Return-Path: Received: from [24.25.9.101] (HELO ms-smtp-02-eri0.southeast.rr.com) by logan.com (CommuniGate Pro SMTP 4.2b7) with ESMTP id 310215 for flyrotary@lancaironline.net; Sat, 10 Jul 2004 08:37:33 -0400 Received: from EDWARD (clt25-78-058.carolina.rr.com [24.25.78.58]) by ms-smtp-02-eri0.southeast.rr.com (8.12.10/8.12.7) with SMTP id i6ACaUNs009822 for ; Sat, 10 Jul 2004 08:36:31 -0400 (EDT) Message-ID: <001501c4667a$8942dbc0$2402a8c0@EDWARD> From: "Ed Anderson" To: "Rotary motors in aircraft" References: Subject: Three candidates for Turbo Failure Date: Sat, 10 Jul 2004 08:36:34 -0400 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_0012_01C46659.01F80350" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1409 X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1409 X-Virus-Scanned: Symantec AntiVirus Scan Engine This is a multi-part message in MIME format. ------=_NextPart_000_0012_01C46659.01F80350 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable I agree Dave. =20 This is of course all speculation about what actually happened to John's = turbo. But, at the moment, probably about the best we can do. Having = two failures that soon and with the same results - loose compressor = wheels - certainly indicates its probably just not a random factor, but = has some underlying common cause. I agree if the compressor maps we are = looking at are even close to the Mazda compressor parameters, it is = unlikely that John was in the surge zone. If he had been according to = what I have read, there would have been indications like rapidly = fluctuating pressure surges and sound and vibration associated. = Although with the engine at the back the sound may not have been heard = and the vibration could have been dampened - which a composite construct = is good at. So, apparently no evident to support the surge hypothesis. It does appear that whatever is causing it, the stress appears to be = concentrated on the compressor wheel. Surges stress the entire rotating = assembly but does particularly stress the compressor wheel. I need to = sit down again and calculate the Pressure ratio and flow on the = compressor chart after adjusting for the pressure ratio creep caused by = altitude. But, it would seem to take a considerable increase in = pressure ratio to move him into the surge zone. The worst case = situation for surge appears to be in boost with low mass flow, in other = words operating toward the left on the mass flow axis of the chart. The = further over you are the more like to enter it. But again - no = indication that is happening. Overheating could indeed weaken the metal, however, its my opinion that = with the water cooling - overheating should not be a problem. I mean = almost all aircraft turbos I am aware of do not have water cooling and = survive for much longer, hundreds perhaps close to a thousand hours. So = you would think that with water cooling the Mazda turbo would do even = better - if overheating were the cause. Can't rule it out, but I would = put it further down the list. My prime candidate of the three would be overspeeding of the rotating = assembly.=20 One of the main reasons I lean toward overspeeding is that the turbine = housing of the Mazda turbo is designed and tailored for low engine rpm = boost. The Mazda SAE papers made a point of that. Higher engine rpm = would increase the exhaust mass flow which would tend to spin the = assembly even faster. Operating at altitude with the pressure ratio = creep also increases the speed required of the rotating assembly to = maintain the same differential boost pressure. The twin scrolls = driving the turbine wheel are separated with one having a waste gate and = the other not. That means the one scroll never has its exhaust gas = diverted by the waste gate. Again a design feature to promote boost at = low exhaust mass flow. Then as I have speculated, I believe there is a = chance that the BOV can add to the risk of overspeeding. So given those = factors and possibilities my assessment would lean toward overspeeding. However, that being said, I would expect to see damage that would = support that assessment. If no damage or little rotational damage that = would seem to indicate low rotation speed when the wheel separated. So = if overspeeding were the cause, I would expect to see the compressor = wheel and housing fairly well beaten up. =20 The turbine wheel is generally friction welded to the shaft whereas the = compressor wheel is held on with a nut. I believe John indicated in one = case the wheel was just laying in the compressor housing and the in last = case I believe he indicated it was wedged in tightly. I don't recall = how the nut is secured (other than torque). I'll have to go down to the = shop later and take a look. =20 I don't recall John mentioning how damaged the compressor wheel