Honda B20B Belt PSRU

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

PMD

Well-Known Member
Joined
Apr 11, 2015
Messages
1,056
Location
Martensville SK
Efficiency is won back in auto Diesel's by maxing out torque to operate at lowest possible RPM. The reduced pumping/parasitic loads at say ~1200RPM compensate for the reduced efficiency of the pilot/pulsed injection. The pulsed injection is more about NOx/noise and spreading peak cylinder pressures on the crank than about efficiency of each charge The reduced RPM compensates for the combustion efficiency loss. Your modern European Tdi will be squawking at you to change gear at <1500RPM in order to capture this benefit.

Of course none of this is useful in aviation engines running at high RPM MCP or close to for most of their life. I know for a fact that Continental/Theilert diesels don't bother with pilot shots in their custom ECU's but not sure about Austro. Either way, you are likely right in that best aviation power and BSFC prefers rapid and complete injection BTDC at the expense of harmonics/crank loads and noise.
All of course is - exactly as you have pointed out - relative to engine RPM. Historically, diesels were first call "constant pressure" engines, and the very low speed stationery engines could NOT harvest the rapid pressure rise of gasoline or even diesel combustion, but the injection event could be carried deep into the combustion stroke to raise cylinder pressures when crank and rod geometry was able to extract far more work. This is what rate shaping in higher speed engines is also intended to do, but as you said, you can best optimize that by simply moving your peak torque RPM to a lower value. It is not just about "compression ratio" per se, as spark ignition engines must deal with an aspirated charge that limits cylinder pressure due to detonation. You can still make it work very well as RPM goes up (thus giving more mass flow through the engine from normal aspiration). You can improve that with supercharging, but once again, detonation of the aspirated or forced induction charge limits the cylinder pressure. Diesels, on the other hand, can exchange RPM for MAP and move to lower speed where rate shaping CAN be more beneficial than in direct comparison with an SI combustion cycle. This is possible since there is no aspirated charge to detonate. As Billski had pointed out, though, doing this can run NOx way up because along with very high cylinder pressure comes very high cylinder temperature - exactly what make it easier to oxidize nitrogen. Another note: the old style diesel "clatter", the big bang that caused so much of a shock wave that the outside of cylinder walls would erode by cavitation (implosion of vacuum bubbles created by said shock wave) was pretty much due to some of the charge in the cylinder detonating instead of burning. Modern HPCR injection is so able to control combustion that the clatter is gone - along with the old-school diesel rise in pressure at a stage of the power stroke when geometry left little torque to be harvested.

BTW: I was not aware that the Thierlert/Conti engines had no pilot injection event. Thanks for bringing that up, now I need to find out why ( and how they prevent impingement onto the piston crown).
 

TFF

Well-Known Member
Joined
Apr 28, 2010
Messages
17,353
Location
Memphis, TN
This is an Enstrom piston helicopter belt. 205-225 hp from a Lycoming. Transmission is fixed. Engine has regular engine mounts. The Jack strut has a bearing on the transmission shaft that is ball bearing. The bearing on the crankshaft is a ball bearing but is mounted in essentially a spherical bearing so it can articulate. Belt is 8” wide with about 6-8 v ribs per inch. The Jack strut is fixed. Bottom pulley is shimmed in line with the top with belt tensioned.
 

Attachments

  • 00879CA2-B37D-4736-A82D-5F0A3D03CF9D.jpeg
    00879CA2-B37D-4736-A82D-5F0A3D03CF9D.jpeg
    50.4 KB · Views: 41

Geraldc

Well-Known Member
Joined
Nov 12, 2011
Messages
843
Location
nz
Cantilevered belt drives often have shaft deflection issues which cause belt tracking issues. Ben Haas outlined the problems he had with his drive.
This design keeps the cantilever length to a minimum.
Only top pulley shown. Bottom pulley is supported at each end.1630577260055.png
 

