New RO Membrane Troubleshooting

[schematix]

Last week i replaced my RO membrane after observing my TDS had tripled over the past 2 months... I got 3 years out of it, not too bad for a $99 investment.

Since i bought some cheap Taiwanese parts originally, I decided this time i'd buy some better quality filters and membrane. I got a 1 micron pre-filter, 0.5 micron carbon filter, and another 0.1 micron carbon (for chloramine) filter. I also replaced the membrane with a Filmtec 100GPD and the correct flow restrictor for a 100 GPD.

Per the instructions i ran the system for an hour to flush it out. I pulled a sample at the end and was surprised to see the TDS was 20 PPM, which was higher than anything i'd ever seen before. I verified with 2 different meters and they were both reading close. This corresponds to a rejection of about 87%.

Today i went to run it again and gave my inline TDS system a check. Initially when I turned it on and the water was reading about 60ppm, and then quickly (less than a minute) had dropped to 4ppm. This i expected. However, by the time i had collected a gallon i was back up to 15ppm. This corresponds to only 93% rejection according to my inline meter.

I've never observed the TDS go up as the water was was collected, in fact, always the opposite.

I've got 75 psi at the inlet to the membrane and water temp is almost 80F.

Anyone have any thoughts what might going on?

 

[Buckeye_Hydro]

Could be a lot of things... If you'd like our help troubleshooting it, please give us a call when you are in front of the system.
Let's get this figured out!

Russ

 

[Buckeye_Hydro]

Not 100% sure yet - until we talk, but I think we are going to want to change to order of your prefilters around. The highest capacity carbon block should be the first thing after the sediment filter.

Check the spec's on that chloramine carbon block - I have my doubts that it is a 0.1 micron - more likely it is a 1 micron. If it is a 0.1, then we need to change your sediment filter(s).

Russ

 

[schematix]

Sorry I don't know how i got this so mixed up but here's what i really have:

Stage 1: PreFilter - Sediment Polypropylene 1 micron
Stage 2: Carbon - Chloramine carbon block 1 micron
Stage 3: Carbon - 0.5 micron
Stage 4: RO Membrane 100 GPD Filmtec
Restrictor - 550 ml/min

No tank, no polishing filter, no resin.

Incoming water is currently running about 230 ppm, 80 degrees and 75 psi at the RO membrane inlet.

 

[Buckeye_Hydro]

OK. Your prefilters are in the correct order, and should do a good job with your chloraminated city water.

Filmtec's are good membranes. Top shelf.

Your restrictor, given your pressure and temperature, is going to result in a recovery higher than spec. This will result in a lower rejection rate, and may shorten the life of your RO membrane (depending on your feedwater quality).

How, specifically, are you measuring the permeate (RO water) TDS? Pulling RO water out of a collection bucket of some sort?

Russ

 

[schematix]

Your restrictor, given your pressure and temperature, is going to result in a recovery higher than spec. This will result in a lower rejection rate, and may shorten the life of your RO membrane (depending on your feedwater quality).

I sized the restrictor based on 2x rejection. 100 GPD = 263 mL/min. Multiplying by 2 puts it at 524 mL/min and I rounded up to 550.

What do you suggest?

How, specifically, are you measuring the permeate (RO water) TDS? Pulling RO water out of a collection bucket of some sort?

I measured 2 ways. I have my old school pen style TDS meter and i also have an inline TDS meter.

Obviously the in-line meter reads right in the line via a tee. The in-TDS measures coming right out of the house supply, and the out-TDS measures coming right out of the membrane.

For the pen samples i took water samples of 100-200mL into a beaker, at the same time i checked the inline TDS meter, and then measured with the pen.

There was only a few ppm difference between the samples. Close enough for me.

 

[Buckeye_Hydro]

I sized the restrictor based on 2x rejection. 100 GPD = 263 mL/min. Multiplying by 2 puts it at 524 mL/min and I rounded up to 550.

What do you suggest?

Well, spec recovery (not "rejection") would be closer to using a 4x or 5x multiplier.

And assuming you have a Filmtec 100 gpd membrane - that is spec'ed at 50 psi and 77F. Your pressure is higher, and your temperature is higher, so the membrane will produce more than 100 gpd. How much more? Using the calculator (see link on our homepage) that membrane will produce about 168 gpd if you use a 4x multiplier.
168 x 4 = 672 gpd = 1767 ml/min

Our 150 gpd capillary restrictor would be a decent fit, or at least the 100 gpd restrictor.

