gidge348
Senior Member
Joined: September 2010
Posts: 426
|
Post by gidge348 on Dec 21, 2013 19:50:05 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. agree +1 It may also be worth a talk with the people at the hard chroming facility for their take on it..... Never hurts to get everyone's ideas? Ian...
|
|
wolfdragon
Senior Member
Joined: April 2011
Posts: 287
|
Post by wolfdragon on Dec 22, 2013 18:25:31 GMT -5
Hard chroming to build up the shaft diameter will be a good way to go, the process is the standard layers of copper and nickel and then a super thick chrome which gets ground to the desired OD As for the the rest of the hot section components, if insulation is the idea, then www.jet-hot.com/ is worth a call, these guys are getting used to odd jobs aside from automotive and aircraft exhausts
|
|
Deleted
Joined: January 1970
Posts: 0
|
Post by Deleted on Dec 23, 2013 5:09:33 GMT -5
Hi Ash Ex RAAF electroplater here, the best way forward for the shaft is a Sulfamate nickel layer ,dont even consider chrome plating internal jet engine parts, the coating will fail. As for the Hydrogen embrittelment, the shaft must be heat treated within 4 hrs of plating at 191 C for at least 24 hrs, lathe or grind the nickel back into spec.This is basically the way we fixed TF30 main shafts (ex F111 engine) Do not use multi layers (Cu,Ni,Cr)as they will fail as each layer expands and contracts differently. Also hard chrome is brought on as ONE layer
|
|
gidge348
Senior Member
Joined: September 2010
Posts: 426
|
Post by gidge348 on Dec 23, 2013 8:55:25 GMT -5
|
|
ashpowers
Veteran Member
Joined: February 2011
Posts: 207
|
Post by ashpowers on Dec 23, 2013 11:38:59 GMT -5
Hi Jetturbine,
I've looked into the sulfamate nickel process and while I have done electroplating before in the past my concerns are with the heat treatment that would be required following. I wouldn't be inclined to purchase all of the equipment to electroplate this so I would send it off to be done but I know I wont get it back in time to heat treat it.
I only need a few tenths at best to produce the interference fit onto the shaft. My question is do you think I could do this with eletroless nickel plating? If no-go on that then I'll probably just opt to lock the inner races to the shaft via a sleeve between the bearing sets. This would allow me to put the shaft under tension as well via the compressor nut, through the compressor wheel, front shaft collar, frontal bearings inner race, bearing sleeve, rear bearing inner race, and the rearmost bearing butted against the shaft shoulder. This would be a much easier assembly/disassembly not having to press the bearings.
|
|
|
Post by racket on Dec 23, 2013 15:21:46 GMT -5
Hi Jetturbine
Is that you Chris ..............I've been trying to find you after a computor "upgrade ??" lost all my contact addresses :-(
What have you been up to ...........hows Cameron ?? ............lost his email address as well .
I'm living back on the Mid North Coast again
Cheers John
|
|
Deleted
Joined: January 1970
Posts: 0
|
Post by Deleted on Dec 24, 2013 16:36:49 GMT -5
Hi Ash With EL Nickel you will run into the same problem, need to get rid of the hydrogen within 4 hrs (even though a lot less). After the initial bake at 191c it could then be taken up to ±340c in an inert atmosphere to get a hardness even surpassing Hard Chrome Another process to do the shaft would be brush plating, this would solve the oven thing, I would assume expensive and harder to source.
|
|
Deleted
Joined: January 1970
Posts: 0
|
Post by Deleted on Dec 24, 2013 16:45:12 GMT -5
Hi Jetturbine Is that you Chris ..............I've been trying to find you after a computor "upgrade ??" lost all my contact addresses :-( What have you been up to ...........hows Cameron ?? ............lost his email address as well . I'm living back on the Mid North Coast again Cheers John Yeah John back again, well not really ,never left...just lurking been following all of major projects especially yours....wow what a monster!!! have to come down and see it one day. Your only 5 hrs away these days. Cameron has gotten married and bought a house ,he has just completed rebuilding his lathe , so he's been busy I changed jobs and have been busy since... just got my mancave finished as well so will be picking up where we left off will give you a call soon Cheers chris
|
|
|
Post by racket on Dec 25, 2013 14:58:36 GMT -5
Hi Chris
Thats great news :-)
Yep, just down the road .
