Gravity Filtering With a Sawyer Squeeze

That may sound like an oxymoron, but there’s a good reason to try it. Forcing water through a filter more slowly might make it last longer. Does decreased pressure increase lifetime? With a brand-new Sawyer Squeeze, I started my thru-hike of the Pacific Northwest Trail intending to find out.

It’s been a long time since I first got frustrated with my Sawyer Mini on the PCT, and I’ve learned a lot since then. Here’s what I’ve written previously.

At first glance, gravity filtering seemed to require more items than I expected.

1. 72-ounce (2.1L) Cnoc Vecto. If you want to speed things up, use the 3L version.
2. Trekking pole. Because there isn’t always a place to hang the Vecto.
3. Mesh storage bag. The drawstring loop attaches the Vecto to the pole’s strap.
4. Sawyer Squeeze, outlet cap, homemade inlet stopper, back-flush syringe.
5. 12″ gravity tubing and inline adapters. These come with the Squeeze.
6. Cleaning coupling. This doesn’t come with the Squeeze.
7. 32-ounce pouch. For collecting filtered water.

I soon realized that the only extras required for gravity filtering are items 5 and 6.

Pre-trail testing

Before leaving for the PNT, I measured the filter’s flow-rate with and without the gravity tubing. I also filled a 24 ounce bottle by squeezing water through the filter. Each measurement was repeated three times, and the results below are the mean of those three values.

I doubted the accuracy of the fluid-ounce markings on the Vecto, but wanted to use them to calculate water pressure at the filter inlet. Measuring the distance of each mark from the filter was straightforward, but calculating the amount of water remaining in the Vecto was more complicated.

1. I filtered water into a 64-ounce measuring bucket. I assumed its markings were accurate, and timed how long it took for water to reach each mark.
2. Then I timed how long it took for the level in the Vecto to drop past each marking.
3. By combining the two sets of timings, and interpolating, I calculated the actual amount of water in the Vecto at each mark.

Pre-trail gravity filtering results.

• (A) The reference line shows how much water would still be in the Vecto if its markings were accurate. The green and red lines show how much is actually still inside.
• (B) The 12″ gravity tubing significantly increases water pressure at the filter inlet.
• (C) Shows how much has been filtered from the Vecto at each of its markings.
• (D) Higher inlet pressure causes faster filtering, but it would still take 45 seconds to fill a 24 ounce bottle. In comparison, filling that bottle by squeezing water through the filter took 19 seconds. This implies a squeezing pressure of 2.06 pounds per square inch, more than twice that of gravity filtering.

No squeezing

On the PNT, I never squeezed water through the filter. No squeezing, only gravity. The entire point of the experiment was to see if I could extend the filter’s lifetime. Whenever possible, I completely filled the Vecto before filtering. Even if I didn’t want that much water, I did want to speed things up. Afterwards, I simply threw away what I didn’t need.

Several times, when it was raining, I had to sit in my tent to filter water. Using the gravity tubing wasn’t practical, so I connected the filter directly to the Vecto. With my elbows on my knees, I held everything high enough for gravity to do its thing; its slow, slow thing.

During prior thru-hikes, squeezing water through the filter, I often noticed that at first, the water leaving the filter contained lots of air bubbles. This was caused by putting away the filter “dry” after use. Gravity filtering doesn’t create enough pressure to clear this air from the filter, which can slow things down even more. To clear the air, I attached the full Vecto, gravity tubing, and filter – outlet cap still in place. Then I shook the filter from side to side, or knocked it against my leg, with just enough force to produce a stream of air bubbles within the Vecto. When the bubbles stopped, I stopped.

No spilling

The four 32-ounce (0.95L) Sawyer Squeeze pouches that I took with me on the CDT managed to survive their journey. So, because I like testing things to destruction, I took them with me on the PNT. I used them for clean-water storage only, but still found their limits pretty quickly.

• Day 3. Splat! A full pouch fell off the bench I was sitting on, and hit the ground broadside. It only dropped about 18″, but that was enough to burst a seam.
• Day 9. Fatigue. A pouch sprang a leak at one of its many creases.
• Day 21. Fatigue. Another weak-point in the tired, worn-out plastic.

Pouch number four survived the trail and is now in the trash retirement. Once I was down to my last pouch, I mostly used a soda bottle to collect filtered water. The drawback is that air needs to escape as the bottle fills up, so it can overflow if you don’t keep an eye on it. Pouches can be tightly screwed onto the filter outlet, and there’s no risk of spillage.

A full 32 ounce pouch will dangle from the end of the gravity tubing for a few minutes at most. It’s heavy enough to slowly pull the inline adapters from the tubing, and the Vecto will then empty its contents onto the ground. This can be prevented by shortening the trekking pole.

