I just received a shipment of tubes and heat-pipes for my next series of geyser pump /heat pipe experiments. I am going to attempt to geyser pump a 30-tube collector array. I will start with a standard Apricus 30-tube header acting as an evaporator for a heat-pipe. The condenser for the heat-pipe will be the exterior surface of a parallel array of geyser-pumping lifters. Last August, I was able to pump a 10-tube collector with a single lifter, but the pump rates were insufficient for a high efficiency system.
I’ve also purchased a lot of Raspberry Pi and Arduino boards and sensors to instrument the experiments. I will try to be more diligent in my posting of developments this summer.
I have decided to test just the effectiveness of the heat-pipe header’s ability to transfer heat to a condenser pipe inserted through it. I have built several prototypes that use geyser pumping to transfer the working fluid through the heat-pipe condenser, but the results were confusing, so I’m going to decouple the problem by eliminating the geyser pumping in lieu of an electric pump that will pump the tank water through the heat-pipe condenser and back to the tank.
The copper coil heat exchanger was a little too wide for any of my containers, so I built a soft tank from 10′ of mylar coated bubble wrap and a garbage bag.. It’s well insulated and leak tight. I’m going float Styrofoam balls on the surface for added insulation. I got this idea from Tom Gocze almost 30 years ago. He still sells them at the link below. So far I the tank hasn’t exceeded 140F, so I haven’t really tested the high temp limits.
This is the first attempt at using the header and vertical stand pipe as a heat-pipe with a 10-tube collector. It seems to be functioning well, but there is a significant ΔT across the steam pipe header. In full sunlight the dead end of the header runs at about 270F and the exhaust side runs steadily at 230F. The geyser pump exhaust runs at about 160F and is moderated by the tank temperature. The ΔT across the heat-exchanger is about 20F indicating a good pump rate from the geysering.
My next experiment will be to insert some kind of wicking agent into the header pipe to more effectively draw liquid to the dead end. My theory is that the dead end of the heat pipe is over heating because the liquid is evaporating before it reaches that end, thus the high ΔT across the header.
As I was starting to design the straight-through header I mentioned in the previous post, I was preparing a presentation about heat-pipes to an inventors group I meet with weekly. While playing with a naked heat-pipe and handling it while heating the evaporator end in a cup of hot water, I was amazed at how effective it was at transferring heat. I did some googling and found formulas for describing how effective it is, and I was even more amazed. I dawned on me that I could create a heat-pipe that extends upward from the header. I tried this with my little 4-tube show model. I just attached an elbow on one end and plugged the other end of the header. I attached a vertical 3/4″ x 12″ pipe onto the elbow and capped the end with an evacuation port. I added enough water to half fill one of the two header pipes, evacuated it from the top port and set it in the sun. The temperature of the top of the vertical pipe was only a few degrees below the temperature of the header pipe. I insulated it, but left the top 2 inches of the vertical pipe open to the air. The temperature of the naked pipe got up to 280F in partial sunlight.
I realized that I might be able to transfer the heat from the header to a vertical nucleating riser using this idea. I built one the next day and it worked great. The only problem was there was a large temperature difference between the two ends of the header. The capped end was getting much hotter because the liquid dropping back down into the header from the condenser was not flowing all the way to the capped end, so it was overheating. I have read that horizontal heat pipes requires some sort of wicking or capillary agent to move the liquid. One promising idea is to line the header with copper screen. Even with the large delta T the system was pumping heat to the heat-exchanger and storage tank.
I have only a few digital thermometers to determine the temperatures in the systems. I need better temperature data.
The vacuum tube geyser pump worked well for about a week, Then, one day, it locked up. The header pipe approached almost 350F, but the vapor condenser remained relatively cool. The system pressure stayed well below one atmosphere, approximately 12 In. Hg vacuum. My theory is that rather than nucleate boiling, the fluid in the header just evaporated, and the header emptied to its ends where a surface of water maintained an equilibrium evaporating just enough into the header to maintain slightly positive pressure. The rest of the system lost enough heat to maintain a vacuum in the vapor condenser. I was able to start the pumping again by drawing a little bit of vacuum or letting a little air in to the system, and then it would pump for the remainder of the day, but would not start pumping the next day.
The solution to this would be to modify the header so that it has active artificial nucleation sites, but the holy grail I am seeking is to use the headers without modification, but I plan to test the nucleation theory by building a straight through 3/4″ I.D. header with a 3/4″ O.D. nucleator.
To build a straight-through header, I will have to build wrap-around sockets for the heat-pipe condensers. I’ll build it with removable high-temperature insulation and install thermal sensors at several points on the header. I’ll start with a 4-socket header, and if it works, I’ll try to pump a 10-tube slave collector. If that doesn’t work, I’ll try building a 10-tube geyser pumping collector with a straight-through, nucleating header.
Solar domestic water heating is a non-starter. Even though it makes practical sense, it’s not sustainable in a market driven by artificial financial incentives. Systems that should be installed for less than $3,000 are inflated to $10,000 so that participants can maximize incentives. I played in this arena for almost 10 years before I grasped the irony of this vicious cycle.
Instead of pursuing the impossible U.S. SDHW market, I have decided to direct my attention to larger, higher-value systems. I think I have identified one that may optimally exploit the advantages of this system: space heating in northern tier states that have brutally cold ambient temperatures, high cost of electricity, no access to natural gas, and mostly clear daytime skies. My parents live in just such a situation. They burn about 13 cord of wood every winter to heat their home near Grand Marais, Minnesota. They live in an area that will never have access to natural gas, and because coal plants are being retired in Minnesota, the electrical rates are high.
