snips....
OK Doug I hope this makes you happier.
The clamping force for the CVT goes through the gray pins into the green plate
and then
in to the housing through the large needle thrust bearing. You may have to use
as
many as three back to back here. Other types of thrust bearings will work
as max RPM is only about 8000. Thrust load is about 1000 pounds.
The 100 pound modest clamping force for the first stage is provided by the
three
Belleville spring packs. The one bearing running at 100,000 RPM is the thrust
bearing on the
half inch input shaft so a ceramic bearing is called for here. The planet disk
bearings
are only going about 20,000 RPM or less with a 33 pound thrust load. They are
a slip
fit on the gray pins.
This arrangement reduces the thrust on the 12 mm high speed bearing from 1000
pounds
to 100 pounds.
Paul Lamar
Paul
Yes the bearing arrangement now looks a lot better and when we need the
planetary unit we can come back to it. I'm not totally convinced we need the
unit at this point in time.
Perhaps we have to examine our designs premises. We have always assumed we
have to start with 100K rpm from the turbo. In a two turbo test this is not
the case. In the one turbo case with or without the electric boost, we will
be designing or finding a turbo that may run at lower speeds to bridge the
gap between two applications and perhaps simplify reduction devices. The
electric boost may help bridge this gap as well though this is not the
original design application Mazda had in mind. These are unknowns and IMHO
they won't be known until I can get some test data. I realize that car
turbos commonly spin up to 150K rpm, and those rpms come about as a result
of compensating for small size and other design variables of which cost is
likely very important. It won't be particularly easy to come up a
multi-purpose turbo but we are designing the back end to it as if we have it
already.
In any case for testing, and I am assuming in application design as
well, we need a CVT. Wouldn't it be better to spend our efforts on getting a
working CVT together that I can use in a few months? I got my dyno
information package now and have begun to build the test stand. You are a
busy guy and I hate to see you spinning your wheels :) trying to get a fix
on a device that is a moving target at this point in time. I understand you
have tremendous emotional attachment (more than 30 years worth) to the
rotary turbo compound concept and I respect and admire you for it. It is an
elegant idea and won't go away. I'll be the first to admit that there are
probably many design factors that I haven't thought about as yet, which you
may have already dismissed as unfeasible. I need to know though. For
instance. Can a CVT be run at turbine speeds? They don't need to in current
auto applications but....
The CVT prototype I am now designing is about 90mm in outer diameter. The
inner part of the toroid wheel has a radius of 12mm and the outer radius at
37mm so we have a maximum reduction of about 3:1. I would think if
connected into the e shaft with a large 10:1 toothed pulley or even a simple
planetary we don't need more than a 2:1 CVT. Less is better for fine control
of rpms. Ie; a slight adjustment of the CVT can mean large rpm fluctuations.
This is the advantage of the putting the planetary in front of the CVT. I am
still thinking only about getting through the primary testing though.
For ease of explanation lets say the torque transfer wheels are about 32mmų
or about 100mm in circumference. Correct me if I am wrong but each transfer
wheel in the high aspect ratio would revolve about 2.33 times and travel
about 233mm for each revolution (74 x pi) on the input torid wheel. At 40k
the velocity would be about 155 M/sec. The bearings would be at 93,200 rpm.
For turbine bearings.... No problem.
Using the train analogy again; when the high speed train in France
reaches 550km/hr the wheels moves along the rail at about 153 M/sec.
At higher speeds:
At 80K the CVT transfer wheels would move along at about 310 M/sec. and the
transfer wheel bearings are at 186,400 rpms.... Within the realm of bearing
possibilities.. just. The unit and the bearings are tiny as we don't have
that much torque to transfer if at these higher speeds. Thrust bearings on
the unit can be much bigger though.
One point I need cleared up. Is the clamping pressure required a function of
the amount of contact surface area? If for instance we double the contact
area is the required clamping force halved? In either case we are speaking
presently about point contact on the profile tapered wheels or at best a 1
mm square area per wheel if we profile the wheel to be slightly flat on the
outer diameter. 2mm squared for transfering a maximum torque of 4 ft/lbs at
40K. The output of the CVT is seeing 2 times that torque though isn't it?
Half the rpms and about 7.9 ft/lbs of torque. Half that if we are running at
80K into the unit.
The CVT transfer wheels side thrust loads concern me and would increase as
the clamping pressure increases. They can be at some very oblige angles at
the outer edges of their range. Their bearing loads and friction increase
the more we clamp. I would think our best design point would be to achieve
no more clamping pressure than is needed to provide the proper contact
friction to induce the rotational movement. We may find that this clamping
force requirement is different in a high speed CVT design environment. Ie:
higher speed increases the (stickiness?) of the oil or foaming tendencies
lessens it. We can't say for sure now. Unfortunately we can't test it by
spraying special oil in front of train wheels at high speeds:)
I am off to Hiroshima within 20 hours and away from my computer for four
days.
Doug in Japan
For hard steel on hard steel the friction coefficient is independent of the contact area.
Paul Lamar
The Rotary Engine NewsLetter. Powered by Linux.
ACRE NL web site. http://home.earthlink.net/~rotaryeng/
Copyright 1998-2004 All world wide rights reserved.