If your water is soft, your tighter restrictor may not cause scaling.

Russ

 

[mabrungard]

As Russ mentions, it is the scaling potential that ultimately defines what the bypass (wasting) rate needs to be. Those of us with ion-exchange softened water, can run much lower bypass flow.

Russ, isn't this TDS problem more likely to be an initial conditioning result? I expect that the TDS will improve as the membrane is hydrated and conditioned by the incoming flow. In addition, I sometimes see TDS fluctuation from my aged membranes too. But until I see consistently high TDS readings, I just live with the variation. I don't feel those initial readings are concerning.

 

[Buckeye_Hydro]

Neither the flux nor the rejection should be taken too seriously until after the first 6 to 24 hrs of run time on the membrane. Both will change. That said, the rejection on new residential Filmtecs is typically very good (high 90's).

 

[schematix]

Should I let the system run for 4-6 hours to see if it stabilizes? The membrane package said to run for an hour....

 

[Buckeye_Hydro]

Those times you're referring to Schematix are really for two different things.

The 1 hour "rinse up" time is to rinse anything, including preservatives from the manufacturing and initial shipment of the membrane.

The 6 to 24 hour period of time is when the entirety of the membrane will hydrate and performance will stabilize.

Russ

 

[schematix]

OK so back to my initial question of why i am seeing the TDS creeping up from really good (4 ppm) to so-so (20ppm) after the initial discharge (60-80ppm), could possibly be related to it being a brand new membrane? If i run it for a few more hours it may start to go back down and stabilize there?

 

[Buckeye_Hydro]

I doubt you'll see an improvement from 20 to 4 ish by letting it break in. It will likely improve some, but we don't see that much improvement. Let me go back and read from the beginning of the thread.

 

[Buckeye_Hydro]

First - are you 100% certain the RO membrane is fully seated (push in all the way)? You may have some waste water making its way out the permeate port.

If it is inserted all the way, rarely we see a crack develop in the walls of the little cup into which the membrane seats. Sometimes you can see this crack if you pull the membrane out and look inside the housing.

Also - check your waste water to purified water ratio and let us know what you see there.

Russ

 

[schematix]

First - are you 100% certain the RO membrane is fully seated (push in all the way)? You may have some waste water making its way out the permeate port.

If it is inserted all the way, rarely we see a crack develop in the walls of the little cup into which the membrane seats. Sometimes you can see this crack if you pull the membrane out and look inside the housing.

Also - check your waste water to purified water ratio and let us know what you see there.

Russ

All good suggestions. I will check those ideas out.

 

[schematix]

ok...

1. I tried to run the membrane for a couple more hours, but it went to a steady state of 20 ppm after about 2 gallons. It had the initial spike from sitting, then within a few hundred mL was down to 4mL. By the end of 2 gallons it was at 20 ppm TDS. Again, confirmed with 2 different meters and also verified distilled water was reading 0 ppm.. both agreed.

2. I took the membrane out of the housing and inspected the housing carefully with a flightlight. There were no signs of any cracks. There was a minor scuff on one of the two membrane o-rings so i gave it a shot of the silicone lube I use on keg o-rings, and it slid right back in... should have done that the first time!

3. I re-ran the system and it repeated the same increasing TDS behavior.

4. I used beakers with volume marks to determine that i'm getting a rejection ratio of only 1.4. The flow restrictor i have is rated for 550mL/min, and it tested at 650mL/min.

I am moving towards the conclusion that when my water is this warm that i need an even less restrictive flow restrictor. Do you guys agree?

Right now I am getting 93% rejection. I'd like to be more around 97%.

 

[Buckeye_Hydro]

4. I used beakers with volume marks to determine that i'm getting a rejection ratio of only 1.4. The flow restrictor i have is rated for 550mL/min, and it tested at 650mL/min.

You mean 1.4 waste water to 1 permeate? If so, we can change this ratio and improve your rejection.

BTW - this is RECOVERY, not REJECTION RATE
RECOVERY: Percent of the feedwater that is recovered as permeate [RO water].
REJECTION: Percent of the feedwater TDS that is NOT in the permeate.

