X-Virus-Scanned: clean according to Sophos on Logan.com Return-Path: Received: from mail-iy0-f180.google.com ([209.85.210.180] verified) by logan.com (CommuniGate Pro SMTP 5.4.0) with ESMTPS id 5053101 for flyrotary@lancaironline.net; Fri, 15 Jul 2011 15:20:43 -0400 Received-SPF: pass receiver=logan.com; client-ip=209.85.210.180; envelope-from=fluffysheap@gmail.com Received: by iyh42 with SMTP id 42so1294107iyh.25 for ; Fri, 15 Jul 2011 12:20:06 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; bh=iSF9tRNHsT6NJtJdJVT8sIGFnUjuxIPAodNAz8qlWTM=; b=mGIKJln+9D1UP100L2b/aVwB57JrZ/fW+x4GTCRZTP7xdBNwQkQ0AuAGHlMhsmRM1n fhkCyutcRUGYH9bdyXanyEMh2uwCvBtwp58ICKrjC2vUba89qLz4R70lLbMGY29MOliO cnjX4c0zxUWN+p4ISaDf/laCDRW5uS6MHieSc= MIME-Version: 1.0 Received: by 10.231.139.169 with SMTP id e41mr3521819ibu.3.1310757604983; Fri, 15 Jul 2011 12:20:04 -0700 (PDT) Received: by 10.231.156.82 with HTTP; Fri, 15 Jul 2011 12:20:04 -0700 (PDT) In-Reply-To: References: Date: Fri, 15 Jul 2011 13:20:04 -0600 Message-ID: Subject: Re: [FlyRotary] Re: Flex Plate Failure - Improvements From: William Wilson To: Rotary motors in aircraft Content-Type: multipart/alternative; boundary=0016e64610e46e836c04a820899a --0016e64610e46e836c04a820899a Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: quoted-printable I don't think the problem is power, by itself. The auto transmission was designed to accommodate the high-output engine*. Drag racers often use the stock flex plate, of course the duty cycle for drag racing is very different, but they have plenty of power. Comparing the RX-8 to the earlier rotaries, the engine produces less torque= , but it operates at substantially higher RPM. Since flex plates don't generally break in the RX-7 or Cosmo (I have never heard of it happening), the improvements are probably to handle the higher RPM - which could point to either gyro loading or to warping as the problem (but probably not cracking as a root cause, IMO). It seems like holes always have cracks wit= h them and usually it is not a big deal, but when subjected to additional stress - either impact, flexing, thermal or whatever - the cracks can go from microscopic hairline to broken part in relatively short order. I'd be inclined to say the light steel flywheel is the better bet, even independent of cracking (I am not convinced that it would not suffer the same problem). Aluminum flywheels have a reputation for being a little mor= e problematic in the cars. At least in the US, auto transmissions are an also-ran, and with the overall declining parts availability for non-Renesis rotaries, the ability to get brand new flywheels in a variety of weights & materials seems like an advantage, compared to dealing with flex plates. The overall fact that the flex plate is not a weak part in the car but does seem to be a problem here means that there is something going on that is no= t well understood. Understanding it resolves the whole discussion. * Historical note. Mazda didn't design their own auto transmissions, they bought them from JATCO like most other Japanese manufacturers. Since most of the automatic transmissions were sold in Japan, and all RX-7s and Cosmos from 1986 on were turbo in Japan, the auto trans design was actually built for the high output engines, and adapted for non-turbo in the US! This explains why the flex plate is "turbo size," and some other quirks in the 2nd gen cars, like why the auto trans driveshaft in the US has a turbo styl= e U-joint on one end, and a NA style U-joint on the other end. On Fri, Jul 15, 2011 at 12:41 PM, Tracy wrote: > I'm still in analysis mode on this myself. > > So far the possibilities are: > > 1. The flex plate may have been warped when installed ( I bought