|
|
Post by finiteparts on Dec 15, 2013 23:04:32 GMT -5
Hi Ash,
The forward end of that turbo has a thrust bearing just like a standard turbo because the forward ball bearing can only absorb a thrust load in one direction. So you are really not getting a very good trade by putting just one ball bearing in...when you use two ball bearings, you control the thrust loads in both directions. The thrust bearing has considerable viscous drag, so if you eliminate it by using two ball bearings you get a substantial reduction in viscous losses. There are several ASME papers that looked into the bearing losses in turbochargers that will shed some light on the magnitudes of each...I will try to find them and share some numbers.
As for the journal vs ball bearings, I have a paper by Koyo bearings and they measured the frictional power loss in a floating bush bearing and in a hybrid ceramic ball bearing in an accelerating rotor at about 100,000 rpms. It shows roughly a 30% reduction in these losses and the benefit from the ceramic bearings getting better as rpms increase. This actual data counters the claims by some manufacturers that there is not much difference between different bearing frictional losses at higher rpms. I will find that one too and share some values.
As for roller bearings, they are often used in larger engines because the shaft thermal growth would be to excessive for a ball bearing to absorb...also, the rotor often acts as the inner race, but this may not be a good choice for a turbo rotor...bearing surfaces go through loading and unloading cycles as they rotate due to unbalance forces, shaft bending, etc. and thus these surfaces are fatigue limited. I would suspect that the shaft material used for turbos with journal bearings does not have the material properties to act as an inner race and would suffer spallation quickly.
Now, one of the main benefits for using the journal bearings is that they are not fatigue limited (thus very long operational lives...usually only fail due to rubs at start/stops or at large shaft whirls) and they do not have the speed limitations like ball bearings (sort of). Ball bearings have speed limitations due to centrifugal forces acting on the rotating parts...when ball contact forces get to high, the bearing races fatigue quickly. Since the journal bearings distribute the forces over a larger area, fatigue is usually not an issue....but, there is a speed limit on floating bearings due to self-excited instabilities. There are many papers on journal bearing supported rotors not being able to get up to operating speed due to excessive vibrations caused by self excited instabilities reaching what is called a limit cycle. These instabilities can be caused by seals, aerodynamics of the compressor or turbine, surface finish, bearing clearance, viscosity changes due to high temperatures in the bearing (caused by fluid shear), etc...so you point about trial and error is so true!
|
|
|
|
Post by racket on Dec 16, 2013 3:29:21 GMT -5
Hi Guys Heres a graph that might be interesting  . Most of the benefits of balls vs brass seem to be with smaller turbos , but as we "idle" at normal SI engine peak boost level theres not much to be gained from the better "advertised" rotor acceleration rates , our DIY engines on brass bushes can go from idle to max rpm in a second ...........RC turbines can need up to 3 seconds to do the same despite running on hybrid ball races . LOL, I'm sounding like an advocate for brass bushes :-) Cheers John |
|
ripp
Veteran Member
I'm sorry, I don't speak english, so I torment you (and myself) with a translation program,Sorry
Joined: January 2013
Posts: 237
|
Post by ripp on Dec 16, 2013 16:01:20 GMT -5
|
|
ashpowers
Veteran Member
Joined: February 2011
Posts: 207
|
Post by ashpowers on Dec 16, 2013 18:05:45 GMT -5
John, The #8 bearings from the RR A250 engine, do you know which direction the axial load is supposed to go? Another configuration I've been considering to use is duplexing this bearing both on the front and rear of the shaft for a total of 4 bearings. The rigidity and load capacity of doing it this way would be very stout. The shaft is almost completely supported along its length with only a very small gap between the two pairs. The inner races of the bearings will also, in effect, "thicken" the shaft and improve on its rotordynamics. The image below shows what it would look like. Lube oil is pumped in through the two ports on the right of the bearing housing that narrow down to smaller orifices. This oil is delivered into a small annular space between the bearing tube and the bearing housing which is also double o-ringed for each entry point. These lube-filled annuli will act as a fluid dampening system and the orings not only containing the lube oil but also serving to positively locate the bearing tube back to a centerline position. 