Building an internal combustion engine from hardware store parts

[MacGyver] [Lou Wozniak] is on a mission to build an internal combustion engine using only hardware store parts. What you see above is his third attempt at it. Depending on your hardware store this may have ventured outside of what they sell because [Lou] switched over to using gasoline. But the first two attempts were powered by a propane torch fuel canister.

Unfortunately it still isn’t running. But the demo below makes us think that he’s really close. Timing is always touchy and that seems to be what is causing the problems. He makes use of a lot of plumbing fixtures. At the right you can see the parts (including a peanut butter jar) which make his carburetor with a valve pointing straight up as the choke. The fuel and air mixture moves down through the pipe to the cylinder and valve assembly where it is ignited by the black grill igniter module. His custom cut plywood gear moves with the fly-wheel. It triggers his improvised spark plug by using a bit of wire to pull on the leaf switch.

We feel like he’s so close to getting this up and running. If you have any advice on where he might be going wrong [Lou] welcomes your input.

64 thoughts on “Building an internal combustion engine from hardware store parts

    1. He might be able to make an hour glass shape by squeezing the piping(maybe use something more malleable) in the middle to get a venturi effect to properly atomize the gas.

      Might be a good idea to also make a timing mark and use a time gun(assuming tools are allowed), while making the wire that triggers the spark more stable (a sheet of metal).

      Plus, as others point out, the fiywheel is wood and needs to be heavier but knowing the liter of the engine and compression would help (some metal standardize weights drilled in).

      The gas flying out the intake suggests that the check valves he is using is not giving him proper compression and that’s probably the main issue. Maybe make a more robust camshaft to mechanically trigger the valve, either a gear off the fly wheel or the up and down of the “camshaft”.

  1. Basically a derivative of a “Henry ford” engine.

    First, the flywheel is too light. Second, he is blowing gas back out of the “carburetor”. This means his intake valves is not working as is should. It looks like the is using non-spring check valves, they are slow closing.

  2. There are like 5 or 6 problems at once here – and that’s going to make this very hard to troubleshoot.

    “Blow By”: I’ve seen those 1 way valves at the hardware store, and they just might not do the trick for gas vapor or any sort of compression. Plus the second valve is held closed during the compression stroke with another valve for the exhaust. This may be too large a volume to fill and once it IS filled it’s going to blow by that valve. If they are allowed to clang around you run into a timing issue where the valve is pulled open by the intake stroke and is still flapping when you move into the compression stroke – thus forcing gas out of the “tank” as we saw.

    I suggest finding a way to spring load those valves and setup a small combustion test with just a pipe, igniter, two valves and some starter fluid.

    I’m also curious why he moved to a 4 cycle – that piston flywheel combo doesn’t look like it can continue to freewheel through a exhaust stroke… Just put fuel in and compress-ignite every time.

    Good luck!
    – Kris

    1. I agree here. Very likely the flywheel is too small. See steam engines! Also what is the compression ratio? What is the boreXstroke? Also, I think having the intake/exhaust gases mixing should be changed. This can cause a whole slew of problems, including danger! The turbocharging crowd tells us that there is an awful lot of gas expansion happeninrg after leaving the chamber. In your case I think much of you intake is actually exhaust… this is just from speculating tho.

      I dont think ud need ay help plumbing this:

      I know you like the jar, I do too. It would be very easy to prototype with a “float” carb like from an old techumseh small engine. It could be (re)-made with hardware store parts, or after that variable is ruled out, back to that lovely bubbler :)

      There were some very very old cars that had even simpler carbs. They weren’t reliable by today’s standards, but hell, white knuckled men dirt track raced them! Good luck my friend. My father was/is a vintage mechanic, and I cant wait to show him this! haHAH!

    2. I agree that the flywheel is way to small.
      I think the closest attainable goal is essentially a hit-and-miss engine. That’s the design I’d try to copy.

      And I’d move the spark-timing microswitch over next to the wooden gear itself and let a pawl on the wooden gear hit the microswitch lever. The timing can be set more precisely that way.

    1. Would a magneto be any easier / more reliable? Lots of small scooters / mopeds have used them over the years. Just a coil, a magnet, and something to spin the magnet round. You could just add the magnet to the flywheel, mount the coil nearby. If the flywheel’s angle on it’s axle is adjustable then that makes timing easier.

      You can get very strong magnets nowadays, should make it more powerful. And it’s a pretty easy system to diagnose, no switches needed, no electrical contacts at all.

