AI Hydra Fan Connector Repair - Significant Corrosion

[_AV]

My previous electronic equipment repair projects

Radion XR15:

Radion Lights Board Level Repair

I enjoy an occasional board level repair. It gives me a new angle to this hobby, which is often very heavy on DIY. In this thread I'm going to post some of my more interesting and challenging repairs, picture heavy, including high magnification microscope imagery. Recently, I got ahold of a...

www.reef2reef.com

Radion XR30:

Radion XR30 complete no light repair

My previous electronic equipment repairs Radion XR15: https://www.reef2reef.com/threads/radion-lights-board-level-repair.774331/#post-8197536 A fellow reefer mailed me their Ecotech Radion for repair service as it's long been out of warranty. The unit turns on briefly before shutting down and...

www.reef2reef.com

Imagine a problem with a connector breaking off the board. Usually not a big deal. Worst case - a couple of torn traces and a new connector. Unless, of course, we are talking about the reef gear. In which case, we have to expect things breaking in new and unusual ways!

This is certainly the case with this AI Hydra light. It was also a mail in service request from a fellow reefer. So I am not exactly sure how this connector came off the board, but... well, take a look at this:


What??? The entire fan circuit is missing. Every. Single. Component.
The area is covered in a thick layer of corrosion to the point you can't tell what is what:



Let's clean up the corrosion and see if there is anything left below it:



Now it is clear that at least some traces and pads corroded away entirely. Both pads from the connector are no longer connected to the board circuit. Looks like there should be a diode, a resistor and some other component.

This is what the fan control circuit should look like:


The third missing component turns out to be a small mosfet. This makes sense, something needs to help the controller switch the fan on and off after all.

Let's get started on restoring the board traces. This time, instead of using the jumpers, I decided to use copper to lay out the new traces.


As you can see, it's not all that bad. Just a couple of runs from the connector pads to link them to the diode and the mosfet, while carefully avoiding unrelated traces in close proximity.
Anyway, we are ready to layout the missing components:


At this point, the fan should work once plugged in. But all of this exposed copper will corrode very quickly again. And then the corrosion will push the new components off the board again.
To solve this, I'll put a good layer of conformal coating over the entire area:


This coating needs to be cured to harden it:


Once hardened, the conformal coating not only protects against corrosion, but also helps to hold my traces to the board.


Now it's time to test if the new fan circuit actually works:


My leading theory on why the damage to the board was so significant is a sequence of events with the old fan failing and shorting first. The short caused a part of the circuit to burn out. You can actually see a small hole in the board right next to the diode where a part of the trace used to be. The short on the board made the black coating to flake off exposing copper on several traces. Copper usually corrodes violently, which might have been enough to push all the circuit components off the board.
Either way, the light is now working and the board circuit is corrosion resistant.

 

[Ron Reefman]

nice work

 

[VO3434]

Very nice and thank you for the repair!! Working great!!

 

[A. grandis]

Very nice and thank you for the repair!! Working great!!

Your dog (avatar) smiles. Love it!!!

 

[Sherief Singer]

@_AV, I have the same problem with all 3 of my 26HD lights over a 6 month period.... all seems to be in the fan circuit area. if you don't mind me asking, where did you order the components from? also, do you have any idea what the capacitor and resistor values or SMD codes are? Are the caps just for filtering/necessary?

Thanks again!

 

[_AV]

I order most of my components from digikey/mouser/newark to ensure they are original and of high quality. It's getting much harder now with major shortages everywhere.

I don't know hydra 26 specific circuitry details without looking closely at it. There is no capacitor in the fan circuit. The resistor is marked 1003, which suggests that it's 100 kOhm.

 

[_AV]

Finally got around to record a fan circuit repair video for a 26 HD. Hope this will be helpful. Please like and subscribe to keep me motivated

 

[hhaase]

If I had to venture a guess, something (probably Q1) had been reworked at the factory with aqueous flux and they didn't properly clean it. You can see the corrosion migrating along the traces. The corrosion is consistent with an aggressive flux as well. Once the corrosion starts, it's hard to stop.

 

[_AV]

There's no doubt there are assembly issues across all brands.
Then there is water damage and... also, let's call it, pet damage as well. Both contribute significantly to the number of corroded units I receive for repair.

Here's an example of one of the worst corrosion on a Radion light I've seen:

This corrosion was a surprise the the owner.
It was a no repair:

 

[Biologic]

Finally got around to record a fan circuit repair video for a 26 HD. Hope this will be helpful. Please like and subscribe to keep me motivated

I just tried to attempt to clean with 95% alcohol the fan circuitry of a used AI Hydra 26. Mine was in much better condition. I used a very fine extra soft tooth brush sweeping the pads and semi-conductors. I did not get a response from the fan. I tested the fan with a 12 volt power supply. Spun up with no issues. I still have some corrosion. Do you think I could reflow what solder I have here by touching a soldering iron to the pads to reset the various places with corrosion?

The first pic is the before, the second pic with the white molex connector on the board facing the 12 o'clock position is the after.

