Sunday, July 29, 2007

July 29, 2007 - Treadle Pump Test at Jim's 160 Feet Long and 25 Feet of Height

Update by Jim: On Saturday, July 21, Larry and I decided to bring the pump to my house where I have a pond that we could use as a source of water and a fairly long and steep hill to run pipe to simulate how the pump will be used in the CODEP project in Haiti. Some things we've learned: Saturating the piston-seal leather in melted paraffin did not work very well. The leather did not seal very well and there was a lot of friction as it rubbed on the cylinder wall. So far it appears that the best lubricant is Crisco; not the liquid vegetable oil but the solid product, much like old fashioned lard. We first set up the pump beside the pond and connected the output to 160 feet of 1 inch polypipe that was run up a hill for a distance of 160 feet and to an elevation of 25 feet. The following video shows a view from the exit end of the pipe looking downhill where Larry is operating the pump. Effort was quite easy and rate of flow was about 5 gallons per minute.

Video - 160 Feet Long/ 25 Feet of Height

Sunday, July 15, 2007

Latest Tests, Mixed Results, Mostly Good

Well, the pump described in the last post was hooked up and wrung out a few days ago, with mixed results. I have mosly good news though.

The new plumbing worked just fine, as did the new check valves. I was holding my breath about the new valves. The easy up-and-down mast worked too. The new plywood braces were plenty rigid. That leaves the paraffin valves. Results there were less than stellar. Without going too much into the gory details, the paraffin seals were almost impossible to prime, and once primed, the upper seals, the ones that should provide a seal on the vacuum stroke, were excessivly leaky, and wouldn't hold their prime. After a half hour fighting with the balky seals, I retreated to the older molded seals, primed the pump, and everything else worked.

Except the pulleys. The pulleys that hold the cable that raises the treadles on each vacuum stroke have been something we've sort of neglected till now. We were using a pair of old brass pulleys that I had lying about from an attempt to hang a bird feeder where the squirrels couldn't get at it. That endeavor failed by the way. Anyway, the pulleys were too small for the task we had applied them to, and they didn't survive. All the vigorous jumping up and down I did while trying to use the paraffin based seals didn't help any.

The pulleys have since been replaced with three inch diameter pressed steel pulleys with a ball bearing in the center, intended to help raise a garage door. If you click on the photo, you should be able to see the new pulleys and get a better look at the metal bails that hold the mast.

We still have a working pump. In fact, plans are to take the pump to Jim's house, plumb it to his pond, and try to pump the pond dry. No, really, we're going to take advantage of the water his pond can supply to work the pump for as many hours as he and I can make ourselves pump, so as to identify any part that is prone to premature failure. We're also going to play some games involving feeding the pump sand and mud to see how the seals and check valves handle that. We've got some strainers to test as well. Until now, we've been pumping clear water.

There's been some preliminary talk about shipping this pump to Haiti. Should all this testing we've planned work out with no big setbacks, I hope this can happen. The sooner the better. The sooner we can manage to get a pump on the ground in Haiti, the sooner we'll know what works, what doesn't and what needs to be changed.

Saturday, July 14, 2007

Treadle pump as of July 12

The pump has been rebuilt again. I think this is the fifth time. Fortunately, it's starting to look a lot more practical. At least I think so. This iteration incorporates some new features I have talked about in earlier posts, and some I haven't.

The seals, which are sticking out in front of the cylinders to dry are molded using melted paraffin. The pump cylinders have new homemade check valves. The valves are installed both on the same side of the cylinders, which allowed us to reduce the amount of plumbing required, and simplified the routing. The plywood braces have been cut down, to minimize weight and save materials. A cleat has been added on the left end of the pump, to act as a treadle stop, and keep the treadles from beating the plywood brace on the end of the pump to bits. A lot of the extra holes have diasappeared. We needed them earlier so we could change the position of things easily. The design has firmed up now to the point that those adjustments are not needed any more. One last feature that was Jim's idea was to mount the mast with metal loops that allow it to be raised and lowered without tools or any hardware. That's the mast lying across the top of the pump. If you look closely, you can see the metal bails that it slides through. The one on the left rotates with the mast to a vertical position, and allows the mast to slide down and seat on a bracket between the pump bases that you can't see, but is there. In the lowered position, the mast projects past the vertical supports and provides a carry handle on each end. How handy. More tests coming soon.

