Quarter Scale Merlin V-12

[Blogwitch]

I just can't believe how many anal sphincter twitches are involved in machining bits for this engine

Do you have an attrition rate amongst the parts you are making or are they all made perfect every time ?

If they are all first timers, then my hat is off to you sir, but even if you do screw up occasionally, my hat is still off because of the great progress you have made so far.
I for one would be pulling what hair I have left out.

John

 

[Ken I]

Looks like a good call on the drive pin.

Just a suggestion - in my other hobby - restoring old jukeboxes - I found they used a lot of taper pins.

They work really well as they cause the shaft to swell slightly giving the same sort of interference fit from a slip fit.

They never come loose or work loose. You can also pein the small end which guarantees it wont come out without grinding it off.

This prompted me to use them in some of my industrial machines and they have never failed me (other than in one instance where a toolmaker "grew a brain" decided I was a cranky old fashioned git and used dowel pins - once the taper pins were installed no problem).

Regards,
Ken

 

[dsage]

Terry:

All I can say is OMG.
Your workmanship and attention to every detail is above and beyond.

I certainly hope you are keeping this log safe elsewhere and not just here on HMEM. One never knows how long it will be available here. Certainly no longer than some bean counter figures it's worth it. Your detailed logs and pictures are priceless and very much appreciated. I get a lot of good ideas and encouragement to make MY work better each time I read this.

Thanks

Sage

 

[mayhugh1]

John,
I make mistakes and re-make parts. Even triple-checking isn't a 100% guarantee at my age. My wife kids me all the time about all the prep I sometimes go through to make a single cut. Of course she's occasionally ripping the stitches out of her latest sewing creation while doing so.

Ken,
Thanks for the suggestion on the taper pins. I guess I don't make enough use of them.

Dave,
I do keep a copy of this log for myself. I'm hoping it's useful to others who purchased the castings but found the project too daunting to even begin. Richard Maheu put so much time and effort into such an historically important engine that it would be a shame for his beautiful castings to end up languishing in estate sales.
- Terry

 

[Blogwitch]

Terry,

It was so refreshing to read your reply, there is hope for us all then.

I always think that a notebook and pencil are two of the most critical pieces of tooling when coming to machining a part. Plan everything out first, then cut once, hoping all your calculations are correct. 99 times out of 100, they are.

John

 

[Ken I]

Terry,
But wait there's more......

Another aspect of taper pins is when the angular alignment is critical and for whatever reason you are out by a whisker - you can force the wheel part in the "right" direction whilst reaming the taper - I've fixed up some horrible jukebox cam alignment issues by this brute force method.

Perhaps that's why they used them ?

WTH it works.

Regards,
Ken

 

[bazmak]

I grew up on taper pins.Very expensive to buy and require skill to fit
Time consuming and if I remember cutting to length with recommended projection small end filed chamfer and big end rounded for ease of removal
The rolls Royce of accurate remove/refit precision fittings.I still have a few taper reamers.I remember the days when costs became more important
and roll pins became more fashionable and cheaper.I remember fitting a large camshaft with varios cams,fixtures and fitting all of which had to be able to be removed and replaced accurately.It was a days work but they always fitted
taper pins

 

[Blogwitch]

If I remember correctly, there were three versions of taper pins, Imperial was 1 in 48, Metric is 1 in 50 (and the type I use nowadays) and the third was one used in precision aircraft fitting, which was 1 in 20, and the type I fitted when I was in the industry.
They did have one difference though, it came with a recessed cup washer and with a thread on the end. You reamed until the pin (not thread) was level at the small end, then fitted the cup washer and torque loaded the nut, then the thread, through the castellated nut was drilled through and it was then split pinned. The small taper pins were destroyed on removal, and a new pin had to be fitted each time, larger ones got a second chance as they had dedicated extraction systems.
These were used on torque tube control rods, holding on the bearing ends and anywhere on control systems that reqired a very rigid but removeable fitting, such as holding on helicopter rotor blades etc, bolts were not used as they could easily come undone and their hole tolerances weren't good enough.

Sorry to butt in, but a bit of general info that might be of use to someone sometime.

John

 

[bazmak]

I remember a taper pin with threaded end and a flat,on my pushbike pedals
we called it a cotter pin ?? Before my engineering days, I have seen split pins called cotter pins,which I believe is incorrect

 

[Blogwitch]

Baz,

I think it is just the same thing as split pin, depends where in the world you come from.

