Another Radial - this time 18 Cylinders

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I think it's something you need to call about to get. I might give Gary a call myself.
 
Peter,
I didn't change the o-rings in the Super Tigre for compatibility with gasoline. I, maybe erroneously, assumed they were compatible with either. I haven't noticed any o-ring related issues with the carb on the H-9, but since I didn't need to change the settings once the carb was dialed in, I might not have noticed any minor swelling if there was any.
I've noticed in the local hobby store that fuel components like tank stoppers that look like white silicone are sold for use with alcohol only; and they also have black ones, presumably nitrile or Viton that are sold for either gas or alcohol. I don't know why they wouldn't supply just the black stuff, whatever it is.
The Perry website doesn't discuss fuel type as far as I know, and frankly the pump stuff was unfamiliar and confusing to me and was the real reason I called. When I spoke to Gary I told him I was planning to use gasoline, and I wanted to make sure there wasn't an issue with the carb's plastic body. That was when he told me that he would supply my carb with an alternate idle disk that was calibrated for gas. I believe their inventory is probably a bunch of modular parts that are assembled per order. I wasn't real sure of the Venturi size I wanted, and Gary told me if I ended up needing a larger carb I could just purchase another throttle barrel instead of buying a whole new carb. I forgot to ask him about gasoline compatibility of the o-rings in his carb, but I think it's reasonable to assume that if he sells idle disks designed for gas then he is using o-rings compatible with either fuel.
By the way, I might be interested in that 40NA carb that you're looking at. What is its Venturi size? I looked it up on the Tower Hobbies website, and I just had to smile when I read it was listed as end of life with them. - Terry
 
Rather than hijacking Terry's wonderful build thread I am starting another one to discuss the last entries in this one about carburetion.
gbritnell
 
By the way, I might be interested in that 40NA carb that you're looking at. What is its Venturi size? I looked it up on the Tower Hobbies website, and I just had to smile when I read it was listed as end of life. - Terry

Hopefully I didn’t confuse matters, Terry. My intention was to show the style of OS’s 4-S downdraft methanol carbs because you were happy with the ST & possibly evaluating alternatives. Basically just pointing out OS makes progressively larger suite of 4S engines all of similar carb flavors (62, 72, 81, 95, 110, 120, 155..) I believe the 56-FS-a is still the current '56' version, at least it shows up in catalogs. I recently bought a 56-FS-A piston liner set for my radial project, unfortunately not the carb yet or I’d measure for you.
http://www3.towerhobbies.com/cgi-bin/wti0001p?&I=LXRUF7&P=ML
http://www.osengines.com/engines-airplane/osmg0956/index.html
http://www3.towerhobbies.com/cgi-bin/wti0095p?FVPROFIL=&FVSEARCH=OSMG2485

And just to drive you crazy, OS also makes a GF40 4-Stroke gasoline version & correspondingly different looking carb.
http://www3.towerhobbies.com/cgi-bin/wti0001p?&I=LXCZLF&P=7
http://www.osengines.com/engines-airplane/osmg0800/index.html
It’s designed around bigger displacement though; 2.4 cid, 1.57”B x 1.25”S (vs. 56A @ 0.57 in3 0.945"B x 0.811"S). How that sizing might relate to your displacement & double-row net effect is way above my pay grade. Sorry, wish I could offer useful insight but I'd be blowing smoke. At least with the Conley’s sounds like you can swap & tune among his orifice barrels. If everything works, you are home free. If the orifice is basically correct but other operational misgivings, at least you will be armed with target size.

I agree with George, sorry, did not want to divert your great post.
 
Seems like I've studied too many RC carbs. The bore of the OS 40NA found on the FSa-56 is 6mm. It has a spraybar and needle that should be roughly equal to a 1mm rod through the middle. Carb sizing is directly related to how much air the engine uses. If the fuel draw is solely due to pressure drop, without muffler pressure is in most RC applications, the venturi must be smaller.

Two needle type carbs are problematic when the air fuel ratio needs of the fuel is not as designed. A glow fuel carb on gasoline will end up rich in the midrange.

The carb on the GF-40 is a Walbro. The smaller GF-30 has a design like their diaphragm regulated glow carbs with a separate diaphragm pump.

