ALL Power Labs: Tools for Power Hacking

Gasification Home:

Gasification Basics

How to Build a Gasifier

Technical Resources


overview - gallery - reactor options - manfacturing - news - GEKwiki - GEKforum - test reports - downloads - purchase

GEK Test Report #1: Air Preheating Configuration (Zero Tar)

all flanges and end plates all vessel tubes

The above pictures show a new air preheating scenario for the GEK nozzle and constriction (Imbert) type downdraft reactor. We ran this configuration at the Maker's Faire in San Mateo, CA, May 4-5 to drive a Lister 6/1 diesel genset in dual fuel mode. The result was a tar free gas over about 10 hours of operation. The fuel was walnut shells at 15% moisture content.

This air preheating system uses heat exchange tubes in the form of stainless steel natural gas lines, 48" in length, 5/8" ID, spiral wrapped around the reactor. When the reactor is bolted down into the gas cowling, these heat exchange tubes fill the circular volume where the product gas rises upwads towards the outlet. The heat exchange tubes are mostly perpendicular to the direction of gas flow, unlike the previous "J tube" scheme. SS natural gas lines are well optimized for repuposing as heat exchange tubes. They are thin and corregated (thus lots of surface area), easy to bend, non-corrosive and cheap.

Somewhat to our surprise, this aggressive air preheating system resulted in a gas that appeared completely tar free. This is likely the result of higher combustion and reduction zone temps and thus improved tar consumption and conversion. Condensate too was significantly decreased, owing to increased H2O conversion in the reduction zone from the higher temps.

Below you can see the clear gas hoses we use for the product gas piping. After one hour of running, there was absolutely no brown or black discoloration in the tube. I did not have a paper tar test filter rig hooked up at the Faire, but from experience I know this lack of deposition in the clear gas tube equates with a tar level below measurement for typical sample times with the paper filter. The only liquid visible is the typical water condensation, also completely clear.

(With the GEK I try to use clear gas tubes, filters, condensate vessels and hoppers wherever possible so as to be more transparent and honest about real world performance, and make the processes more visible for learning.)

The product gas "smoke" was similarly nearly transparent, with only a slight white coloration. Nothing like the usual cream colored smoke. After a full day of running at 1-2" water vac, there was a slight brown tinge in the tubing, but nothing like the usual black stains and soot.

The swirl burner similarly showed an unusually clean gas signature. After half of day of running, the burner had zero soot deposition. The inside of the burner still looked identical to the original raw steel.


In addition to preheating the air, the heat exchange tubes also cool the product gas. I was surprised to see gas exiting the top of the gas cowling at 150-210F. Yes that is correct, 100C or less. Compare with typical no air preheating or gas cooling configurations which have gas exit temps in the 200-400C range. Hot gas is much more difficult to process downstream than cool gas. Cyclones, filters, plumbing and pumps are all complicated by high temp gas.

This success of this "gas cooling via air preheating" suggests the potential of eliminating the radiator all together. In the next run I'm going to double the length of the air preheating lines, put fins on the cyclone, increase the size of the granular filter, and see if we can do without the radiator (or in other words, incorporate the radiator into the gasifier). Efficiency will also continue to increase the more we recycle this "waste" heat back into the gasifier.


The internal geometry for this run was the "textbook" dimensions for Sweedish inverted V hearth designs for a 2.5" reduction constriction. This is a very small constriction, intended for engines around 5-10hp. I did not do any special nozzle burn path tests or other diagnostic configurations, so i'm sure what I had could be improved upon. The fuel was walnut shells in 1/4 - 1/2 shell form. Moisture was measured at 15%.

In the drawing above you will also see an ejector venturi based tar recycling system. This is built into the GEK downdraft nozzle and constriction reactor, but we were not using it during the current test. The inlets in the pyrolysis zone were capped off so we could test the air preheating alone. A run in the near future will explore and document this tar recycling system.