Lendo

Well-Known Member
Joined
Feb 6, 2013
Messages
935
Location
Brisbane
Yep! Shaft deflection seems the issue with Tracking on a belt drive. I would suggest a belt tensioner might also be a problem if not set-up right. It is also a problem with gears as well. The early Powersport drive broke teeth on the Spur Gear, which then required the gear to be supported both sides, then no problems. Space to put support on an internal Spur gear (within the Ring Gear) was the main issues Powersport had to solve, which they did elegantly. I didn't see that on other internal Spur gear designs, which would suggest to me that it could be a problem down the track, no matter how good the supporting bearings.
Bottom line everything should be supported both sides.
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,168
Location
Saline Michigan
Efficiency is won back in auto Diesel's by maxing out torque to operate at lowest possible RPM. The reduced pumping/parasitic loads at say ~1200RPM compensate for the reduced efficiency of the pilot/pulsed injection. The pulsed injection is more about NOx/noise and spreading peak cylinder pressures on the crank than about efficiency of each charge The reduced RPM compensates for the combustion efficiency loss. Your modern European Tdi will be squawking at you to change gear at <1500RPM in order to capture this benefit.

This is thread drift. Let's not confuse road vehicle fuel economy with thermodynamic efficiency at airplane power levels. Yes, adjusting camming, induction, calibrations, and gearing to make useable torque at lower rpm and then running lower rpm has always been able to go further down a road on the same fuel. During WWII, early gear changing was taught so your ration of fuel would go further. This is fuel economy of road vehicles that travel at modest power settings. In detail, the effort is not necessarily to run lower rpms, the effort is to get the vehicle to run on the fuel use island (min fuel per horsepower) as much as possible, then to run on the best fuel use line as much as possible the rest of the time. Too low RPM costs fuel almost as much as too high RPM. Understand that this is all fuel economy at modest power operation.

Of course none of this is useful in aviation engines running at high RPM MCP or close to for most of their life. I know for a fact that Continental/Theilert diesels don't bother with pilot shots in their custom ECU's but not sure about Austro. Either way, you are likely right in that best aviation power and BSFC prefers rapid and complete injection BTDC at the expense of harmonics/crank loads and noise.

BINGO. Get into airplanes - where we rarely run less than 50% and usually more like 75% power - and all these low power adjustments are meaningless. Then we must talk thermodynamic efficiency. At airplane power levels, with any fixed RPM and manifold pressure, putting in fuel at or as close to TDC as you can makes the most power on that fuel flow. Here is why. All air heated by burning fuel after the piston leaves TDC can not expand as much as if it had been heated at TDC. We get power by having force applied through rotation. The further down the piston is when the energy is added, the more of that energy goes out the exhaust system instead of powering the crankshaft. Unavoidably. It really is that simple.

The starting point for all of this is how thermodynamically efficient the engine is in the first place. Putting fuel in significantly after TDC is about reducing NOx and NVH, but in an airplane, it makes your fuel bill higher.

Billski
 

tspear

Well-Known Member
Supporting Member
Joined
Feb 12, 2014
Messages
1,229
Location
Outside Boston
@wsimpso1

I was invited a long time ago to a webinar from the EPS (a.k. flat V8) diesel guys.
One comment stuck with me, and you explanation on the thermodynamic side finally made it all fit together.
They said a continuous injection of fuel after TDC allows for a more efficient total system by lowering the structural requirements significantly. The end result is a more efficient total package. They also said something along the lines of optimal fuel burn requires significantly more structural strength and resulting structural mass, combined with significantly higher transmission losses to handle the stress of the vibrations caused by the power pulse. Therefore, for airplanes where weight is the enemy, optimizing for minimal vibration reduces structure on the engine, and also transmission losses and more than offsets the loss of fuel burn efficiency.

Do I understand it correctly (not asking if their comment pencils, just my understanding of what they stated)?

Tim
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,168
Location
Saline Michigan
Cantilevered belt drives often have shaft deflection issues which cause belt tracking issues. Ben Haas outlined the problems he had with his drive.

Absolutely.

Put a shaft of fixed length between two bearings (simply supported beam), put a fixed belt load on a sprocket at mid-length, and the shaft deflects a certain amount and the angle of the sprocket to the belt DOES NOT CHANGE. Put the same length out as a cantilevered beam, with the same load, and the deflection in the direction of belt pull is 16 times bigger, and the sprocket is now at an angle to the belt. It takes a lot of shortening the length and beefing up of the shaft to make up for a fundamental 16 times bigger deflection. The inherent angling of the sprocket remains, but is reduced by reduced shaft length and more beefiness.