Sounds like we've all been busy this last year or so , but slowly getting on top of things here , the house needed, ..........LOL,still needs , a lotta work , but I can see an end to it if my back holds out long enough to finish it .
Looking forward to a chat :-)
Cheers John
|
|
syler
Member
Joined: January 2014
Posts: 39
|
Post by syler on Jan 18, 2014 21:17:54 GMT -5
Very impressive machining! Is that what your business is? Your engine looks like a giant hollow point bullet. I'm wondering; with all your machining capability, do you think you would be able to make your own axial flow compressor section? Or, perhaps fabricate another compressor stage to raise your overall compression ratio? Better yet, would you consider starting over on the compressor section and just make 3 or 4 axial compressor wheels?
|
|
|
Post by finiteparts on Mar 15, 2014 23:13:27 GMT -5
Hi Ash, ... 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. Ok, sorry for the delay on these numbers, but I finally got settled in to the new house and had time to do some digging. I am still looking for the ASME papers that broke out the specific contributions of each bearing to the mechanical losses. Due to copyrights, I can't post them...but, I will provide the paper numbers so if you have access toASME, SAE and others at a college library, you can get a copy. Though not the specific contributions of each bearing component, F. Payri, et. al. shows some trends in their paper, "Experimental Methodology to Characterize Losses in Small Turbochargers", ASME GT2010-22815 since they cannot show numbers due to proprietary agreements. They show that the mechanical efficiency: 1.) increases with increased rotational speed, 2.) decreases with an increase in axial load and 3.) increases with an increase in oil temperature (due to viscosity reduction). P. Podevin, et. al. show in their paper, "Influence of the Lubricating Oil Pressure and Temperature on the Performance at Low Speeds of a Centrifugal Compressor for an Automotive Engine." (Applied Thermal Engineering, 2011, Vol.31, pp 194-201) that the power loss curve increases steeply for a floating journal bearing arrangement as shaft speed increases. For this turbo, the power loss they measure with a inline torquemeter set-up, goes from 145 W at 30 krpm to 475W at 110 krpm. This equates to a mechanical efficiency of sub 30% at 30 krpm to around 85% at high speeds...notice, that is way different than the typical 97% we put in the equations for mechanical efficiency when we guess! He states a neat theory on why the mechanical losses increase with an increase in oil supply pressure (around a 3% reduction by going from 2 to 4 bar oil pressure). The increase in pressure causes an increase in flow through the bearing...thus less heat per volume flow is picked up in the oil and thus you don't get the corresponding drop in oil viscosity. Another interesting take on bearing selection was given in Colin Rodgers paper, ""Low-Cost" Microturbines via the Turbocharger Route" (ASME Turbo Expo 2011, GT2011-45220). Apparently, there are a bunch of people trying to make microturbines for power generation or hybrid car battery charging, so he takes an indepth look at the differences in the two turbomachines. He had developed some turboalternator ideas for the US Army, so he had already run through the design methodology previoulsy. He states, "The friction losses in the sleeve bearings and in the thrust bearing can result in mechanical efficiencies in the 70% range in small T/Cs. IN one T/C test the author witnesses an overall turbine efficiency including sleeve bearings of 58% at 220 krpm." He mentions ball bearings or air bearings as a more efficient option. Here is the paper published by Koyo bearings... eb-cat.ds-navi.co.jp/enu/tech/ej/img/No157E/157E_05.pdfI did some more investigation on this topic and found some other good papers, which unfortunately again, due to copyrights, I can't share.. The first paper is on IHI's ball bearing turbo development (SAE870354, "Development of High Efficiency Ball-Bearing Turbocharger", K.Miyashita, et.al., 1987). The interesting points from this paper include a table of the DN values for their range of turbos, cage vs uncaged bearings, bearing material selection and bearing life. The DN values range from 1.7 million for the small turbos (RHB3) to 1.4 million for the truck turbos (RHC9). This is quite a high DN value for any bearing. The topic of the bearing material is touched on and they discuss that at the higher temperatures reached at the bearing nearest the turbine, cheaper bearing steels loose their hardness rapidly and are not suitable for long life operation. They end up using M50 bearings to keep an acceptable life. With these races, they show that the bearing life is a function of shaft revolution and drops as you go to higher speeds, but they maintain over 3500 hrs life if the turbo was to operate at full speed over it's entire life. If a more realist duty cycle is used for the lifing, they show over 10400 hrs life or 500000 km at 50 km/hr. They also show that the use of a cage for the bearings reduces the life of the bearing. The efficiency