Another Radial - this time 18 Cylinders

[mayhugh1]

Scott,
Thanks for the tip. It's on my to-do list for this weekend. - Terry

 

[dr_kingston]

Hi Terry,
I am very inpressed with your build. What did you use to bond aluminium stock to particle board etc.
Regards Dale

 

[mayhugh1]

Dale,
It's a quick set 2-part epoxy distributed through Lowe's home building centers called Devcon Super Gel. When the machining is finished it releases cleanly when heated to 190F. - Terry

 

[mayhugh1]

After enough procrastination, I decided to tackle the starter adapter problem. Basically, I could see two options. The first option is to leave the existing design unchanged and create a new starter adapter that acts upon the prop instead of the spinner hex. Currently the prop is sandwiched between the front and rear prop hubs using six bolts. The rear hub is keyed to the stronger 1/2" diameter portion of the crankshaft. The prop assembly is held against a machined offset on the crankshaft by the spinner which is threaded on the weaker 3/8" portion of the crankshaft. The rear of the spinner has a machined hex that engages the starter adapter, but when it does so it places excessive stress on the small diameter portion of the crankshaft due to the high torque required to turn the engine over.
A second option, which is my interpretation of George's suggestion, is to re-design the front hub with an integral large diameter threaded boss for the spinner. Although a 3/8" nut must be threaded onto the end of the crankshaft to secure the prop assembly against the machined offset, the weaker portion of the crankshaft will no longer see the starting torque. In addition, since the prop assembly is keyed to the crank, there'll be no net force except for the inertia of the nut trying to unthread it, and so its torque requirements are minimal.
I decided to go with the second option since it's what I should have done in the first place, and I don't want another starter adapter. The only potential issue with this option is the difficulty in keeping the spinner running true with respect to the crankshaft after it is re-threaded. I originally finish-turned the spinner while it was threaded onto the front section of the crankshaft to true it, but now that the engine is assembled, that's no longer an option. Wobbly flywheels and spinners drive me nuts.
My first step was to re-thread the spinner. I used a left-hand 3/4"x20 thread, since I happened to have a tap for that particular thread. Its major diameter must also be large enough to clear the 3/8" nut that will be threaded on the end of the crankshaft. I indicated the rear surface of the spinner's hex in my lathe before boring out the existing 3/8x24 thread, and the tailstock was used to support a tap holder in an attempt to keep the axis of the new thread coincident with the axis of the original thread. In retrospect, I probably should have turned the threads instead of tapping them.
Next, I turned a new stainless steel front hub with a 5/8" long .75" diameter boss. This boss was threaded for a close fit to the spinner. I cut the left hand thread using an inverted single point tool and a reverse-running spindle while continually checking the fit with the spinner. I got an extremely close fit when I was done, but the spinner ended up with some annoying run-out after all. I wrote a program to re-skim the spinner while it was threaded on the front hub. The spindle motor (or its controller) in my CNC lathe began acting up during this last step and thankfully allowed me to finish before it finally gave up the ghost. - Terry

 

[gbritnell]

Hi Terry,
I'm sorry I didn't get back to you but it seems you have solved the problem most exquisitely. I like you hate having runout on rotational parts. No matter how accurate the remaining parts are it just takes away from the whole picture.
gbritnell

 

[petertha]

Hi Terry. Sorry for 'old post' question, but I’m working on a similar design issue & remembered your pictures. Maybe you have a section view handy or can confirm how intake tube is retained in the housing. My hunch: slide intake tube through metal ferrule & through an O-ring, insert tubing end into case, it’s loose enough to allow positioning of flange end, bolt to cylinder head, then screw down ferrule, squishes O-ring, contact area seats/seals against tube?
Your Post 77
http://www.homemodelenginemachinist.com/showthread.php?t=21601&page=8
.. compression nuts that are used to support the intake pipes in the fuel plenum in the rear crankcase. Initially, it wasn't clear to me from the Chaos photos how these were used. They didn't seem to be pipe threads nor did they use ferrules or flares. I eventually decided they must compress a rubber o-ring through which the intake pipes pass. So, I threaded the nine bosses in my rear crankcase section with simple 1/2" fine threads...