Greg
 
For the past few days the outside temperature has been in the mid-40's F with a very high humidity due to our rainy weekend. I was tempted to haul the engine outside and play with it anyway because I'm really anxious to see how it now runs with a reasonable carb setting. The only reason I didn't was my concern about condensation in the crankcase and on my 12L14 crankshaft. The weather people expect the current conditions to continue for at least another week, but I decided I just couldn't wait that long.
Without telling my wife, I rearranged the furthest corner of my shop away from the computers to create some space where I set the engine up for a brief run. I only wanted to run it for a minute to see if I might now have some semblance of throttle response. With 18 cylinders I expected to quickly fill the enclosed area with fumes, but my real concern was staying out of the prop while moving in the limited space around the engine to make adjustments. The last thing I want is an accident where I end up with a bent crankshaft - neither mine nor the engine's. The ideal place would have been in front of an open door, but without some serious rearranging I just didn't have enough space in front of a door for both me and the engine.
After cleaning the soot off the plugs from last week's run, I opened the high speed needle 3/4 turn from fully closed; and then, after manually priming the engine, I spun it with my drill starter. It started right up and ran great. It appeared to be running on all cylinders as far as I could tell by putting my hand near each exhaust tip. I think I tested all of them, but I'm not entirely sure. There wasn't much light where I was working, and I had to carefully stretch to reach around the nearly invisible spinning prop.
The throttle seemed to smoothly control the engine, and I was able to adjust the high speed needle another 1/8" turn for a peak rpm of about 2500 with the throttle about 3/4 open. The engine responded nicely to the needle change without being overly sensitive. I realize the high speed needle really should be set with the throttle wide open, but I'm not ready to go there yet. I want to break-in the engine with several short runs between idle and a few thousand rpm with complete cool-downs in between before revving the engine up fully.
I noticed the engine seemed to run best with zero indicated spark advance. I wasn't expecting this, and so I need to verify my timing calibration. I didn't attempt to adjust the idle disk just yet. So far, the engine is capable of idling as low as 800-900 rpm, and its low speed idle sounds really great. The idle stop needs to be raised some because right now the engine stalls when the throttle is pulled fully back.
Against the agreement I had with myself I ran two tanks of gas through the engine. A 10 ounce tank of fuel still lasts about only about two minutes even with the new carb setting. I filled the tank for the third time, but the engine didn't sound right while being cranked. Then I noticed the Hall indicator led on the rear bank distributor was constantly ON. Later testing verified the Hall device in this distributor was dead. This is the first Hall sensor that has ever failed in any of the four engines I've built. I don't believe it was damaged by a direct discharge inside the distributor since it's well protected in there. Instead, I think it may have been destroyed by a transient on the power supply line. I installed 10 Volt MOV solid state transient protectors on the supply lines of all my TIM-6 ignitions to protect the semiconductors from transients created on the supply line by the high voltage section. I didn't use them on the T-18's CDI modules because I was cramped for space inside the enclosures; and then, well, I forgot about them. It might just be a coincidence, but the tach drive signal also comes off this particular sensor; and so I'll need to take a look at a possible power supply sequencing issue or maybe some other interaction with the tach electronics.
Another minor issue (again) was oil flow rate. With the drip feeder valve opened only one turn while running, the crankcase filled with oil to the point where it began running out the front bearing. At this drip feed setting the engine pumped the entire contents of the oil tank into the catch bottle during the four minutes of running. I had the same issue with the H-9, and it seems a much lower drip rate will be required for this larger engine. The drain-back channel between the front of the engine and the sump inlet at the rear of the engine is as big as I could safely make it, but it's still enough of a restriction that the pressure pump can easily get ahead of the scavenger pump unless the oil flow from the tank is severely restricted.
The oil in the catch tank was very dark due to blow-by contamination, but I was thrilled that even though the shop was filled with exhaust fumes, there wasn't the slightest hint of oil smoke. After draining the sump I found an obvious build-up of ring material on the magnetic drain plug. In fact, after the four minutes of runtime I had more material on the drain plug than I remember accumulating during an hour of 'motoring-in' the engine while testing the oil pumps a few months ago. This reinforces my suspicion that motoring an engine without the pressures of combustion does little to seat the rings.
So now I have some repair work to do. The good news is that I think the engine is basically running well, and only some minor tweaking remains. From all early indications the Perry carb will likely be acceptable. I need to look into the timing advance and make sure I haven't mis-calibrated the indicator. And, of course, the major issue is that I need to replace the Hall sensor on the rear row distributor and figure out why it failed. I'm reasonably sure I can replace the sensor without pulling the distributor, but the tedious process of re-synching the timing will have to be done. The most important task is to figure out why it failed and then correct the problem so it doesn't happen again. - Terry
 
Wonderful report Terry! I bet your feet can't touch the ground yet.