comparison of the two types of bearing systems shows a 20% improvement at low gas flows and a 5% improvement in the high-flow gas range...notice, that is not "essentially equal". Notice that this equates directly to a gain in turbine efficiency. The final point to take from this paper is their no lubrication oil test. They left it operate without oil for 30 mins at 100 krpm and saw "...neither large shaft vibration nor abnormal conditions of the bearings was observed." The second paper is on Nissan's development of their ball bearing turbos (SAE 900125, "Development of a Ball Bearing Turbocharger", M.Aida, T. Umaoka, et. al., 1990). Their squeeze film damper looks almost exactly like the one you showed previously, with the small lands that proved the squeeze film location, except that theirs is made up of separate parts, not one nice outer sleeve. They show a neat plot where they calculate the optimal bearing size...since the bearings basic rated life increases with bearing size, but at high speed bearing size decreases life due to higher centrifugal loads, there exists an optimal size were the life reaches its maximum. They also use M50 bearings for the same reason as the IHI guys. They show how they optimized the oil jet location and flow by testing for minimal flow with max life. They had an odd preloading arrangement that had a center spring so that it could take a thrust load in either direction without locking a bearing...it was set at 30N and the plot shows a rapid life decrease when the preload varied in either direction from this. The friction loss was measured from the reaction torque of the bearing housing, so I question if it gave accurate results...but it still shows a benefit from the BBs up to the max speed...no numbers were given though. They did a contaminated oil test since the film thickness in a BB is on the order of 1 micron while a plain bearing is several microns...with no noticeable wear. Now for the good stuff! The other two papers didn't have the ceramic bearing technology available to them at the time...so now we jump to a modern application on heavy-duty on-highway trucks. Caterpillar and Honeywell have gotten together again (Cat sort of got Garrett into the turbocharger business). Robert Griffith, Seth Slaughter and Peter Mavrosakis discuss the development in "Applying Ball Bearings to the Series Turbochargers for the Caterpillar Heavy-Duty On-Highway Truck Engines" (SAE 2007-01-4235). So why go BBs in trucks... "While these conventional bearings provide a low cost solution, they do create significant mechanical loss....The ball bearing system provides higher load capacity, improved rotordynamics,...and has proven more robust to contaminated and/or marginal lubricant conditions." The cost prohibitive nature of ABEC 7/9 grade bearings is questioned in light of other key customer needs. They used an enthalpy balance technique to measure the loss in the rotor system and state "The end result...a test rig capable of power measurements with an accuracy and repeatability to the hundredth of a horsepower." The show a plot of the bearing power Loss as a function of rotor speed and at 120krpm, the journal brg losses around 3.75 hp while the BB losses around 1.3 hp. This shows up as an increase in turbine efficiency by switching to BBs and at a turbine pressure ratio of 3:1, it equates to around 2 points in efficiency...and much larger at lower turbine pressure drops. The reduction in mechanical loss shows up as a "...significantly reduced.." heat generation as compared to journal bearings once the oil flow rate is optimized. Now here is something really cool! They shut off the oil flow to measure a time to failure comparison...the journal brgs failed in 3.5 minutes...but the BBs failed in 78 mins!!! They also did a contaminated oil test like the IHI and Nissan guys and found that the journals reached their max radial eccentricity in 6 hrs. Their max eccentricity is were there is a hard failure imminent (housing rubs). The BBs contaminated lube resistance was 6X the journal brg systems. The final outcome was to show that the cost effectivenss of the BB system was better due to increase BSFC of the engine, reduced warranty, increased reliability/durability. etc. So finally, my statement that, "...it all comes down to cost...a brass ring is WAY cheaper than a precision ball bearing..." is not totally true, but if you read this paper put out by ABB www05.abb.com/global/scot/scot267.nsf/veritydisplay/497e1a757dd0753b852577c200537ff9/$file/singlestage%20high-pressure%20turbocharging.pdf See section 2.3, where the author states, "...the use of plain bearings, supported by a squeeze oil damper in the bearing flanges, remains the most reasonable option in terms of costs and operational reliability [7] ..." it was pretty close.
|
|
|
Post by racket on Mar 16, 2014 19:50:38 GMT -5
It interesting how the real world impinges on laboratory results .............under controlled conditions balls are good , in the real world , lets use brass :-)
|
|
|
Post by finiteparts on Mar 16, 2014 21:22:25 GMT -5
"Laboratory results"? Is it really a laboratory? Maybe a mad scientist lab! Hehee!