Russ

 

[ajdelange]

I am moving towards the conclusion that when my water is this warm that i need an even less restrictive flow restrictor. Do you guys agree?

When the water gets warm it viscosity goes down and so it
•flows more rapidly through the restrictor resulting in lower pressure on the brine side
•flows more rapidly through the membrane thus increasing recovery rate

These two effects would, I would think, cancel one another to some extent at least but if you replace the fixed restrictor with a needle valve you would then be able to control the pressure on the brine side and thus the recovery.

I don't see how decreasing the recovery would improve the rejection unless there is some second order effect I'm not aware of. In my own system I can, by adjusting the restrictor valve, set recovery pretty much anywhere I want (up to about 50%) without any change in rejection that I have noticed.

Concerns with recovery rates are usually centered on concentration of the brine side critical salt (the one that is most likely to precipitate i.e. calcium carbonate unless the water has been softened in which case it is another salt and usually not a concern).

Right now I am getting 93% rejection. I'd like to be more around 97%.

From everything I know about RO you should be getting more like 97% from a new membrane. We have to allow that just as with any other product some bad RO membranes are going to slip through the QC process. Given that you have checked all other reasonable explanations I'm thinking it may be a defective cartridge. I'd start talking to the supplier about a replacement.

 

[schematix]

You mean 1.4 waste water to 1 permeate? If so, we can change this ratio and improve your rejection.

That is correct. 1.4 units of waste water per 1 unit of permeate. The permeate has about 7% of the feed water TDS. This is about 1/3 higher than the 3-year old membrane it replaced.

Regarding the flow restrictor, that seems like a different issue. Is it worth the effort to get another one or is 1.4 going to be fine?

Thank you both for your suggestions. I think AJ is right that this membrane is probably defective.

 

[Buckeye_Hydro]

1.4 : 1 is likely too tight. If pressure and temperature was at spec, that flow restrictor is to be teamed with a 50 gpd membrane. You need more concentrate, relative to the permeate.

This is a VERY easy, quick, and cheap ($4) fix.

 

[The_Bishop]

Hey, crazy question but can I use a needle valve in replacement of my flow restrictor to fine tune my rejection/recovery rate?

 

[Buckeye_Hydro]

CAN you? Sure.

SHOULD you? Probably not the best/practical idea on the residential scale system you have. Much more practical on commercial (higher capacity) systems where you can have a flow gauge on the concentrate.

Russ

 

[ajdelange]

Hey, crazy question but can I use a needle valve in replacement of my flow restrictor to fine tune my rejection/recovery rate?

Not a crazy question at all which is why I suggested doing just that in #18. Larger systems have these and are, as mentioned, often equipped with flow meters for permeate and concentrate. While these are nice to have what is more important with an adjustable restrictor is a pressure gauge so you can see what the consequences of throttling down are. You don't want, obviously, to exceed the pressure rating of the membrane. For an inexpensive under the sink system there might be implications with respect to the pressures tolerated by other parts of the system.

Closing down the valve reduces the flow out of the valve and increases the pressure on the concentrate side of the membrane. This results in improved recovery (bigger numerator, smaller denominator) but the rejection rate should stay pretty much the same regardless of recovery rate unless you want extremely high recovery and achieve it by feeding some permeate back to the pump inlet. In systems that do this rejection is compromized,

 

[Silver_Is_Money]

Here is a crazy question: I purchased my RO unit from among the low end of the choices made available to me. When I reach the juncture of replacing the various filters and membranes can I elevate its performance by choosing the right filters and membranes by which to accomplish this end, or must I replace exactly in kind?

 

[Buckeye_Hydro]

IF your system takes proprietary filters, then your stuck buying them from a vendor selling that brand. If the system takes standard-sized filters, then yes, you can buy higher quality filters, as well some other things a budget-system my not have included, like a pressure gauge and a TDS meter.

Russ

 

[schematix]

Here is a crazy question: I purchased my RO unit from among the low end of the choices made available to me. When I reach the juncture of replacing the various filters and membranes can I elevate its performance by choosing the right filters and membranes by which to accomplish this end, or must I replace exactly in kind?

That's exactly what I did. And now my system with brand new name brand everything has higher TDS than my 3-year old cheap stuff was putting out.