was. I = would assume that if it separated from the shaft while spinning at = 80,000 + rpm it would really tear up the wheel and housing. On the = turbine side I have seen photos where the wheel actually tore through = the cast iron housing as hard as that might be to believe. =20 John do you have any photos of the wheels and housing after the = incidents?? I am still looking at the BOV and any possible role it may play. I = still appears to me that suddenly releasing the back pressure across the = compressor wheel when the BOV opens could open the possibility that the = compressor could be stressed. If the back pressure across the wheel = suddenly drops and lessens the load while the turbine end is still = getting the boost mass flow that there is the possibly of overspeeding = the rotating assembly.=20 I, of course, could be wrong and the BOV effects not only be benign but = actually helps prevent damage in an aircraft operating environment as = you and others have suggested. If aircraft turbo systems had blow = off valves then we would have another data point, but of the several = aircraft systems I looked on the internet had either sized the A/r of = the turbine housing or use a waste gate type of bleed for the exhaust.=20 Like you say it could be a combination of factors, but I do agree with = the three you mention as to being the most likely culprits. Ed Ed Anderson RV-6A N494BW Rotary Powered Matthews, NC ----- Original Message -----=20 From: DaveLeonard=20 To: Rotary motors in aircraft=20 Sent: Saturday, July 10, 2004 12:42 AM Subject: [FlyRotary] Re: BOV more/less air flow Ed, Great write up! In general I very much agree with you on all = points but I want to free flow some thoughts that may modify people's = feeling about a BOV. I am even less of a turbo expert than you, but = here are some thoughts to consider. First, I think we have to figure out exactly what is causing John's = turbo's to fry. IMHO we have 3 general candidates 1) overheating, 2) = turbo overspeading, or 3) compressors into the surge region. While it = is possible that there is a combined cause, and probable that more than = one factor plays a partial role, most likely ONE factor is primarily = responsible for JTD (John's Turbos' Demise). IMHO that one factor is overheating. We know that heat is a factor = while according you the map you posted we are probably not in surge or = overspeed region most of the time. Also, turbo overspeading and = compressor surge are likely to present with very different symptoms like = surging pressures or a disintegrated turbo. In contrast, overheating = weakens the metals involved and would behave much like what John = Describes. In this case it seems to be turbine axle and possibly the = blades that are not up to the task. There are 2 ways to improve the situation. Reduce the heat, of course, = and reduce the forces acting on the metals. The BOV can be very = instrumental in reducing the force acting on that axle, and possibly = also slightly reduce the heat transfer. My BOV (which I assume is = fairly standard) is not an on/off valve, but a spring loaded valve that = will open more or less depending on the pressure in the plenum, the = ambient pressure, and the pressure differential across the throttle = body. There is no way it will instantly drop the pressure = significantly. In normal operation it would never go from full closed = to full open suddenly, but rather just start leaking more and more air = as the combination of those three pressures changes. The only time it = would suddenly open is when the throttle is suddenly closed. While on one hand it does seem very inefficient to compress air only = to have it bleed out before entering the engine, I believe it is far = more inefficient and potentially harmful to have the compressor = constantly fighting against a large pressure differential. That = pressure differential is the largest determinant of the force on the = axle. Further, if the turbine wheel is free to turn a little easier (faster) = it places less back pressure on the exhaust and less heat will be = transfeffed. It is true that the faster RPM increases the radial = stresses on the blades, but radial stresses are tolerated much better = than the axial stresses caused by high back pressure. (Since I have = decided that heating is the problem and not overspeeding.) There is no doubt that we both agree on the BEST solution: a larger = turbo with an even larger waste gate placed perhaps a little further = from the engine. But until I change my set-up to fit such a turbo and = waste gate I think the BOV is a helpful addition. Further, the BOV does provide some measure of protection against over = boost like what happened to Todd. Dave Leonard ------=_NextPart_000_0012_01C46659.01F80350 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
I agree Dave. 
 
