Play in turbo shaft - disastrous or live with it?
Discussion
Hi, I have a MY00 UK Impreza Turbo which I belive has the T4 turbo. When I had the intake pipe of recently I noticed there was noticeable play in the turbo shaft. The car has 135K miles on it so I don't won't throw money at it unless I have to. Can anyone tell me if this is likely to lead to some kind of turbo failure that could destroy the engine by throwing pieces of metal into it, or whether it will slowly get worse and I can live with it?
The car doesn't use much oil and pulls like a train!
Thanks
Doug
The car doesn't use much oil and pulls like a train!
Thanks
Doug
dougisaacs said:
Hi, I have a MY00 UK Impreza Turbo which I belive has the T4 turbo. When I had the intake pipe of recently I noticed there was noticeable play in the turbo shaft. The car has 135K miles on it so I don't won't throw money at it unless I have to. Can anyone tell me if this is likely to lead to some kind of turbo failure that could destroy the engine by throwing pieces of metal into it, or whether it will slowly get worse and I can live with it?
The car doesn't use much oil and pulls like a train!
Thanks
Doug
They most certainly do not have a T4. They have a TD04, very different.The car doesn't use much oil and pulls like a train!
Thanks
Doug
And what exactly is noticable play ? Unless there is a specific problem, leave it alone. They are generally very reliable, and even if it does go, replacements can be bought for under £100 as they are so common place and reliable.
dougisaacs said:
Hi, thanks for your response. when I held the induction impeller and tried to move it up and down I could feel about a mm of play in the shaft. So as far as your concerned it isn't likely to destroy itself and be ingested by the engine!
It isnt a very accurate test, especially if you arent familiar with what they should feel like. If indeed it was 1mm like you say, the impellor would already be hitting the housing, and you make no mention of that. So I'd guess it isnt hitting.Unless the turbo has been removed, it is almost physically impossible to get a good look or feel at the impellor due to the turbo mounting position either.
And it would be pretty much physically impossible for anything to happen to the impellor, where parts of the turbo would actually pass through the intake tract, then intercooler, and then the engine
The CEL indicated a knock sensor fault and the bolt holding this on had seized and I managed to shear it off in the block. So then I had to take the inlet manifold off to get decent access to remove the rest of the bolt. That was fun!
I do seem to get a fair bit of oil in the intercooler and after idling for a while and then putting my foot down there is a bit of blue smoke so figured the turbo oil seals might be leaking a bit. So when I found this 'bit' of play in the turbo shaft I thought it might be fooked. I don't know how long the turbos normally last but as I mentioned, this has done 135k.
I do seem to get a fair bit of oil in the intercooler and after idling for a while and then putting my foot down there is a bit of blue smoke so figured the turbo oil seals might be leaking a bit. So when I found this 'bit' of play in the turbo shaft I thought it might be fooked. I don't know how long the turbos normally last but as I mentioned, this has done 135k.
I'd have just bolted the knock sensor elsewhere than that hassle !
The TD04's are very reliable. Generally any oil visible is from the breather system, not the turbo.
Although a lot does depend on what sort of life the car has had. But until it lets go, unless its making any noises, or performance is bad, there is little reason to tamper with it.
The TD04's are very reliable. Generally any oil visible is from the breather system, not the turbo.
Although a lot does depend on what sort of life the car has had. But until it lets go, unless its making any noises, or performance is bad, there is little reason to tamper with it.
stevieturbo said:
I'd have just bolted the knock sensor elsewhere than that hassle !
The TD04's are very reliable. Generally any oil visible is from the breather system, not the turbo.
Although a lot does depend on what sort of life the car has had. But until it lets go, unless its making any noises, or performance is bad, there is little reason to tamper with it.
Thanks stevieturbo, exactly what I'll doThe TD04's are very reliable. Generally any oil visible is from the breather system, not the turbo.