180-degrees around the bearing tube there is a small port for each bearing pair that allows feed oil to enter into the bearing annulus. Bleed air coming from behind the compressor wheel is delivered to the outer bearings in the group via channels through the bearing housing. The lube and cooling air is vented out of the engine through the larger port located between the two oil feed ports. Doubling up on the bearings, especially at the rear will make for a very stiff bearing support system. The rear bearings would be allowed to move freely in an axial direction within the bearing tube to accommodate for thermal expansion on the shaft. The P2/P3 interstage pressure that builds up behind the compressor wheel would be allowed to vent out of the engine and serve as a cooling airflow and motive force to move the lube through the bearings in the same stroke. This will also reduce the axial loading that the thrust-carrying bearing would have to endure. The big question here is if a fuel/oil mix could be used and the fuel pump itself serve both as the combustor feed as well as the bearing feed. A deaerator would be necessary to collect and separate the mix of bleed air and fuel/oil to recover the fuel/oil mix back into the main fuel supply but that would be the extent of the lubrication system's complexity. I think it would just come down to a process of making small adjustments to the bleed air quantity (by way of the clearance between the radial seal behind the compressor) and the fuel/oil jetting to the bearing group. A temperature measurement of this exhausted air/fuel/oil mix would give us a good indication of the environment the bearings are working within. I think the best approach to this though would be to use a production turbine lube which will have considerably greater viscosity than that of a fuel/oil mix. I tend to lean more towards using a thicker dampening fluid since the orings are going to act like stiff springs that can compound the rotordynamics in an undesirable way. The lube needs to be thick enough to dampen the vibrations but not too thick and cause the bearings to absorb excessive loads. I think we can take from the engineering that turbo manufacturers have already done though - they are all designed to be used with automotive engine lubes and there is a pretty decent range of viscosity in them. None of the turbo manufacturers state what oil should be used or recommended as that is more a matter of the specifics of the engine itself, but it would be nice to know what they are using in their development. I've mentioned this before some time ago but perhaps integrating vibration detection into the engine would be helpful in the process. Vibrations within the engine could be monitored over its operating range and would allow us to determine the mode type occurring within the group at various speeds. This would at least give us a starting point for refinement of the system and at the least, indicate to us what RPMs to avoid, LOL.    |
|
ripp
Veteran Member
I'm sorry, I don't speak english, so I torment you (and myself) with a translation program,Sorry
Joined: January 2013
Posts: 237
|
Post by ripp on Dec 17, 2013 2:12:12 GMT -5
|
|
|
|
Post by racket on Dec 17, 2013 3:19:23 GMT -5
Hi Ralph
10,000 euros for that Jetcat ....OUCH .............I don't think I've spent that much on all of my turbines added together over the past 20 years
With the RC turbines the shaft is pretty long compared with a turbo , this can help with some of the shaft dynamics , the lightweight axial turbine wheels are also a positive when it comes to the dynamics as well as balancing the axial air/gas thrust loads .
But having to change out bearings after 25 hours on the RC turbines is a bit of a negative due to the lubrication used .
Cheers
John
|
|
|
|
Post by racket on Dec 17, 2013 3:31:28 GMT -5
Hi Ash
LOL....the Allison bearing is a deep groove , so it can go either way , ............heh heh , they did for me :-)
It would be possible to use fuel for the lube , it is done on some "expendable" turbines.
After all the problems with FM-1, I'm loath to make suggestions about bearing systems , I just don't enough about the "unknowns" , plus theres the accuracy requirements for the components .
