What is the lowest TDS you can have to let the water sensor do its job?
I have a cheap Chinese type

Thanks
Randy

This comes up from time to time on these pages. I wonder, though, if anyone has performed any rigorous testing. I can tell you that when my machine was new, I fed it a diet of about 5 ppm TDS from my previous RO and it worked just fine, despite a number of folks here and on CG telling me it wouldn't. My current RO has a bypass, which I set to achieve about 25 ppm TDS, as I think I slightly prefer the taste to that of pure RO (I say 'I think' as I've not done blind tasting). I've never tried distilled water, but suspect that it would work fine for at least a while in an older machine, as it would dissolve what scale there may be in the boiler. Maybe someone getting a new machine would be willing to run an experiment with water of various hardness. It shouldn't be hard to mix up some batches of distilled and mains water of various proportions...or maybe someone has already done that?

I've not done any rigorous testing, and I would think it would depend on the controller, probe-ground geometry, probe condition, etc.

Based on the probe sensitivity for a Gicar 3d5, per this informative post from Assaf Litai (AssafL) : A Question for Electrical Engineers on Surges and Control Boards, TDS values as low as 3ppm may work.

Note that the particular controller in that post lets you select a high, medium, or low sensitivity. At the high sensitivity of 1 mega ohm, and assuming 5 cm from the probe to the boiler wall a cell constant* of 1/5 cm⁻¹, your water would need a conductivity of only 5 micromho/cm, which would read maybe 2 to 4 ppm on a TDS meter. At the low sensitivity it would be up in the neighborhood of 30 micromho/cm ( 15 - 25 ppm.)

Different controllers may have very different circuits with different sensitivities. See this recent post from JP (Pat) Boyt (jpboyt) about a couple different sensor schemes: Rebuild of Quick Mill Andreja Premium.


Edit addition: The cell constant is a factor related to the surface area of the probes and the effective distance between them, and is complex to estimate for a probe and boiler wall sensors. My choice of 0.2 cm⁻¹ is a little arbitrary, so my result here could easily be low by a factor of 2 or more.

Showing my work:
we want
1 megaohm resistance across 5 cm inverse effective cell constant
so
resistivity = 0.2 megaohms ・ cm

conductivity is the inverse of resistivity, and
1 / (0.2 x 10⁶) = 5 * 10⁻⁶

so
conductivity = 5 micromhos / cm