Although a lot does depend on what sort of life the car has had. But until it lets go, unless its making any noises, or performance is bad, there is little reason to tamper with it.
stevieturbo said:
I'd have just bolted the knock sensor elsewhere than that hassle !
wow, brave man, i'd try and leave the knock sensor exactly where the manufacturer intended it to be!!(don't come running home crying when your KCS no longer reponds and kills your engine because you've moved the fundamental or Nth order knock sensor oscillation frequency out of the detection window........... ;-)
As for the turbo play, in a plain bearing turbo, there are actually 2 bearing surfaces per journal, (housing-oil-collar-oil-shaft, so each oil shear only occurs at 50% of the rotational velocity) and these 2 parts both require a dynamic oil film to operate, so if the engines off, and you havent got oil pressure, then you won't have a properly located shaft anyway. On a brand new plain bearing turbo, typically you would expect to get a max of 0.5mm play (radially) which dynamically will be much much less (typically <0.1mm radius orbit of mass centroid) when running
If you have 1mm of radial play, then your turbo is definately worn (as you would expect, given your mileage) As to if this is likely to cause a terminal failure, it depends. You could remove the compressor housing and check for the tell tail signs of radial contact, if none, then you are ok for a bit.
What tends to happen is that as the bearings wear, the orbit of the shaft tends towards lower orders of the rotational velocity, and these lower orders contain significantly higher proportion of energy (as the shaft has a longer time to move sideways, so moves / accelerates more before the oil film dynamics restore the shaft position towards the centre). Eventually, these shaft motions increase to the point where 1 of 3 things happens:
1) the shaft of the turbo fails due to repetative stress fracture (typically at the friction weld between turbine and shaft)
or
2) the dynamic oil film can no long resist the sideways movement, and the oil film is squeezed out completely, causeing a metal-metal contact, leading to total bearing failure seconds later (this can also typically happen at the axial thrust bearing first, especially on high boost cars where the axial loads are massively increased)
3) somehow the bearings and shaft withstand the motion increase, but the compressor wheel (almost always due to the tighter clearance than at turbine end) touches the housing, which is not directly terminal, but the now out of balance wheel will throw itself apart in short order
(remember, turbo's sping at something like 160 to 200krpm, so cycles acrue pretty dam fast)
If the car hasnt had the boost wound up (and hence the turbo's max rpm), then to answer the original question, is my turbo going to fail?? the answer is yes, eventually, but not probably immediately.....
ETA: if it does fail, typically this does not distroy the engine (on a gasoline engine that is), as the larger compressor wheel parts tend to get caught by the intercooler matrix, and the smaller bits just get swallowed and spat out. You can get pistion / head damage on cars with very short exh manifolds and close coupled turbo's where the engine ingests parts of a failed turbine (which are a hard material compared to ally compressor) by sucking them back into the exhaust port on the overlap etc. (unlikely on a scooby due to the long run between turbo and exhaust valves) Typically, anything ending up in the sump, (like the remains of the bearing pack etc) just either sits on the floor of the sump, or ends up in the oil filter (take filter off, cut open, wash out in petrol (carefully! lol) and look for small brass / copper coloured bits, that would have come from turbo bearing pack) (if you are worried and have a filter thats done a reasonable distance on the engine, that might be worth doing anyway)
Edited by anonymous-user on Tuesday 31st August 01:38
Thanks alot for the replies guys. Unfortunately I just changed the oil filter last week so cannot inspect it for bits.
I didn't accurately measure the play in the shaft, or see if it moved in and out, just up and down, so I think I'm going to stick with the original plan to leave it, especially as it sounds like failure is unlikely to destroy the engine.
In terms of what was the cause of the knock error reading, all I know is the old sensor was very rusty and the block quite oxidised where it mounted so I replaced the sensor flatted off the block with some wet and dry, put some grease on both to try to stop future corrosion and I've had no error codes since and I've driven about 500 miles, and given it some beans on track day!
My fingers are crossed!
I didn't accurately measure the play in the shaft, or see if it moved in and out, just up and down, so I think I'm going to stick with the original plan to leave it, especially as it sounds like failure is unlikely to destroy the engine.
In terms of what was the cause of the knock error reading, all I know is the old sensor was very rusty and the block quite oxidised where it mounted so I replaced the sensor flatted off the block with some wet and dry, put some grease on both to try to stop future corrosion and I've had no error codes since and I've driven about 500 miles, and given it some beans on track day!
My fingers are crossed!