 

[mayhugh1]

Peter,
It's pretty much what you said if you're referring to the threaded nuts with the hole through their center as ferrules. There is a rubber o-ring sitting in a recess at the bottom of the threaded bore. The fuel tube goes through this o-ring, and the nut is screwed down on top of it. The nut compresses the o-ring around the tube and against the recess to create the seal. As you said, the other end of the tube is mechanically stabilized by its bolted flange, and this end 'floats' in the compressed o-ring. - Terry

 

[mayhugh1]

For the past five months I've been working on a background tachometer project for this engine. It's been one of the most frustrating projects I've been involved with because it has required so much time dealing with questionable documentation and flaky vendors.
The project began when I purchased a really cool miniature (1-1/4" square) aircraft tach in 'unknown condition' through eBay from someone in Bulgaria. I was hoping to make it the center piece of the control console I'm now designing for this engine. 'Unknown condition' on eBay usually means 'run away fast', but the meter's face looked OK in the photo so I decided to go for it. It wasn't very expensive, and I actually ended up paying less for it than the shipping charges to get it here.
I've modified d'Arsonval meters in the past to alter their functionality for special applications, and so I only needed a working movement. The attraction of this particular one was that it was the right size, and I wasn't going to have to make a new face for it. When I received it there was no electronics - just a meter with four terminals in a four inch long housing where the original electronics had probably been located. I shortened the housing to the actual depth of the meter, which was about one inch, since I planned to add my own circuitry off to the sides instead of to its rear.
Some ohmmeter checks showed the meter probably contained a pair of coils which was puzzling enough, but I also found the meter was relatively unresponsive to dc current. After some web research I learned the meter was an air core movement, something totally new to me, and required complex quadrature signals instead of dc current to drive it. My guess is that my 'unknown condition' meter had been cannibalized for its more valuable electronics. After some more investigation I discovered there have been special purpose ICs manufactured to generate the needed signals to drive such meters since they were commonly found on automotive dashboards for many years. The current move seems to be toward stepper motor based-movements. A review of several on-line data sheets and an exhaustive Web search revealed the DIP version of the one that I probably needed (CS8190EN16) had been long discontinued and no longer available from common US distributors. However, a guy in Hong Kong claimed to have ten of them for sale on eBay. I ordered five parts at a good price from him hoping to get at least one that worked.
With help from the data sheet/application note I soon developed a schematic. The IC requires its own 12 Vdc source, and since I plan to run my engine from a single 6V lead acid battery I had to include a small dc/dc converter module in my design. Actually getting one of these converters in my hands grew into its own months' long saga created by several vendors who continually lied about having access to them. After finally receiving one, though, I breadboard it separately from the meter circuitry so I would have more flexibility later on with the final packaging.
I made up a Hall-triggered test board to drive the meter circuitry as well as a nine magnet trigger disk that I could spin with a drill or mill spindle to test and calibrate the tach. This test circuitry duplicated the trigger circuitry I would be interfacing to in my already completed ignition modules so I wouldn't have to involve and possibly damage them in my early breadboarding attempts.
It was a good thing I had included a socket on the breadboard for the driver IC because after receiving them from Hong Kong I found that not a single one worked. Each presented different symptoms telling me I had probably purchased poorly QC'd counterfeit parts or else floor sweepings from someone who had worked at the original fab plant.
The discontinued part was available from several other Chinese direct suppliers, including some others also advertising on eBay. I decided to try once more with a different vendor. A month later when five more parts from the second vendor showed up, I found these ones almost worked. When powered up, the meter zero'd as it was supposed to, but all five of the IC's were totally unresponsive to any trigger input. I had troubleshooting access to a portion of the internals of the ICs through two key pins, and it seemed that all five parts suffered from the same internal defect. An internal block of circuitry known as a current mirror that connects the input signal conditioner to the F/V converter did not seem to be functional. This was really suspicious, and the real frustrating part was that the section of the data sheet that explained the operation of this portion of the circuitry and its effect on the selection of several external components that I had to add contained so many errors and ambiguous statements that I didn't trust it. I just couldn't be certain whether the problem with these ICs lay with me, the data sheet, or all the parts I had purchased.
I considered trying to breadboard the surface mount part because it is still in current production and available from reputable distributors, but I felt that with all the uncertainty over the past several months I needed to stick with a socketed part. After many false leads and lots of phone calls I eventually located a small Canadian automotive parts supplier who had been stuck with a batch of the DIP parts after dashboard designers quit using them in new designs. After being assured they were truly in stock in North America, I ordered another five parts and paid a healthy small quantity premium to get them.
While waiting on the third batch of IC's, I machined a housing for the meter; and I made a new breadboard designed to fit around the meter inside the housing. I also developed a new schematic using the functional meter driver portion of the last batch of ICs along with my own front-end circuitry just in case the third batch of parts didn't work either.
Since they had no experience in dealing with small orders to individuals, this last vendor made a paperwork error that triggered a customs and security snafu that ended up delaying delivery of the parts for nearly a month. Despite what I had been promised, the parcel had actually been shipped from India instead of Canada.
When I finally received them, and with very low expectations, I tested each one in my original breadboard. To my amazement and after nearly five months of frustration, I was finally rewarded with a rational movement of the tach's needle in response to my test generator. While wiping tears away I packaged my new circuit board and meter movement inside the enclosure I had made. I hauled the tach, battery, and test driver circuitry over to my mill spindle where I was finally able to run and calibrate the tach. As a final step I re-tested the tach with it connected to one of my ignition modules while actually firing a spark gap. I'm pretty sure I've never experienced a better example of Murphy's Law than I have with this 'little' side project. - Terry