Cheers
John
|
|
ashpowers
Veteran Member
Joined: February 2011
Posts: 207
|
Post by ashpowers on Dec 19, 2013 12:54:24 GMT -5
Hi John,
For some reason I was thinking that was an AC bearing. I've come across some really high performance angular contact bearings over on Barden's site in their catalogs - bearings that fit our dimensions and speed/load ratings. I'm sure they come with a hefty pricetag as well though. Hopefully by reducing the pressure behind the comp wheel and taking the axial loads off of the Allison bearings will be enough to keep them happy....
|
|
|
|
Post by racket on Dec 19, 2013 23:17:05 GMT -5
Hi Ash
The Allison bearings normally go to ~52,000 rpm , I was running them to ~125% , but they seem to survive OK .
LOL, ....yeh the Barden bearings would be a bit pricey , hopefully you can pick up some time expired C20 bearings , they'd have to be pretty tough being tucked away under the heat shield in the base of the flametube , they'd be at least as good a quality as anything fitted to a turbo .
Cheers
John
|
|
ashpowers
Veteran Member
Joined: February 2011
Posts: 207
|
Post by ashpowers on Dec 20, 2013 12:15:42 GMT -5
Hi John, I have 8 of the C20 bearings - all of them still have some "miles" left on them too. Got them as freebies from two different companies that rebuild the C20 engines.  I can tell they must have a tough life - they are a good bit discolored, almost black, but spin very freely. I think I'm going to commit to the quad bearing arrangement I've drawn up and give it a shot. Time go get this engine together finally. -Ash |
|
|
|
Post by Johansson on Dec 20, 2013 13:43:27 GMT -5
That´s the spirit! Cheers! /Anders |
|
|
|
Post by racket on Dec 20, 2013 17:04:11 GMT -5
Hi Ash
Yeh , its time to start cutting metal :-)
Cheers
John
|
|
ashpowers
Veteran Member
Joined: February 2011
Posts: 207
|
Post by ashpowers on Dec 21, 2013 6:22:13 GMT -5
There is a small unfortunate reality concerning my TV turbine shaft.... I machined it down to a bearing fitment of 20mm without any interference. I have a local hard chrome plating company literally within 5 minute walking distance. Rather than waste my current turbine I am considering to have him plate the shaft to build it up a bit. But after thinking about that it dawned on me the possibility of having him do the same to other hot section components. It offers great corrosion resistance and I'm entertaining the idea of having him plate out several of the hot section parts. Would this allow me to use "substandard" materials for the hot section and get away with it? Items like the NGV haseplate, turbine heat shield, jetpipe, etc, all could possibly be hard chrome plated to improve their thermal characteristics?
|
|
rexhunt
Member
Joined: August 2012
Posts: 32
|
Post by rexhunt on Dec 21, 2013 10:54:28 GMT -5
Hi Ash,
Isn't there something about when you get metal chromed it can build up hydrogen underneath and fail sooner than normal?
Regards,
Josh
|
|
|
|
Post by finiteparts on Dec 21, 2013 11:54:44 GMT -5
Hard chroming hot parts is a good method to reduce the fatigue capability, which is probably not what you are trying to achieve. Differences in thermal expansion coefficients will often cause the plating to flake off and initiate surface cracks. The surface cracks in the coatings will propagate into the base metal, thus reducing fatigue life. Josh is correct...the potential for hydrogen embrittlement is also a concern on hot parts. The propensity for hydrogen embrittlement is a function of the chemical composition, such as nickel content, etc...but it should be considered especially if the coating or plating is done in a non-inert enviroment. The use on low stress parts may be ok, but the fact that when you are calculating fatigue life and you use a plating/coating, you must apply a knock down factor might keep me from using them. You might find that by using such coatings you are actually hurting the part life.
If you want to increase the thermal capability (I am assuming you mean hot corrosion, yield capabilty, LCF/HCF capability, etc) it might be better to brainstorm better cooling methods.
|
|
I can tell they must have a tough life - they are a good bit discolored, almost black, but spin very freely.