 

[_AV]

I just tried to attempt to clean with 95% alcohol the fan circuitry of a used AI Hydra 26. Mine was in much better condition. I used a very fine extra soft tooth brush sweeping the pads and semi-conductors. I did not get a response from the fan. I tested the fan with a 12 volt power supply. Spun up with no issues. I still have some corrosion. Do you think I could reflow what solder I have here by touching a soldering iron to the pads to reset the various places with corrosion?

The first pic is the before, the second pic with the white molex connector on the board facing the 12 o'clock position is the after.

The original picture looks pretty bad. While you certainly can reflow the joints and it should help keep the corrosion away, since you mentioned that the fan is still not working, I'm willing to bet that the original corrosion on the connector damaged the actual fan circuitry (any combination of U1, Q1 and D1). You can test it with a voltmeter to be sure, but in my experience when the connector is this badly corroded, it likely causes a dead short and the circuit is not really designed to handle that.

 

[Biologic]

The original picture looks pretty bad. While you certainly can reflow the joints and it should help keep the corrosion away, since you mentioned that the fan is still not working, I'm willing to bet that the original corrosion on the connector damaged the actual fan circuitry (any combination of U1, Q1 and D1). You can test it with a voltmeter to be sure, but in my experience when the connector is this badly corroded, it likely causes a dead short and the circuit is not really designed to handle that.

I do have a multi meter, though I am not great at using it. I have a very basic understanding. With the unit off and unplugged, I set it to 1 ohm, I tested the resistance across the terminals of the connectors. I was able to register that current was moving through the terminals. When I touched it to another trace around the fan circuitry, I found current was moving through.

I would have to individually test U1, Q1 and D1 as well?

Also looking at the terminals where the fan connector is, it still looks like I have some cleaning to do. I used a wooden tooth pick as to limit my possible damage I could do to the board. The point is not as sharp and its softer than the metal dental picks that I could use.

 

[_AV]

I do have a multi meter, though I am not great at using it. I have a very basic understanding. With the unit off and unplugged, I set it to 1 ohm, I tested the resistance across the terminals of the connectors. I was able to register that current was moving through the terminals. When I touched it to another trace around the fan circuitry, I found current was moving through.

I would have to individually test U1, Q1 and D1 as well?

Also looking at the terminals where the fan connector is, it still looks like I have some cleaning to do. I used a wooden tooth pick as to limit my possible damage I could do to the board. The point is not as sharp and its softer than the metal dental picks that I could use.

Yeah, if I were repairing this, I'd be removing the molex and likely replacing it after cleaning and tinting the pads. On the last picture, the left pin is severely discolored, which suggests a thick oxidation layer, which likely does not conduct reliably anymore.
With a multimeter, and the molex removed, I'd be testing voltages instead of resistance at the fan pins, inputs and outputs of Q1 and U1. I believe the hydra fan is 5v (don't test it with a 12v power supply). So I would be looking for 5v on light start up and then again when it warms up sufficiently at high intensity.

 

[Biologic]

Yeah, if I were repairing this, I'd be removing the molex and likely replacing it after cleaning and tinting the pads. On the last picture, the left pin is severely discolored, which suggests a thick oxidation layer, which likely does not conduct reliably anymore.
With a multimeter, and the molex removed, I'd be testing voltages instead of resistance at the fan pins, inputs and outputs of Q1 and U1. I believe the hydra fan is 5v (don't test it with a 12v power supply). So I would be looking for 5v on light start up and then again when it warms up sufficiently at high intensity.

so what I should do is test with the board on So I can see if there’s power being delivered to the pins themselves?g

 

[_AV]

so what I should do is test with the board on So I can see if there’s power being delivered to the pins themselves?g

exactly

 

[Biologic]

exactly

oh I should add, it runs lights up perfectly fine, just no fan. I could add a 24 volt muffin fan off my apex if I cannot get this resolved.

should I stop while I am ahead? I could potentially cause more harm than good.

thank you for your inspiring YouTube channel. I was an early subscriber.

 

[_AV]

oh I should add, it runs lights up perfectly fine, just no fan. I could add a 24 volt muffin fan off my apex if I cannot get this resolved.

should I stop while I am ahead? I could potentially cause more harm than good.

thank you for your inspiring YouTube channel. I was an early subscriber.

Thank you for being a subscriber. Much appreciated.

I think it all depends on your comfort level and the skills. I don't think there is a whole lot of risk in measuring a life fan circuit. But I do understand that it still requires precision probes and a steady hand.

Adding a separate fan is certainly an option.
You can always mail it to me for a repair as well

 

[Biologic]

Thank you for being a subscriber. Much appreciated.

I think it all depends on your comfort level and the skills. I don't think there is a whole lot of risk in measuring a life fan circuit. But I do understand that it still requires precision probes and a steady hand.

Adding a separate fan is certainly an option.
You can always mail it to me for a repair as well

I am going to give it the old college try. If that doesn't work, it's coming to you. I really need a microscope. haha!