Corvair Aircraft Engine Conversion

[Stuart_Snow]

Here are some photos of the Corvair aircraft engine conversion for my Pietenpol Air Camper project. The Air Camper was designed by Bernard Pietenpol in 1929 and used a model A ford engine originally. Later Pietenpol was the first to use a Corvair engine in his design. My engine is based on the conversion design and parts developed by William Wynne at flycorvair.com and Sport Performance Aviation flywithspa.com.  It is a bored out .030 2.7 Liter (164 C.I. originally) 100 HP engine. 3.0 liter versions are also popular using other piston/cylinder combinations and a 3.3 stoker option is available using SPA's billet crankshaft.

My engine uses original cylinders with forged pistons. A 5th bearing  and crankshaft for prop loads designed by SPA. A modified oil assembly with high volume oil pump and oil filter adapter. Reworked and modified heads by SPA include welded on intake tubes, deep hardened seats and stainless steel valves. All the original corvair fan and blower equipment goes away and traditional aircraft cooling ducts get used. The ignition is a modified distributor which has points and a crane electronic unit and two coils. The advance curve is modified to come in early and total advance is 32 degrees.

These modifications have been tested and proven by Mr Wynne over the past 30 years. The corvair has proven to be a very robust and reliable flight engine and has a reputation for smooth running. I call it my Tonawonda Tornado.

[Moses Ludel]

This is so exciting, Scott...The robust nature of the Corvair flat six air-cooled engine is immediately evident to anyone familiar with Posche/VW engines.  SPA's exotic aircraft-upgrade components are impressive, especially the crankshaft!  Eliminating the Corvair OEM cooling and fan system is a major gain, the stock engines were notorious for tossing their circuitously routed drive belts.

The ignition for aircraft use obviously needs safeguarding!  So, a modified stock distributor runs both a breaker point component and the Crane electronic conversion?  Two coils, each ignition circuit operating independently to assure reliable spark?  In the event that one circuit fails (either the breaker point or the electronic pickup), the other primary ignition circuit continues to fire its own coil?

There must be a reasonable following for these Corvair aircraft engine conversions if SPA makes these precision components.  Do you have any data on how many SPA modified engines exist? 

The Corvair engine cores should still be available, G.M. built over 1.7M units between 1960 and 1969!  Which year cores are the "best"?  Your '64 would be the larger displacement 164 cubic inch version (rather than the earlier 1960-1963 145 CI engine)...Corvair was at its prime by the later era, the redesigned '65-up chassis handled well, and the car was very quick when in 1965-66 the turbocharged Corsa models boasted 180 horsepower at 4,000 rpm from the "pancake" engine!

Moses 

[Stuart_Snow]

I don't have the exact numbers of currently flying engines but I can say the "Corvair Movement" has become quite popular in the experimental aviation world. Perhaps 500? I'll have to do some research on that. The engines best suited for conversion are the 64-69 164 CI ones. They use a GM 8409 crank which had longer throw to increase displacement from the early 145CI. The case has small indentations inside for clearance. Preferred heads are the 95 and 110 HP non-smog heads which have a quench area. this helps create the turbulence needed in the combustion chamber to ward off detonation. Smog heads do not have this quench area and are not used.

The ignition is not dual as you would normally expect on an aircraft engine running two plugs and dual mags. It is dual in the sense that there is a primary crane and secondary points unit in the distributor which can be switched by the pilot. The two coils use a coil switcher if one fails. This type of setup has proven to not be a problem for reliability.  Since this type of setup uses extremely high voltage racing coils and plug gaps are at .035 this produces a very efficient ignition spark given the relatively small bore size of the corvair. Plug fouling is unheard of with he corvair but does happen with magneto aircraft engines using narrow plug gaps. Most aircraft engines have very large bore size which requires the two firing plugs for efficient combustion.

Successful corvair pilots are using traditional aircraft carburetors like the Marvel Schebler or Stromberg that provide a rich mixture at full power.  Detonation is always the enemy that can creep up if things are not set up right which is the reason for the forged pistons. I intend to run 100 Octane in mine since I have flat topped pistons and am using the 110 HP heads.  Other piston/head combinations provide more margin for detonation using lower 93 octane fuels.  The corvair has the advantage of 6 cylinders, is smooth running and can still fly well on 5 cylinders.