This is of course all speculation about = what=20 actually happened to John's turbo.  But, at the moment, probably = about the=20 best we can do.  Having two failures that soon and with the same = results -=20 loose compressor wheels - certainly indicates its probably just not = a=20 random factor, but has some underlying common cause.  I agree if = the=20 compressor maps we are looking at are even close to the Mazda compressor = parameters, it is unlikely that John was in the surge zone.  If he = had been=20 according to what I have read, there would have been indications like = rapidly=20 fluctuating pressure surges and sound and vibration associated.  = Although=20 with the engine at the back the sound may not have been heard and the = vibration=20 could have been dampened - which a composite construct is good at.  = So,=20 apparently no evident to support the surge hypothesis.
 
It does appear that whatever is causing = it, the=20 stress appears to be concentrated on the compressor = wheel. =20 Surges stress the entire rotating assembly but does particularly stress=20 the compressor wheel.  I need to sit down again and calculate = the=20 Pressure ratio and flow on the compressor chart after adjusting for the = pressure=20 ratio creep caused by altitude.  But, it would seem to take a = considerable=20 increase in pressure ratio to move him into the surge zone.  The = worst case=20 situation for surge appears to be in boost with low mass flow, in other = words=20 operating toward the left on the mass flow axis of the chart. The = further over=20 you are the more like to enter it.  But again - no indication that = is=20 happening.
 
Overheating could indeed weaken the = metal, however,=20 its my opinion that with the water cooling - overheating should not be a = problem.  I mean almost all aircraft turbos I am aware of do not = have water=20 cooling and survive for much longer, hundreds perhaps close to a = thousand=20 hours.  So you would think that with water cooling the Mazda turbo = would do=20 even better - if overheating were the cause. Can't rule it out, but I = would put=20 it further down the list.
 
My prime  candidate of the = three would be=20 overspeeding of the rotating assembly. 
 
One of the main reasons I lean toward = overspeeding=20 is that the turbine housing of the Mazda turbo is designed and = tailored for=20 low engine rpm boost. The Mazda SAE papers made a point of = that. =20  Higher engine rpm would increase the exhaust mass = flow which=20 would tend to spin the assembly even faster.  Operating=20 at altitude with the pressure ratio creep also increases the = speed=20 required of the rotating assembly to maintain the same differential = boost=20 pressure.   The twin = scrolls driving=20 the turbine wheel are separated with one having a waste gate and the = other=20 not.  That means the one scroll never has its exhaust gas diverted = by the=20 waste gate. Again a design feature to promote boost at low exhaust = mass=20 flow. Then as I have speculated, I believe there is a chance that the = BOV can=20 add to the risk of overspeeding.  So given those factors and = possibilities=20 my assessment would lean toward overspeeding.
 
However, that being said,  I would = expect to=20 see damage that would support that assessment. If no damage or little = rotational=20 damage that would seem to indicate low rotation speed when the wheel=20 separated.  So if overspeeding were the cause, I would expect to = see the=20 compressor wheel and housing fairly well beaten up. 
 
The turbine wheel is generally friction = welded to=20 the shaft whereas the compressor wheel is held on with a nut.  I = believe=20 John indicated in one case the wheel was just laying in the compressor = housing=20 and the in last case I believe he indicated it was wedged in = tightly.  I=20 don't recall how the nut is secured (other than torque).  I'll have = to go=20 down to the shop later and take a look. 
I don't recall John mentioning how = damaged the=20 compressor wheel was.  I would assume that if it separated from the = shaft=20 while spinning at 80,000 + rpm it would really tear up the wheel and=20 housing.  On the turbine side I have seen photos where the wheel = actually=20 tore through the cast iron housing as hard as that might be to=20 believe.  
 
John do you have any photos of the = wheels and=20 housing after the incidents??
 