I know I'd be walking on cloud 9 even with the minor issues you've mentioned.

Congrats and great work.

Anxiously waiting a video.

Ron
 
That is a great feeling...getting an engine running.

You can only make it better from there!

Great job...looking forward to hearing her run!
 
Just stumbled upon this amazing thread. Funny to find it right at the moment of truth when all this incredible workmanship comes to fruition.

I found this thread by accident, not being a hobby machinist but being fascinated by radial engine design.

A little while ago, I found the plans online for Karl Erik Olsryd's 9 cylinder radial engine which I subsequently modelled in Pro/Engineer.

[ame="http://www.youtube.com/watch?v=XkcKMMa_kO8"]www.youtube.com/watch?v=XkcKMMa_kO8[/ame]

And you can download the entire CAD model of it for free at Grabcad.com

There are some nice rendered images of the engine for those of you who don't have access to CAD software.

http://grabcad.com/library/olsryd-9-cylinder-radial-engine--1

I have been wanting to model an 18 cylinder radial ever since completing the 9 cylinder radial but I came to realise that the Olsryd engine had too many design problems for it to be converted into an 18 cylinder design.

This engine however, is in a totally different league of sophistication, having a pressure fed lubrication system and the ingenious bronze main bearings and built up crankshaft. It is an absolute work of art both in design and of course, the execution of this design into a real physical working engine. Incredible skill and attention to detail.

I would really like to get a set of engineering drawing for modelling my own 18 cylinder radial in Pro/Engineer.
 
Terry,
I just got caught up on your thread, wow. Its amazing that you had so little trouble firing it up for the first time. The 10 oz. fuel tank is going to seem to small once you get all the sorting out done. I must admit not making any chips recently but drywall dust instead, no fun.
Art
 