The aircraft is mostly spruce and plywood with fabric covering but uses a lot of chrome-moly fabricated parts and fittings. I'm enjoying learning how to weld the thin walled chrome-moly tubing with my small victor torch. Your welding tutorials using low pressure and small tips are helpful as heat management is critical.

I certainly did a lot of research before I was convinced that using the corvair is a reliable option. I have found that successful pilots using corvairs are the ones who follow Mr Wynne's  formulas  to the T since he has spent his whole life doing the testing and development and is willing to teach others.

It sure was rewarding to take a grimy junkyard hulk and turn it into a running flight engine.  I'm thankful for people like yourself and Mr Wynne who are willing to teach and pass on knowledge.

 

[Moses Ludel]

Scott...Fascinating stuff...See my comments below:

13 hours ago, Stuart_Snow said:

I don't have the exact numbers of currently flying engines but I can say the "Corvair Movement" has become quite popular in the experimental aviation world. Perhaps 500? I'll have to do some research on that. The engines best suited for conversion are the 64-69 164 CI ones. They use a GM 8409 crank which had longer throw to increase displacement from the early 145CI. The case has small indentations inside for clearance. Preferred heads are the 95 and 110 HP non-smog heads which have a quench area. this helps create the turbulence needed in the combustion chamber to ward off detonation. Smog heads do not have this quench area and are not used.

This makes sense, detonation is harmful in any case.  These engines should run extremely cool at altitude;  however, detonation is more a matter of compression ratio, fuel type and combustion chamber/port design.  I understand the need for the right cylinder heads:  Jeep 4.2L inline six engines have detonation issues that got progressively worse in the 'eighties.  The 4.0L (later MPI/EFI type) cylinder head conversion has helped alleviate combustion chamber issues on the late smog era 4.2L engines.  With the stock 4.2L head, the Mopar EFI conversion recommends 92-octane or higher fuel.  The stock 4.2L head (carbureted era) is prone to detonation.  Rigid A/F ratios with EFI actually make the '81-'90 engine more prone to detonation than with the OEM carburetor system.

The ignition is not dual as you would normally expect on an aircraft engine running two plugs and dual mags. It is dual in the sense that there is a primary crane and secondary points unit in the distributor which can be switched by the pilot. The two coils use a coil switcher if one fails. This type of setup has proven to not be a problem for reliability.  Since this type of setup uses extremely high voltage racing coils and plug gaps are at .035 this produces a very efficient ignition spark given the relatively small bore size of the corvair. Plug fouling is unheard of with he corvair but does happen with magneto aircraft engines using narrow plug gaps. Most aircraft engines have very large bore size which requires the two firing plugs for efficient combustion.

 This is what I surmised:  one spark plug per cylinder, the distributor has the breaker point and Crane electronic primary circuits operating separately with their own coils.  Sensible way to safeguard the ignition in the event that the electronic system (or breaker system) might fail in flight.  Both circuits are not employed at the same time, you're switching between the two in the event of a failure or for periodic testing, and the same spark plug gets fired.   I see your point about a magneto, which would also have other inherent challenges.  (Did any automotive aftermarket manufacturer offer a mag for the Corvair engine?)  The modified stock ignition makes good sense.  Crane is the former Allison product, much better and more reliable than Pertronix.

Successful corvair pilots are using traditional aircraft carburetors like the Marvel Schebler or Stromberg that provide a rich mixture at full power.  Detonation is always the enemy that can creep up if things are not set up right which is the reason for the forged pistons. I intend to run 100 Octane in mine since I have flat topped pistons and am using the 110 HP heads.  Other piston/head combinations provide more margin for detonation using lower 93 octane fuels.  The corvair has the advantage of 6 cylinders, is smooth running and can still fly well on 5 cylinders.

Are these "vintage" carburetors?  Is jetting an issue, or do Corvair flight buffs have this down to reasonable blueprint specifications/tuning for the Marvel Schebler or Stromberg carburetors?  Are these carburetors readily available?

The aircraft is mostly spruce and plywood with fabric covering but uses a lot of chrome-moly fabricated parts and fittings. I'm enjoying learning how to weld the thin walled chrome-moly tubing with my small victor torch. Your welding tutorials using low pressure and small tips are helpful as heat management is critical.