Don't get me wrong...I am definitely a proponent of rolling element bearings, especially due to my project goals...but if you don't mind lugging around a battery, tank of oil and oil cooler, then journal bearings may be the way to go. But we should also caution that journal bearings are just as treacherous and potentially costly in their design. They exhibit more complex rotordynamics especially when you have a full floating sleeve with two separate oil films...
Both styles of bearings seem deceptively simple, but the devil is in the details. Rolling element bearings have no intrinsic damping (very stiff) so their reaction forces can be really large if the system damping is not designed correctly. They can experience ball skidding or overloaded contact stresses that cause spallation and then the bearing chews itself up. Preloads are tough to get correct, and even if you get it correct at one condition, as the bearing speed and load go up, the bearing contact angles and stiffness change, causing the preload requirements to change. Press fits can effect bearing life, fluid film thicknesses,heat effects on race hardness, etc...lots of things pop up! Think of it as anything that changes the local stress applied to the ball/race interface will effect the life.
Hydrodynamic bearings (floating/semi-floating journals) have very non-linear responses to input conditions...what do I mean by nonlinear? A linear response is what you learned in algebra...y=Mx+B..it's a line on a graph...when you plug in some value of X, you get a predictable response. Non-linear on the other hand do not necessarily have a single solution for each X. They can respond in multiple ways and this can result in odd whirl/whip orbits or in driving the rotor-bearing system into a completely unstable rotor mode. For example, they usually suffer from cross-coupling that causes the shaft reaction forces to respond 90 degrees to the input force (maybe not 90 degrees, that was just for the example)...this means that if I push down on the rotor, it shoots out to the side! They are also sensitive to the oil viscosity since this is what makes up their stiffness. They say that near the upper limit of oil temp, the viscosity is roughly that of water! Because of their non-linear response, they are susceptible to sub-synchronous vibrations that are not an issue for rolling element brgs. Sub-sync vibes can be quite energetic and damaging...and due to there nature are particularly hard to identify the source and solve.
Hydrodynamic bearings have their issues to, but the amazing thing about them is the system damping. They can really knock down a vibrational input if done right! I read a paper on the design of floating journal bearings and they mentioned that if you get the ring speed ratio (RSR - the ratio of the floating journal bearing (the ring) angular velocity to the shaft angular velocity) designed at a wrong condition (you design it by setting the inner vs outer gap, or more precisely, balancing the torques applied by the inner and outer films on the floating journal) the supply holes can act as a centrifugal pump and reduce your oil film thickness on the inner journal, thus the stiffness, load capacity, viscous shear rate, damping, etc! So it has to be right, both for hydrodynamic and rolling element bearings.
The moral of the story here is that hydrodynamic bearings are just as twitchy as rolling element bearings! When we play around with the rotor-bearing system we have to accept the reality that we are playing Russian roulette with our hardware. After reviewing my prior posts, I want to make sure no one thinks that I am telling them that if they go the rolling element bearing route that all is great and easy! I just wanted to make sure that the complete story on ball bearing turbos was being represented. I think the same can be said of the hydrodynamic bearing route.
|
|
|
Post by racket on Mar 16, 2014 22:37:10 GMT -5
Hi Chris
Yes, both have their place , I'd love to be able to successfully run ball bearings but the development costs are just too high for us DIY'ers , balls start to lose their appeal when one is forever buying new rotors .
I've a very good Aeronautical Science text I use .....Analysis and Lubrication of Bearings by M C Shaw and E F Macks , its a first edition circa 1949 , it delves into both high speed rolling element as well as hydrodynamic sleeve bearings as used in aircraft supercharging and the early jet engines , being written in 1949 it assumes the reader "knows nuthin'" , so it explains things in fairly basic terms , ...............it makes interesting reading when taken in context with more modern Papers on the subject , nothing much has changed other than we can measure things more accurately now ,....... they had the basics understood more than 60 years ago .
Theres some handy data in the book on actual rpm of the bush with varying clearance ratios when using a 2" dia journal ............just about the same diameter journal as I'll be using in my SKULD666 engine ;-)
There'll be a number of us here on the Forum who will be very interested to hear how you manage the bearing requirements for your project .
Cheers John
|
|