View Full Version : Force plate treadmill summary

10-29-2007, 12:16 PM
The original question was posted:

Dear colleagues:
We are in the process of acquiring an instrumented treadmill with built-in
force plates. We are in between two vendors: 1) the Bertec (two plates
placed right/left with two individual split belts) and 2) the Kistler (two
plates placed front/back with one belt). I would really like to get some
feedback about the Pros and Cons of each piece of equipment. This will
help us tremendously to obtain the best treadmill.

A summary of the responses is provided.

Shing-Jye Chen, Ph.D.
Co-Director of the HPER Biomechanics Laboratory
School of Health, Physical Education & Recreation

The following messages are given by the colleagues, and I categorize them
into Kistler and Bertec treadmill and others.

I. Kistler:

"Dewitt, John K. (JSC-SK)[WLS]"
10/03/2007 09:34 AM
Hi sj,
We have a Kistler Gaitway treadmill in our lab and have used it in a
variety of investigations, including parabolic flight. It is relatively
simple to use, and is accompanied by manufacturer's software that is easy
to use and kicks out custom reports, which is useful in a clinical
setting. It is also possible to obtain raw force transducer and speed
sensor data, which we use for processing using Matlab.

The pro of the design is that the front-rear force plate setup allows for
both walking and running, and steps can be decomposed as right/left based
on the foot striking the front plate and pushing off the rear plate. If
the subject drifts too far forward or backward of the midline, some
footfalls may be difficult to discern. However, the problem there is
mainly in the Kistler software, and if you use custom software (i.e.
Matlab), you can get rid of that problem.

We have had some issues in the past with treadmill vibration that affect
the data. In addition, in some cases the GRF profiles have looked odd. I
believe that the odd GRF (a depression in the propulsive hump) is caused
because during some portion of foot-ground contact, forces from both
plates need to be combined. Slight calibration differences between the
plates can cause this issue. I believe it can be overcome with analysis.

I hope this helps.


"Andrew Kwarciak"
10/04/2007 01:12 PM

Dr. Chen,
We have an older Kistler system (within an HP Cosmos treadmill) and we've
been having difficulty getting accurate measurements of known weights. The
errors in our measurements are non-linear and the belt seems to affect the
measurement more than it should. We are also having difficulty obtaining
data from each plate separately. Our system cannot properly measure the
forces on each limb during dual support stance. These issues may have been
resolved in the latest equipment; however, I would make sure to ask plenty
of questions and ensure that the software can provide the output you need.
Hope that helps.

Good luck,

Andrew Kwarciak, M.S. Senior Biomedical Engineer
Human Performance and Movement Analysis Laboratory
Kessler Medical Rehabilitation Research & Education Center
1199 Pleasant Valley Way
West Orange, NJ 07052
Tel: 973-243-6903
Fax: 973-243-6984

Paul Bussman
10/04/2007 02:38 PM

Hello Dr Chen,

The current version of Gaitway uses a better quality treadmill from
HPCosmos with a more robust frame and drive mechanism. Some of the major
hardware differences between the HPCosmos and Trotter versions are:

Belt speed - old version 19.8 km/h - current version 22.0 km/h with
optional 30.0 km/h
Grade - old 20% - current version 24% with optional -24 to +24%
Belt reversal - old version not available - current version available as
an option
Dimensions - old version 1400 mm by 500 mm - current version 2100 mm by
820 mm
Weight - old version 320 kg - current version 364 kg

The current version of the Gaitway software is significantly improved from
the version that was sold with the Trotter treadmill. The old Trotter
version Gaitways can and have been be upgraded to the current version of
the Gaitway software.

Best regards,

Paul Bussman
Kistler Instrument Corp.
75 John Glenn Drive, Amherst, NY 14228-2171, USA
Tel +1 716 691 5100 (direct 716 213 5781)
Fax +1 716 691 5226
II. Bertec

"Bing Yu"
10/03/2007 01:44 PM

It is quite clear that Bertec treadmill allow you to measure ground
reaction forces on each foot. We already bought one from Bertec. I can
not see any logic and use of the setup of force plates in Kistler

Bing Yu, PhD
Associate Professor, Director
Center fof Human Movement Science
The University of North Carolina at Chapel Hill

"Lee E. Johnson"
10/04/2007 12:28 PM
Hi Shing-Jye,

Caught your post to biomech-l. We have long been interested in
instrumented treadmills as part of a total solution for gait analysis.
Kistler was one of the first ones into this market but because their
treadmill was limited to measuring vertical forces we have not done an
integration of that product. The Bertec instrumented treadmill measures
shears as well as vertical forces and, as a result, we are in the process
of integrating their product with The MotionMonitor. We are still early
in the process so we have not had a chance to do any accuracy testing but
they have a great reputation for engineering and product quality so we are
confident that any deficiencies will be quickly corrected.