So, the chromoly tubing is gas welded, a proven and time-honored method.  Where do you wind up metallurgically?  Is the tubing 4130 annealed (not heat treated) to start?  What are the requirements around annealed versus heat treated materials?  For off-road racing roll cages, rules have gone back and forth around annealing and heat treating.  Annealing or normalizing occurs locally in weld areas.  Strength and ductility are important.  What are your guidelines for tube welding, annealing, etc.

I certainly did a lot of research before I was convinced that using the corvair is a reliable option. I have found that successful pilots using corvairs are the ones who follow Mr Wynne's  formulas  to the T since he has spent his whole life doing the testing and development and is willing to teach others.

No room for error when flying!  Structurally, the Corvair engine looks more robust than European air-cooled engines...In that regard, Subaru engines enjoy a good reputation, is there any interest in using Subaru water-cooled engines?  They would be plentiful.  Would water cooling be problematic for a small aircraft?

It sure was rewarding to take a grimy junkyard hulk and turn it into a running flight engine.  I'm thankful for people like yourself and Mr Wynne who are willing to teach and pass on knowledge.

I'm pleased to do so.  In my case and likely Mr. Wynne's, researching and finding reliable answers is a steadfast goal.  One of the issues with the "information era" is too much lay level, anecdotal information without professional experience or an academic basis.  Mechanics, welding and metallurgy follow scientific and engineering principles. 

When we're working with vintage and restored parts that have blueprint requirements, parts are often unavailable and must be restored.  A vivid personal experience with metallurgy and misinformation involved a Ford T89N cluster gear with chipped teeth.  The material was case hardened 8620.  Teeth were badly chipped, there were no replacement/NOS parts available, and by design there was no such thing as "good used". 

At the local welding supply, I threw out the question whether anyone had experience with building up gear teeth, and a "professional" weldor at the counter who likely had certifications at pipeline or other professional welding quipped up, "Well, I'd just use a MIG and ER70S-2 wire, build up the teeth and shape them!"

 There were several doomed problems with his approach:  1) the case hardening would go away in the welding process, Rockwell Rc would drop to the level that the teeth and adjacent metal would have no stamina, 2) the HAZ (heat affected zone) around the welds would be annealed (low hardness) and likely allow the formed tooth to break off, 3) the ER70S-2 is a mild steel common structural filler that is not a good metallurgical match for 8620.  8620 is a low carbon nickel chromium molybdenum alloy steel, and any attempt to re-heat treat the 8620 piece with ER70S-2 repair filler would be futile.  The 8620 requires a carburizing case hardening process.  ER70S-2 filler is up to 85,000 tensile as welded and does not benefit from heat treating.

This began research on my part that led to my close work with Weld Mold Company's filler materials and Nevada Heat Treating.  Ultimately, the gear repair involved first normalizing the gear to eliminate case hardening.  After bead blasting and thorough washing/flushing, I TIG/GTAW welded the broken teeth with a 100% 8620 compatible Weld Mold filler that matched the metallurgical properties of the 8620 base metal.  I shaped the teeth in this normalized (softer) state before returning to Nevada Heat Treating for case hardening the finished gear repair.  The entire piece behaved like 8620 during the heat treating process, which resulted in an "as new" gear that matched the original case hardened piece.  Heat treating matched the Rc hardness and case hardening depth of the original gear.  You can see the results at:  http://www.4wdmechanix.com/Gear-and-Transmission-Case-Restoration?r=1  I partnered and earnestly sought information from industry professionals and a heat treating metallurgical engineer before tackling this job!

These pictures show the damage to the reverse gear teeth on a T89N cluster gear.  This transmission has an exceptionally poor reverse sliding gear arrangement that leads to gear damage like this if the vehicle is not at a complete standstill during reverse gear engagement.  The replacement cluster gear is obsolete, all NOS gears sold out long ago, and there was no option but to repair the gear as seen in the lower photo:

  

I have a current repair project that involves heat treated parts and dealing with the risk of annealing.  I will let members know when the project's how-to video is available!

Moses

 

 

[Stuart_Snow]

9 hours ago, Moses Ludel said:

Jeep 4.2L inline six engines have detonation issues that got progressively worse in the 'eighties.  The 4.0L (later MPI/EFI type) cylinder head conversion has helped alleviate combustion chamber issues on the late smog era 4.2L engines.  With the stock 4.2L head, the Mopar EFI conversion recommends 92-octane or higher fuel.  The stock 4.2L head (carbureted era) is prone to detonation.  Rigid A/F ratios with EFI actually make the '81-'90 engine more prone to detonation than with the OEM carburetor system.