  3. 1) The spark igniter is too erratic, timing wise, to generate a spark at the right time .

    2) the check valves should be right at the base of the cylinder, right now there is too much volume in the ‘head’ area relative to the piston, lowering the compression he gets.

  4. Awesome ! a gasoline powered pipe bomb !

    Another Darwin Award entry.

    I can almost hear the Looney Tunes theme in
    the background, followed by a charred Wile E. Coyote!

    At least wear safety goggles !

  5. air fuel in the correct ratio. Compression, and ignition. All that with the correct timing where IC troubleshooting begins
    Not that I would say that a 2 cycle engine would be easier, although it could be simpler. Respectfully fro reading Lou’s page he doesn’t understand how a 2 cycle works, but sorta does does. Lou;”Gasoline needs to be pulled in with an intake stroke, but I don’t have that in this 2-cycle engine”. True that the 2cycle doesn’t have separate intake stroke, that function is taken care of the compression and power strokes of the 2 cycle engine.

  6. I like this project if for no other reason than that the guy has the sense to use a transmission and knows how it works. People who do motor and engine powered projects where they compensate with more power or bigger props are choosing to design things the least intelligent way possible, pretty much ignoring centuries old science and encouraging staggering inefficiencies..

    Looks like timing is off, probably because loose mounting on the gear, and there is inadequate compression to produce enough torque..

  7. Same thing as before with this guy. Even primitive hit-and-miss engines had machined metal components to some degree of precision, not a bunch of wooden gears and bent up coat hangers. Go find a hardware store with some machine tools and the air/fuel ratio will be easy.

    1. Lots of machinists have built engines. Don’t get me wrong, doing so is impressive. But, If he can build one using just hardware store supplies and without all those specialized skills and tools then he will have accomplished something unique and new. He could even be leading the way for other non-machinists to further his or come up with their own designs. I think you have missed the point somewhere.

  8. Hundred if not thousands of engines are operation with intake valve what is essentially a check valve & the valve timing with that is automatic. One check valve on the exhaust with spring to assist closing item as another mentioned may work there if Lou is going to stick with a 4 stroke engine. The hardware store in my town sells ignition coils My opinion that is they way to handle the ignition. Using light rather than a mechanical switch, a way to manually adjust the advance should be designed in just in case if the engine wants to run. IMO Lou should return to propane gas for a fuel to run the engine although the engine probably will have to be started with a liquid fuel. Often gas fueled oil field engines are adjusted with this simple valve.

    When looking inside it while operating the opening has an infinitely variable diamond shape through it’s operating range. The only thing I have seen that’s similar is inside an old school GM cruise control unit.
    Two stroke or four stoke I believe Lou is going to figure out a way to create piston rings. With piston rings I believe he would have better luck with engineering a two stroke engine. No matter what he does a better flywheel would help. Hardware stores sell engine powered equipment, IMO it would be fair to grab a flywheel of a junk engine to use.

    Those ductile iron fittings most likely will fail under the pressure spikes they will see. Hopefully they will simplely split, but even at that they may jet hot gas at anything in range Expect to see those check caps to try to see if they can achieve escape velocity.
    Lou get yourself a quality oil squirt can like this fill it was gas or starting fluid liquid if you dare. Use it to prime to engine and as manually controlled fuel injector to get the engine to speed where it will run on the propane. Man I don’t miss having to start a 503 Fairbanks Morse single lunger. That’s 503 CID manual start Good luck, have, stay safe LOU.

  9. Assuming the piston even holds a seal (there’s no rings, right?), compression can’t be more than 2:1. If it doesn’t hold a seal, AND he atomizes the gasoline, he’ll be picking copper out of his face in the next episode.

    With the addition of the gears, he essentially has no flywheel anymore.

    That being said, given that he’s gone to a four-stroke for some reason, let’s make it a V-8!

  10. Valves are so far away that compression volume is huge. I doubt that sufficient amount of gas/air mixture gets in this way per cycle. Moving valves right next to the cylinder head might help, about where the spark plug is now.

    Using some off-the-shelf carburetor initially might help to root out worst problems with the design.

    While it is obvious that this isn’t going to be good efficiency engine, it still is quite interesting idea to use non-engine off the shelf parts for building one. (Even if actual engine parts are available off-the-shelf, just a different store.)

  11. I race 2-stroke bikes, I have spent more hours with the bike in pieces in the pits than I have riding it. Here’s the lessons I learned

    1) they are ridiculously simple but also hell of a temperamental without all the nice custom designed components to make them stable. So I second the “2-stroke will be simpler but more difficult” comments.