I am designing a solar heating configuration that would use a small evacuated tube/heat pipe collector that is optimized for geyser pumping. In the development of the Copper Cricket we balanced the pumping function with the heat collection/loss function, but in this new system I intend to maximize the pump function in the pumper unit so that it can drive the heat transfer fluid through an array of standard Apricus evac, tube solar collectors. At this point, I envision about 100 tubes being pumped by 6. The tubes will be mounted nearly vertical to capture the maximum winter insolation, while being almost hidden from the summer sun. I may even build a short awning over them to shade them from falling sleet and summer sun. I plan to mount the tubes two feet above the roof so that they will be above most standing snow.
One of the homes is already heated with a hydronic in-floor system, the other has a forced air system with an easy-to-modify plenum that could house a hot water radiator.
I have two applications for this space heating solution within 100 yards of each other. My goal is to design the systems and install them in time for the 2016/2017 winter. Both of these homes are on the shore line of Lake Superior and both have southern exposure.
Since early October, I have been working on a test unit in my front yard that integrates geyser pumping with evacuated tubes. I had planned to write a series of posts describing the process and challenges in the chronological order, but instead I have decided to now post the current status and work backwards. My hope is that more of you will share ideas and collaborate with me on the development. I continue to believe that this is an important technology, one that should not die.
Apricus Tubes and header with geyser pump. Notice double Copper Cricket on my roof.
This is the second revision of this system. The first required a check valve to limit back flow. I added a vertical loop of pipe on the entrance side of the header to increase the mass against the acceleration of the steam. It worked.
Also, I added a small loop on the exhaust side of the header so the steam would have to push a slug of liquid up the lifter and not be able to by-pass the liquid.
This unit pumps almost perfectly. The only problem has been that every few days it locks up and doesn’t pump. The header reaches over 350F, but it must be completely empty of steam or liquid. The exhaust temperature reaches only ~200F or less. I imagine there is a surface of liquid in contact with the 350F void that probably has a very small, balanced amount of steam pressure that refuses the entry of any additional fluid. If I either add a little bit of vacuum or release some pressure the system jumps to life with lots of gurgling and internal bashing of liquid. It’s interesting also that the fusible plug on the right side of the collector doesn’t release. Even though the header is at 350F, the pressure in the system is only a few psi, and the heat doesn’t reach the fuse plug installed just a few inches above the header.
First Prototype Schematic
Vacuum Tube Geyser Pump Schematic
The Copper Cricket was, and still is, a simple, efficient and reliable solar water heating system. Most of those I visit these days are systems on homes in the process of being re-roofed. I help remove the collector and return to replace and recharge it. Once in a while a Cricket is out of sorts. The most I have had to do is flush the heat exchanger, replace the fuse plug and recharge it.
I had been considering integrating the Copper Cricket pump (geyser pump) with evacuated tube solar collectors. I envisioned several ways of doing this, but had not built one. In 2014 Raymond Lam of Silk Road Environmental contacted me because one of his clients had a non-functioning Copper Cricket, and he didn’t know what to do with it. I met him at the Cricket’s home in Portland and recharged the system. It had over heated while the home-owners were away on an extended summer vacation. Ray was excited about the system and spent the next several months trying to build a working model using the evacuated tubes and headers that he imports from China. After hitting a wall in development he hired me to troubleshoot his prototype. I visited him at his shop and determined pretty quickly that his prototype had no chance of working. He didn’t quite understand the functioning of the system. Instead of trying to get his to work, I showed him a sketch from my notebook. I said I thought this would work, but I hadn’t tried it yet.
We rummaged through the fittings and fixtures in his warehouse and then went to the plumbing store to buy the rest of what we needed. We started to assemble the prototype that afternoon. I have to say, I was amazed at how quickly and easy it was to assemble his 20 tube evacuated tube collector.
That night in my hotel I sketched the details of the prototype we would complete the next day. By 1pm we had completed the fabrication of the system and filled it with water. It pumped immediately, and I was able to track the rapid flow of heat through the uninsulated system. I suspected that the system would not hold a vacuum due to some of the odd fittings we had to use, so we let it run with the evacuation port open to the atmosphere. I didn’t try to measure the pump rate, but it was clearly fast enough to move fluid through the system, and very little steam was escaping the evacuation port which indicated that the steam was condensing internally.
Ray has decided to move forward with a proprietary revision of the system, so I agreed not to post photos of the one we built at his site, but I told him I was I would continue to move forward on my own and publish the geyser pump solar water heater as an open source technology.
Friday, I picked up 6 complete Crickets from Solar Assist in Eugene, Oregon. They were about to scrap them. I now have 4 heat exchangers, too. Also on Friday, someone in Eugene called and said he had a Copper Cricket in his garage and was wondering what he should ask for it to sell to his neighbor.
Homeowners are removing the collectors to re-roof and can’t justify the expense of re-installing them. It’s a strange statement about the change in values from 20 years ago. Many of the homeowners removing the systems are not the original owners. I can see their point. In Eugene, we still pay less than 6 cents per kWh. A new tank now costs about $400 and a typical re-install and fill costs $600-$800. At about $100 savings per year, that’s a 10-year payback. That’s about the payback of a new system when these were sold with tax credits and rebates in the ’80s.
Installers now urge homeowners to buy heat-pump water heaters. They have a much easier installation, and we’re told that they have at least a 10-year life-expectancy.