While doing my engine build I'm trying to measure as many parts as I can on our CMM both new and old to get any information I can. Below is a scan of a used rotor bearing on the side opposite of the gear. I did scan on both sides of the bearing to get taper and x/y centers and stuff but the second scan went screwy on me and I don't know when I'll be able to do another one. So here it is I'm sure some of you guys will find this interesting:



I'm sure anybody who has looked at a badly used one has noticed wear spots facing the combustion chambers. In the plot the big spike at the top is the puzzle seam of the bearing which was facing an apex seal. You can clearly see 3 spiked areas about 30 deg appart. On this side of the bearing the total roundness was already way in the hell out while the side closest to the rotor gear was much worse which didn't make any sence to me. I'm guessing that the spiked area farthest right was probably the chamber making the most power or made the most power the longest followed by the one at the bottom. The spike all the way to the left looks pathetic and I'm guessing went to hell first.

Also when I baught the car it had about 80 psi exactly through both rotors. I have some scans of the RX-8 stationary gears and they are extremly precision made.

To maintain confidentiality of our equipments accuracy the measurements have been truncated.

 

ok. Pretend I've never seen that before. can you tell me what the red circle is and what the screwy circle is also? I want to asume that the red is the inner part of the bearing and the black one is the outer part (against the gear?)....

so that's the rotor bearing which holds the e-shaft.

what's the Puzzle seam?

the three spiked areas at 30* are where the e-shaft wears out the bearing? or is it the gear?

what does it mean when you say: " On this side of the bearing, the total roundness was already way in the hell out?"

pics would be awesome.

 

The red circle is what the bearing is suppose to be. The center is determined by the X/Y origin I setup on the machine. I did not call out the exact coordinates to know how far out of center it was.

Roundness is how much of a perfect circle it is, so 0.0 roundness would be a perfect circle. If it is .001" out of round that means the diameter fluxuates .001".

The bearing runs through the center of the rotor, if you have it sitting flat on its side you have one side with a gear which rides around the stationary gear. The other side has veins for oil and cooling. The bearing is also one piece but it has a seam in it which is not a straight line, it has puzzle like teeth which interlock.

I measured this with the rotor fitting flat on the granite block and the gear inside the rotor facing up. The scan then followed around the part of the bearing closest to the table and then the part near the top. Any time you are building an engine you need to measure your journals and bearings like this to find taper which is how cone shaped it is. You may measure in one spot but that does not mean the whole journal or bearing is the same as that measurement.

The scan above is only the bottom scan of the bearing as it sat on the table the top scan went all screwy on me so I could not use it as a measurement.

The seam of the bearing was facing an apex seal. Rereading my post I said 30 degrees and what I was thinking was 3 evenly spaced, sorry. You really need to know what a rotor looks like in order to get the idea of this. The apex seals are 120 deg appart with the combustion chambers in the center. So looking at the graph the very top was an apex seal. Going clockwise down to 3 o'clock was a combustion chamber. About 4:30 was another apex seal with a combustion chamber at 6 O'clock another apex at about 7:30 a combustion chamber a little above 9' oclock.

When I looked at the bearing I noticed it had 3 badly worn spots which aligned with the combustion chambers which makes sense because that would be the higest load point is when the chamber fires. The little sqiggle marks comming away from the rest of the line is where the bearing is worn. The farther the marks get from the center the more badly that area is worn. So at 3 & 6 oclock in the diagram it has sqiggle marks that are far away from the red line. Since these areas are badly worn the where the stronger chambers and put more load on the bearing. Up at about the 8 oclock position you see the other marks and they are very faint. This to me looks like that particular chamber was weak for some reason, it either didn't make the same amount of power as the other ones or the seals worn out before the others or something.

 

oh. awesome explanation. Thanks.

what else are you measuring? I'm interested in what you find.

 

Everything I can get time to do that a critical or hard to get measurement needs to be taken.

 

I've already scanned the front and rear RX-8 stationary gears I'm installing. The eccentric shaft pushes the limits of the CMM because you have to measure the journals in two parts and merge the measurements which doesn't work so well. If we get our roundcom before I start assembling I'll measure it on that.

I need to install new bearings in both rotors so I want to measure those after they are installed, corner seal hole roundness, depth and location. I also want to do the apex seal grooves so that I'll have a measurement accurate to atleast a tenth so that I can lap down the new apex seals to precise clearance. I've already checked the rotors for warpage and flatness and both are are only about .0004-.0005" so they are plenty straight. I'm not to worried about scanning the rotor housings because I can check them for flatness on the granite table using a height gauge. I do want to scan the insides which will be complicated because in order to get a graph I'll need to measure most of it sections of circles. I want to do this to mainly compare to results I'll get later when I take the engine out at about 40,000 miles ore so.

I may try to do some other stuff but I don't see what else I could possably measure other than maybe the flatness of the steels where the rotor seals make contact but with a straight edge I cannot see light and can't fit a 1 thou feeler gauge through so they are good.