several > that were) I think the standard should be close to zero runout or wobble= on > the flexplate in view of this failure. > > 2. The flexplate could have had cracks when installed (I also bought > several that had cracks and they were not visible unless you looked close= ly) > > 3. The dang things (RX-7 flexplates) are not strong enough for boosted > 13Bs or 20Bs. > > I do NOT think the problem is fore & aft travel of the input shaft. The > RD-1x drives have had thrust bearings from day one. There is no thrust l= oad > at all on the damper, flexplate or crankshaft imposed by the drive. Thi= s > was not the case with the Ross drive I used at first. > > I have over 1000 hours on the same flexplate on my NA 13B. I have only 5= 5 > hours on the 20B. But I'll be checking both. I did inspect both > flexplates VERY carefully before installing. > > If I found that a change were necessary, I would consider two > possibilities. > > 1. Use a modified aluminum racing flywheel used with an auto > counterweight. (Or possibly a lightweight steel one if there is such a > thing) > > 2. Use an RX-8 flexplate with an adapter plate made of aluminum or steel > sheet to go between the flexplate and damper. The RX-8 version looks > stronger than the RX-7. I briefly studied this when good RX-7 flexplates > were getting hard to find. I did not develop a formal procedure for doi= ng > this but it looked like it would be pretty easy. Anybody with spare time > out there want to take a closer look? Hint: The damper bolt pattern wo= uld > be offset 45 degrees from the RX-8 torque converter bolt pattern. > > Tracy Crook, RWS > > > On Fri, Jul 15, 2011 at 1:47 PM, Bobby J. Hughes wrote= : > >> ** ** ** >> >> Bill summed up what I=92ve been thinking the last few days. I=92m not ch= anging >> anything until ****Tracy**** makes a recommendation. It could be fore an= d >> aft movement on the shaft, flex plate out of round, worn rubber isolator= s or >> a combination of all these items. We may need a scheduled maintenance pl= an >> to replace items like the rubber isolators or other items that wear and >> loosen tolerances.**** >> >> ** ** >> >> ** ** >> >> I will however be inspecting for this problem before most flights. **** >> >> ** ** >> >> Bobby Hughes**** >> >> (playing with fiberglass again)**** >> ------------------------------ >> >> *From:* **Rotary motors in aircraft** [mailto:flyrotary@lancaironline.ne= t] >> *On Behalf Of *Bill Bradburry >> *Sent:* Friday, July 15, 2011 10:03 AM >> *To:* **Rotary motors in aircraft** >> *Subject:* [FlyRotary] Re: Ok Here are the important flywheel >> measurements**** >> >> ** ** >> >> ** ** >> >> David,**** >> >> ** ** >> >> I can=92t help but think that you may be fixing something that is not >> broken. I realize that the flexplate came apart on your plane, but that >> doesn=92t mean that the flexplate is the cause of the problem. That fle= xplate >> holds up quite well in the car with a heavy torque converter attached to >> it. Think of the forces when a car under heavy acceleration changes gea= r! >> As long as the forces are along the plane of the plate it is very strong= . >> Not so much in the forward/rearward directions. In fact that would caus= e >> metal fatigue and cause it to break just like yours did. I think that w= hat >> ever is causing the endplay is what you should be looking at. I suspect >> that if you check out the transmission on the car you will find that the >> endplay of the input shaft is controlled by thrust bearings or some such >> device. ****Lynn**** may know the answer to this??