get this going too. I suspect it is similar to the Mars Aqua on/off functionality. It's a 2 channel fixture, right?

 

[AbjectMaelstroM]

I'm interested in figuring out the right way to get this going too. I suspect it is similar to the Mars Aqua on/off functionality. It's a 2 channel fixture, right?

Yes, see above update lol. I'm down to the busted HAT with USB only 5v. Ordering another hat.

 

[Michael Lane]

That looks terrible!

It looks like it may have been a short directly to GND. It looks like there's a bit of charring on the 12v trace. Can you describe how everything was connected when this happened?

 

[AbjectMaelstroM]

That looks terrible!

It looks like it may have been a short directly to GND. It looks like there's a bit of charring on the 12v trace. Can you describe how everything was connected when this happened?

So it may have been a combination of factors exacerbated by by incompetence/lack of attention to detail. First, I'll show you how I had it wired in my mind, then I'll follow up with how it was actually hooked up:

This was the connection between the pi and the viparspectra during the incident. I created an Inlet in Reef Pi and assigned Pin 22. Went into equipment, created the fixture...life was good. It toggled on and off. Pwm didn't work so I thought that there was a bad connection. As I picked up the connection point (forgot to power off the light) I smelled smoke and saw a bright orange spot growing on the HAT.

Now, I think the above should have worked. However, I didn't have it plugged into light 3 and this is where my lack of attention to detail bit me in my "third point of contact". It was actually plugged into PH2 (light 1,2,3 were already populated and I completely zoned out on it).... Which would make what I thought was PWM pin 7 actually be 5v pin. :mad:

Now that I've realized all that, I really hope I didn't smoke the PH board along with it. Yet to be seen.


Hopefully that makes sense.

 

[AbjectMaelstroM]

Something else, not sure what the implications are, but the brown wire is putting out 8.8-8.9v (which is supposed to be the on/off) when the fixture is off.

I had a buck converter hooked up to the 12v on the hat, no power supply was plugged in. V+ is connected to what I believe to be 12v on the vipar driver. IO22 PIN connected to what I believe is on/off on the vipar driver. What I noticed was that when vipar is plugged in, the buck converter is being powered (power led on).... Again buck converter power adapter is NOT plugged in, so it's being reverse powered by the viparspectra.

At this point I'm in the weeds. The only thing I can think of is that the wire coding is not what I think it is.

 

[Ranjib]

Something else, not sure what the implications are, but the brown wire is putting out 8.8-8.9v (which is supposed to be the on/off) when the fixture is off.

I had a buck converter hooked up to the 12v on the hat, no power supply was plugged in. V+ is connected to what I believe to be 12v on the vipar driver. IO22 PIN connected to what I believe is on/off on the vipar driver. What I noticed was that when vipar is plugged in, the buck converter is being powered (power led on).... Again buck converter power adapter is NOT plugged in, so it's being reverse powered by the viparspectra.

At this point I'm in the weeds. The only thing I can think of is that the wire coding is not what I think it is.

You can not verify the wires by directly powering the light and using a multimeter to,probe the signals ?

 

[Michael Lane]

Yep. That 8v behavior sounds familiar. The regulator isn't stable until there is sufficient current, so you'll see ~8v when the lights are off. It would jump to 12v once the lights are on.

I'm still surprised the ULN burned up, so I'll keep thinking on that. In the meantime, I've included a ULN2803 in your package. It's pretty easy to solder if you want to try repairing your board.

I'll follow up again when I get a few more minutes.

 

[AbjectMaelstroM]

You can not verify the wires by directly powering the light and using a multimeter to,probe the signals ?

Doing it now.

OFF
Grey - 0.000v
Blue - 0.021v
Brown - 8.22v
White - 0.021

ON - 1%
Grey - 0.000v
Blue - 0.067v
Brown - 11.72v
White - 3.030 v

ON - 5%
Grey - 0.000v
Blue - 0.187v
Brown - 11.72v
White - 3.030

ON - 15%
Grey - 0.000v
Blue - 0.353v
Brown - 11.72v
White - 3.029

So: Grey GND, Blue PWM, Brown 12v, White ON/OFF? Or am I wrong?

 

[Michael Lane]

Sounds correct. Can you also measure the frequency on the blue PWM wire?

 

[AbjectMaelstroM]

Sounds correct. Can you also measure the frequency on the blue PWM wire?

No idea how to do that. Just set the MM to Hz (only allows AC) and then black to ground red to blue PWM wire? If so showing me 0.00.

 

[Michael Lane]

No problem regarding the frequency measurement. Some meters can measure DC frequency, but it's not too important for this. I was mostly curious what the default PWM frequency is.

It sounds like the wiring went a bit backwards and the signal from the ULN was shorted to ground over the yellow wire. I don't think the ULN2803 will be very helpful in directly turning the light on and offer in this configuration.

I'll bet the on/off control is based on an open or short between +12v and on/off. You can probably simulate this by using a bit of wire to bridge those 2 positions.

How far down does the light actually dim? Does it go down smoothly to 0%?

It might be a bit overkill, but I think something like this would work. It's based on the circuit for the Mars Aqua adapter. The on/off part is controlled by Q3 and Q1 based on a threshold voltage. It might need a bit of tweaking depending on the PWM frequency and actual dimming range.