But I also use 'cotter' when referring to a straight flat headed pin that has a hole drilled through it at the opposite end to the head that takes a split pin to hold it in position.

Your bicycle cotter pin is in fact a straight pin with an angled flat on it that holds against a flat machined on the main crank pin. Same as here in the UK.
When I was a kid, they were treated like gold dust as everyone seemed to have well worn ones on their bikes, which soon gave you a kerplonk type of motion as the cotter was so worn, it couldn't hold against the crank's flat face.

That is why we have to be careful, someone at the other side of the world just might misunderstand what is going on in a post, purely because of a different phraseology in the same language.


John

 

[mayhugh1]

What I would call the 'main' countershaft in this engine is really just referred to as the 'countershaft' in Merlin parlance. Inside the wheel case, it's located between the main shaft and the starter countershaft. When the engine is running, an integral gear on the countershaft is driven 5X the crankshaft speed by the main gear on the main shaft. A slip-clutch on the rear of the countershaft, behind the bearing plate, will eventually drive the supercharger another 2X the crankshaft. During starting, the starter countershaft drives the countershaft through an over-running clutch.
Since the countershaft will spin some 18k rpm at a maximum engine speed of 3600 rpm, its bearings deserve some consideration. The maximum speed of a ball bearing is limited by several factors including its roller and cage types, the lubricant being used, and its pitch diameter (Dm). Manufacturers empirically spec the maximum speed of their bearings using an NDm factor which is the product of a bearing's pitch diameter and its maximum rpm. The pitch is calculated by averaging the bearing's i.d. and o.d. in millimeters. The single-row, open-cage, oil-lubricated countershaft bearings that I'm using have a typical NDm of 300,000. The pitch diameter of the rear (largest) bearing is roughly 15, and so this corresponds to a maximum speed of 20K rpm. The supercharger itself may rev to 36k rpm, and for it I've ordered a pair of full ceramic bearings.
The recess for the front countershaft bearing in the front flange of the wheel case was bored earlier during the wheel case machining just after the locations of the various shafts were determined. This particular recess left very little supporting material behind the shoulder for its bearing which I wasn't comfortable with. I machined a backup insert and pressed it into the recess behind the shoulder from the front side of the wheel case. This plug will become sandwiched between the front face of wheel case flange and the rear face of the crankcase when the two are assembled. If there were ever to be a revision to the wheel case casting it would be a simple matter to add additional material here.
I started the countershaft fabrication with the machining of its over-running clutch. This clutch was made up from a commonly available one-way bearing and the gear blank I used earlier to establish the distance between the countershaft and the starter countershaft. I've used these bearings in the past to make drill starter adapters for two other engines, and it seems they've always been a bit of a hassle for me to deal with. The outer race of these bearings is a thin-walled shell and is designed to be pressed into an accurate bore in a supporting sleeve. These bearings use the shaft that they ride upon as their inner race, and so the shaft must also be accurately machined and polished.
In both my previous applications I ended up pressing the bearings backwards into their sleeves due to confusion about their required directions for rotation. Actually, I even managed to mis-install one of them twice. Removing these bearings often damages them, and so I purchased a spare for this project - fully prepared to repeat the same mistake. Two errors must have somehow cancelled along the way, though, because I finally succeeded in installing one of them correctly on the first try, I think.