I am still looking at the BOV and any = possible role=20 it may play.  I still appears to me that suddenly releasing the = back=20 pressure across the compressor wheel when the BOV opens could open the=20 possibility that the compressor could be stressed. If the back=20 pressure across the wheel suddenly drops and lessens the load = while=20 the turbine end is still getting the boost mass flow that there is the = possibly=20 of overspeeding the rotating assembly. 
 
 I, of course, could be wrong and = the=20 BOV effects not only be benign but actually helps prevent damage = in an=20 aircraft operating environment as you and others have = suggested.  =20   If aircraft turbo systems had blow off valves then we would have = another=20 data point, but of the several aircraft systems I looked on the = internet=20 had either sized the A/r of the turbine housing or use a waste gate type = of=20 bleed for the exhaust.
 
 Like you say it could be a = combination of=20 factors, but I do agree with the three you mention as to being the most = likely=20 culprits.
 
Ed
 
Ed Anderson
RV-6A N494BW Rotary Powered
Matthews, NC
----- Original Message -----
From:=20 DaveLeonard=20
Sent: Saturday, July 10, 2004 = 12:42=20 AM
Subject: [FlyRotary] Re: BOV = more/less=20 air flow

Ed,=20 Great write up!  In general I very much agree with you on all = points but=20 I want to free flow some thoughts that may modify people's feeling = about a=20 BOV.  I am even less of a turbo expert than you, but here are = some=20 thoughts to consider.
 
First, I think we have to figure out exactly what is causing = John's=20 turbo's to fry.  IMHO we have 3 general candidates 1) = overheating,=20 2) turbo overspeading, or 3) compressors into the surge region.  = While it=20 is possible that there is a combined cause, and probable that more = than one=20 factor plays a partial role, most likely ONE factor is=20 primarily responsible for JTD (John's Turbos'=20 Demise).
 
IMHO=20 that one factor is overheating.  We know that heat is a factor=20 while according you the map you posted we are probably = not in=20 surge or overspeed region most of the time.  Also, turbo = overspeading and=20 compressor surge are likely to present with very different symptoms = like=20 surging pressures or a disintegrated turbo.  In contrast, = overheating=20 weakens the metals involved and would behave much like what John=20 Describes.  In this case it seems to be turbine axle and possibly = the=20 blades that are not up to the task.
 
There are 2 ways to improve the situation. Reduce the heat, = of course,=20 and reduce the forces acting on the metals.  The BOV can be=20 very instrumental in reducing the force acting on that axle, and = possibly also=20 slightly reduce the heat transfer.  My BOV (which I assume is = fairly=20 standard) is not an on/off valve, but a spring loaded valve that will = open=20 more or less depending on the pressure in the plenum, the ambient = pressure,=20 and the pressure differential across the throttle body.  There is = no way=20 it will instantly drop the pressure significantly.  In normal = operation=20 it would never go from full closed to full open suddenly, but rather = just=20 start leaking more and more air as the combination of those three = pressures=20 changes.  The only time it would suddenly open is when the = throttle is=20 suddenly closed.
 
While on one hand it does seem very inefficient to compress = air only to=20 have it bleed out before entering the engine, I believe it is far more = inefficient and potentially harmful to have the compressor constantly = fighting=20 against a large pressure differential.  That pressure = differential is the=20 largest determinant of the force on the axle.
 
Further, if the turbine wheel is free to turn a little easier = (faster)=20 it places less back pressure on the exhaust and less heat will be = transfeffed.  It is true that the faster RPM increases the radial = stresses on the blades, but radial stresses are tolerated much better = than the=20 axial stresses caused by high back pressure.  (Since I have = decided that=20 heating is the problem and not overspeeding.)
 
There is no doubt that we both agree on the BEST solution: a = larger=20 turbo with an even larger waste gate placed perhaps a little further = from the=20 engine.  But until I change my set-up to fit such a turbo = and waste=20 gate I think the BOV is a helpful addition.
 
Further, the BOV does provide some measure of protection = against over=20 boost like what happened to Todd.
 
Dave=20 Leonard
 
------=_NextPart_000_0012_01C46659.01F80350--