posted on June 08, 2011 19:56
By Khiem Dinh
Khiem Dinh is an engineer for Honeywell Turbo Technologies at the time of this writing. All statements and opinions expressed by Khiem Dinh are solely those of Khiem Dinh and not reflective of Honeywell Turbo Technologies.
We like turbos at MotoIQ. Why? In the phrase made famous by Tim Allen, "more power!" In order for a turbo to cram air into an engine, it has a pair of wheels and a shaft spinning at incredible speeds. Depending on the size of the turbo, they can spin anywhere from 100,000 rpm (big turbos) to 300,000 rpm (small turbos). 300,000 rpm equals 5000 revs a second. Put another way, if you rolled a compressor wheel on the ground, it would go about six football fields a second. That's about 1.7 times the speed of sound. Wrap your head around that! The rotating assembly consisting of the shaft, turbine wheel, and compressor wheel encounter axial and radial loads that want to push the wheels into the housings and rub parts together. To keep the rotating assembly supported and freely spinning, turbos currently either use a journal or ball bearing system. There are variations on each theme, but we're going to go over the most common designs. First, we need to discuss the loads they encounter.
The bearing system of a turbo has to resist radial and thrust loading created by the compressor and turbine wheels. Radial loading is probably easier for most people to visualize; the rotating assembly will want to wobble due to various loadings so the bearings keep the rotating assembly from moving around excessively. The wobble is known as shaft motion in industry speak. If there is too much shaft motion, the compressor and turbine wheels may end up rubbing the housings destroying the turbo.
|Thrust and radial loads act on both the compressor and turbine wheels. Thrust loading is along the axis of the shaft and is created by pressure differentials between the compressor and turbine sides. Radial loads act perpendicularly to the centerline of the turbo and are a cause of shaft motion.
Thrust loading pushes the rotating assembly back and forth and is created by pressure differentials in the compressor and turbine housings. An approximation of the thrust load on the rotating assembly can be calculated by taking the difference between (compressor housing pressure x compressor wheel area) and (turbine housing pressure x turbine wheel area). For example, let's say you have compressor and turbine wheels of equal area at 3 in^2. There is 15 psi of pressure in the compressor housing (boost) and 30 psi in the turbine housing (exhaust back pressure). The net thrust load is 15 psi x 3 in^2 = 45 lbs of thrust pushing from the turbine side towards the compressor side. The bearing system has to be able to support the 45 lbs of force or parts will start rubbing against each other leading to expensive things breaking. Take a 45 lb plate from the gym, stack it on the end of the shaft and that’s the thrust load the bearing system has to support.
The journal bearing system is the most common type of turbo bearing system because it is cheap and it works. The journal bearing system typically consists of two types of plain bearings: cylindrical bearings to contain radial loads and a flat thrust bearing to manage thrust loads. Plain bearings in turbos use oil to keep parts from rubbing each other based on the hydrodynamic principle. In hydrodynamic lubrication, the fluid (oil in a turbo) stays between the moving surfaces and prevents metal-to-metal contact.
|The flat plate that looks like Pac Man is the thrust bearing. The one shown is know as a 270 degree thrust bearing. There are also 360 degree thrust bearings that do not have the slot and can handle greater thrust loads. The round thing above the thrust bearing is the thrust collar which slides into the slot of the thrust bearing. The two cylindrical bronze looking parts are the journal bearings. The part between them is a spacer used to locate the journal bearings on the shaft of the turbo.
Thursday, June 09, 2011 11:29 PM
Thx for the great info KD, always an interesting read :-)
Friday, June 10, 2011 8:24 AM
Awesome read! Granted some of the terminology is interchangeable with an amateur porn film ;)
Friday, June 10, 2011 8:53 AM
^^^ KD is no amateur..... ;-)
Friday, June 10, 2011 9:35 AM
Nice read Khiem. This is why compressor surge is so damaging to journal bearing turbos, since the loads exceed what the oil film can support and there is direct part to part contact at a most inopportune time, correct?
Wes: I wonder if someone's measured Peter North's radial and thrust loads...
Friday, June 10, 2011 9:36 AM
KD, an awesome balance of tech, fluid dynamics and the trademark MotoIQ irreverance. Great insight on a subject that escaped me. I got pissed offf with turbos when they made the beauty of F1 quiet.
Lately after driving so many awesome turbo'd cars on track (leatest was a TechArt Porsche @700hp+, as well as the phenomenal Turbo Panamerica) it's time to drop the rock . . .
