I can give you two first hand accounts...
I was part of a massive Maxtec 305 recall. If you’ll look, you’ll notice Maxtec 305 was replaced with a new model, I think it’s 305F now or something like that.
Also, early on, I got a bad batch of AI’s, but that was a fluke. I certainly don’t think it’s the norm.
Today, at least a few manufacturers are stating that you should replace all sensors at once. I talked to Dive Soft recently regarding the Liberty. Because that machine is so precise, (as I understand it) it wants to calibrate cells that are exactly the same. Apparently there are very minor variances in cells prior to calibration. My optima doesn’t care. It’ll take a very broad array of cells and work seamlessly, but if I put three different cells in a Liberty, it freaks out on the dive, voting out dissimilar cells constantly.
When I took rEvo class, Paul Raemakers said to spread out cell replacement. Recent classes have told me to buy all in the same batch. I think that bad batches are rare, and we test for cell linearity before and during a dive. We have protocols to ensure our wellbeing and I’m not afraid of a bad batch of cells. But I do think it’s unlikely but possible.
The old Maxtec was one of the things my instructor talked about. But basically said that problems really haven't occurred recently. All were a while ago.
When you say you test for cell linearity during the dive are you talking about an oxygen flush at 20ft? It seems I was told checking cells on a dive is pretty standard in training, but I'm finding that's not true.