    2)There are three things that will prevent a motor from working: compression, spark and fuel mix.

    2a) Compression. You have to have it. If your spark source is loose or your valves not working you won’t fire. I would suggest trying out reed valves, they should be easy to make and are designed for this. Also you need some form of ring on the piston. And your exhaust is critical. Especially for two strokes. Two stroke motors have a tuned cavity in the exhaust which basically is tuned to produce resonance that helps draw the exhaust fumes out giving better combustion and reducing the load on the motor. For instance on a race bike, denting the cavity will throw the engine performance.

    2b) Spark. I have tried using grill lighters to ignite propane in potato guns, they are unreliable at best, so for timing critical applications you’re gonna have a bad time. They do not produce a hot enough spark (i.e. not enough electrical energy) and you will really battle with petrol. You can use an ignition coil (i.e. a big fat inductor) and plug from a bike/moped but you have to drive it with a high frequency signal. And you have to find a hardware store that sells that. The spark gap is also a significant factor.

    2c) Fuel Mix. my race bike will go from running to not running with a change in jet diameter of about 2-3mm. So the motor is very sensitive. You have to make sure you’ve done at least a rough calc on the fuel air ratio and the volumes and know what is needed for your fuel. For instance if the volume of air between your carb and the piston inlet is greater than the piston, you’re never gonna get fuel into the piston, especially with all the blowback.

    I would also suggest that you need to relook your bore/stroke combination. While longer stroke makes it slower and easier to operate I think you may find in this case it actually exceeds the operational limits. I’m pretty sure if you look up some basic combustion engine theory you’ll be able to find some figures on what will work.

    I’m also an EE student and have designed and built many things from scratch and my experience is you cannot troubleshoot like this. You need to unit test, so start with something like can my spark actually ignite petrol and work out a way to test that in isolation, test systems one at a time and slowly add working things together and retest. There are way too many factors here for you to tell what’s working and what isn’t.

    1. Best answer by far. To expand on the “two stroke is simpler but more difficult”; there’s fewer moving parts and you get more power out of it, but EVERYTHING is tied together in a critical way. People spend ridiculous amounts of time & money making sure the exhaust is shaped properly. There’s software out there to help with building tuned exhausts.

      Oh, and watch out for the shrapnel from your glass fuel tank.

    2. I agree with the last paragraph – you need to unit test – treat the system as a state machine and test each state statically before moving on to a dynamic test.

      I am surprised you have moved to a liquid fuel – I think that makes your job harder due to the need for some sort of air/fuel combiner.

  12. This is a bad idea….

    Safety! There is a strong chance that the copper cylinder will explode in your face. Provide a way to contain the shrapnel. Protect your remaining good eye!

    Safety! You don’t want the fuel going into undesired places (like your clothes). Consider the possibility that a backfire could rupture your “carb”. Have a fire extinguisher nearby. Make sure someone (not you) can put out the fire when your clothes catch fire. Propane may be a safer fuel in many respects. Make sure you have good medical insurance, and some income while you recover. Check on your fire insurance too.

    You should reconsider this project…. Things can go very badly, extremely quickly.

    Placing the intake and exhaust valves so far from the cylinder suggests that the entire gas volume within the cylinder oscillates between the cylinder and the plumbing – a fresh air/fuel charge never really gets into the cylinder. Mount the valves to opposite sides of the cylinder (i.e. remove the T-fittings). This will also increase the compression ratio.

    Ignition timing is critical. A better way to generate a spark on demand is required. The grill igniter takes time to generate a spark – something faster is needed. The floppy wire link driving the igniter switch needs to be rigid – replace with a rod. Better yet, mount a cam on the gear wheel and place the switch in contact with the cam.

    1. Would a lower compression ratio and using a less energetic fuel make it safer? A set of goggles I think would be vital, and maybe some of that plastic the Mythbusters use, made into a transparent box to put round it, with a few holes where adjusting screwdrivers might need poking in.

      Yep there’s a reason they don’t make engines out of copper, but is it something you can get away with if you’re only tinkering?

  13. Needs compression and stiffer valve springs…. especially the intake. Might want to retard the timing to fire after top-dead center, so you can be a bit sloppy on the timing.

    1. In 2-stroke engines the hollow casing around the piston sort-of acts like a piston itself. Both surfaces, top and bottom, pushing and pulling, of the piston have essential roles to keep it working. It’s all horribly complicated and flukey, a system that only works while dynamic. Thinking of it as separate states like a 4-stroke doesn’t work. The exhaust / intake / combustion aren’t entirely separate in a 2-stroke.