Replacing the failed Hall sensor took much longer than expected because testing its replacement uncovered one timing foul-up after another. The whole experience left me wondering if I had been asleep when I originally set it up, and just how the engine had been able to run at all.
I started disassembly of the rear row distributor by temporarily replacing its blue Delrin cap with a transparent test cap that I had made earlier. After rotating the crankshaft to bring the rotor into firing position under tower #1, I pulled the spark plug from cylinder #1 to verify its piston was at TDC of its power stroke. Unfortunately, the piston wasn't even close to where it was supposed to be. I then discovered that I hadn't fully tightened the SHCS that is suppose to secure the trigger disk and its keyed rotor to the distributor shaft when I timed the engine months ago. Because of this 'screw-up,' the timing for half of the engine's cylinders had been slowly shifting as it ran. I couldn't believe I had pulled this stunt again. I made the exact same mistake three years ago when I initially timed my H-9. In the process sheet I wrote up for myself to do the timing on this engine I even made a note to not forget to tighten that screw this time.
I rotated the loose trigger disk by hand to exercise the sensor, but it still indicated bad. After removing the dead sensor from the distributor I noticed the shrink tubing that I was depending upon to insulate its leads from the distributor housing had been badly abraded. This might have happened when I inserted the sensor through the bottom of the distributor during assembly. I remember that it really should have gone in from the top, but I had already soldered on its connector making that impossible. It looked probable that the sensor's Vcc lead might have, at least intermittently, shorted to the distributor housing. I re-tested the sensor while it was off the distributor hoping the problem had just been the shorted lead; but I finally had to accept it was really dead.
I think it's very unlikely that at least one of these two screw-ups on my part wasn't somehow responsible for the failed sensor. One possibility is that the shifted timing allowed the rotor to end up between two tower electrodes when the CDI capacitor fired. The very long discharge path that resulted would have allowed the coil's secondary voltage to soar uncontrollably until, possibly, a breakdown occurred somewhere in the CDI module sending a killer spike back to the sensor. If the path had included Vcc, the missing MOV might have saved the sensor. I think it's more likely, though, that the spike was coupled back to the sensor's signal lead through the SCR used to discharge the cap. This scenario is pretty ugly because a similar result can routinely occur in an engine with a fuel-fouled plug or an excessively wide gap. As one of the photos shows I had made the rotors extra wide in order to encourage such misfires, if they occur, to dump the coil energy to an adjacent tower electrode before the coil's secondary voltage can rise to its open circuit maximum.
Another possibility is that the rotor actually did arc to the distributor housing. It's possible, but not likely, that the impedance below the portion of the housing containing the sensor allowed the housing to briefly rise to a high potential that eventually reached the sensor through its insulation-abraded lead.
After replacing the sensor I re-phased the trigger disk to cylinder #1 and finished the reassembly of the distributor. Establishing the timing of one of my distributors is complicated by its timing advance scale. Two separate but interacting adjustments must be made so that, at precisely TDC, the sensor is triggered by a magnet when the rotor electrode is directly under the correct tower electrode. The extra complication is that both of these events must occur while the distributor is rotated so the timing indicator on the crankcase is pointing exactly to 0 degrees on the advance scale on the distributor housing.
I took a series of photos of the reassembly to show how the sensor is shrouded from the electrical storm in the distributor cap and also to point out a few of the culprit components. After verifying the rear row sensor led was now blinking as it should when the crank was rotated, I noticed it's relationship with respect to the blinking front row sensor led looked suspicious.
So, I put a test cap on the front row distributor and rotated the crankshaft 180 degrees until cylinder #1 was at TDC of its intake stroke. This should have placed the front row rotor electrode directly under tower #10's electrode. But, strangely, it was advanced by 30 degrees even though the SHCS on this distributor had been properly tightened.
I believe what happened was that when I timed this distributor I was confused by its opposite direction of rotation, and I probably timed it with the distributor rotated so its timing pointer was at the wrong end of the distributor housing's advance scale.
After re-timing this distributor I then noticed its sensor light seemed to occasionally miss a blink as the crankshaft was turned. I think I had actually noticed this earlier when starting the engine, but at that time I thought it must have been my imagination since it was very infrequent, and the engine seemed to run OK.
It turned out that the trigger disk in that distributor was just a bit too far away from the Hall device, and occasionally it would not fire on a particular cylinder. This distance is set by a machined spacer. When I made the distributors I made a pair of spacers for each one that were different by .01" in thickness. Since the required spacing is determined by the strength of the magnets as well as the sensitivity of the Hall device, I selected the proper spacer during the distributor's assembly and test. This distributor somehow ended up with the thicker spacer. Since I still had all the spacers I solved the problem by replacing the spacer with the thin one.
Despite my ugly twin's efforts to scuttle the timing on this engine, it somehow started and ran as long as it possibly could. All these timing issue could explain why the engine started so easily the very first time the prop was bumped by hand, but then needed the drill starter for all subsequent starts. I now also understand why, during the last session, it didn't like any additional timing advance. Half of the cylinders were already advanced 30 degrees, and the other half were, well, who knows where. After straightening out the final issue I carefully, once more, verified the timing of both distributors.
Without knowing for sure the exact cause of the sensor failure, but being reasonably sure it was related to a voltage spike, I decided to shotgun its protection. I added 12V unipolar transient protectors (LittleFuse SLD10U-017-B) between the sensor's Vcc line and ground and also between the sensor's signal line and ground. These particular devices, which are new to me, are essentially very fast, heavy-duty Zener diodes. Their clamp voltage is much better defined than that of an MOV, and they are used to protect sensitive electronics in automotive applications. I ordered mine from Mouser.com without checking their physical dimensions. When they arrived I was taken back by their huge size. Because of the limited space inside my ignition module enclosures, it was difficult to find room for them, but there is now a pair protecting each sensor.
The next time I try to start up this beast I plan to have a video camera sitting in front of it. - Terry

877.jpg


878.jpg


879.jpg


880.jpg


881.jpg


882.jpg


883.jpg


884.jpg


885.jpg


886.jpg


887.jpg


888.jpg
 
I've been reading this thread now for a couple of days (in complete AWE) but the one aspect of the Hodgson design I cannot fathom is how the cam rings are kept running concentric with the crankshaft. On the Olsryd design that I am familiar with, the cam ring has a centre bearing that runs on the crankshaft. This keeps the crank and the inner ring gear on the camshaft concentric

On this Hodgson design, the camrings are "centreless." There is no centre bearing

So how is concentricity maintained????
 