Pressing these bearings into their sleeves isn't at all satisfying, and I'm always left wondering if I created latent damage. The diameter of this bearing's thin outer shell is a whopping two thousandths over its recommended bore size which is a lot of interference for a 5/8" diameter bearing. One end of the bearing has a large beveled lead, and so there may be a preferred but undocumented direction for pressing. The bearing is rather fragile outside its sleeve. If the pressing operation isn't done correctly it can be easily damaged, and so an installation tool is recommended. I made mine with a full end-face contact area and a close fit to the bearing's i.d. If the bore is a bit too small or too large, the sprags may not rotate properly for maximum lock-up. A shrink-fit can be tricky because the bearing is pre-greased, and it contains some plastic cage parts. The cold pressing operation on this particular bearing required everything my two ton press could give, and in my shop that's normally a bright red flag when dealing with parts this small. However, the bearing seemed to function properly after all the dust had settled.
The ring gear for the clutch sleeve was bored for a .0005" interference fit on a machined shoulder on the aluminum bearing housing, and it was augmented with Loctite. Three shallow grooves were machined on the sleeve to collect and spread the Loctite during the pressing operation.
The shaft, itself, was turned from Stressproof 1144. In the past, I've used hardened drill rod for the shafts used with these bearings since the shaft will serve as the bearing's inner race. However, there's always a chance that even a simple shaft will distort during the quench cycle of its heat treatment. Since the countershaft also contains some complex features including an integrally machined spur gear as well as a key slot for its slip-clutch, I decided to use pre-hardened 1144. This material is probably also a good choice for the narrow teeth of the integral 12 tooth gear that will be heavily loaded by the momentum of the supercharger during engine speed changes.
The portion of the shaft that will be in contact the bearing rollers was finish-lapped on the lathe using 600 and 1000 grit paper backed up with a flat metal bar. This lapping was followed up with metal buffing compound. Since the bearing is free to slide on the shaft, a slip-on retainer was machined to limit its aft movement. The stock shaft design had no provision to limit the bearing's forward movement. So, I added a stepped spacer behind the inner race of the shaft's front bearing to prevent the clutch from contacting the outer race of this bearing. The countershaft's rear bearing is retained in the bearing plate, and this stepped spacer also removes the shaft's thrust clearance.
In order to verify the mesh of the countershaft's newly machined integral gear with the driving main gear on the temporary main shaft, I machined the main gear. Unfortunately, although I had a gear cutter for the proper DP and number of teeth, its 20 degree pressure angle didn't match the 14-1/2 degree pressure angle of the cutter used to machine the countershaft gear. Surprisingly, as far as I could tell even under a magnifying glass, the resulting gear pair appeared to mesh beautifully. I'd probably leave well enough alone if it weren't for the 18k rpm at which the 12 tooth countershaft gear is going spin as well as the huge loads that it will experience from the momentum of the supercharger during engine speed changes. So, I've ordered yet another expensive gear cutter. The next step is to machine the rather involved slip-clutch for the rear of the countershaft. Its purpose is to absorb some of the gear train stress that will be created by the supercharger. - Terry