Some clogging

Occasionally, the water I collected wasn’t as clear as it looked. The Vecto contained a cloud of tiny particles, and after hanging it on the trekking pole, the particles began to sink to the bottom. I attached the gravity tubing, then sat with my thumb over the outlet for about 10 minutes. Once the debris settled, I coaxed it towards the gravity tubing by gently flicking it a few times. I removed my thumb from the outlet for a fraction of a second, flushed the debris, and finally attached the filter. It would have been quicker to let the debris clog the filter, then back-flush. But as I said, the whole point of the experiment was to extend the filter’s lifetime.

I performed a back-flush an average of once every two days, i.e. about half as often as I’ve needed to during prior thru-hikes. It wasn’t that the filter clogged more slowly on the PNT, just that it wasted less of my time by being slow. While gravity was pulling water through the filter, I was busy eating a snack, emptying debris from my shoes, scouting upcoming waypoints in FarOut, or putting up my tent.

Usually, while I was busy with some other task, filtering would finish successfully in the background. The result was an almost-full collection bottle, an empty Vecto, and a “dry” filter. If I subsequently wanted to back-flush the filter, I had to first ensure that I cleared the air from it, as described earlier. Then, with the filter full of water, I knocked it against a tree-trunk (or the edge of the sole of my boot) for 30 seconds. Finally, the back-flush.

Filter preparation

After the PNT, the gravity filtering part of the experiment was over. All that remained was to restore the filter as much as possible. I started by giving it 30 seconds of Knock-on-Wood (KW) treatment. Essentially, the same back-flush procedure I used on trail, except this time I knocked the filter against a block of wood.

Next, the Sawyer Squeeze got the Vinegar (VG) treatment. I pushed air through the filter, screwed it onto a vinegar bottle, squeezed some of the contents through, and left it for 30 minutes. Afterwards, I squeezed some more vinegar out and left it for another 30 minutes.

Finally, I put the filter through three rounds of HK treatment.

Hot-water-plus-Knock-on-wood (HK) treatment

1. Put 24 oz. water in a Pyrex bowl
2. Drop the filter into the bowl without its inlet stopper and outlet cap.
3. Microwave the bowl at maximum power until it reaches 140℉. It takes about three minutes – use a thermometer to check.
4. Remove the filter from the hot water, insert the inlet stopper, cap the outlet.
5. Give the filter 30 seconds of knock-on-wood treatment.
6. Back-flush the filter with a syringe of 140℉ water.
7. Top up the bowl with a syringe of cold water and drop the filter back in.
8. Microwave the bowl at maximum power until it reaches 140℉. It takes less than 30 seconds – use a thermometer to check.
9. Repeat steps 4 thru 8 as needed. (I did it 18 times.)
10. Squeeze 24 oz. of cold water through the filter.

Results

The table below shows how long it took (min:sec) to fill a 24 oz. (0.71L) bottle by squeezing water through the filter as quickly as possible. I used the average of three measurements, and converted the time into gallons-per-minute (1 US gallon = 3.79L). The charts show how flow-rate drops, assuming that I drink two gallons of water per day while on trail.

Comparing my latest Sawyer Squeeze to previous filters.

• Row 1. This Sawyer Squeeze is faster than the previous one.
• Row 2. After the PNT, this Sawyer Squeeze retained 42% of its initial flow.
• Row 3. Vinegar treatment had no benefit.
• Rows 4 thru 6. The first round of HK treatment made a significant improvement.
• (A) Shows flow reduction with use. This Sawyer Squeeze has filtered no more than 130 gallons, i.e. about half as much as the previous two filters.
• (B) Shows relative flow reduction with use.

Conclusions

I spent 65 days on the PNT, and gravity filtered two gallons per day at most. In that time, the flow-rate of my Sawyer Squeeze dropped to 42% of its original value. The HK treatment described above is the most efficient way I know to unclog a filter. A single treatment, approximately 40 minutes in duration, restored the filter to 70% of its initial flow. Two more treatments provided diminishing returns, and I stopped once the filter reached 76% of its original value.

The Sawyer Squeeze I used on the CDT was significantly slower at the same point in its lifetime. It retained less than 65% of its original flow after filtering the same amount of water. A likely explanation for this difference is gravity filtering, although there are two possibilities I can’t rule out: I drank less per day on the PNT, and its water sources were cleaner.

The only times I squeezed water through this filter were pre-trail, when the filter was clean, and post-trail, when I wanted to compare it to previous filters. Gravity filtering causes lower inlet pressure, and should reduce the probability of debris getting permanently stuck. In other words, it should improve back-flush efficiency. For every 10,000 particles trapped in the filter membrane, the PNT back-flush consistently dislodged 25 particles more than the CDT back-flush. That small increase resulted in a significant difference over time.

In future, I’ll use gravity when I can, and squeeze when I have to. My Vecto and filter will last longer, I’ll have more free time, and last but not least, I’ll have fewer opportunities to complain about frozen fingers.