Mattsta,
The cam ring concentricity is maintained with a circular groove machined into the rear of the cam ring assembly. A matching circular tenon is machined into the bronze crankcase bearing. The cam ring rides on this tenon which is concentric with the crank bearing. A Delrin bumper-like retainer, bolted to the bronze bearing, keeps the cam ring in place against the bronze bearing. - Terry

889.jpg


890.jpg


891.jpg
 
Today was a beautiful 70F degree day here in central Texas making it a perfect opportunity to test the new timing on my engine. I set up my ancient flip video camera outside my shop to record the very first start-up. I apologize for the poor quality of the video. It seems I only use this camera every couple years to record a new running engine, and I find myself having to relearn its operation. Evidently, I still don't know how to use it in bright outdoor lighting.
There really isn't much to add to the video. The noise you hear in the video before the engine is started is the fuel pump. I manually choked the engine, set the timing at 10 degrees BTDC, and rotated the prop a fraction of a revolution by hand. As the video shows, the engine took off running. The carb settings haven't been changed since the last start-up, and the video included the engine idling at 900 rpm and running as high as 2500 rpm at about half-throttle. Again, I didn't push the engine or attempt any re-tuning of the carb since I want to wait until I accumulate more short runs. I plan to do the remainder of the tuning while keeping an eye on the spark plugs. I expect the final carb settings will be a compromise with the top rear cylinders running a bit lean in order to keep the lower front cylinders from running overly rich.
I ran 3/4 tank of gasoline through the engine during the video, and then I re-filled the tank and started the engine a second time off-camera. I had to use the drill starter, though, for the re-start. I suspect the hot soak creates an intake lock near the intake/exhaust flange which is heated by the exhaust. The high velocity starting flow created by the drill probably helps cool the area.
My preliminary temperature measurements, though, showed the flange temperatures were only on the order of 130F. These measurements were made just after shutting the engine down since the plug wire EMI makes my non-contact thermometer pretty much useless on a running engine.
The engine appeared to be consistently firing on all cylinders, and there was no oil smoke indicating all 54 cast iron rings were working as they should. The only oil I'm currently catching on the firewall appears just after the first start-up when the prop wash blows oil that has seeped down along the pushrods from the lower lifter bushings. If the engine is allowed to sit a few days, there is also some oil from one or two lower exhaust pipes whose cylinders were left standing with open exhaust valves. If I were starting the engine build over, I would include pushrod tubes.
I included a photo of the oil in the catch bottle. It's color is black due to the bluing on the cylinder walls being polished away by the piston rings. It will likely take at least a quart of flushed oil for the color to clear up, and it will be at this time when I'll close the oil loop and begin recirculating the oil. The oil spray visible on the interior wall of the bottle is created by the crankcase pressure pulses which work to reinforce the scavenger pump operation. I believe this action also helps to prime the scavenger pump even when the sump is drained between runs. The small amount of oil in the catch bottle after about four minutes of runtime probably indicates my new 1/4 turn open oil feed setting Is now a bit too low.
I have to mention that I really like my little tach. It was a real PITA to get functional, but it turned out to be well worth all the effort.
Over the next week or so I'll put more short runs on the engine and slowly start raising its rpm. By monitoring the plugs I'll be able, over time, to optimize the carb settings and learn whether the current carb is properly sized for my maximum target 3500 rpm. I also need to go through the engine and re-tighten any loose fasteners. After these loose ends are tied up I'll make a better quality video to wind up my little T-18 build. - Terry

[ame]https://www.youtube.com/watch?v=x6xtmo8hFRk[/ame]

892.jpg
 
Truly fabulous engine Terry! Thanks so much for posting the video. Next best thing to being there!
 
Terry,
I have the worst case of goose bumps I've ever had. What a sound!!!
Fantastic journey, fantastic destination.
Thanks so much for sharing.
gbritnell
 
Terry,
I'm with them it sounds wonderful ! great job.
Art
 
An absolute work of art, sounds great as well. You gave me a fright when you leant forward to check something, I thought you might get a haircut from the prop.

Paul.
 
Terry! You owns!
Thanks for sharing! You are a great inpiration and helpfull to we who also building Hodgson.
And ...what a sound:eek:
Ove
 

Latest posts

Back
Top