M.E.

Department Water Brake Dynamometer Unit 2012

Abstract

California State University Sacramento Mechanical Engineering departments students designed and
built a water brake engine dynamometer power absorption unit in 2002 for their capstone design
project. The absorption unit may have worked, but history of its operation after completion of the
project remains unknown until 2009 when the cell was attempted to put back into operation. Defects
were found in the sealing and bearing mechanisms in late 2010. Water had continually leaked past the
seals during operation and past the bearings seals corroding the bearings. The bearings eventually
seized, and the impeller spun on the inner race. The impeller shaft showed abnormal wear.
A design fix using mechanical face pump seals and thin section bearings was then created that involved
minimal parts additions and simple modifications. All machine work was completed in California State
University Sacramentos student machine shop. The modifications were tested against high inlet
pressure water of 55[PSI] and at a maximum dynamometer speed of 5800 RPM during engine testing.
The power absorption unit was dissembled and checked for indications of water leaking past the seals,
proper bearing functioning, or signs of the impeller making contact with the case. All the
aforementioned items checked out okay, showing that the implementation of the modifications to be a
success. More testing will be conducted to verify the components work over a much longer period of
time.

2|Page

M.E. Department Water Brake Dynamometer Unit 2012

Table of Contents

Abstract ......................................................................................................................................................... 2
Introduction and Background ....................................................................................................................... 4
Dynamometers and the Water Brake Design ........................................................................................... 4
CSU Sacramento Mechanical Engineerings Water Brake Unit ................................................................ 5
Bearing and Seal Redesign .......................................................................................................................... 10
Modified Dynamometer Assembly ............................................................................................................. 14
Operation, Inspection, and Conclusion....................................................................................................... 17
Works Cited ................................................................................................................................................. 18
Appendix A Unmodified Dynamometer Original Drawings ..................................................................... 19
Appendix B Unmodified Dynamometer Solid Model Renderings............................................................ 24
Appendix C Modified Dynamometer Drawings ....................................................................................... 26

3|Page

M.E. Department Water Brake Dynamometer Unit 2012

Introduction and Background

Dynamometers and the Water Brake Design

Dynamometers date back as far as the late 18th century. They are used to measure a force, moment, or
power of some input device. They usually consist of a power absorbing unit (PAU) and a sensor or
calibrated measuring device that can directly, or indirectly, measure the horsepower being developed by
the input device. Uses of dynamometers vary across a broad spectrum. The medical field uses them for
testing peoples hand strength and other body, or muscle, functions. Other uses include, but are not
limited to, maritime equipment, electrical motors, the automotive sector, and aeronautical and
aerospace applications.
Dynamometers can be divided into two categories; driven and driving. For example, driving
dynamometers can be used to spin an internal combustion engine without combustion. This can help to
analyze the oiling system without combustion bearing loads, or look at the inertial and dynamic effects
of the moving masses. Driven, or passive, dynamometers allow another device to move, or power them,
and provide some type of load to the engine to test a variety of different things and/or conditions. For
the purposes of this report, and the project, driven dynamometers are the only area of concern.
The automotive sector uses dynamometers widely, for everything from measuring the power an engine
is making, to testing brakes, transmissions, or other systems. One of the simpler, more compact designs
is of the water brake variety. Through some mechanism, the water provides a load on the device that is
inputting power to it. That load is usually developed through an impeller that spins with the engine
being tested. The shearing forces of the water between the case and impeller, along with the change in
momentum required to move the incoming water, provides the necessary force. Many long standing
engine dynamometer companies have utilized variations of the water brake design for their customer
units. Land and Sea Inc., Superflow, Go-Power, DTS, and Stuska are among the most recognized water
brake dynamometer manufacturers in the industry. Depending on the application, the unit can be
configured for a direct shaft input, chain drive, hub, or wheel drive.

4|Page

M.E. Department Water Brake Dynamometer Unit 2012

CSU Sacramento Mechanical Engineerings Water Brake Unit

Roughly a decade ago, a mechanical engineering capstone project team from California State University
Sacramento (Sac State) undertook the task of modeling, casting, and assembling a water brake cell. The
project was overseen by now retired professor Joseph Harralson. Conversations with Professor
Harralson revealed that the Sac State dynamometer is based on one of Stuskas small engine water
brake absorption units. At the time of the project, Stuskas smallest offering went under a different
name, but the current analogous unit is known as the XS-19. Specifications for the unit from Stuskas
water brake brochure can be seen in Figure 1:

Figure 1(Stuska Dynamometers)

5|Page

M.E. Department Water Brake Dynamometer Unit 2012

The Sac State dynamometer has very little available documentation. Professor Harralson provided the
original 2D part drawings, but there is little information known beyond those. The condition of the
dynamometer cell in April of 2009 was not ideal. Additionally, the test stand was not in a condition to
safely operate an engine with the cell. After some repairs and reconfiguring, a Honda CBR600F4i with
Sac States Hornet Racing Formula SAE intake, exhaust, and electronics was successfully started and
run against the water brake dynamometers load. It was quickly discovered that the seals to keep the
water from entering the impellers supporting bearings had failed. Furthermore, the impeller felt as
though it had excessive drag, indicating that the bearings were in poor condition. The cell, before it was
disassembled to fix the problems, is shown in Figure 2:

Figure 2

6|Page

M.E. Department Water Brake Dynamometer Unit 2012

The original inboard seals can be seen in Figure 3

Figure 3

7|Page

M.E. Department Water Brake Dynamometer Unit 2012

The bearing, which resides behind the seal, cannot be seen from the perspective in Figure 3, but the
results of bearing failure can be seen below in Figure 4. The corrosion discovered on the shaft highlighted by the yellow box - shows that the inner race of the bearing was spinning relative to the
shaft, as opposed to spinning with it.

Figure 4

The corrosion clearly indicates that the seals had ceased to function; allowing water to enter bearing
cavity. Upon inspecting the original drawing files (available in Appendix A), the original seal specified to
be used was National Oil Seal PN 471276 from Timken. The specifications for the seal are below in
Figure 5

8|Page

M.E. Department Water Brake Dynamometer Unit 2012

Figure 5(Timken)

Most engines to be tested with the Sac State dynamometer cell are from motorcycles and under 600ccs
in swept volume. Many of those engines have a mechanical redline of up to 15,000 RPM. In the case of
the Honda CBR600F4i, the projected test maximum RPM is 13,000. If motor testing is conducted in the
1:1 6th gear, the output sprocket on the motor is the same as the input sprocket on the dynamometer,
and taking into consideration the 1.822 primary reduction in the Honda motor, the dynamometer will be
spinning at 7135 RPM. That is well above the 2,500 RPM limit for the specified seal. During operation of
the dynamometer, the case may also fill completely with water and begin to build pressure inside. It is
very possible, given the city water feed in the Internal Combustion Laboratory has a static pressure of 60
PSI , that the pressure inside the dynamometer case has far exceeded the recommended 4 PSI design
specification of the seal. Both of the aforementioned conditions are extremely likely to have caused the
seal failure and subsequent bearing failures. The bearings used also contained seals, which must have
experienced some pressure to fail despite those seals.

9|Page

M.E. Department Water Brake Dynamometer Unit 2012

Bearing and Seal Redesign

Obviously, a different seal was needed in order to prevent water from entering the bearing area and
eventually causing catastrophic corrosion within the bearing itself. Research showed that Depac Dyno
Systems, a company that makes measurement and automatic load control units for water brake
engine dynamometers from many different companies, had a recommended upgrade package for
Stuska dynamometers. That package specifies the replacement of the original Stuska seals with PACSEAL #237 from Grainger 5NC16(Depac Dyno Systems).
The PAC-SEAL #237 proved to be too big for the Sac State dynamometer application, but PAC-SEAL
offers many more seals in different sizes. They all center around a carbon on ceramic face seal system.
The mechanical face seal was decided to be the best option due to the relatively high operational limits
of that class of seals. Figure 6 shows a basic cross section view of a mechanical shaft seal. As a side note,
the Grundfos document on mechanical shaft seals contains an abundant amount of information about
the seals including formulas for calculating leakage rates, face heating rates, and many other things.

Figure 6(Grundfos Management A/S)

Finding a seal that would fit in the confines of the already manufactured Sac State dynamometer proved
difficult. A type 304 stainless steel body seal of PN 9281K352 (analogous to Pac-Seal Type 21 and
industry standard number 185) was selected from McMaster Carr due its inner sealing surface diameter
being very close to that of the dynamometers impeller shaft OD along with having a working length that
could possibly fit in the area available inside the dynamometer. The seals are relatively inexpensive at
$18.14 a piece, especially when compared to the $114 for the aforementioned PAC-SEAL. The design
limitations for the selected seal specify a maximum operating pressure of 250 PSI and a shaft speed of
no more than 5000 FPM. A basic calculation to find the surface speed of the impeller shaft is shown in
1 []
[] = []
12 []
[] = 7135 1.25