Good luck and let me know which way you go.

Lee E. Johnson
773.244.6470 voice
773.244.6473 fax
Innovative Sports Training...The Total Solution in Motion Capture

From: Murat K. Berme [mailto:murat@bertec.com]
Sent: Friday, September 28, 2007 7:52 PM
Subject: RE: Concerns and questions regarding the Bertec dual-belt
instrumented treadmill-HPER Biomechanics Lab at UNO

Concerning walking naturally on the split-belt treadmill:

The University of Delaware looked at how a subject’s gait is affected by a
split-belt treadmill. They found that after 5-6 minutes the gait returns
to normal. Please peruse the article for yourself through the following


Furthermore, the split-belt nature of the Bertec Fully Instrumented
Treadmill is great for perturbing a patients gait and/or balance (request
our overhead support structure with harness, if you will be doing this).
Also, Amy J Bastian (PhD PT) of the Kennedy Krieger Institute used a
split-belt treadmill (not instrumented) to discover some fascinating
things about the gait recovery of stroke survivors. She was actually able
to return the survivor’s gait to normal by running the left and right
treads at different speeds! We found her research through the following
CNN.com article:


Concerning shear forces:

The Bertec Fully Instrumented Treadmill (FIT)’s treads are part of a
closed looped system. Other “instrumented” treadmills have the treads
directly over the force plates: this means that the force plate is being
rubbed by the tread. The Bertec FIT only measures external forces (i.e.
your subjects’) to the closed looped system. I have attached a force
diagram explaining this. This means that yes, the Bertec FIT accurately
measures shear forces.

Concerning Center of Pressure (CoP):

The Bertec FIT was tested for CoP accuracy along its edge using static
calibrated weights. I’ve attached the results, which show accuracy of +-
2mm in the x-direction and +- 3mm in the y-direction. I believe that this
particular test was performed inaccurately, and the actual accuracy is
higher (+- 1mm in the x, and +- 2mm in the y). I will have this test
rerun next week, and I will explain why the results are different.

But in the meantime, +- 2mm in the x-direction and +- 3mm in the
y-direction is pretty good. Please note that the “CoP accuracy” pdf
attachment is not drawn to scale.

Concerning Sampling rates:

Bertec’s force plates (including the ones built into the Bertec FIT) use
Bertec designed 16-bit technology to acquire and condition the signals
coming from the transducers. This data is sampled at 1000Hz; therefore,
providing a useful bandwidth of 500Hz. It is worth noting that when the
Bertec FIT is used with a motion capture system (integrated as easily as a
force plate is) the signal is converted back into analog for the motion
capture’s A/D board. At that point, you can sample the data via the
motion capture system at whatever rate you prefer.

Concerning Natural Frequency of the system:

The incline system tested to have a natural frequency of 200Hz. The
treadmill (force plates) themselves have a vertical and shear natural
frequency of 240Hz and 220Hz, respectively.

Concerning Patient Harness:

Bertec offers a static overhead support structure with a harness for
patient safety. It is very similar to the one used for our BalanceCheck
product and can be seen in some of the pictures on the following page:


However, more interesting is our FIT installation at the Denver Children’s
Hospital, where we have partnered with Robomedica to provide a Body Weight
Support System over top of our Treadmill. Their system can be powered and
run off of the same rack-mount control system as the Bertec FIT, giving
you a cleanly integrated solution. Their product website is below:


Concerning Cross-Over Steps:

Placing both feet on one side of the treadmill at the same time will
result in force measurements that represent the superposition of both
feet’s Ground Reaction Forces (GRF). If this is an occasional occurrence
within a trial, then you can easily dismiss the data from those few steps;
however, if the subject consistently walks with a cross-over then you will
not have accurate GRF for each foot, unless the subject is jogging or
running. We are currently designing and building a “mini” quad-belt
treadmill for animal research, so it would be possible to design a
“front-back” treadmill, but I’m not sure that there is any literature
supporting this kind of treadmill design.

Concerning the end of this email:

Here it is. Please don’t hesitate to contact me. I will be sending you
some data of the Bertec FIT in use. Also, I’ve attached a copy of the
Bertec FIT’s manual (rough copy), which addresses the issue of the origin
of the force plates coordinate system.

Seven (7) year warranty covers all hardware and electronics.
Dr. Cenk Guler (614-430-5421 Ext: 223) is the lead designer for the Bertec

III. Others

"Christopher MacLean"
10/04/2007 11:43 AM

You might want to contact Steve Swanson at Frappier Acceleration to
inquire about their new 3-D force treadmill..

Christopher L. MacLean, Ph.D.
Director of Biomechanics

Paris Orthotics Ltd.
3630 E1st Avenue Vancouver, BC V5M 1C3

T 604.301.2158; 800.848.0838 Ext.158
F 604.301.2151
E cmaclean@parisorthotics.com