      “Complicated” and “horribly” are the 2 important words there. I think you may be right, but working out exactly why would be a headache.

  14. I am pleasantly surprised at the comments for this project. I usually get the feeling that HAD readers are all software geeks, but this one has brought out quite a lot of people knowledgeable in IC engines. And very few fussing about the danger or sneering at the build. Yay!

  15. A combustion engine needs really high compression, in a car the average is 175 psi at the top of the compression stroke. The “Carburetor” also needs to changed heavily. To get it running at first I would suggest grabbing a carburetor from a little dirt bike or motorcycle. Your valve overlap also has to be minimal, I can see form the video that your getting a lot of backwards air flow through your “carburetor”, way to much air, also pretty dangerous considering your sitting right there. Now for the cylinder/sleeve. Very thin for harnessing an explosion. If you get an early explosion when this is hot( its called a knock in a car and is so common that your car monitors this and knows how to act when it happens), and the piston is just before or on Top dead center(depending on rpm), a sleeve like that could easily exploded. I am a part time mechanic and motor enthusiast, and would love to help you with this.

  16. Thank you for all your comments. I tallied all 117 suggestions, so far, from hackaday, video comments, and emails. Here is what you said and my responses.

    15% Check valves are too slow to close. Evidenced by the gasoline splashing back and fire coming out of the intake manifold. Try cam controlled valves –Lou: I plan to replace the exhaust valves with one cam-controlled gate valve.

    15% The flywheel should be more massive. Maybe try a wheel and tire –Lou: I am probably being stubborn, but I have seen a good fire send the flywheel around through all the cycles. I think it will be fine. I plan to focus on getting fire every time, and then look at the flywheel.

    15% The compression is too low. The valves should be moved closer to the cylinder to decrease space and increase compression. –Lou: I wrapped foil around the piston and got a nice increase in compression. I also plan to move the valves closer to the head, as suggested, to increase the compression ratio.

    13% The spark timing is off. Get rid of the string and micro-switch set up. You need something a lot more stable. Try timing gun. –Lou: See next comment. I think the grill igniter is the main problem. If not, I will switch to a cam driven spark.

    12% Grill igniters do not have reliable timing. When you push the button there’s a short, unpredictable delay until the first spark. –Lou: I totally agree. I just went out and bought a coil, points, and condenser, at an AUTO PARTS store, but don’t tell anyone.

    12% You aren’t getting the right air/fuel ratio Your fuel isn’t atomizing properly with the bubble system, so there is an inconsistent fuel/air mixture with each cycle. –Lou: I was totally ready to give up on my bubble carb, but then someone posted this:

    5% This looks dangerous. –Lou: We are not having fun until someone gets hurt  I will wear goggles, at least.

    3% The stroke is way too long for the air to properly compress and combust. –Lou: Maybe. I will work on other stuff first and then look at this.

    3% The spark may be too weak. –Lou: See comment above about replacing grill igniter.

    2% Build a better starter motor. –Lou: Maybe. I hope I don’t have to.

    1% a choke cuts the amount of air through the carburetor to enrich the mixture — your ‘choke’ just lets more air into the cylinder and less fuel, thus leaning it out. –Lou: my “choke” allows me to change the fuel/air mixture. I think that is all I need.

    1% This can’t be done without precisely machined parts. –Lou: You are right, so far.

    1% Hundred if not thousands of engines are operation with intake valve what is essentially a check valve & the valve timing with that is automatic. –Lou: I plan to stick with a check valve on the intake side, for now, at least.

    1% Just buy an engine at the hardware store! –Lou: I wish had thought of that, a year ago, when I started this project!!

    1. Re the last item: Please do not give up until you have convinced yourself 100% it can’t be done. IMHO physics is on your side, theoretically it should be possible to create a very low efficiency ICE using this method. I agree with a previous poster that you should be doing more unit testing – that is making sure that each individual state is working before expecting it to work as a whole – you are probably doing just that to a degree but not documenting it – that may be useful for others to comment. Good luck, I think this is a very worthwhile project – but yes be careful.

    2. All I can say is that there is very little difference between an internal combustion engine and a compressor. If you turn over an engine, air is forced through the components in a very specific sequence. If you’re building an engine and simply turning it over does not cause air to _only_ enter the intake and _only_ exit the exhaust port, you have a major design problem.