In theory there should be no issues with this since i had a standard housing system. Just bad luck i guess.

Hopefully will be resolved this week though.

 

[The_Bishop]

Not a crazy question at all which is why I suggested doing just that in #18. Larger systems have these and are, as mentioned, often equipped with flow meters for permeate and concentrate. While these are nice to have what is more important with an adjustable restrictor is a pressure gauge so you can see what the consequences of throttling down are. You don't want, obviously, to exceed the pressure rating of the membrane. For an inexpensive under the sink system there might be implications with respect to the pressures tolerated by other parts of the system.

Closing down the valve reduces the flow out of the valve and increases the pressure on the concentrate side of the membrane. This results in improved recovery (bigger numerator, smaller denominator) but the rejection rate should stay pretty much the same regardless of recovery rate unless you want extremely high recovery and achieve it by feeding some permeate back to the pump inlet. In systems that do this rejection is compromized,

I understand that, I just didn't know if a needle valve would have the fine resolution of control necessary to really fine tune the flow.

My system's a bit of a hot-rodded frankenstein setup, with a booster pump, a permeate pump, and two 75 GPD membranes in series being fed at 90 PSI. It'll generate roughly 200 GPD at 4-5 TDS, with about a 2:1 recovery ratio. I'm able to get away with this because the feedwater to the system is dead soft from my home water softener. It's been about a year so far and no problems.

It's a bit overkill but it not only provides brewing water but household drinking water and ice via an 11 gallon capacity pressurized storage tank.

Also, I hate waiting for water when I want to brew.

 

[Buckeye_Hydro]

If you want to try it, http://www.buckeyehydro.com/precision-needle-valve/

 

[The_Bishop]

Sorry for the thread hijack, but I've got a booster pump related question:

I'm running an Aquatec 8800 pump with a 3/8 feed line. As things are now, the pump is maxxed out. If I increase the reject flow the pressure falls off pretty rapidly. How much more water does the high capacity pump deliver?

 

[Buckeye_Hydro]

Sorry for the thread hijack, but I've got a booster pump related question:

I'm running an Aquatec 8800 pump with a 3/8 feed line. As things are now, the pump is maxxed out. If I increase the reject flow the pressure falls off pretty rapidly. How much more water does the high capacity pump deliver?

What flow restrictor are you using with your two 75's in series?

 

[Buckeye_Hydro]

We use the high capacity pump on systems up to 400 gpd @ 100 psi.

 

[The_Bishop]

Right now it's a 100GPD. Item 0218, from you.

 

[ajdelange]

I understand that, I just didn't know if a needle valve would have the fine resolution of control necessary to really fine tune the flow.

Yes, definitely but you will want some flow and pressure gauges in order to be able to see what the system is doing.

My system's a bit of a hot-rodded frankenstein setup, with a booster pump, a permeate pump, and two 75 GPD membranes in series being fed at 90 PSI. It'll generate roughly 200 GPD at 4-5 TDS, with about a 2:1 recovery ratio.

I run a similar system. The heart is a skid with prefilters and a pump capable of 150 psig, flow meters and needle valve. I run it between 2:1 and 1:1. I kluged up pressure tank, atmospheric tank, pressure pump, float and pressure switches and a control box. The brewery and lab are plumbed (with plastic pipe). Fed with softened water this system delivers a little over 1 lpm and I have run it for over 5 years with softened water feed and have had no failures, except for the pressure gauge, in that time. I have not had to change filters nor membranes and still get 4 -5 ppm tds (about 97% rejection). The surprising thing to me is that my silica is at 28 mg/L and I thought that would gum things up by now but it hasn't (touch wood).

I'm able to get away with this because the feedwater to the system is dead soft from my home water softener. It's been about a year so far and no problems.

There shouldn't be any problems. It's easy enough to calculate your brine salt concentrations from your feed water chemistry and rejection data. Increase the concentrations by about 50% to allow for 'polarization' (ion concentration in the channel is higher close to the membrane) and, I'm sure, you will find that the critical salt is well below saturation at 2:1 recovery.

Given that you have pretty much put this together your self you could try adding a second needle valve and plumbing its output back to the input to the pump. You'd want a flow gauge on this line too. By doing this you can increase recovery even further which may be of interest since you obviously like to tinker. It might be possible to get a water that meets some brewing need at 60 or 70% recovery (though you would obviously need to supplement calcium).