I'll keep an eye out for this when I start running my rebuilt 4.2 by making sure my EGR is working properly for one thing and use good cooling and proper fuel and carb setup.

9 hours ago, Moses Ludel said:

(Did any automotive aftermarket manufacturer offer a mag for the Corvair engine?) 

No magnetos that I'm aware of for the corvair but the Model A conversion for the Pietenpol uses a novel system that runs off the crankshaft front (now rear). Model A' conversions are still popular back East with Pietenpol purists. Lots of low end torque and a lovely sound. I needed a little more horsepower for my warm western climate and altitudes which the corvair delivers.

The biggest hang-up with using the distributor and primary/secondary ignition has not been reliability but getting builders to understand that they need to set their timing with a light for the full 32 degrees before flight. This is done in steps with the engine not running for safety as the adjustments get made then started and rechecked. Initial run-in on the test stand is normally done with just the initial 8 degrees of static timing at about 1/3 power.  Each distributor is custom curved by Mr Wynne on his Allen machine to have the full advance come in much earlier than was the case for the car version.  One of the nice advantages of this type of ignition is starts are easy especially hot starts compared to some fussy aircraft engines.  

9 hours ago, Moses Ludel said:

Are these "vintage" carburetors?  Is jetting an issue, or do Corvair flight buffs have this down to reasonable blueprint specifications/tuning for the Marvel Schebler or Stromberg carburetors?  Are these carburetors readily available?

These could be considered vintage carbs as they were found on most aircraft right before and after WWII and were used for years on 200 C.I.  and smaller sized motors. They do get re-jetted to match the Corvair. Their is at least one certified aviation carb shop that specializes in rebuilding and re-jetting these. They can Also be bought directly from SPA. There seem to be a good supply since they were used for so many years. I will be using a Marvel Schebler MA3-SPA. This will require carb heat and a mixture control. 

9 hours ago, Moses Ludel said:

So, the chromoly tubing is gas welded, a proven and time-honored method.  Where do you wind up metallurgically?  Is the tubing 4130 annealed (not heat treated) to start?  What are the requirements around annealed versus heat treated materials?  For off-road racing roll cages, rules have gone back and forth around annealing and heat treating.  Annealing or normalizing occurs locally in weld areas.  Strength and ductility are important.  What are your guidelines for tube welding, annealing, etc.

The tubing comes in condition N (normalized). I have been working with very thin material so far .032-.035 on the tubing and about .090 on some of the other stock. Later I will get into heavier thickness on the landing gear. A normal weld starts with slowly bringing the piece up to temperature. Setup can be important so as not to sink off needed heat. I'm using RG-45 and RG-60 filler rods.  Weld process is normal working the puddle and watching for oxidation. I normally run a very slightly carburated flame. The critical part is to slowly work the flame over the finished weld and not allow any cooling breezes to shock the metal causing cracks. Complex cluster welds can build up stresses which can be relived by bringing the weld back up to a dull cherry glow and them letting it slowly cool.  No other heat treating is required other than relieving stresses.  Over the years a lot of amateurs have built very strong airframes as the chrome-moly materiel seems to be very forgiving if it is allowed to cool slowly. I am amazed at how strong and light the tubing and welds are. I think the process for race chassis must be more complex than this. Strength and ductility are the goals we want in aircraft also. The repair you did on the Ford gear is amazing. The Weld-mold product looks interesting and I'll have to learn more about it.

Subaru conversions have been around a long time in experimental aviation but haven't seen the kind of success or following that the corvair has. I'm not sure why since there are other engines that are water cooled as well like the 4 stroke versions of the Rotax which are very popular now. The Honda Fit engine has seen some success as well if you like an engine that cruises at 5000 RPM which its very happy to do with no problem. My corvair will operate in the 2800 to 3300 range.

[Moses Ludel]

Scott...The Corvair engine has a lot of bottom end torque and delivers power without the high-revving needs of other engine types.  This kind of power is similar to my favorite war bird engine, the supercharged V-12 Rolls-Royce Merlin that found its way into the legendary P51 Mustang.  We live near Reno, and I've spent time in the pit area at the Reno Air Races, poring over P51 engines and photographing their artful attributes.  The prop size and torque are so obvious that my wife will attest to my running out our shop door at Yerington, Nevada years back because I heard a P51 approaching Mason Valley from 5-10 miles away.  The exhaust and prop note is highly distinctive, easy to identify...