Redesigned Plumbing


Somewhere in an earlier post, I remarked that we were going to have to get the loopy green plumbing under control. We have. Instead of drilling the pump bases on opposite sides, we decided to drill and tap for the check valves on the same side of the pump, with the holes spaced 36 degrees apart. 36 degrees happened to be the magic number that placed the valves close enough together to fit inside the bases, but didn't cause them to interfere with each other when they were screwed in. Some geometry was involved. The photo shows the result on a partially assembled pump. I guess this is a good time to say that the pump has been torn completely to pieces and is being rebuilt from the ground up. That's one reason I've gotten a little behind in the blogging. Anyway, the result is greatly simplified plumbing. It's going to be quite a bit cheaper too, because there's a lot less of it, and because there are no short radius curves, so a cheaper type of tubing can be used it, even though I'm not using it in the photo. One thing at a time, please. The astute observer will also note that the new check valves have been installed. The new pump will incorporate a number of new features. Details later. Oh, I should also say that the 2x4 in the middle and the little pieces of plywood back by the pump cylinders won't be part of the finished pump. They're just there to get the spacing right.

Sunday, July 8, 2007

Shop Made Check Valves

For some time we've been operating the pump with commercially manufactured check valves. They work great, but they're expensive. One of the first posts here was about an attempt to make check valves for an early pump using standard PVC fittings, and kid's toy glass marbles. These did in fact work, but clearly we had to make improvements. We're also trying to keep the cost of the pump as low as possible, and four of the most expensive parts are the check valves. The problem was finding a source of some sort of sphere that would serve to replace the glass marbles. There are lots of companies that make all sorts of spheres, balls and bearings out of every material you can imagine, from the mundane to the exotic. They're also very happy to sell them to you, in lots of thousands. We needed eight or ten for our purposes. Fortunately, we were able to locate Salem Specialty Ball Co. in Canton, Connecticut. They were very helpful, and very accomodating to sell us Delrin balls in sizes and quantities that met our requirements, for a reasonable price. Thank you, Salem Specialty Ball Co., and special thanks to Deloris. In case you're wondering, Delrin is an acetal resin similar to nylon. It is said to have excellent abrasion resistance, which is important if you're trying to pump water that may carry considerable quantities of sand and mud, does not absorb water (it shouldn't swell), and is advertised as an excellent choice for check valves. It's also cheap in quantities, which we hope to be buying, if one day this pump is well accepted. So there. As usual, I've provided a photo showing four check valves ready to be installed, and some of what's inside. You can click on the photo to enlarge it and get a better look at what I'm talking about.
The valves are made from one 1 1/4" coupler and two 1 1/4" x 3/4" flush reducers. A 7/8" delrin ball fits nicely in the throat of one of the reducers, and stops flow in that direction. When flow reverses, the ball would travel back to the other reducer, and stop flow in that direction too, except that it's stopped by what I'm going to call a spider, which is what those two complicated looking star shaped objects are in the front of the photo. One is made out of aluminum, which I didn't like too much. The other is made out of a piece of the PVC that both spiders are sitting on. The PVC came from a short section of PVC pipe that was slit down one side, and heated in a 300 degree oven for about five minutes. At that temperature, PVC is as pliable as putty, and can be flattened between two boards and held till it cools. After that, it can be cut, filed and drilled. The spiders look complicated to make, but they're not. You cut a square the right size, and drill a half inch hole in the center. Then you use a half round bastard file such as you see in the photo to file the half round recesses in the four sides of your PVC square. It's easy, and it only takes a few minutes. To the left you see a Delrin ball sitting on it's spider, inside a spacer cut from a piece of pipe to provide space in the valve for the ball to travel back and forth, and retain the spider. The second photo shows a cut away copy of the valve so readers can see how all this goes together inside. In quantities, the cost of this valve should be less than three dollars each. I'll let you know how they work as soon as I can get them installed in the pump.

Paraffin Molded Seals



Early on in this project, while we were studying treadle pumps in general, and leather seals in particular, we ran across several references to making seals by soaking leather in melted paraffin(candle wax) and pressing the still hot leather into a mold. In an earlier post I wrote about making molded seals using the pump pistons and a section of PVC pipe after soaking the leather in warm water. Since then we've made molds for making paraffin seals, which you can see in the first photo.

The base and and piston are made out of old countertop material. They could be made out of anything, even wood. The rings are short sections of the same four inch PVC pipe that's being used for the pump pistons. There's a large 45 degree bevel at the top of the rings to help ease the leather into the mold without tearing it up. Today, I finally got around to molding seals. Five inch diameter leather circles were soaked in melted paraffin, and placed in the molds. The bolt through the center of the mold was used to tighten down on the leather and force it into the mold. It worked like a charm. The only problem was that I made the rings a little too short. I improvised by taping the rings together with masking tape, and using one mold at a time. I'll make taller rings later. The second photo shows one of the seals out of the mold, and one of the seals still wrapped around the piston, waiting to be trimmed. I trimmed the edges of the seals with the razor knife to the left, using the top of the piston as a guide. I have no idea how well these will work. We've read that these seals will be self lubricating, and less prone to rot and swell from absorbing water. We'll try these seals the next time we take the pump out for a walk, along with the new check valves we made today. That's what the next post is about.

Pump Video as of June 23