**** >> >> ** ** >> >> That msg from Al Wick really resonates with me. ****Tracy**** had to >> install a thrust bearing to prevent the drive shaft from hitting the bot= tom >> of the pilot bearing and causing an eccentric shaft failure on the engin= e. >> Maybe you now need to add a cage similar to what Al was discussing that = will >> trap the shaft end play completely. That, to me, makes more sense than >> adding a bigger flywheel. If you strengthen the weak spot, you will the= n >> discover where the next weakest spot is because that will be the next >> failure! **** >> >> ** ** >> >> This is a real opportunity for you and the rest of us with the same >> drive. Please study this a little more before you begin the fix!**** >> >> ** ** >> >> Bill B**** >> >> ** ** >> ------------------------------ >> >> *From:* **Rotary motors in aircraft** [mailto:flyrotary@lancaironline.ne= t] >> *On Behalf Of *David Leonard >> *Sent:* Friday, July 15, 2011 12:42 AM >> *To:* **Rotary motors in aircraft** >> *Subject:* [FlyRotary] Ok Here are the important flywheel measurements**= * >> * >> >> ** ** >> >> I finally got to the hangar and measured out my S5 turbo flywheel and th= e >> remains of my S5 flexplate. The results are very encouraging, even as g= ood >> as the possibly could be to make the swap, but I only had ten min at the >> hangar before I had to leave to pick up my daughter from daycare, so the= y >> are one try and slightly sloppy... I make no guarantees. The height >> measurements of the flex plate were particularly suspect because, well, = mine >> is all warped and broken. >> >> *Ring Gear Diameter* - Perfect fit. As we suspected, same size, same >> tooth spacing. >> >> *Ring Gear Fore-Aft Placement* - As measured from the underside of the >> big nut. (for all further reference, aft means aft in the CAR). The fle= x >> plate puts the aft side of the ring gear 1.290 inches from the front of = the >> big nut. The flywheel puts it 0.913. So the flex plate puts the ring ge= ar >> (appx) 0.377 in. further aft. This is actually fortuitous because it ju= st >> so happens that the ring gear thickness is .365 ... So all we have to= do >> to fix the discrepancy is weld and extra ring gears onto to the flywheel >> ring gear and there will be teeth that the starter can reach:(bottom of = the >> page.. the cost $80) >> http://mazdatrix.com/flywheel.htm >> It may also be possible to fix the discrepancy by changing the starter..= . >> will look into that. >> >> *Damper Plate Placement* - On the flywheel, the mounting surface is prou= d >> (or aft) of the ring gear. On the flex plate, the mounting surface is >> recessed (or forward) from the ring gear. And it all pretty much works o= ut. >> That position on the flywheel will vary somewhat depending on which flyw= heel >> you have and how much wear you have on the clutch slip surface. But on m= y (I >> assume well-worn) flywheel the placement was 0.965 while the flex plate = was >> 0.950. Pretty damn close and given measurement error it might me meant = to >> be right on. Or, any difference could be made up by trimming down the >> spline on the damper plate or removing or adding spacers between the two= . >> >> So in my mind, a flywheel is the way to go. Stout, easy to get, pick yo= ur >> weight, only mods seem to be drilling the mounting holes and welding on = a >> second gear ring. >> >> -- >> David Leonard >> >> Turbo Rotary RV-6 N4VY >> http://N4VY.RotaryRoster.net >> http://RotaryRoster.net**** >> > > --0016e64610e46e836c04a820899a Content-Type: text/html; charset=windows-1252 Content-Transfer-Encoding: quoted-printable I don't think the problem is power, by itself.=A0 The auto transmission= was designed to accommodate the high-output engine*.=A0 Drag racers often = use the stock flex plate, of course the duty cycle for drag racing is very = different, but they have plenty of power.