1
= 2334
12

10 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

The 2334 [FPM] is well under the specified maximum of 5000 [FPM] giving the seal a safety factor of
2.14 for its maximum operational rotational speed. The dynamometer could spin to 15278 [RPM] before
reaching the limits of the seal in that aspect. Furthermore, the dynamometer should theoretically
operate at a lower pressure dynamically than the incoming waters static pressure. Assuming the worst
case scenario of the dynamometer case filling completely during operation, water heating from
operation and localized pressure build up from other unknown effects, 100 [PSI] seems possibly inside
the case. That is still well under the specified 250 [PSI] limit, giving a safety factor of 2.5.
Fitting the seal proved to be much more of an issue than anything else as the space between the
impeller and the case did not allow for the original bearing to work with the new mechanical face seal.
Additionally, the impeller shafts design had three steps in it, as can be seen in Figure 4 and in the
original design drawings below in Figure 7

Figure 7

The mechanical face seal selected required a working length of 1.062 in order to maintain the correct
spring force of the moving portion seals face on the non-moving sealing face. Additionally, the nonmoving part of the sale required a 1.875 ID on the case to fit its 1.875 OD to seal properly along the
outside of the seal. The ID of the moving seal portion also required a 1.25 OD on the impeller shaft to
work as designed. As can be seen in Figure 7, the original shaft did not have any section with a 1.25 OD.
The internal sealing and bearing areas of the case also did not meet the necessary requirements to allow
the seals to work properly. Furthermore, the impeller must maintain alignment, because there was very
little distance between it and the walls of the case; a necessity to create the load from the shearing
action of the impeller and the water. The original design called for 0.071 of clearance on either side
which can be seen in Figure 14 in Appendix A.
Both halves of the case, and the impeller were then carefully measured using the granite flat table to
work from to verify that they matched the original designed dimensions. The faces of the case where
the separate where used as the datum planes to take all measurements from. . In order to easily make
design changes a solid model was then created in Solidworks to match its real life counterpart. Views of
the finished model can be found in Appendix B Unmodified Dynamometer Solid Model Renderings.

11 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

A cross section view of the model assembly in Figure 8 shows the area available to make changes. The
distance between the where the bearing sits to the impeller was 1.15.

Figure 8

This was not enough to keep the original bearing in its original position and place the new seal. Several
ideas were thought of for a valid solution. The first viable solution included the use of thin section
needle bearings in the outboard sections of the case with thrust element to be placed on the ends of the
case to locate the impeller relative to the case. The area to do this is highlighted in green in Figure 8. It
was determined that the needle bearing solution would be too complicated and possibly not locate the
impeller properly with the thrust bearings being that far outboard.
A final, viable solution to the bearing and seal packaging problem included the use of thin section ball
bearings that could both handle radial and thrust loading without using much space. A deep groove ball
bearing with PN 61806_2RS1 from SKFs bearing division was selected. The bearing has a width of 7
[mm], an ID of 30 [mm], and an OD of 42 [mm]. The OD required some changes to the case, because the
12 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

area that the bearing would rest in was much bigger. Specifications for the bearing can be found in
Figure 9 which shows the bearings limiting speed to be 9500 RPM, which is under the 7135 RPM
required to test the Honda CBR 600F4i engine. Additionally, the shafts OD in the area that the bearing
would rest did not match size required by the bearing. An adapter needed to be made for the shaft to
fit, and locate the bearing properly.

Figure 9

An adapter was designed to properly fit, and locate the non-moving portion of the seal in the case
halves. The drawing for the case sleeve can be found in Figure 21 in Appendix C Modified
Dynamometer Drawings. The sleeve was originally designed to press in with 0.001 interference fit with
the case and were to be made of brass due to the ease with which it can be machined. It was also do
less damage to the case if something were to go wrong when pressing it in. Unfortunately, only enough
brass was available for one of the case sleeves, so the second was made from a 6061 alloy.
Another adapter was designed to locate the bearing on the impeller shaft, and provide the necessary
1.25 OD for the moving part of the mechanical face seal to seal properly. Its drawing can be seen in
Figure 22 in Appendix C Modified Dynamometer Drawings. Brass was the originally specified material,
but it proved to be too soft to remain dimensionally stable during the machining process for the
designed dimensions. Instead, 6061 aluminum alloy was used which proved to machine out to the
design requirements on the lathe.

13 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

To fit the adapter on the impeller shaft, the shaft itself need to be modified. Those modifications
involved turning the shaft down to a uniform diameter to allow the adapter to slip on with a 0.002
clearance fit. They can be seen in Figure 23 in Appendix C Modified Dynamometer Drawings. An image
of the impeller after machining was completed on the lathe is shown below in Figure 10. The adapter
was positively located on the shaft with Loctite 680 Retaining Compound. Additionally, the case
needed be modified to properly fit bearing in the cavity and retain the correct working length for the
mechanical face seals. The drawing for the modification can be seen in Figure 24 in Appendix C
Modified Dynamometer Drawings.

Figure 10

Modified Dynamometer Assembly

After new components were made, and old ones modified, the dynamometer needed to be
reassembled. The case sleeves of Figure 21 were first attempted to be pressed in. The casting of the
case did not allow for an easy press fit. The cases were setup on the Herco milling machine in the
student machine shop after which the bores where the sleeves would reside were cleaned up, and
slightly enlarged to allow for slip fit. The case sleeves were positively located through use of the same
Loctite 680 Retaining Compound used on the impeller shafts. Sleeve fitment in the case along with the
mocking up of the stationary part of the mechanical face seal can be seen below in Figure 11.

14 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Figure 11

Another image of the case sleeve in place with the entire mechanical face seal stationary and moving
components being placed for a clearance check can be found below in Figure 12.

Figure 12

15 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

The bearings were then placed in the case, again using the retaining compound. The mechanical face
seals were then installed on each side of the impeller. Final assembly just before the two halves of the
cases were mated can be seen below in Figure 13. The purple compound on the mating face is Loctite
510 flange sealant compound. The two halves were mated mechanically using 28 grade 5 nuts and
bolts of the appropriate lengths.

Figure 13

16 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Operation, Inspection, and Conclusion

The dynamometer cell was tested on the new test apparatus created by another Sac State Mechanical
Engineering Department Capstone project team. Initial hand rotation of the impeller detected no
interference issues. Water was the plumbed to the cell and run through it without movement of the
dynamometer impeller. Pressures recorded during testing reached as high as 55 [PSI]. No leaks were
detected anywhere around the sides of the cell where water would run through the seals and then the
bearings.
The next test involved running the motor against the load of the dynamometer to high RPMs to test the
seals and bearings dynamically. The motor was run as high as 10500 RPM with sustained loading.
Unfortunately, it could not be run any higher, because the motor did not make enough power to do so.
The loading valve proved to be very sensitive and could not effectively vary the load on the motor at
that RPM without bogging the motor, or allowing the motor to unload and bounce off the rev limiter.
After the tests were completed, the case was dissembled and sides checked for any indication that
water may be have seeped past the sealing mechanism into the bearing cavity. No such evidence was
found; indicating that all modifications related to the sealing mechanism was a success. The bearings
had no excessive play or abnormal friction throughout their rotation. Additionally, the impeller had no
signs of contacting the case demonstrating proper location of the impeller relative to the case, thus
maintaining the correct clearance.
More testing will be needed to ensure that the modifications do not fatigue prematurely. The unit will
be disassembled when another 10 hours of testing have been completed to verify all components are
working properly.

17 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Works Cited

Depac Dyno Systems. Go Power Dyno Support. October 2008. December 2011.

Grundfos Management A/S. "Mechanical Shaft Seals." 2009. Grundfos. PDF Document. 9 November
2011. <http://www.mechanicalseals.org/index.php/mechanical-seals-documents/17brochurecatalogue/61-grundfos-mechanical-shaft-seals>.
Stuska Dynamometers. "Stuska Dynamometers: Stuska Water Brakes." n.d. Stuska Dynamometers Web
Site. Document. 30 January 2012. <http://www.stuskadyno.com/pdf/waterbrake.pdf>.
Timken. "Products and Services: Catalogs and Literature." n.d. Timken Company. Document. 16
December 2012. <http://www.timken.com/en-us/products/Documents/Seal-SpecificationGuide.pdf>.

18 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Appendix A Unmodified Dynamometer Original Drawings

Figure 14

19 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Figure 15

20 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Figure 16

21 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Figure 17

22 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Figure 18

23 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Appendix B Unmodified Dynamometer Solid Model Renderings

Figure 19

24 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Figure 20

25 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Appendix C Modified Dynamometer Drawings

Figure 21

26 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Figure 22

27 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Figure 23

28 | P a g e

M.E. Department Water Brake Dynamometer Unit 2012

Figure 24

29 | P a g e