All of this assumes that there is no deflection of the bearings. Harrumph. Bearings and cases also deflect small amounts and those amounts go with the size of the loads on them. On the simply supported shaft, load in each bearing is one-half of the load from the belt. On the cantilevered arrangement, the loads are usually several times of the simply supported arrangement. The bearings are thus beefier to carry the load, but usually deflections end up about the same. Same for the cases supporting the cantilever. So what? They both deflect about the same... In the simply supported shaft, they slightly shorten the distance to the sprocket. In the cantilever shaft, one bearing is deflecting up and the other is deflecting down and their being close together makes for more angle on the shaft that we then add in the bending deflection. The belt ends up with even more anglular error to the sprocket.

Can you shim some angle into one of the shafts so belt is well aligned under load? Yeah, you can make it perfect for some intermediate load, say 1/2 of max torque. Then it only has half the alignment error at full torque, and half the alignment error at zero torque. Maybe instead, you set it perfect at 75% of max torque, then you only have a quarter the error at max torque and 3/4 of the error at idle... Maybe that will give a long enough belt life for the application. Maybe the belt still tracks off the sprocket or into the side walls of the sprocket and dies a quick and smokey death. What does the belt and sprocket supplier say is max belt misalignment?

Most of the other guys making belt PSRU for V-6/V-8's use simply supported shafts against belt loads and only cantilever the prop. Maybe they did that on judgement alone or maybe they worked out the beefier cases, bearings, and shafts to get to reasonably small belt alignment errors, and decided that they had a lighter better cheaper design with simply supported shafts. And the guys who went with cantilevered designs did have belt alignment issues. Hmm.

I hope that our OP is within the acceptable range on belt alignment and bearing life so he can fly this airplane without dead stick landings. But without some mechanical engineering calcs and reference to allowables for the equipment chosen, I would not count on it.

Billski
 

PMD

Well-Known Member
Joined
Apr 11, 2015
Messages
1,056
Location
Martensville SK
Billski: Please excuse my intrusion into your last post.

ALL of the things you mentioned are also major considerations in the design of gear drives of any kind. While gears are somewhat tolerant of radial deflection that remains parallel they can be quite intolerant of uneven radial deflection (bending of one or both shafts). I have unfortunately a fair bit of experience with a medium duty truck gearbox that counts on the pilot bearing in the flywheel to keep the "cluster gear" alignment. Even the tiniest bit of wear of that bearing will clean the teeth off of the gear at the other end as the load shifts from the length of the helical gear (i.e. shared by three teeth) to loaded at one end - and it will quite nicely clean a few teeth off of that shaft. Doesn't bode well when for the gearbox and would be a genuine disaster in a PSRU.

Add in harmonics to drive line paranoia and you can imagine that when things get into sync, the actual loads on a gear or belt and shaft(s) can be many times what the torque being delivered (and calculated) can be.

Also, BTW, why over time, I have become far more tolerant and receptive to the concept of poly V belt drives.
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,168
Location
Saline Michigan
Therefore, for airplanes where weight is the enemy, optimizing for minimal vibration reduces structure on the engine, and also transmission losses and more than offsets the loss of fuel burn efficiency.

I suspect some important details were missed or others exaggerated...

Having been in the business of applying engines to vehicles, I can tell you that peak internal engine loads are somewhat higher with TDC injection vs tailored schemes, but not hugely so. When the crank has some angle on it and then the piston gets big pressure, it makes a second torque spike on everything and the combination is about as bad as with a single fuel squirt near TDC. Margins on this stuff are not usually cut that thin. With common rail diesel, we keep getting higher rpm and that is driving crankshaft size up more than anything. Crankshaft journal size is set as much by getting crankshaft resonance above 2x max firing frequency as by strength or bearing capacity. Not much changes there, and no one builds to small margins in diesels anyway for other reasons.

Then everyone puts a rather large inertia on the crankshaft with either a manual tranny flywheel and clutch or with the flex plate and about 3/4 of the torque converter. This tends to reduce the amplitude of the speed changes. Next in line is then a springy element that effectively smooths out most of the firing and 2x firing pulses from the combined downstream inertia of the gearbox, shafting, final drives, and wheel/brake/tire sets. What vibe remains after the isolator goes to the transmission and looks like a firing order sinusoid of modest proportions and another one at twice firing order. If the powertrain builder also includes a bifilar pendulum damper, a big fraction of the remaining firing order gets smoothed out too. As a tranny guy who owned the torque converter isolator (among other things) I never heard anything about the new schemes making our lives easier...