 

[mayhugh1]

The Quarter Scale documentation, initially released in 2003, showed a 32DP 50 tooth gear on the rear of the countershaft for driving the supercharger. In 2007 a drawing for an optional slip clutch was added to the documentation, and in 2013 its design was revised. Gunnar's engine was built according to the original documentation, and a photo of his wheel case clearly shows a simple driving gear. John's documentation is later than Gunnar's, and he included a slip clutch; but it's one of his own design. The Quarter Scale notes explain the purpose of the clutch but also mention that it's optional, untested, and may not be required. I don't know when that particular note may have been added to the documentation, but it's possible that testing eventually showed the clutch was required after all, since a design revision showed up so very late in the project's development.
When the engine speed is changed, the 10x geared supercharger follows, but its angular momentum creates stress on the gear train. The magnitude of this stress is a result of the applied torque which is a product of the rate of change of the speed and the moment of inertia of the supercharger's impeller assembly. Without a clutch in a full-size Merlin, enthusiastic use of the throttle during combat could raise the drivetrain stress to destructively high levels.
One of the highly stressed components in the Quarter Scale version is the 12 tooth gear on the countershaft. In order to achieve the rpm multiplication for the supercharger, the number of teeth on this particular gear was reduced to as few as practicable. Rather than using a separate gear, it was integrally machined into the 1144 alloy countershaft for maximum strength, but its small diameter and few number of teeth still resulted in a tooth profile that was at the lower edge of good design practice.
A slip clutch can absorb the transient energy dumped into the gear train by either the throttle or the supercharger's angular momentum and dissipate it as heat rather than dangerous gear strain. Although adding a clutch to the gear train may seem like a no-brainer, it doesn't come without cost. The clutch needs to be inserted into the gear train behind the starter which means it will add its own angular momentum to the high-revving rear end of the system. Also, its high speed rotating parts are potential sources of imbalance.
The first photos show exploded and sectional views from my SolidWorks interpretation of the clutch design. The provided documentation was an assembly drawing that isn't really clear about some of the hub and spacer details. But, basically the clutch is a steel ring gear sandwiched between a brass backing plate and a flexible brass friction disk. A spacer partially fills a gap between the two and establishes the gripping force on the ring gear. The Quarter Scale drawing lacks the dimensions to determine this gap, and so it must be empirically determined during assembly. Its value, in any event, is dependent upon the particular brass alloy used as well as the surface finish of its various parts.
There are several potential sources of imbalance in the clutch including the TIR of every component, and so the challenge became to work out a machining strategy to minimize them. The narrow widths of the parts greatly complicated their work-holding, but each was indicated-in before every machining operation. Other than zero, I wasn't sure what the final TIR goal should be, but I figured I'd know it when I saw it.
The steel hub was the first component to be machined since I planned to use it as a mandrel to machine the brass frictional parts. The ring gear was the most tedious part to machine because I had not been concerned with its final size and shape when I initially machined its huge starting blank. The numerous holes in both the backing plate and friction disk have several functions. They reduce the momentum of the rotating parts, create spring in the friction plate, and they help air cool the frictional components. A copper alloy might have been a better choice for the frictional surfaces, but soft brass was the closest material I had on hand in the diameter needed.
The spacer was the real PITA part. It had to be very flat and very thin. I wasn't sure exactly how thin, though, because I had no idea about how the breakaway torque might vary with its thickness. The pre-assembly gap that I could measure was evidently 4-5 thousandths different from its assembled value which I couldn't measure. It turned out that only 1-2 thousandths separated a usable spacer from an unusable one. I had considered machining the spacers from thin shim stock and then stacking parts to build up the spacer, but I couldn't figure out how to align their screw holes during assembly. Individual spacers were turned on the lathe and parted off from a pre-drilled Stressproof blank before being surface ground to their final thickness. It took several spacers to zero in on the correct thickness.
I selected a breakaway torque of 1-2 ft-lbs for the clutch as this felt like a safe maximum for the gear train. I cobbled up an adapter for my torque wrench so I could measure the torque for each spacer that I tested. A bit of freshman physics allowed me to estimate the supercharger's maximum angular acceleration while protected by the final clutch. This calculation showed the clutch would limit the supercharger's acceleration from 20k rpm to 30k rpm to about 1 second which probably wouldn't be noticeable on a model.
After machining all the parts, I trial-assembled the clutch on the countershaft so I could measure the TIR. The full ceramic bearings that I ordered for the supercharger had arrived, and so I used them to support the countershaft in a vee-block for the measurement. The TIR was disappointing with an unacceptable .002" measured at the gear teeth. Most of it appeared to come from a .005" lateral wobble created by one or both of the brass parts. Even though they had been turned on the hub used as a mandrel, neither had come out uniform around its center hole. Evidently I had not fully cleaned out the sharp corner adjacent to the hub's flange, and the close-fitting parts distorted when the mounting screws drew them against the flange for machining. I was able to correct the backing plate, but I had to make a new friction disk.
The second assembly looked better. Both the wobble and TIR were less than a thousandth. I was eventually able to find a relative angular position between the backing plate and the friction disk that reduced the errors to less than .0005". Witness marks were scratched into the two disks so they could be reassembled with the same relative orientation.
When all the measurements were completed, I performed a final calculation just to see if the clutch had really been worth the effort required to add it. The computed moment of inertia of the clutch was less than half of what I had estimated for the supercharger, and by definition its angular velocity will be one-half as well. Therefore the angular momentum of the clutch is less than a quarter of that of the supercharger which is a relatively small price to pay for the protection it provides. As will be seen later, the torsional load of the supercharger adds unbelievable complications to the design and machining of the main shaft.
I'm anxious to play more with the ceramic bearings, and so my next step may be to work on the supercharger's bearing assembly. - Terry

 

[mayhugh1]

