Sunday, August 19, 2007

Going Somewhere?

Yes, we are. Or I should say the treadle pump is. After such good results from the testing described in the previous posts, it was decided that the pump is ready for field trials.

We've located a group going to the CODEP project (, in early September and they have agreed to take the disassembled pump as part of their luggage. A member of the group helped to number the parts and disassemble the pump so he'll know how to reassemble it when it gets to Haiti. As you can see, the pump and hoses are packed in three suitcases and a duffel bag. The only part that has to be made there is the mast, which is 60 inches long, and won't fit in the bag. We still have to collect a few parts and fittings, so there won't be problems attaching the pump to the hoses on site when it arrives. We also have to burn a CD with all the pictures Jim took during the disassembly process. So, the next post will hopefully show the treadle pump being reassembled and used in Haiti.
Update: I've received word that the suitcases made it through the various airport security checks, and customs in Haiti. So far, so good.

Tuesday, August 7, 2007

Pump Plumbing as of 8/6/07

In earlier posts both Jim and I have written that there were issues with the green vinyl suction plumbing that had been used to plumb the pump's valves. The plumbing worked just fine. The problem was that it was near impossible to install, and impossible to remove in order to maintain the check valves. The green hose was just too tight and inflexible for the fittings we have been using. We tried black poly pipe too, which was easier to install, but required additional fittings and hoses, because it was so rigid. This increased parts counts and costs. It was no easier to remove than the green vinyl.

Finally, we hit on ordinary automotive type rubber heater hose. It's flexible enough to install and remove easily, readily available, and cheap enough. The only problem is that it collapses under vacuum. This little conundrum was solved by making a coil out of galvanized 14 ga steel wire. Almost any wire of sufficient stiffness will do, even if it wants to rust. A wire coat hanger would work. The coil was made by winding the wire around a 3/4 inch steel rod. Again, what one uses for a mandrel is not critical. As long as the result fits inside the heater hose fairly closely, and prevents it from collapsing under the vacuum produced by operating the pump, it's all good. Only the hoses on the vacuum side of the pump need to have a coil installed.

We tested our cheap and dirty vacuum hose arrangement on 08/06/07. It works. We'll be posting the results of the testing later, probably with more videos. We're still looking for a cheaper substitute for the pretty stainless steel hose clamps we've been using to secure the hoses. They're kind of expensive. There are a number of possibilities, and as soon as we know what we're going to do, we'll post it here.

Saturday, August 4, 2007

Test of Treadle Pump to Longer Distance and Higher Elevation

With the success of pumping a distance of 160 feet at a height of 25 feet and with an improved strainer, we wanted to try for the maximum distance and height. We ran out of pipe long before we reached the pump’s limitations. The previous 160 feet run was with 1 inch Poly Pipe which was all that we had on hand. So we decided to extend it with standard ¾ inch garden hoses. By using all the hoses I had and borrowing some neighbor’s hoses, we came up with a total of 260 feet. Add to this the 10 feet of suction hose into the pond, we had a total of 270 feet. We also wanted as much height as possible, so we continued to run the pipe up a long hill. At the very end, in order to add more height, we threw a rope over the limb of a tree and hoisted it up an additional 30 feet over the height of the hill itself. To determine the total elevation, I waited until dusk so that my cheap laser level’s red dot could be seen and used it to shine onto lower trees. By making several steps down and measuring the distance from the dot to the base of each tree, then repeating several times, I was able to find the total elevation. The elevation of the end of the run as it was hoisted up to a tree limb was easy to find in that I simply measured the length of rope needed to hoist the pipe all up to the limb. The total was 60 feet from the surface of the pond to the open end of the garden hose in the tree.

We were expecting the pump to be much harder to operate with this increased length and elevation. Instead, it was about the same; just the strokes per minute were slower. It is not necessary to use leg muscles to force the treadles down. Your body weight provides the force. At the end of each stroke, you just have to raise your body weight high enough to step up on the other treadle. To accommodate people of different weights, one can just move their feet position further away from or closer to the pistons to change the mechanical advantage. We measured the volume of water per minute pumped and knowing the total lift, we could calculate the work required. Rate was 4 gallons per minute which is about 50 watts. According to several studies by others, this an easy output for people to maintain almost continuously. Larry operated the pump for 2 hours on a very hot day and was not excessively fatigued. Realizing that we probably weigh more than the average person who will be using this pump, we decided to add several additional inches to the treadle length to accommodate people that weigh less. We feel that except for some minor changes in the way we interconnect the piping so that it is more easily disassembled for cleaning, we have arrived at the final pump design. The following video is about 2 minutes long and shows the pump being tested at the stated pipe length and elevation. There is a short video following the first one that shows Larry operating the pump at this length and elevation after he has been at it for 2 hours. These test results were very encouraging. Our next step will be to get a prototype pump to our CODEP project in Haiti and begin creating drawings and instructions for how copies can be made by others.


Prior to the testing in the last video, all of our tests had been done using potable water. That's with no mud, and no sand or rocks. It occurred to us that when we started pumping pond water, we would need a strainer of some sort to keep such material out of the pump. Our check valves were designed so they would probably tolerate particles smaller than 1/8", allowing them to pass through. One of the first things we did with the new valves was to allow the pump to ingest a lot of sand. As expected, it clogged things up and required the disassembly of the pump to clear out the sand. What this taught us was that our Delrin ball check valves were as robust as we could expect, as they experienced no damage for the abuse they received, and that as expected, particles of sand or rock slightly larger than 1/8 " would jam the ball in the valve open. It also taught us that the green suction plumbing we were using had to be removed with a hacksaw. Not good. Plumbing changes are underway as I write.

The photo shows two of our homemade strainers. The thing at the bottom was an idea we abandoned. It's a lateral for a swimming pool sand filter. It most likely would have done a good job, but there aren't just lots of pool supply stores where this pump is going, and they're expensive.

The other two strainers are just pieces of PVC pipe. One was drilled with 1/8" holes and the other slotted with a bandsaw, although it could have been done with a hacksaw, or a carpenters coping saw just as well. Both strainers were tried, and both worked. The second photo shows what happened to the strainer with holes after a few minutes of pumping. The same particles that would have stopped the check valves clogged the strainer. What we finished with was a slotted strainer about twice as long as long as shown in the photo, to provide more slots and accordingly reduce restriction to flow. We pumped for two hours, with no appreciable clogging. There will be more on our test results with the next post, and another video.