Another approach might be to keep the first side of the circuit the same, and just bypass straight to the PWM pin. That might help with on/off stability.

 

[AbjectMaelstroM]

No problem regarding the frequency measurement. Some meters can measure DC frequency, but it's not too important for this. I was mostly curious what the default PWM frequency is.

It sounds like the wiring went a bit backwards and the signal from the ULN was shorted to ground over the yellow wire. I don't think the ULN2803 will be very helpful in directly turning the light on and offer in this configuration.

I'll bet the on/off control is based on an open or short between +12v and on/off. You can probably simulate this by using a bit of wire to bridge those 2 positions.

How far down does the light actually dim? Does it go down smoothly to 0%?

It might be a bit overkill, but I think something like this would work. It's based on the circuit for the Mars Aqua adapter. The on/off part is controlled by Q3 and Q1 based on a threshold voltage. It might need a bit of tweaking depending on the PWM frequency and actual dimming range.

Another approach might be to keep the first side of the circuit the same, and just bypass straight to the PWM pin. That might help with on/off stability.

Yeah, I'll play around with it tomorrow.

 

[devlzluv]

Maelstrom, replace the uln and keep the fresh board ML sends you once you figure it all out. That way the burned out one will still be your test. In about 2 weeks im going to start doing mine, if you haven't figured it out by then, I will try and help and maybe all of us can figure it out together. Fingers crossed though that no more mishaps happen for you, hate when things are frustrating.

Also found this for you if you want to take a look at it.

My Reef pi build - Well, my Viparspectra PWM light controller :D

Where is your uln2803? That I think is your problem. You are trying to run the relays directly off of the raspberry pi outputs which draws too much current thru the pi output pins causing the undervoltage error. The uln2803 is part of the power controller for a few good reasons. It isolates...

www.reef2reef.com

 

[AbjectMaelstroM]

Maelstrom, replace the uln and keep the fresh board ML sends you once you figure it all out. That way the burned out one will still be your test. In about 2 weeks im going to start doing mine, if you haven't figured it out by then, I will try and help and maybe all of us can figure it out together. Fingers crossed though that no more mishaps happen for you, hate when things are frustrating.

Also found this for you if you want to take a look at it.

My Reef pi build - Well, my Viparspectra PWM light controller :D

Where is your uln2803? That I think is your problem. You are trying to run the relays directly off of the raspberry pi outputs which draws too much current thru the pi output pins causing the undervoltage error. The uln2803 is part of the power controller for a few good reasons. It isolates...

www.reef2reef.com

Yeah I just got reading that thread. Good find. Haha, as for the pi, I have another one of Mike's HATs but the 5v backpower circuit was damaged in a drilling accident so it can only be powered by 5v usb on the pi itself. Works fine for this particular situation. I'll try repair the other with the spare ULN.

After looking through that thread and seeing how the relay is implemented (I could be wrong), but it seems as Mike thought and it is activated by bridging 12v and on/off.

I'm at work for 5 more hours, but when I get back I'll give it a run. I have a spare 8x channel 12v relay on hand somewhere. So the plan would be to leave PWM as is, but run the on/off and 12v to the relay and activate it via "outlet" using ULN IO pin.

 

[Michael Lane]

This thread just pushed me over the edge to grab one of these lights since they claim to be able to dim down to 0%. If it works out, I might consider swapping out all of my Mars Aqua lights.

 

[AbjectMaelstroM]

This thread just pushed me over the edge to grab one of these lights since they claim to be able to dim down to 0%. If it works out, I might consider swapping out all of my Mars Aqua lights.

Yeah that was one of the biggest draws to this over Mars Aqua. I did notice that their 1% is pretty dang bright so I wouldn't be surprised if the 1% is actually 10% or more. But, I guess I'll see tonight.

I do like the color of these, but I also have a bunch of leds coming, mainly more RB, 395 and 420s. Their 12k looks more like 6500 and 6500 looks more like 4-5000k.

 

[Michael Lane]

OK. Keep us updated on the actual dimming ability. I cancelled my order until then.

 

[ludnix]

Great build thread! I hope you have good luck with the next hat.

I just recently got the vipars sepctras as well and I agree that their 1% seems a lot closer to 10%. It is very reminiscent of the pwm ELN meanwell drivers I have used in the past for dimming LEDs. They don't seem allow much control on the lower end of that dimming range despite there being 100 "steps" in the dimming.

 

[AbjectMaelstroM]

OK. Keep us updated on the actual dimming ability. I cancelled my order until then.

Will do, I'm hoping im wrong but I'm ready (with more leds to change out on the wite channel) in case I'm not.

BTW, did you pull lenses on your MA lights? I'm seeing a bit of hotspotting with these. How high do you have yours mounted?

Great build thread! I hope you have good luck with the next hat.

I just recently got the vipars sepctras as well and I agree that their 1% seems a lot closer to 10%. It is very reminiscent of the pwm ELN meanwell drivers I have used in the past for dimming LEDs. They don't seem allow much control on the lower end of that dimming range despite there being 100 "steps" in the dimming.

Thanks!