It's a bit overkill but it not only provides brewing water but household drinking water and ice via an 11 gallon capacity pressurized storage tank.

In addition to brewing I have found that having 50 gal (or more - I can set how high the system fills the atmospheric tank) RO water on hand beneficial for things other than brewing.
•Brewing water for me
•Brewing water for anyone who wants to bring jugs
•Brewery Rinse water
•Feed for lab DI water maching
•Lab faucet for glassware pre-rinse
•SO's CPAP machine
•Source of water for cooking, drinking, toilet flushing... when plumber or well guys shut off house supply

Most recently when the HVAC installer said he needed 70 gal pure water to mix with glycol for a new geo heat pump system my response was "Got a hose?"

Also, I hate waiting for water when I want to brew.

That, in a nutshell, is why I put the system in. I've had more that one brew day dawn on which I realized that calculations were done, grain was ground and ready to roll but that I had forgotten to collect water (used to take me a couple of days with my old 10 gpd system). This realization was followed by trips to clean out the entire stock of DI water from several drug stores.

And putting the system together was a fun project.

 

[The_Bishop]

Thanks for the pointers, AJ.

What are you using for a non-pressurized storage tank? I've been considering a non-pressurized storage tank and using a delivery pump to feed the 'household' end of things. The main reason I haven't is I'm not sure if there'd be an issue with the storage tank going 'stagnant' if the water sits too long.

I've got pre and post filter gauges to see if there's any restriction from the filters and a pressure gauge tee'd into the pressurized feed going into the first membrane housing. I should consider some flow gauges... Right now measuring flow consists of a graduated beaker and a stopwatch.

 

[ajdelange]

I found some company (I'd say which if I remembered) in California that make semi custom plastic tanks - semi custom in the sense that they have several stock sizes (round and rectangular) but will spin-weld fittings onto them anywhere you want them. I ordered a 100 gal rectangular (50 would have been plenty) with cover. The cover is vented through a fine mesh screen to keep bugs from being able to get in. There are two pairs of NPT ports on one end for high and low float switches which shut off, respectively, the RO unit (prevents overflow) and the pressure pump (if it runs the atmospheric tank dry trying to fill the pressure tank). I also have a level sensor on the outside which I can position at 1/4 full or 1/2 full by means of two brackets which I welded on myself. Most of the time the tank is kept 1/4 full which is why I said 50 gal would be plenty.

I, too, worried about stagnation and so I had them put a fitting on the bottom which connects, through a valve (plastic) to a line that runs to a drain. We go away for the summer and I do drain the atmospheric tank before we go. I've had the system since before we started coming up here for the summer (so I guess it's actually more like 7 years I've had it) and never had any problem with stagnation but then when I was there year round I would be draining it for brewing from time to time. I'd guess that if you are using it for brewing and drinking water you would probably be similarly safe. Atmospheric tanks are, AFAIK, a common part of whole house systems.

I'll also relate that in my youth a friend decided to have a party and serve punch from the crock that was part of the Sparklets water dispenser. What we found in there when we opened it up was most impressive and thoroughly disgusting but everyone had been drinking from this thing for years with no noticeable bad taste nor smell.

 

[Buckeye_Hydro]

Right now it's a 100GPD. Item 0218, from you.

The Aquatec 8800 is spec'ed for use with up to 120 gpd membrane. You can make it work with the 150 gpd membrane if you go with a 100 gpd restrictor.

If you want more pump yet, then the High Capacity Pump is the next step up. Much less expensive than a motor and pump like we put on commercial RO's.

Russ

 

[Buckeye_Hydro]

When we install storage tanks for RO in commercial applications, say a 500 g or 1000 g tank, we often install a UV and pump to recirc the tank water.

 

[The_Bishop]

Thanks, Russ. Doing a bit of planning now.

 

[schematix]

I received a replacement membrane earlier this week and gave it a whirl. It's putting out about 50% more TDS than the last one, nearly 30ppm! That's a rejection of less than 90%.

I'm out of ideas here.

 

[Buckeye_Hydro]

Assuming you bought a good membrane brand, I suspect: Internal crack in the RO membrane housing. Not visible when you look inside - only opens when membrane is seated.

Russ