RO water and TDS meter

[Larso]

Hi, just bough a TDS meter which arrived a few days ago. I have an ion exchange water softener in my home and an RO unit for drinking water. The water straight from my water softener came in at 283 and the RO water came in at 21 when tested with my new meter. Does this mean my RO filters are useless or would it still be OK to treat the water from the RO unit as 'RO'? I've been usign the recommendations on the brewing water chemistry primer I'm just wondering if, based on the TDS value, that the primer would be irrelevant?

Thanks

L

 

[ajdelange]

A TDS of 21 is low enough that the water is useable with the primer. However it represents a rough rejection of only 93%. I may be time to replace membranes and filters in the RO unit especially if it has been a while since this has been done.

 

[BuckeyeFS]

"Rejection," or the percentage of feedwater TDS that is sent to the waste water (a.k.a. "concentrate") line varies with a number of variables. Among them are water temperature, net water pressure, and RO membrane specifications.

Colder water temperatures will yield better rejection.

Higher net pressure will yield better rejection. So if you need cleaner water, consider NOT delivering your RO water to a pressurized storage tank. Use an atmospheric (unpressurized) reservoir instead. Also consider increasing your feedwater pressure with a booster pump. Improvement in rejection is most pronounced when feedwater pressures low in the operating range are improved.

Some example data:

Factory spec rejection varies for different membranes. Here's some data for membranes we most commonly deal with:

Also - if you want accurate readings, be careful with how you measure TDS. Remember that you are measuring in ppm - parts per million - and nearly any contamination in the sample (e.g., dust, your finger) will skew the reading. Also, make sure you understand TDS creep so you don't let that affect your reading. When we field troubleshooting calls about presumed RO and RODI system malfunctions, a substantial portion of the time the problem is inaccurate TDS measurements.

Russ

 

[aiptasia]

TDS meters also need to be calibrated before use. Did you purchase the calibration solution for yours?

Assuming you did, and the probes on it are new, yes your R/O unit could use a little help. Depending on how hard your source water is, it may indeed be time to replace the pre-filter and the membranes on your R/O unit. It might also help to add Cation & Anion (cati/ani) chambers after the membrane to remove any remaining dissolved solids. Many cati/ani resins are rechargeable using muriatic acid (pool acid) and will last for years.

 

[BuckeyeFS]

...it may indeed be time to replace the pre-filter and the membranes on your R/O unit.

If you do some experimenting with your TDS meter, you'll find that the TDS after the prefilters is similar to the feedwater TDS.

Nearly all the change in TDS in an RO system is accomplished by the RO element.

Russ

 

[BuckeyeFS]

Many cati/ani resins are rechargeable using muriatic acid (pool acid) and will last for years.

This process requires caustic as well as acid.

 

[ajdelange]

Nearly all the change in TDS in an RO system is accomplished by the RO element.

True enough but if his poor rejection is because of chlorine/chloramine breakthrough in his carbon filter then he'd better replace that too.

 

[BuckeyeFS]

We try to get people out of the knee jerk approach that when their permeate numbers or DI numbers are bad, they change the prefilters.

Also - chloramine won't harm the membrane.

 

[mabrungard]

Recognize that the rejection rate for differing ions varies for any RO or nanofiltration process. Monovalent ions like sodium and chloride are smaller in diameter and pass through the membrane pores easier. The larger divalent ions like Ca, Mg, and SO4 do not pass as easily. They are rejected at a higher rate than the monovalent ions.

In the case above, the ion-exchange softener is taking out the divalent ions and inserting monovalent. The rejection rate of the RO machine suffers a bit from that. But the good thing about this move is that ions like Na and K are much less likely to precipitate onto the membrane and prematurely clog it. So there is definitely a benefit to pretreating a very hard water via ion-exchange prior to RO feed. You do end up with a little more sodium in the RO product water though. Its a small price to pay. Na at a concentration of less than 40 ppm is generally OK, although 20 ppm is better. If the TDS reading is correct, there is little chance that the Na is more than about 10 ppm.

You are good to go regarding the likely RO water quality.

Membrane processes are very much affected by water temperature. Cold feed water will significantly reduce the efficiency of a RO unit. This is due to the increased viscosity of the cold water. Sure, the rejection rate improves slightly. However, the water production efficiency suffers more. Warming the feed water is very helpful. Keeping your RO unit indoor is an important feature if you live in a cold region.

 

[mabrungard]

We try to get people out of the knee jerk approach that when their permeate numbers or DI numbers are bad, they change the prefilters.

Also - chloramine won't harm the membrane.

Um, chloramines will definitely harm some membranes. There are some membranes that are more resistant. However, its most likely that the membranes in Home usage are affected by chloramine. The relatively low flow rate through the typical activated carbon prefilter is likely to remove all chloramine from the RO feed water.

 

[BuckeyeFS]

Have you found a practical, cost effective means to warm RO feedwater? We've found the more practical approach is to add pressure to compensate.

 

[BuckeyeFS]

I had to dig a little but found the reference I was thinking of re chloramine and residential membranes: The chloramine (excluding any free chlorine/ammonia) tolerance reported by Filmtec was 300,000 ppm-h.

 

[mabrungard]

Have you found a practical, cost effective means to warm RO feedwater? We've found the more practical approach is to add pressure to compensate.