      As for the comments that it’s unsafe, I wouldn’t get too worried. Copper is so ductile that it will most likely split if the pressure is too high. Do wear some safety gear though – it only takes one unexpected outcome to ruin the fun.

  17. i think its the carburator, if you dont stack the wood right, you cant get a fire going.
    if the air fuel mixture can be finely adjusted, and the design of the carb is appropriate it should be fine
    a narrow bottle with the air coming in from a really thin spray nozzle (adjustable. they make them from brass) at the bottom could generate finer bubbles and more of them. a taller bottle would allow the growth of these bubbles and make a larger body of air fuel mixture.i believe that if you can make it rich by default it should easier to reduce the volume than increase.

  18. You’re also going to have to find a way to cool the cylinder, if you plan on running this thing for any length of time. Depending on what method you use, you need to design your carburetor setup to work with it. Air cooled engines use a slightly richer mixture than a water cooled engine,

    1. Once I get it going, I was thinking of wrapping 1/4 copper tubing around the cylinder, and running water through it. I could also have the cylinder splash into a pan of water. I wish I had that problem :-)

  19. You might need a larger diameter flywheel. Also I see that you have a slight balancing issue because your flywheel has no counterweight to balance out the weight of the piston.

  20. My 2 cents:
    a) more flywheel
    b) that ignighter has a timing delay that will screw it up, you need something that fires much faster with the switch. (and once, can you get it to hit a mechanical bbq lighter?)
    c) bring the ‘spark plug’ much closer to the “piston head”
    d) watch your compression ratio, remember that all the pipework from the head of the piston is part of the ‘dead space’

    “Engines need 3 things to run,
    fuel, compression, spark” (but this is more of a repair guide)
    To keep running they need oil and cooling…..

  21. part of your problem is the spark. and the flywheel is too light. add balanced weight to the flywheel. Also, you need to find a way fire that thing more acurately. Keeping with hardware store part list, I would mount an egg shape the cam lobe that pushes a mechanical grille lighter as it turns. it will take a little work to time it properly but once set, it will provide a pretty accurate spark time and time again.

  22. I don´t know if they have allready said that, but why don´t you put the spark sensor fixed next to the gear with a bar or knot that triggers the spark each turn? I don´t know how you will calculate the position, but the spark timing depends on the inercy of the engine (or flying well), at first sight I would recomend to make it PMS or past that (so it doesn´t counter explode). I would also recomend to put a biger flywheel cause of the 4 strokes (you need to keep the system spining 2 turns) but beware of counter explotions…
    Also, try putting a flapper on the air entrance so the carburator doesn´t spit fuel (or check the valves, I think you probably need better ones or advance the exaust valve so you loose compresion a bit of compresion but makin the air always buble the carb (if the is fuel on the carb entrance, it will be swalowed with no air to bubble).
    Also, without lubrication I don´t think you will get compresion to much… But if you make it 2 strokes remember to put the ingnition to make each turn (not each 2 turns like 4 strokes) so put the sensor on the first wheel or 2 sensors on the 2nd wheel (if you keep the ratio 2/1)
    Hope it will help you and good luck with that!
    And if you want to explain a bit more of the compresion chamber would be great.

  23. I would say heavier flywheel and possibly use a “reed valve” out of a putty knife and some sort of drilled out flat surface. Also decreasing the volume of the “head” by reducing the diameter of the intake and exhaust would increase your compression ratio a bit.

  24. in a 4 stroke engine the valves need to be timed, it is impossible for check valves to work. You could make a 2 stroke by sealing the bottom of the connecting rod to cylinder (some kind of o-ring compression fitting) and you would need to have a “4 way T” looks like an X, below the travel of the piston and another one and the bottom of the piston travel so that both sides are open to let exhaust out one side and fresh air fuel in the other side. and only one check valve after the carb

  25. The stroke should be no more than twice the bore and try using model glow plugs and nitromethane(model fuel ),it would eliminate your timing issues.Lube the cylinder to help seal better to achieve enough compression.

  26. to much pipe work to get enough compression.. inlet not sealing hence fuel coming out of inlet.. try a spring behind the check valve as in fords 1893 engine.. you could also try some heavy wire wool in the intake to atomize the fuel.. the gas cooker spark is a good idea but I don’t think its duration is long enough for petrol

  27. I think the crankshaft/cylinder part is way too heavy. I heard some ignition in your video but it seems there’s not enough power to even lift it. Is it possible to separate those two parts ?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s