This post will be uncharacteristically short (for me). Due to the forum's current software issues I can only view it using my iPad, and its interaction with Safari will allow me to upload only a single photo per post. I've tried the HMEM app but lost patience trying to use it. I'm confident, though, the administrators will eventually resolve the problem while we remain patient.
The 14-1/2 degree gear cutter that I needed to machine the new main gear arrived much earlier than expected. This cutter is for the main shaft gear that will drive the integral 12 tooth gear on the countershaft for the supercharger's first 5x rpm multiplication. The 20 degree pressure angle of the main gear that I cut earlier didn't match the pressure angle of its driven integral gear on the countershaft, and so it was used only to verify the center distance between the two shafts.
Since I ended up cutting two 32DP 62 tooth gears using cutters with two different pressure angles, I thought it was worth a photo to compare the tooth profiles of the two actual gears. Compared with the 14-1/2 degree gear, the teeth on the 20 degree gear have a bit more meat down near their root. One of the reasons that 20 degree pressure angles have become popular is that even though they're typically noisier, they can transmit more power.
So, it may seem illogical to have used 20 degree gears throughout the engine and then switch to a 14-1/2 degree pressure angle for the strength-critical 12 tooth gear on the countershaft as well as the main shaft gear driving it. The issue I ran into, though, was that 20 degree 32DP gear cutters seem to be available only in 2-1/4" diameter compared with 1-3/4" for the 14-1/2 degree cutters. The collateral damage created to the inner bearing for the over-running clutch on the countershaft by this larger diameter cutter would have greatly shortened its life if it even functioned at all. The extra tooth strength didn't seem worth the risk especially after I added the protection of the slip clutch. A better solution might have been to make my own small diameter cutter as I did to solve similar problems on my two radial builds. However, the rpm involved in this particular application concerned me since excessive wear created by an imperfect profile might wash damaging debris into the numerous open bearings inside the wheel case.
Although there will be another handful of gears inside the magnetos, the supercharger's driven gear was the last gear machined for use inside the wheel case. Even though it will become part of the bearing cartridge in the supercharger, I was able to verify that it meshed correctly with the slip-clutch by installing a temporary Delrin bushing in the bearing plate. The pair worked smoothly together with the same one degree or so of backlash that I've been using for the rest of the gears.
The next step will be to start wotking on the bearing cartridge for the supercharger. - Terry

 

[Parksy]

Your work always leaves me gobsmacked Terry. Very impressive. I always wonder what occupation a person is or has done and whether such beautiful pieces of work is due to a lifetime of honed skills in the field or whether it is just a hobby.

 

[xpylonracer]

Some great engineering work Terry, also your written description matches the quality of your machining skills, thanks for posting.

Emgee

 

[dalem9]

All I can say is ! WOW !!!! You are amazing .Wish I had a little of your talent .Thanks for sharing !

 

[mayhugh1]

Thanks all for the very nice complements.
Parksy, I was an electronics engineer during my working days, and instead of taking a busman's holiday, I took up machining as a hobby to keep busy after I retired. I spent most of my professional career solving problems using results interpreted through the eyes of various instruments. I think what I enjoy most about machining is that I now get to use my own eyes. - Terry

 

[mayhugh1]