Beginning with normalized tubing makes sense, you can work and weld it without risk of cracking.  You get to be your own metallurgist with the stress relieving process.  Nice!  A book I like that includes a small section on oxyacetylene aircraft frame welding is Richard Finch's Welding Handbook.  I have the 1997 edition, there is a later (2007) edition available.  Finch is an EAA Counselor and devotes a small section to aircraft 4130 tube welding with oxyacetylene method.  Perhaps you have a copy of his book?

As for the Model 'A' engine, which may be a purists' delight, it's a technological dinosaur.  I guess a carefully blueprinted "mythical" Model 'C' four-banger would make some kind of sense, that's the last iteration of the inline four built for 1933-34 U.S. Ford cars.  Here's the scoop on the alleged Model C engine, actually a Model B type according to this source.  The late "B" engine is clearly a better starting point than the 1928-31 Model A prototype, especially for use in an airplane: http://www.fordgarage.com/pages/modelcmyth.htm.  Maybe the Pietenpol builders use a contemporary aftermarket crankshaft, rods and improved pistons?

I've seen the Rotax engine in ultra-light planes, you must be referring to a larger version for planes like yours.  Rotax engines enjoy a solid reputation and have been widely used in a variety of applications.

In my recent brake tubing flare and bending how-to series at the magazine, I glanced at quality tools available for aircraft work.  Had I been covering 37-degree aircraft flares, this tool was attractive and affordable:  http://www.cleavelandtool.com/Parker-Rolo-Flare-Tool/productinfo/TRF37/#.WuN25qQvyUk.  You likely are familiar with Cleaveland Tool.

The HD video series on tools and hands-on work would be useful for your automotive and Jeep projects:

http://www.4wdmechanix.com/video-series-how-to-flare-automotive-brake-tube-fuel-lines-and-cooler-tubing/

http://www.4wdmechanix.com/how-to-fabricating-restoring-and-repairing-hydraulic-brake-lines/

You're well versed on this form of flight, especially home-built aircraft and its safety measures!  Have you flown for a long time?

Moses

 

 

[Stuart_Snow]

Moses,

May was a busy Month with a daughters wedding and a trip to the Kaibab National Forest on the North Side of the Grand canyon.  Both events were a delight. I was able to tow the CJ-3B on the trip with a new addition to my Mopar fleet. A 93 Dodge Ram 250 club cab  5.9 Cummins.  It has the poor mans dually conversion and has 50,000 miles. It belonged to my Father -in law and wasn't driven often. I'll have to read through the Cummins forum to get up to speed on this early generation of the 5.9.  Exploring the Kaibab with the CJ-3B has been a dream for many years. Conditions were dusty but I saw some great country and wildlife and the Jeep comes alive going down old narrow trails.

I share your enthusiasm for the Merlin v-12 and P-51. Nothing sounds better than a Mustang on a high speed pass. Corvair powered planes have the  unique growling sound of a 6 cylinder boxer engine but not quite as impressive as the mustang. I've always wanted to go to Reno but have not made it yet. This will be my first year for going to OshKosh in August.

Work on my Corvair powered aircraft continues. I have been working on the Elevator bell-crank assembly and was doing some welding today. I struggled a little as I was trying to weld dissimilar size material. I think I need better regulators as I'm constantly adjusting my flame setting. I will look for Richard Finches book and also check out the Cleveland tools site.

Some more good news is my  4.2 block from my 77 CJ7 project  is back from the machine shop and all the new parts. I cant wait to start working on that.

I started flying back in 96 and most of it has been in Cessna's. The open cockpit Pietenpol will be a real throwback to the old days of flying. She is a good honest flying ship with no surprises but it is a tail dragger and its a light weight and high drag machine which bears watching in the energy management department.