Friday, June 10, 2011 10:06 AM
Wes/mxpop, I've found that knowledge on thrust loading and shaft motion is important. Proper lubrication is also critical for durability.... just sayin
Steve, yup, compressor surge creates really big pressure spikes which in turn creates really big changes in thrust loading. It's pretty much like whacking the shaft with a hammer. In the picture of the thrust bearing, you can see little pads on it which supports the loading. The three little holes are the oil feeds for the three pads. So compressor surge will overload the oil film and you'll get metal-to-metal contact between the thrust bearing and collar. The more it happens, the more those pads wear away and the thrust bearing losses its capacity to handle loads. And then your turbo blows up.
Friday, June 10, 2011 10:27 AM
I was beginning to wonder when the URL of this site was going to change to driftIQ.com. Thanks for the refreshing non-drift related article. I will send all who ask about BB turbos here.
Friday, June 10, 2011 10:54 AM
^^ What Eric Hsu said, completely, including the thanks.
Friday, June 10, 2011 12:51 PM
"A pro-cyclist can only do 1000W in an all-out sprint" with EPO :)>
I think we MIGHT be seeing more ball-bearing turbos, if it means peak torque can be found much earlier and fatten the torque curve to achieve better mpg at the right production cost of course! The flip side of that is that they'll probably use a longer stroke with higher compression pistons and just utilize a journal bearing turbo that's less expensive and still expected to last 100k+ at stock power levels. As to your point about the oil Khiem, the domestic and asian manufacturers will probably make everybody use the more expensive euro blends to overcome some of the weight/viscosity issues and pass that cost on to the public.
Friday, June 10, 2011 1:14 PM
I'm a total ignorant about turbos, not gonna lie....so this for me it's "somewhat" easy to digest. Thanks for the info Khiem.
But I still love N/A simplicity and power,less stuff that can break :)
Friday, June 10, 2011 1:17 PM
Actually I do have a dumb fucking question I'm sure you guys could answer.
It's known that needle bearings in many application are much stronger than ball-bearings....why not just replace the ball-bearings with the needle ones?
Once again, I apologize in advance if this question is too dumb.
Friday, June 10, 2011 1:58 PM
Needle bearings are not as good at super high speeds.
Friday, June 10, 2011 3:59 PM
Any comment on ceramic versus steel ball bearings? I have only heard of Turbonetics offering ceramic to consumers. Does the cost not outweight the benifit? Is the benifit negligible?
Friday, June 10, 2011 4:40 PM
Some of the larger Garrett aftermarket ball bearing turbos use ceramic ball bearings. The Borg Warner EFRs use ceramic BB. The advantage of the ceramic BBs are reduced friction, greater durability, and I think reduced inertia also. The Garrett Dual Boost turbo used in the Ford Powerstroke has ceramic ball bearings.
Saturday, June 11, 2011 2:25 AM
Good to know, seems there are more out there than I thought :)
Saturday, June 11, 2011 7:29 AM
I was talking to an "un-named" engineer (don't want to get him in trouble, lol) that works in the turbo industry, and he was telling me that the response on the ball bearing turbo's are fudged a bit as compared to the journal bearing turbos (at least in Garrett's case) due to them having different turbine wheels for the same sized turbo. Not sure if this is correct, or if he was just bitter, but I could see the marketing advantage of ball bearing turbos over their journal bearing counterparts.
Saturday, June 11, 2011 10:00 AM
@2_Liter, I'm going to be blunt, it doesn't sound like the guy you talked to has a very fundamental understanding of turbos. First off, what has been 'fudged'? I have never seen any claim from any of the major turbo manufacturers (Garrett, Borg Warner, MHI, IHI, Holset) saying that a ball bearing turbo will responed X percent faster across the board.
Why not? Because there are so many variables. Looking at just a ball bearing system: overall size of ball bearing, size of balls, number of balls, ball material used (steel, ceramic), retainer material (polymide, alloy), material of the inner and outer races. Looking at the journal bearing system: size of journal bearings, clearances, thrust bearing size (270/360 degree, size/number/shape of pads), thrust collar size/mass, volume of oil feed to the system.
Then throw in oil viscosity and the temperature of the oil. It's just a fact that ball bearings have less friction than journal bearings which in turn improves turbo response. See if any journal bearing turbo will do this: http://youtu.be/DbUN9eSjbqE
Saturday, June 11, 2011 12:08 PM
Lol, I was just relaying what I was told because it pertained to the topic. It is not my theory at all. The guy has been in the industry a long time, but he could also have a chip on his shoulder. Who knows, ha ha.
Saturday, June 11, 2011 12:19 PM
I know, that's why I said, "the guy you talked to" :)
Saturday, June 11, 2011 2:17 PM
Ahh, lol. I just wanted to clarify that it wasn't me.
MotoIQ Proudly Presents Our Partners: