New SWG not producing the specified amount of chlorine

[JoyfulNoise]

I say keep an eye on it and continue to make measurements of the overnight gains when you have the chance to do the tests. Perhaps there’s more here to consider with higher water temps possible. Maybe if you can substantially change the salinity you can redo the overnight gain tests and see if there’s a sensitivity to that.

 

[mgtfp]

I say keep an eye on it and continue to make measurements of the overnight gains when you have the chance to do the tests. Perhaps there’s more here to consider with higher water temps possible. Maybe if you can substantially change the salinity you can redo the overnight gain tests and see if there’s a sensitivity to that.

Will do that. At the moment that water temps are still going up and down. Will have another go once the water is reliably above 20°C. I already did do a salinity comparison before and after adding 2 bags of salt. That was before my Taylor supplies arrived from the US and I thought my salt might be on the low side. Same currents and pretty much the same outcomes in the overnight gain tests.

 

[mgtfp]

OK, there is an update. Fluidra service tech came out yesterday. No spiel about phosphates or water parameters in general, didn't even ask for them. Just poked in a salt meter, seemed happy with the result. Ran a few checks on the controller, like making sure that it's not just one of the cell polarities during the reversals that's not working, but all good.

Then he suggested to replace the cell. After taking out the old cell, he mumbled "never seen that before, that goes straight to quality". Turned out that the "loose" plates in-between the ones that are connected to plus/minus were not inserted far enough. I had noticed that also last week, but not really taken in, kind of thinking to myself that it looks odd, but must be a feature, just looked to deliberate to not be intended. . No, double . The loose plates are kind of on a frame that clicks in, and it wasn't pushed in far enough - clicked into a position for a different plate length. Something gone wrong in final assembly.

Anyway, it all started to make sense. Due to the insufficient plate insertion, the outer plates were basically facing directly the center plate, so that in this region (apart from a very inhomogeneous electric field) the potential difference between "neighbouring" plates was five times higher then intended (even though the distance was of course five times higher, but I don't think that the distance is of principal importance in electrolysis, more of an optimisation parameter). And I suspect that this significantly higher voltage is sufficient to overcome the higher overvoltage for O₂ production by electrolysis of water (2H₂O --> O₂ + 4H⁺ + 4e⁻), which is contributing to the total current. (EDIT: This is actually incorrect. Correct explanation much simpler, see post #40. I updated below picture)

In the meantime, I had also measured the voltage supplied to the old cell, not just the current, which turned out to be nonlinear with increasing output-percentage, which didn't look right.

After the cell replacement, I measured current and voltage again. Current was exactly the same, which makes sense, as that's what the controller uses to adjust the Cl₂-output. But the voltage with the new cell was now perfectly linear with increasing output-percentage and about 2V higher compared to the old cell. That means that with the old cell, the controller could create the same current with less voltage, which I take as an indication that there was an additional electrolysis process (presumably oxygen electrolysed from water) creating current. (EDIT: correct explanation, see post #40)

Here an overview with my measurements and my interpretation, hope that makes sense:


Last night, I ran a chlorine gain test with the new cell, and it seems to add up now, but will run another test or two next week or so.

Anyway, happy with my new cell (so far at least - knocking on wood and running a few more chlorine gain tests). And also happy with the Fluidra customer service and the service tech, no complaints there.

 

[kimkats]

What a great read and learning piece this was! Thanks for sharing all of this with TFP!

 

[mgtfp]

What a great read and learning piece this was! Thanks for sharing all of this with TFP!

Was also a big learning curve for me. Sometimes you need things to go wrong to take the time to get back to paper and pencil.

And so much info here. A special thanks to @JoyfulNoise and @JamesW, who's many posts helped so much in understanding what seemed to be going on with my SWG.

By the way: I started this thread in the members only area because I was a bit nervous about my warranty claim. But now that it's through, feel free move the thread to somewhere more suitable, probably more of a Deep End topic.

 

[kimkats]

I am thinking the SWG area so it is more out in the open. The deep end is scary for some people.............ME LOL

 

[JoyfulNoise]

That means that with the old cell, the controller could create the same current with less voltage, which I take as an indication that there was an additional electrolysis process (presumably oxygen electrolysed from water) creating current.

The same current with less voltage is indicative of lower cell resistance (V = I • R). Given that the plates were not set correctly in their “cage”, that means that current was probably “leaking” around the perimeter and finding other pathways to ground. Bipolar electrolysis cells are very sensitive to “edge effects” and so that is why it is important for the plates to be aligned properly and for the pathways to ground to be well shielded. This forces the current to flow between the plates rather than around them. So I don’t necessarily think is was an additional electrolysis process but more likely useful current lost. In effect, some of the current in the “bad” cell was simply wasted and not doing anything useful.

Bravo on the analysis and getting the cell replaced. A very good ending indeed.

 

[mgtfp]

That makes a lot of sense.

What triggered me to think about an additional electrolysis process, was the amount of gaseous activity going on in the cell. And a I am not talking about too many beans for dinner here. When comparing the gas cloud inside the cell (clear cell housing) of the bad Viron cell with my old old Hurlcon cell, then I had the impression that the Viron created a lot more gas compared to the Hurlcon with the the same effective chlorine output (as measured in overnight chlorine gain tests). But in the end, that's very subjective and hard to quantify.

Your explanation is much simpler and straightforward. And simple tends to be right.

On the other hand: Wouldn't any current carried through the water, whether on intended or stray paths, have to be maintained by molecules that discharge (or charge) at the electrodes? As long as only Cl- discharges at the anode, the current will be proportional to the amount of chlorine gas created. When the current is there, but chlorine gas "missing", shouldn't there be something else getting discharged at the anode?

 

[djdonte]

Dang thats a long read. Anytime my SWG had trouble keeping up I found that my CYA was a bit low.

 

[JamesW]

When the current is there, but chlorine gas "missing", shouldn't there be something else getting discharged at the anode?

If the current travels through the water, it should produce heat based on the amount of current and resistance.

It's basically like an incandescent light or heating element.

 

[JoyfulNoise]

There’s lots of other reactions that occur besides oxygen generation and chlorine generation. Pool water is very “dirty” so there’s not just one electrolysis pathway.

 

[JamesW]

Note: This is incredibly unsafe, so Do Not Try This At Home!

However, this shows how passing current through water creates heat.

 

[mgtfp]

If the current travels through the water, it should produce heat based on the amount of current and resistance.

Yes, there will of course always be heat losses, that's a very good point. And if in my case I had needed more voltage to create the same current, I had put it down to resistance effects without further questions, like ions having to travel longer paths along stray field lines.

But what's interesting here is that with the dodgy cell there was less voltage required to create the same current, so the dodgy cell was electrically more efficient (lower resistance), but with reduced chlorine output.

That's where I thought there is a process required that's normally suppressed, but became noticeable in my old cell and was competing with chloride electrolysis.

EDIT: Correct explanation in post #40.

 

[mgtfp]

Note: This is incredibly unsafe, so Do Not Try This At Home!

However, this shows how passing current through water creates heat.

Awesome, so this guy is producing hydrogen gas, chlorine gas and heat. What could possibly go wrong...

 

[JamesW]

He's using AC, so that reduces the amount of electrolysis, but I don't know how much difference that makes.

 

[JamesW]

This person tests for hydrogen and gets no reaction.

So, AC probably makes very little hydrogen, oxygen or chlorine.

 

[AUSpool]

Good to hear it’s all sorted without major drama. I wonder how many chlorinators are out there running at 50% with the owners totally oblivious that something is wrong?

 

[mgtfp]

Good to hear it’s all sorted without major drama. I wonder how many chlorinators are out there running at 50% with the owners totally oblivious that something is wrong?

Yep, very good question...

 

[mgtfp]

Good to hear it’s all sorted without major drama. I wonder how many chlorinators are out there running at 50% with the owners totally oblivious that something is wrong?

And in the end it's really just an assembly issue that can easily happen. In this picture here you can see that the grey frame holding the lose plates has different positions to click into. In my case, the frame was clicked into the first hole, which I suspect is the position for the longer V45 plates:


Unfortunately, I didn't take a photo from the side (hidden anyway behind stickers on the cell housing). But you can see from these two photos where it was wrong.

Old cell - the grey frame is touching the black nozzle thing:




New cell - there is a gap between grey frame and black nozzle:

 

[mgtfp]

I think the Penny just dropped in my brain why less chlorine got created at the same current with the old cell: It's simply because in the blue shaded areas, the current only passed through 1 cell instead of 5. I had that correct in my sketch all along, but was too blind to actually draw the right conclusion from it.

Extreme case is at chlorinator setting 1 where I measured a voltage of only 8.8V with the dodgy cell. Divided by 5 to get the voltage across the sub-cells in-between the lose plates gives a voltage of only 1.76V which is below the minimum voltage required to start electrolysis. That means for setting 1, only the blue shaded areas actually create chlorine (where the full 8.8V are at work and therefore above the production threshold), therefore the amount of chlorine created per current is a factor of 5 smaller than the chlorinator "thinks" it's creating. Only once the voltage divided by 5 is larger than the minimum voltage required for electrolysis, all the sub-cells are actually contributing to the chlorine production, so the amount of missing chlorine per current will become a bit smaller.

But with a dodgy cell like that, there will always be some chlorine "missing", because part of the current will always come from the blue shaded areas, where we only have 1 cell instead of 5. But the controller of course doesn't know that...