 

[ShopShoe]

mayhugh1,

I am simply amazed. After looking everyday for more of your excellent design and machining, I find you have furnished an elegant electronic engineering project with more incredible documentation. In a former life I paid bigtime vendors bigtime bucks and they could not draw a simple schematic like you have just done. Nor could they actually explain how their assembled equipment actually worked without repeating propaganda from their marketing department. As far as Murphy's Law and ebay and vendors: been there and done that.

Thank you for going to the extra effort of writing incredible descriptions and fully documenting everything you do. This is the reason I keep looking for your posts. I'll probably never build a big radial, but I can try to work toward a better, complete, and understandable job when documenting some of the things I do.

Thanks again,

--ShopShoe

 

[kuhncw]

ShopShoe said it very well.

I'll just say, thanks again for all the time and effort you put into your posts.

Regards,

Chuck

 

[mayhugh1]

Hi,
Is anyone else having difficulty seeing the photos in my last several posts? Everything was OK yesterday, but today I can't see my own photos although I can see the photos of others in their threads. I've tried three computers as well as logging out and then back in but still no joy. - Terry

 

[ozzie46]

I can't see them today either. I saw them yesterday.


Ron

 

[Cogsy]

I could see them yesterday but now they are just a list of attachments.

 

[Swifty]

Still just a list of attachments today.

Paul.

 

[mayhugh1]

I've sent a few queries to the web care taker about the strangeness that has started going on with this thread, and I'll likely hold up any further posts until they have a chance to figure out what's going on. I expect it's likely related to the size that it has grown into, and so I don't want to take a chance on making things even worse by adding to it and possibly losing work. - Terry

 

[Scott_M]

I have also noticed that I cannot scroll in this post anymore. My scroll wheel works fine in other posts.