 

I wonder if people will stop believing the that you should not use NEW bearings with a rebuild, and it takes X thousand miles for breakin blah blah..
Seems to me rotaries are about 40-50 years behind piston motors, one of the reasons being r&d isn't widely done, and the few who do, keep it secret.
This is the first time I've ever seen a CMM of a rotor bearing, very interesting. Its interesting what you can start learning with that information. This is pretty common when it comes to rod design in a piston car.

You did that scan with the bearing still in the rotor right?

It would be neat to still do your rotor housings if you could measure the sleeve itself, like mapping the bores on piston cars. Show what a " good condition " used housing looks versus a new one. Again kinda funny, its pretty common in the piston world that you don't install new rings without a hone job, but its super common here...

 

Yes that was with the bearing still installed in the rotor as it came out of the engine. That side of the bearing to the naked eye actually looked ok. The other part where the scan messed up had copper babbit showing through. If the good side was out by more than a thousanth then the bad side is probably out by 2 or 3 thousants. If I reuse that bearing I'm pretty sure it will eat the new apex seals and the rotor housings all to crap very shortly.

I still need to scan the used stationary gears that came out of the engine, I should be able to get both of those in at lunch if we don't have something running. Those only take about 5 min each to setup and run although I do need to setup a scan of the flange on each instead of just taking 10 or 12 points.

 

Oh, that big notch where the seam is, that actually measured down one of the puzzle piece intersections, if did the scan a little higher it would have been a much narrower notch. I think it would be a good idea to burnish the seam with some newspaper or some 1500 grit sandpaper in the direction of rotation since it was obvious a lot of metal in that area rubbed off.

 

Ok, this doesn't do much good for studying the wear of used parts but it does show some of the accuracy of mazda's new parts. Below is some points taken from an RX-8 front stationary gear flange. This graph shows how flat the flange is on the surface that mates to the steel, I didn't do a full scan because I didn't have the time to set it up. I measured this to make sure that the flange would mount perfectly straight, of coarse I would also have to measure the mounting surface on the steel to know for sure.

 

Ok here is some more information, I'm not going to scan the whole report but I will list some numbers. This was with the scan of the front RX-8 stationary gear. I'll be posting the RX-8 stuff in other boards later but for those installing in the RX-7 here we go. The gear was sitting flange down on the table.

Bearing Diameter
1.6941" Top
1.6941" Bottom
.0007 roundness (some lint may have interfered with the scan)
No bearing taper.

For all practicality the bearing was perfectly centered in the part.

The machined part right above the flange which is a tight fit into the front steel:

Diameter:
2.5984" Top
2.5986" Bottom
Roundness < .0001"
Taper .0002

 

to me that picture hardly looks accurate.

 

That is at 500x magnification and as said in my first post, I trucated the full measurements because the accuracy of our equipment is confidential, even if you don't know who's equipment it is. Although anybody who knows anything about a CMM should know how accurate they are.

I'm sure if I scanned it with 2,000 points it would be more accurate but I hardly saw the point on a flange when I had limited time to write the program.

I highly doubt that if I scanned it the flatness would go above .0003" which I think is plenty flat. I think using a cheap torque wrench or the inaccuracy of ones skill with a torque wrench would cause more problems than that.

 

Any chance you could press the bearing out and scan the bore of that same rotor?
Keep it up! I would love to play with that machine all day...

 

Yea, if I can get enough time on the machine, I'll definatly be putting new bearings in so if I can get another program written I'll scan the bore on each. Before I do though I need to see if I can get the top of that one bearing scanned as I really want to see a measurement on how badly it is worn. If need be I can measure it with a dial bore gauge I guess.

 

Very interesting to see where the wear occured. Obviously the (0,0) point of the (X,Y) axis is not set where it should be, it looks like the red circle should be shifted down and to the right.

 

I am a firm believer in new bearings, but could you elaborate on the break in? I always though you should extend the break-in time with new bearings, but it sounds like you disagree. Or did I just mis-read your post?
Also, those results are very interesting. I'm looking forward to your future scans. I would love to see the results of new vs. used housing scans, but I'm not sure how hard it would be to set up.

 

looking back at Karack's post I think he was refering to the runout on the bearings. The runout is off because I did the scan over the seam, I wanted to see how much the seam is in or out. I will try to do another scan that will not go a full 360 deg and skip the bearing seam. I'll try post that here soon.

 

Its just really old school.

 

Well good news our director of operations said as long as a manager or our head engineer is at the shop I can use any of the machinery including the CMM which means I should be able to do stuff on saturdays and occationally a sunday. The only problem is we are working mandetory saturdays until after Daytona.

 

I CMM'd some rotors a while back to get the side seal radius. Here's a scan.

 

I should have noted that the first entry is centering on the rotor bearing diameter. The next is the coner seal at the bearing notch at 12:00. The next two corner seals are then given. All with circularity (ring) and size.

I placed a side seal in groove and it stayed protruded and held by it's pwn springyness. Beeped it and wrote the results down by hand. I only did one seal because I was writing a side thinning program and it would be close enough to guess that the others would be as good as what I had measured already.