Comparing the RX-8 to the earlier rotaries, the engine produces less to= rque, but it operates at substantially higher RPM.=A0 Since flex plates don= 't generally break in the RX-7 or Cosmo (I have never heard of it happe= ning), the improvements are probably to handle the higher RPM - which could= point to either gyro loading or to warping as the problem (but probably no= t cracking as a root cause, IMO).=A0 It seems like holes always have cracks= with them and usually it is not a big deal, but when subjected to addition= al stress - either impact, flexing, thermal or whatever - the cracks can go= from microscopic hairline to broken part in relatively short order.

I'd be inclined to say the light steel flywheel is the better bet, = even independent of cracking (I am not convinced that it would not suffer t= he same problem).=A0 Aluminum flywheels have a reputation for being a littl= e more problematic in the cars.=A0 At least in the US, auto transmissions a= re an also-ran, and with the overall declining parts availability for non-R= enesis rotaries, the ability to get brand new flywheels in a variety of wei= ghts & materials seems like an advantage, compared to dealing with flex= plates.

The overall fact that the flex plate is not a weak part in the car but = does seem to be a problem here means that there is something going on that = is not well understood.=A0 Understanding it resolves the whole discussion.<= br>
* Historical note.=A0 Mazda didn't design their own auto transmissi= ons, they bought them from JATCO like most other Japanese manufacturers.=A0= Since most of the automatic transmissions were sold in Japan, and all RX-7= s and Cosmos from 1986 on were turbo in Japan, the auto trans design was ac= tually built for the high output engines, and adapted for non-turbo in the = US!=A0 This explains why the flex plate is "turbo size," and some= other quirks in the 2nd gen cars, like why the auto trans driveshaft in th= e US has a turbo style U-joint on one end, and a NA style U-joint on the ot= her end.

On Fri, Jul 15, 2011 at 12:41 PM, Tracy <rwstracy@gmail.com<= /a>> wrote:
I'm still in analysis mode on this myself.

So far the possibilit= ies are:

1. The flex plate may have been warped when installed=A0 ( = I bought several that were)=A0 I think the standard should be close to zero= runout or wobble on the flexplate in view of this failure.

2.=A0 The flexplate could have had cracks when installed=A0 (I also bou= ght several that had cracks and they were not visible unless you looked clo= sely)

3.=A0 The dang things (RX-7 flexplates) are not strong enough = for boosted 13Bs or 20Bs.=A0

I do NOT think the problem is fore & aft travel of the input shaft.= =A0 The RD-1x drives have had thrust bearings from day one.=A0 There is no = thrust load at all on the damper, flexplate or crankshaft imposed by the dr= ive.=A0=A0 This was not the case with the Ross drive I used at first.

I have over 1000 hours on the same flexplate on my NA 13B.=A0 I have on= ly 55 hours on the 20B.=A0=A0 But I'll be checking both.=A0 I did inspe= ct both flexplates VERY carefully before installing.=A0

If I found = that a change were necessary, I would consider two possibilities.

1.=A0 Use a modified aluminum racing flywheel used with an auto counter= weight.=A0 (Or possibly a lightweight steel one if there is such a thing)
2.=A0 Use an RX-8 flexplate with an adapter plate made of aluminum or= steel sheet to go between the flexplate and damper.=A0 The RX-8 version lo= oks stronger than the RX-7. I briefly studied this when good RX-7 flexplate= s were getting hard to find. =A0 I did not develop a formal procedure for d= oing this but it looked like it would be pretty easy.=A0 Anybody with spare= time out there want to take a closer look?=A0 Hint: =A0 The damper bolt pa= ttern would be offset 45 degrees from the RX-8 torque converter bolt patter= n.
=A0
Tracy Crook,=A0=A0 RWS

<= br>
On Fri, Jul 15, 2011 at 1:47 PM, Bobby J. Hug= hes <bhughes@qnsi.net> wrote:

Bill summed up what = I=92ve been thinking the last few days. I=92m not changing anything until Tracy makes a recommendation. It could be fore and aft movement on the shaft, flex plate out of round, wo= rn rubber isolators or a combination of all these items. We may need a schedul= ed maintenance plan to replace items like the rubber isolators or other items that wear an= d loosen tolerances.

=A0

=A0

=A0I will however be= inspecting for this problem before most flights.