I think you broke it Terry

Scott

 

[Cogsy]

I expect it's likely related to the size that it has grown into, and so I don't want to take a chance on making things even worse by adding to it and possibly losing work.

I just came across another thread that's having the same issue - http://www.homemodelenginemachinist.com/showthread.php?t=23737 -but it's a very short thread, so I don't think size is the issue here. Seems to be a forum issue. I know this other thread was displaying photos properly 1-3 days ago but isn't now.

 

[Charles Lamont]

All works for me now

 

[mayhugh1]

It looks like someone may have fixed whatever was going on with the photos in this thread, and so I'll make a short post to test out their fix.
Something that I forgot to mention in my post about the tachometer is that the little dc/dc converter I'm using to generate the 12V supply requires a minimum current draw to insure that the converter starts up and regulates properly. This is a common requirement of these converters, and unfortunately my tach circuitry doesn't quite meet the minimum requirement even though everything seems to be currently working OK. Rather than just waste some of the output current in a resistor, I decided to use it to backlight a fuel level window in my fuel tank.
After adding the metal end caps to the polyethylene tank, I liked the improved look but felt it would look even better if the entire tank was enclosed. I turned another cylindrical aluminum cover for the center section and then slotted it to make a window so I can still see the fuel level. I then installed two series connected high brightness white LEDs in the cover to illuminate the interior of the tank above the window.
Peter's sharp eye is probably going to notice that the center section was made much thicker than the end caps, and I expect he'll ask why. The reason is that the polyethylene tank has a 'key' molded into its outer perimeter, and I used the extra wall thickness to cover it up and keep the finished tank cylindrical. Terry

 

[mayhugh1]

While the soldering iron was still hot, I decided to finish up the electrical portion of this build by constructing the control panel. But first, my 9x20 CNC lathe whose spindle died while I was turning my new prop hub was begging for attention. I had been putting off working on the lathe because the packaging of the Mach 3 interface electronics in this machine was designed with so little regard for maintenance that any attempt to troubleshoot it quickly becomes an exercise in frustration. I ran head into its packaging limitations immediately after purchasing the lathe several years ago when it arrived with an intermittent spindle controller. The vendor insisted it was a Mach software issue and pointed me to ArtSoft. I spent over a week tracing out and troubleshooting the circuitry on the lathe's third party break-out board. The problem turned out to be a flaky relay driver, but what really caught my attention was a pair of miniature PCB relays that I mistakenly thought were operating beyond their limits while controlling the 2kW 220V spindle motor. Every time the power was cycled to that spindle I thought about those little relays. I purchased several spares years ago while they were still available and waited for the originals on the control board to fail. After accumulating probably thousands of spindle cycles during these last two radial engine builds, I guessed the inevitable had finally happened. It was a miserable day getting to and replacing both relays, and then further investigation showed the relays actually had been operating within their specs. The lathe is running again, and so my fingers are crossed hoping one of those relays actually was the culprit.
Getting back to the control panel, though, it's purpose is to give me individual power control over the fuel pump, ignition modules, and the 12V converter that supplies power to the tach and, now, the fuel tank. I like having separate power control of the various components because having it can be an aid to troubleshooting. It's also safer to have the ignition switched off when the fuel pump is on and the engine is being manually primed or when pumping left-over fuel out of the tank. I admit I also like having the switches to play with.
While I was searching eBay for candidate meters for my tachometer several months ago I ran across a miniature 6 volt panel meter for sale by yet another overseas vendor. Fortunately, that purchase went without any drama, and I included it in the panel design. I also packaged the dc/dc converter for the tach inside panel's enclosure.
Although the control panel is basically just a handful of toggle switches, I couldn't resist having some fun with it and over-doing the enclosure. When I was finished with the design it pretty much had to be CNC milled from a block of aluminum. All these components I'm making will hopefully come together in a coherent looking assembly on the back engine's firewall. Next up will be the oil tank. - Terry