Stuart

 

 

[Moses Ludel]

Stuart...Congratulations on your daughter's wedding!  Also, there's nothing like the North Rim, we have approached from Nevada via Jacob Lake.  Spectacular country...See below:

36 minutes ago, Stuart_Snow said:

Moses,

May was a busy Month with a daughters wedding and a trip to the Kaibab National Forest on the North Side of the Grand canyon.  Both events were a delight. I was able to tow the CJ-3B on the trip with a new addition to my Mopar fleet. A 93 Dodge Ram 250 club cab  5.9 Cummins.  It has the poor mans dually conversion and has 50,000 miles. It belonged to my Father -in law and wasn't driven often. I'll have to read through the Cummins forum to get up to speed on this early generation of the 5.9.  Exploring the Kaibab with the CJ-3B has been a dream for many years. Conditions were dusty but I saw some great country and wildlife and the Jeep comes alive going down old narrow trails.

I'm writing a quarterly column for Turbo Diesel Register at the editor's request, an offer I could not refuse.  The Gen 1 Ram is a rugged and traditional truck, the 12-valve engine (BT6) is an industrial beast with mild enough torque to keep the transmission alive.  Is it a Getrag 5-speed or A518/46RE automatic?

I share your enthusiasm for the Merlin v-12 and P-51. Nothing sounds better than a Mustang on a high speed pass. Corvair powered planes have the  unique growling sound of a 6 cylinder boxer engine but not quite as impressive as the mustang. I've always wanted to go to Reno but have not made it yet. This will be my first year for going to OshKosh in August.

Enjoy Oshkosh, a world famous event!  We have the Reno Air Races locally.  I filmed at the pits years ago and was fascinated with the Rolls-Royce Merlin engines.  The stock engines look so "period" with carburetors and magnetos!  Amazing that this technology helped turn the air war in Europe.

Work on my Corvair powered aircraft continues. I have been working on the Elevator bell-crank assembly and was doing some welding today. I struggled a little as I was trying to weld dissimilar size material. I think I need better regulators as I'm constantly adjusting my flame setting. I will look for Richard Finches book and also check out the Cleveland tools site.

Two-stage regulators?  A must for lower or even higher pressures...Finch's book will be very helpful, some "Ah, ha!" moments for your oxy-acetylene welding.

Some more good news is my  4.2 block from my 77 CJ7 project  is back from the machine shop and all the new parts. I cant wait to start working on that.

Yea!  A good engine for getting top results from a rebuild.  This will be a nice powertrain when you finish.  Keep us posted!

I started flying back in 96 and most of it has been in Cessna's. The open cockpit Pietenpol will be a real throwback to the old days of flying. She is a good honest flying ship with no surprises but it is a tail dragger and its a light weight and high drag machine which bears watching in the energy management department.

Tail draggers were very popular for short take-off and landings, big at rural Nevada in the day...Avoiding the tip forward is the trick.  You must be a very good pilot by now...

Stuart

Moses

 

 

 

[Stuart_Snow]

On 6/9/2018 at 9:54 PM, Moses Ludel said:

I'm writing a quarterly column for Turbo Diesel Register at the editor's request, an offer I could not refuse.  The Gen 1 Ram is a rugged and traditional truck, the 12-valve engine (BT6) is an industrial beast with mild enough torque to keep the transmission alive.  Is it a Getrag 5-speed or A518/46RE automatic?

I haven't had time to figure out what the transmission is but It has an overdrive. I'm not familiar with the Getrag but my I was able to get 200K of service from the 46RE on my 97 Dodge Ram with a few tweaks. I've completed a full service on the 93 and have discovered a rear main seal leak which I'm not surprised  considering its age and lack of regular driving. I'll try to make a thread in the Cummins forum.

On 6/9/2018 at 9:54 PM, Moses Ludel said:

Two-stage regulators?  A must for lower or even higher pressures...Finch's book will be very helpful, some "Ah, ha!" moments for your oxy-acetylene welding.

No they are not. Mine are old medium duty hand-me-downs. Its time for me to invest in some good modern 2 stage units. 

Good News! I ordered the latest version of Finch's book. Its already arrived an I love it. This is really going to help improve my welding. A lot of good data in here and I've already found some things I'm doing wrong like running my Oxygen pressures too high.

[Moses Ludel]

Hi, Stuart...Glad you got the Finch book, he did a real service here, making welding accessible to many.  The read is easy, the facts are confidence inspiring.  A great book!

You'll have more flame stability with the 2-stage system.  When bottle gas drops down, you won't be fiddling as much with the flame...

Best regards,

Moses