Max_Torque said:
stevieturbo said:
I'd have just bolted the knock sensor elsewhere than that hassle !
wow, brave man, i'd try and leave the knock sensor exactly where the manufacturer intended it to be!!(don't come running home crying when your KCS no longer reponds and kills your engine because you've moved the fundamental or Nth order knock sensor oscillation frequency out of the detection window........... ;-)
As for the turbo play, in a plain bearing turbo, there are actually 2 bearing surfaces per journal, (housing-oil-collar-oil-shaft, so each oil shear only occurs at 50% of the rotational velocity) and these 2 parts both require a dynamic oil film to operate, so if the engines off, and you havent got oil pressure, then you won't have a properly located shaft anyway. On a brand new plain bearing turbo, typically you would expect to get a max of 0.5mm play (radially) which dynamically will be much much less (typically <0.1mm radius orbit of mass centroid) when running
If you have 1mm of radial play, then your turbo is definately worn (as you would expect, given your mileage) As to if this is likely to cause a terminal failure, it depends. You could remove the compressor housing and check for the tell tail signs of radial contact, if none, then you are ok for a bit.
What tends to happen is that as the bearings wear, the orbit of the shaft tends towards lower orders of the rotational velocity, and these lower orders contain significantly higher proportion of energy (as the shaft has a longer time to move sideways, so moves / accelerates more before the oil film dynamics restore the shaft position towards the centre). Eventually, these shaft motions increase to the point where 1 of 3 things happens:
1) the shaft of the turbo fails due to repetative stress fracture (typically at the friction weld between turbine and shaft)
or
2) the dynamic oil film can no long resist the sideways movement, and the oil film is squeezed out completely, causeing a metal-metal contact, leading to total bearing failure seconds later (this can also typically happen at the axial thrust bearing first, especially on high boost cars where the axial loads are massively increased)
3) somehow the bearings and shaft withstand the motion increase, but the compressor wheel (almost always due to the tighter clearance than at turbine end) touches the housing, which is not directly terminal, but the now out of balance wheel will throw itself apart in short order
(remember, turbo's sping at something like 160 to 200krpm, so cycles acrue pretty dam fast)
If the car hasnt had the boost wound up (and hence the turbo's max rpm), then to answer the original question, is my turbo going to fail?? the answer is yes, eventually, but not probably immediately.....
ETA: if it does fail, typically this does not distroy the engine (on a gasoline engine that is), as the larger compressor wheel parts tend to get caught by the intercooler matrix, and the smaller bits just get swallowed and spat out. You can get pistion / head damage on cars with very short exh manifolds and close coupled turbo's where the engine ingests parts of a failed turbine (which are a hard material compared to ally compressor) by sucking them back into the exhaust port on the overlap etc. (unlikely on a scooby due to the long run between turbo and exhaust valves) Typically, anything ending up in the sump, (like the remains of the bearing pack etc) just either sits on the floor of the sump, or ends up in the oil filter (take filter off, cut open, wash out in petrol (carefully! lol) and look for small brass / copper coloured bits, that would have come from turbo bearing pack) (if you are worried and have a filter thats done a reasonable distance on the engine, that might be worth doing anyway)
Edited by Max_Torque on Tuesday 31st August 01:38
TheEnd said:
I'd be interested in what the knock fault actually was.
Disconnected is one fault, but it will also flag up something if it sees a lot of knocking.
If the code reader doesn't give an accurate description, you can end up shooting the messenger.
knock sensor wouldnt be a common fault, but at the same time, not uncommon either. So they can require replacement at times.Disconnected is one fault, but it will also flag up something if it sees a lot of knocking.
If the code reader doesn't give an accurate description, you can end up shooting the messenger.
Airflow meters on the 99/00 model are a different story. They dont last very long at all, and very rarely will any fault codes be raised, or major problems noticed as they are failing. And often they lead to engine destruction if not picked up in time.
stevieturbo said:
Nothing brave about it. There are 2-3 locations where it could be bolted within 2-3" of original, and it will make virtually no difference to its ability to detect knock.
Wow i'm impressed by your thoroughness, not many people either carryout a full structural Vibration analysis of their engine block, or use a high speed datalogging oscilliscope and post process data through a FFT to determine the modal analysis of each mounting location.....;-)Or do you mean you've just stuck the sensor in a different place and so far it's still working?
Max_Torque said:
stevieturbo said:
Nothing brave about it. There are 2-3 locations where it could be bolted within 2-3" of original, and it will make virtually no difference to its ability to detect knock.
Wow i'm impressed by your thoroughness, not many people either carryout a full structural Vibration analysis of their engine block, or use a high speed datalogging oscilliscope and post process data through a FFT to determine the modal analysis of each mounting location.....;-)Or do you mean you've just stuck the sensor in a different place and so far it's still working?
If you cant detect knock with it in a variety of places, you're bloody deaf. Same applies to the OE ecu.
Gassing Station | Engines & Drivetrain | Top of Page | What's New | My Stuff