=A0

Bobby Hughes<= u>

=A0(playing with fiberglas= s again)

From:<= font face=3D"Tahoma" size=3D"2"> Rotary motors in aircraft [mailto:fl= yrotary@lancaironline.net] On Behalf Of Bill Bradburry
Sent: Friday, July 15, 2011 = 10:03 AM
To: Rotary motors in aircraft
Subject: [FlyRotary] Re: Ok = Here are the important flywheel measurements

=A0

=A0

David,=

=A0

I can=92t help but t= hink that you may be fixing something that is not broken.=A0 I realize that the flexplate cam= e apart on your plane, but that doesn=92t mean that the flexplate is the cause of the problem.=A0 That flexplate holds up quite well in the car with a heavy torque converter attached to it.=A0 Think of the forces when a car under heavy acceleration changes gear! As long as the forces are along the plane of the plate it is very strong.=A0 Not so much in the forward/rearwar= d directions.=A0 In fact that would cause metal fatigue and cause it to break just like yours did.=A0 I think that what ever is causing the endplay is what you should be looking at.=A0 I suspect that if you check out the transmission on the car you will find that the endplay of the input shaft i= s controlled by thrust bearings or some such device.=A0 Lynn= may know the answer to this??

=A0

That msg from Al Wic= k really resonates with me.=A0 Tracy had to install a thrust bearing to prevent the drive shaft from hitting the bottom of the pilot bearing and causing an eccentric shaft failure on the engine.=A0 Maybe you now need to add a cage similar to what Al was discussing that will trap the shaft end play completely.=A0 That, to me, makes more sense than adding a bigger flywheel.=A0 If you strengthen the weak spot, you will then discover where the next weakest spot is because th= at will be the next failure!=A0

=A0

This is a real oppor= tunity for you and the rest of us with the same drive.=A0 Please study this a little more before you begin the fix!

=A0

Bill B=

=A0

From:<= font face=3D"Tahoma" size=3D"2"> Rotary motors in aircraft [mailto:fl= yrotary@lancaironline.net] On Behalf Of David Leonard
Sent: Friday, July 15, 2011 = 12:42 AM
To: Rotary motors in aircraft
Subject: [FlyRotary] Ok Here= are the important flywheel measurements

=A0

I finally got to the hangar and measured out my S5 tur= bo flywheel and the remains of my S5 flexplate.=A0 The results are very encouraging, even a= s good as the possibly could be to make the swap, but I only had ten min at t= he hangar before I had to leave to pick up my daughter from daycare, so they a= re one try and slightly sloppy... I make no guarantees.=A0 The height measurements of the flex plate were particularly suspect because, well, min= e is all warped and broken.

Ring Gear Diameter - Perfect fit.=A0 As we suspected, same size, same tooth spacing.

Ring Gear Fore-Aft Placement= - As measured from the underside of the big nut. (for all further reference, aft means aft in the CAR).=A0 The flex plate puts the aft side of the ring gear 1.290 inches from the front of the big nut. The flywheel puts it 0.913.=A0 So the flex plate puts the ring gear (appx) 0.377 in. further aft.=A0 This is actually fortuitous because it just so happens that the ring gear thickn= ess is .365 ...=A0=A0=A0 So all we have to do to fix the discrepancy is weld and extra ring gears onto to the flywheel ring gear and there will be teeth that the starter can reach:(bottom of the page.. the cost $80)
http://mazd= atrix.com/flywheel.htm
It may also be possible to fix the discrepancy by changing the starter... w= ill look into that.

Damper Plate Placement - On = the flywheel, the mounting surface is proud (or aft) of the ring gear.=A0 On th= e flex plate, the mounting surface is recessed (or forward) from the ring gea= r. And it all pretty much works out.=A0 That position on the flywheel will var= y somewhat depending on which flywheel you have and how much wear you have on= the clutch slip surface. But on my (I assume well-worn) flywheel the placement = was 0.965 while the flex plate was 0.950.=A0 Pretty damn close and given measurement error it might me meant to be right on.=A0 Or, any difference could be made up by trimming down the spline on the damper plate or removin= g or adding spacers between the two.

So in my mind, a flywheel is the way to go.=A0 Stout, easy to get, pick you= r weight, only mods seem to be drilling the mounting holes and welding on a second gear ring.

--
David Leonard

Turbo Rotary RV-6 N4VY
http://N4VY.Rota= ryRoster.net
http://RotaryRoster.n= et



--0016e64610e46e836c04a820899a--