Pressure or heat, both add cost to the treatment. When you are dealing with large flows, its not practical to heat the inflow. If there is a source of waste heat in the facility, then incorporating a heat exchanger can be practical. In a home system, the low feed rate may be enough to warm the water to improve performance. The point I was making in the previous post was that this phenomena exists.

Even though we sometimes have to disinfect biofouled membrane systems with chlorine, that temporary application doesn't instantly destroy them. However, it does reduce the life of the membranes. The same thing with the chronic application of low chlorine or chloramine to the membranes...it just reduces their life a bit. Any of the membrane manufacturers will confirm that keeping those disinfectants off the membranes will improve their life.

 

[BuckeyeFS]

Yep. We look at chlorine very differently than chloramine however. Membrane tolerance is only 1000 ppm-h.

 

[Larso]

Thanks for the inputs guys, the 1st stage of filtration in my RO system is very dark/discoloured so I'm guessing the filters need a change. The reason I havent changed them before this is that I've been having problems with my ion exchange softener not doing its job. I recently had some one out to resolve that and I believe that I now have a working softener, time will tell. I'm on a private well with very hard water and a lot of iron so I dont think there should be any chlorine in it as I'm not on a public supply. My RO feed is directly pumped from my well about 40metres deep and the delivery pressure is good as I have a pressure vessel in the house to maintain good pressure throughout the system. My RO unit is under my kitchen sink so would rarely see temperatures below 20C and frequently be well into the 20s.
Also, I didnt know my TDS meter needed calibration? This is the unit I bought http://www.amazon.co.uk/gp/product/B004QC3D20/?tag=skimlinks_replacement-20 .I dont remember the manual saying anything about a calibration procedure but I'll double check. I assume I need a sample with a TDS of zero to do this?

Thanks

L

 

[BuckeyeFS]

Hmm. That is a counterfit knockoff TDS meter, made to look like the TDS3 from HM Digital. I don't know anything about its quality. But assuming it is fine, you can use some calibration solution like this:

You can get TDS calibration solutions at numerous TDS levels and the idea is to get one that is close to the expected TDS of the fluid you'll be measuring. 342 ppm fluid might work well for you. Check the instructions that came with your TDS meter. Some require that you use a specific calibration fluid (e.g., 800 ppm).

If you don't have one, an add-on Pressure Gauge Kit would be a good addition for you. If plumbed correctly it will help you know when your sediment filter or carbon block are clogging and stealing pressure from the membrane.

Russ

 

[mabrungard]

Since we are most interested in assessing the quality of a RO source, wouldn't a good TDS test solution be distilled water? 342 ppm TDS is OK for assessing feed water, but I don't think that is our main interest.

 

[ajdelange]

Distilled water has a resistivity of 18 MΩ-cm. Air is higher than that so air would be a good calibration medium. The meter should read 0. Obviously these meters are calibrated for slope only. If his system is working right the meter should read about 5. If the slope is off by a factor of as much as 2 in either direction it would read between 2.5 and 10 and it won't be off by anywhere near that. More likely that it is, as the thing is new, off by less than 10% (IOW it would read between 308 and 376 in a 342 ppm calibration solution) so that the meter would read between 4.5 and 5.5 if used in a solution whose actual TDS were 5. Close enough for government work. In any case he can compare the reading after installation of new filters and membranes to the previous one. If the reading drops in half, (1-rejection) has dropped by the same amount.

IOW, don't worry about calibration.

 

[BuckeyeFS]

Since we are most interested in assessing the quality of a RO source, wouldn't a good TDS test solution be distilled water? 342 ppm TDS is OK for assessing feed water, but I don't think that is our main interest.

But in practicality, the system user should occasionally be testing the feedwater as well to check the functioning (calculate rejection rate) of the membrane. So 342 could be reasonable. If they are only testing the permeate, as you said, something lower would be more appropriate, or as mentioned above, if the meter is of decent quality, there may be no practical reason to calibrate if you're only measuring down in the single digits.

 

[ajdelange]

Think about this for a minute. Lets say the guy has feed water at 342 and his system takes it down to 5. Let's say his meter reads 0 in air but that the tech that calibrated his meter at the factory was terribly hung over and erroneously set the slope in his meter 20% high (noting as I did earlier that I doubt very much that the adjustment range is that large). He would read 410.4 and 6 and calculate rejection as (410.4 - 6)/410.4 = 98.53%. Further suppose that later that day the postman delivers a bottle of 342 ppm calibration solution and he discovers to his horror that his meter is 20% off and recalibrates so it is spot on. He then goes back and reads his system again and this time he gets 342 and 5. Plugging these into the rejection formula he comes up with (342-5)/342 = 98.53%. Or put another way: the meter reads conductivity and converts it to TDS by multiplying by a constant call it s. We put the TDS numbers into the rejection formula (s*uf - s*up)/s*uf = (uf - up)/uf where uf is the conductivity of the feed and up is the conductivity of the permeate. It doesn't matter whether that constant is correct or off by a factor of 100 as it cancels out. For this application calibration does not matter. If you are trying to measure the strength of the brine to satisfy a sewer authority though it does and the 'instrument' should be calibrated.

 

[BuckeyeFS]

Good point - as the meter error spec is expressed as a % of the reading.