wouldn’t a heat exchanger after the exhaust manifold be able to crack the tar from wood gas?
It would of course it would not work when the engine is cold.
But the problem with tar in engine is mainly when the engine getting cold after been running?
Alternatively, you can start and run on other fuel until the engine is at working temperature.
You dont need to do all that. Just build a gasifier that fully cracks it in process.
Use dry fuel and keep that gasifier flowing and you wont have to worry about tar on a good build.
Jim I can tell you are going to be a good builder. Like many of us coming into this we tried fixing the machines before we ever built one!!
Throw all the perceptions away, what is on the internet is old documentation. The more I learn the more I learn how misleading or wrong 80% of it is. Learn the Physics of gasification first, then learn the science. You can not have the science part with out correcting and overcoming the Physics first.
Once you have moisture content with in parameters of what the machine can process and then get it to flow linear, you can then tune the machine to drive it to full tar and water cracking. You wont even need a filter once to get this point.
To many are trying to devise way to combat tar vs correcting it at the source.
I had a idea of using a updraft gasifier on wood, as I think it would be less sensitive to the fuel that would be used and so reduce the preliminary work of prepare small wood cubes or making charcoal.
An updraft wont work you need a carbon bed to pass the gas through to reduce it to CO and Hydrogen.
The reduction is where you process the gas into a usable fuel. Your engine grade gas is not produced just from primary combustion, this gas at this state is what is called Pyrolysis gas. It is full of un cracked tars, stream and long hydrocarbon chains that must be broken down further to create producer gas.
Ok, I will get co2 instead of CO.
My best advice is to get the imbert spec sheet and build it. For small engine running on very dry wood chips 1x1x3/8" thick use the 2" restriction and drive your geometry to the rest of the specifications.
The Imbert gasifier is not broke, it works just fine. Can it be improved absolutely, but these improvements really have nothing to do with the Imbert reactor. Sort of like a car engine that is designed to run on gasoline. Put grease in the fuel tank and it wont run. Put water in the fuel tank it will run poorly if at all.
Again it boils back to the Physics, and generally this requires automation to over come them. At least for the smaller scale, small engines. The WK gasifier is much larger and fuel flows are much less an issue. But these small engines, you are pretty much going to need to use automation. If you find your self putting efforts into filtration you are taking a wrong approach, nothing can be done with tar after it is produced post reduction. It is carried in the gas as a gas and will stay in suspension until it reaches atmospheric pressure near ambient temperature. The only way to filter tar is via a micron level filter that would need to be regenerative. This is simply just not going to be practical for running engine and besides you can create tar free gas if you get the physics of the process under control and tuned parameters within the means of what the gasifier is capable of processing.
Sniff too much of this gas and you type like me!!
Jim,
Please listen to Matt R. He knows what he is talking about. Your heat exchanger is neither hot enough, nor does it have the white-hot carbon to complete the chemical reaction. You could use a similar design to pre-heat your primary input air. 
Jims idea is not dumb thugh. I was thinking about something similar a while back. Pirolisis gas is way richer thain woodgas.
Jim, we speak of tar cracking but that is not a actual correct term. What we mainly do is oxidise tars in a gasifier to CO2 and H20 then reduce those to CO and H2. Cant achive this with heat alone, at least not with the temps possible wifh exhaust.
Bottom line, hard to do if not impossible. Matt is right.