For a number of reasons I decided to design my own bearing cartridge for the supercharger. I wasn't able to locate a supplier for the specified S3K (R6) bearings that were rated for 36k rpm, and I also had questions about a vague note on the supercharger drawing that specified the front seal as a 'piston ring type seal'. I wasn't sure if the note was referring to the use of a cast iron piston ring or a rubber o-ring, but neither sounded like a high speed shaft seal to me. The stock design also didn't include any provision for supplying oil to the bearings nor was the rear of the cartridge sealed, but a comment in the notes warned that bearing lubrication would likely be necessary.
I designed my cartridge around a pair of full ceramic bearings. 'Full ceramic' means that both the races and balls are ceramic compared with 'hybrids' which have ceramic balls and metal races. Because metal races are subject to micro-welding and corrosion, hybrids require lubrication. Full ceramic bearings do not, and they can be used in sealed environments where the presence of oil is an issue such as inside earlier generation hi-rel computer hard drives. Ceramic bearings are not recommended for impact loads even though they've become popular with performance bike and skateboard enthusiasts. They are capable of handling high rpm and temperatures approaching 180C, and so the Quarter Scale's supercharger seemed like an ideal application for them. A downside is their price. The R6 ZR02 bearings that I purchased from VXB.com were close to $90 for the pair, but when compared with the price of the castings their cost was in the noise.
I designed my cartridge to be run with or without oil, and I sealed the shaft running through it with non-contact labyrinth-like seals. Instead of intertwined fingers, the seals in my design are actually parallel steps separated by .003", and so I admit referring to them as 'labyrinths' may be a stretch. The cartridge was designed to prevent the bearings from being flooded with engine oil which is pretty much opposite to its function in a similar full-size application.
Even though the engine and its bearings have been scaled down, the oil viscosity has not. Automotive viscosities generally aren't an issue in model engines and are even beneficial for oil control. This may not be the case for the scaled-down bearings in the high revving supercharger, especially since the oil will likely never warm up enough in typical use to thin out. The bearing heat created by the thick lubricant could create more problems than the lubricant solves.
There's no metal-to-metal contact between any moving parts associated with the cartridge that must be protected with oil except for the slower-running driven gear, and it will receive splash lubrication from the wheel case oil. Since oil isn't required to prevent micro-welding or corrosion in the ceramic bearings, the only uncertainty is whether a light weight oil may be needed to control the bearing temperature. Considering the Quarter Scale's light duty requirements, I don't expect bearing cooling to be required. To be safe, though, I plumbed the cartridge with an oil line that is accessible from outside the supercharger housing. RF techs might recognize the repurposed microwave connector and length of hollowed out rigid coax used to fabricate the oil line. A few drops of lightweight synthetic oil can be occasionally added to the bearings through this line if testing shows it's required.
Construction began with the machining of the housing and included boring it for a close slip fit to the bearings. I went back and forth trying to decide whether it would be best to use aluminum or steel for the housing. Aluminum would have some advantage in removing heat from the bearings, but steel would be a better choice for dampening high frequency vibrations created in the thin-wall supercharger housing by imbalances in the rotating assembly. In the end, I opted for a hefty chunk of 12L14. The end cap, combined with the supercharger's driven gear is part of the rear labyrinth and was machined from aluminum.
I lapped a pair of aluminum spacers for near zero interference fits between the inner and outer bearing races. I may increase this interference to one or two thousandths later, when the impeller is added, if I measure any radial 'loosening-up'. I initially machined Delrin spacers with slight compression fits but scrapped them after researching Delrin's extremely high temperature coefficient of expansion.
The ceramic bearing races seem significantly looser than what I'm accustomed to with typical metal bearings, but measurement showed that most of this is an illusion created by their lack of lubricant. I inserted a temporary shaft through the assembled cartridge, and with the shaft resting between a pair of v-blocks I was able to measure only a tenth or so of run-out on the housing as it was rotated on the shaft. The friction was low enough that the assembly always settled with the gear cutout in the aluminum end-cap facing up. I estimated the cutout material to be less than a half percent of the total rotating mass; and so without oil, the friction may be low enough that I'll be able to do a reasonable job statically balancing the impeller assembly.
The shaft was also machined from 12L14. My original plan was to separate the labyrinth surfaces by only .001". I wasn't able to adequately control the thrust clearance of the shaft, though, and so I relaxed the spacing to .003". The shoulder washer for the driven gear was ground to maintain a .001" clearance between the rear face of the gear and the inner race of the rear bearing with the end bolt in place. The gear is keyed to the shaft to prevent it from spinning. This was done to avoid adding stress to the inner races when the bolt was tightened. The final thrust clearance of the shaft, and therefore eventually the impeller, came out to be just over a thousandth.
The supercharger documentation called out the rear of the shaft to be splined to the impeller similarly to what was done on the full-scale Merlin. A drawing was supplied for a 30 spline shop-made broach to spline the impeller. I could envision a lot of things going wrong with a little side project like that, including ruining an irreplaceable impeller casting. Instead, I milled a pair of slots on the rear end of the shaft for 1/16" square keys. A third slot was milled for a similar key on the opposite end of the shaft for the driven gear.
Although I probably should get back to the wheel case and start working on the dreaded main shaft, I may next machine the impeller so I can finish up the supercharger. - Terry

 

[Ripcrow]

Piston ring seals are used in turbos on the turbine side at up to 150,000 revs if your seal design doesn't come up to your requirements.