The tranny's job is really not changed much by how fuel is added. Maybe the crankshaft's job is a little harder, but I detailed how there are other factors that make it a wash. Crankshaft guys never said that change made their lives easier either...

For the extant aero diesels, a very similar situation occurs. In carried over automotive engines, you get what they had for structure. No change there. For purpose designed aerodiesels, the crank is still designed to be stiff enough to put resonance safely above 2x firing frequencies, there is still a flywheel that is accelerated with the crankshaft, there is a torsional spring device that isolates engine vibe from the gearset, and then there is still a big inertia on the other end from the prop. The required weight of the engine and gearbox will not be changed much if at all by fueling schemes...

Putting on either the owner-operator hat or the management hat, I would take the substantial increase in efficiency every time in an airplane. Yeah, the engine guys and tranny guys will have to design for the whole schmear either way, but in a clean-sheet design, they would be doing that no matter what scheme you handed them. And if it does add a bit of weight, I bet it is way less than the weight of the added fuel required for one trip from running a lower efficiency engine.

Billski
 

lelievre12

Well-Known Member
Supporting Member
Joined
Jul 15, 2020
Messages
167
Just to add two cents, the Thielert TDI does not have a rubber element between the crank and PSRU. Instead is retains the 'auto' style clutch plate. I guess the theory here is that at critical RPM ranges, the clutch plate will 'slip' and protect the gearbox from overtorque. Its pretty clear that the spring plate which holds the clutch is pretty fine tuned in terms of Newtons of pressure on the plate. I guess they went through quite a bit of trial and error to make it 'slippy' enough.

Here is a photo:
IMG_9399.jpg
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,168
Location
Saline Michigan
Just to add two cents, the Thielert TDI does not have a rubber element between the crank and PSRU. Instead is retains the 'auto' style clutch plate. I guess the theory here is that at critical RPM ranges, the clutch plate will 'slip' and protect the gearbox from overtorque. Its pretty clear that the spring plate which holds the clutch is pretty fine tuned in terms of Newtons of pressure on the plate. I guess they went through quite a bit of trial and error to make it 'slippy' enough.

Here is a photo:
View attachment 115074
No surprises here at all. Let's look really close at the photo:
  • Starting at the top is the flywheel with a perforated wheel for a crank position sensor to read, some spiral grooves and rust marks;
  • Next down is a two sided friction plate with friction material riveted on and some sort of hub riveted in too;
  • Then there is a pressure plate of some sort.
Going back to the flywheel- look at the surface that the friction plate interfaces with. Note the rusted circles and lines on the flywheel that look like the friction element shape. This clutch has spent a lot of time with the friction material NOT slipping. If it was ooching its way a couple degrees at a time on firing pulses during some part of normal flight cycles, the rust would be pretty uniformly distributed. I suspect that this system rarely slips, and that slipping is not part of firing order vibration management. Slipping a clutch also does a lousy job of vibration isolation. There are still excellent reasons to use an overload clutch. In parallel hybrid drives (Toyota, Ford, FCA, and others) it prevents trashing the engine and gearbox when an ABS sensor fails or the car is driven at high power over patches of alternating slick and grippy pavement. Toyota omitted the overload clutch in the second year of Prius production, then put it back in ever since. The other car makers use the clutch too.

Now look down by the bottom of the photo through the large diameter center hole in the pressure plate. You can see two coil springs that appear to be oriented like they go around in a circle. Orient it better for viewing and it would look sort of like this:
1630682263724.png

Those springs are arranged and applied as a torsion spring between the clutch disc and the hub that splines to the input shaft into the gearbox. Yes, no rubber giubo, just steel springs. This is a soft spring element I keep talking about that serves to isolate the downstream elements from the firing pulses on the crankshaft.

No doubt about it, the pictured system has a substantial flywheel inertia on the engine side, a soft element to isolate the primary and secondary vibrations of the crankshaft from the downstream parts, and an overload clutch in case of a prop strike and/or for failure mode management.

This clutch disc may have came off the shelf, and I could very safely bet the spring rate was chosen to put first resonance frequency neatly between firing order at cranking speed and firing order at idle.

Billski
 
Last edited:
Top