 

[mayhugh1]

I continued on with the supercharger since it's so close to being finished, and I began by cleaning up the impeller casting. The impeller was cast with a fairly long top-end spigot that I chucked up in the lathe in order to face its bottom surface. I indicated off its rough bottom and searched for an orientation that minimized the wobble created by the spigot. The best I could do was .010", and since I was planning to leave the metal between the fins as-cast, the impeller's edge thickness would end up with this same variation. Although it wouldn't be visible after assembly, it would create balance issues would have to be dealt with. In hindsight, it was a rookie's mistake to not have first cleaned up the spigot, but I had become so complacent with the accuracy of the circular features in these castings that I got sloppy.
I removed just enough material from the impeller's bottom surface and o.d. to true them up. I then bored its center for the shaft, but after cutting off the excess spigot I saw that its interior was full of soft porous metal. Although most of it was discarded with the spigot, some of the bad metal continued on down about a quarter inch into the impeller.
I installed the bearing cartridge on the front half of the supercharger housing and slid the partially finished impeller on its shaft. As mentioned earlier, the races in these ceramic bearings are quite a bit looser than those in typical metal bearings, and not all of it's due to a lack of lubricant. With the zero pre-load spacer I was using between the inner races there was enough slop in the bearings so the outer edge of the three inch diameter impeller could be nudged over enough to kiss the floor of the housing just .007" below its bottom surface. A .002" longer spacer added a bit of bearing preload and appeared to solve the problem.
At this point I took a couple days off to have five stints implanted into two major vessels in my heart. Last week my cardiologist saw an EKG anomaly during my annual stress test, and he ordered a heart catherization to look into it. Six hours after walking into the hospital for the test, I had been re-plumbed. Twenty-four hours after that I was back in my shop as though nothing had happened. It was a really disappointing surprise since I had no symptoms, have always been careful about what I eat, and have been vigorously working out in a gym four days a week for the last fifty years. Even now, after the procedure, I physically don't feel any different. I have a poor family history, but I guess we all deal with the genetics we've inherited.
Even though the machined bottom of the partially finished impeller ran true inside its housing, the variation in the impeller's edge thickness was a growing irritation I decided I couldn't live with. I was also concerned about the shaft keys working loose in the impeller's porous central interior. So, I decided to stint the impeller and fix both issues.
I set up the impeller on the mill and was able to shim out all but a thousandth or so of the impeller's edge variation. I then bored out the bad metal and replaced it with a steel bushing. Using the same lathe setup I turned the bushing's o.d. for a press fit in the impeller and its i.d. for a close slip fit over the shaft. The bushing was turned extra long so that after installing it I would have a true spigot so I could finally finish machine the bottom of the impeller.
Machining the impeller's bottom was the easy part. The hard part was profiling the rest of the impeller for a close fit to the contour of the supercharger's volute in the rear half of the housing. The drawing spec'd a maximum .010" clearance to the rear half of the housing over the entire profile, but I decided to increase this to .015". I figured the bit of extra clearance wasn't going to make the difference between boost and no boost, while the extra margin would probably be appreciated down the road.
After sawing off the excess spigot, I mounted the impeller on the mill table and machined the top of the newly bushed center to its finished height. I then installed the impeller on a mandrel in my manual lathe and brought the tops of the inlet fins to their finished height and angle. The inlet fins extend into the inlet opening in the rear housing, and their top profile was just a simple 10 degree angle. The reason for finishing the central end of the impeller on the mill before finishing the inlet fins on the lathe was because of the acute reverse angle of the fins. In order to reduce the chances of the lathe tool grabbing and damaging them, I used my last narrow insert designed for non-ferrous metal; and I didn't want to take a chance on it being damaged by the steel bushing.
The remaining profile was not only a bit complex, but being underneath the rear housing made its clearance difficult to determine. A profile for the impeller was provided on the supercharger's drawing, but it was only a starting point, and the clearance would have to be planned for my unique assembly.
I modeled the provided profile in SolidWorks and then printed it out on paper as an actual size cross-section. This allowed me to trial fit the pattern to the contour in the rear housing. Of course, the two were significantly different, and so the fitting process became an iterative combination of changes to the profile as well as to the surface inside the housing. I filed, sanded, and polished the interior of the housing until I had an acceptable match to one of my paper profiles. I then compiled a program for my 9x20 lathe to turn the profile onto a trial workpiece so I could check its fit in the supercharger housing. I used Plastigage to measure the clearance between the installed dummy impeller and the housing. I had to extrapolate the Plastigage measurement to .015" since what I had on hand would measure a maximum clearance of only .009". Once I was happy with the trial workpiece, I used the same program to cut the exact same profile into the actual impeller. The machined impeller and its clearance nicely matched those of the trial workpiece.
Only the diffuser section ring and a bit of mounting hardware remain to be machined before I can finally assemble the supercharger. - Terry