Last summer I did some extensive testing of the Penny and Giles
electrogoniometers. They are neatly packaged and can be used with either
a miniature data-logger or a four-channel amplifier. I was evaluating
them to measure the motion of the ankle in both plantar-dorsi flexion and
inversion-eversion. The following is my summary and the comparison to
Motion Analysis: (Note: ! is supposed to be a degree symbol)

From: Jon Fewster

Date: 13 August, 1992

Re: Penny & Giles Goniometers
__________________________________________________ ___________
After many tests to evaluate the electrogoniometers from Penny & Giles I have
reached some conclusions and have some recommendations. Here are my findings:
The goniometers are small, easy to attach and provide quick analysis of data.
They may be attached to any joint to measure one or two degrees of movement.
The four-channel amplifier we received recently is very helpful for providing
clean signals from the goniometers to our A/D board. The question
remains, are they good enough for our measurements?

1. Using the four-channel amplifier and its LED readout I tested accuracy and
crosstalk. I moved the goniometer in one plane only and measured both
channels for signal and crosstalk error. I measured the actual ROM with a
mechanical goniometer.
a. The accuracy drifted depending on plane of motion and goniometer
configuration (stretched, buckled, or normal). During a ROM from 0! to
160! and back a drift was noted. Up to 5! of error was consistent in these
b. Crosstalk was significant, again varying with arrangement.
Over a range of movement of 180! crosstalk exceeded 120!.

2. Accuracy was compared to Motion Analysis and the actual value from a
mechanical goniometer. The electrogoniometer data conversion was done by the
A/D board and ASYST. The angle scale was not able to be zeroed and the
increments were not the same as the actual angle.
a. Significant crosstalk was recorded. One good example is the
file PRO9.
Range of Motion Signal
Electrogoniometer Angle:114! -> 65.2! ->115! ->67! +/- 25!
Other Channel (error): 87.5! ->93! -> 87! ->93! +/- 3!
Actual Angle: 150!-> 30! -> 150! ->30! +/- 60!
if the actual and measured angles are proportional, the actual
error is: 17.2!
b. The movement was also measured by the Motion Analysis system:
For trial PRO9 the values were recorded as follows:
M.A. 148.37 ->25.8 ->150.25 ->30.40
P&G 114 -> 65.2 ->115 -> 67
Actual 150 -> 30 -> 150 -> 30
The electorgoniometer data needs to be verified for scale before a qualitative
comparison may be made.

3. In response to our measurements Neal Handler of Penny & Giles
responded in a FAX on August 7. He said that our electrogoniomter must be
damaged due to the large errors. He informed us that the standard
allowable crosstalk is 13! over a range of motion of 160!. This is much
less than I have measured. The following calculations assume that if we
purchase a goniometer it will stay within their acceptable values
throughout its lifetime.
Soutas-Little, et. al. provides angles for inversion-eversion and
plantar-dorsi flexion during running in the article "Analysis of foot
motion during running using a joint co-ordinate system." Med. Sci. Sports
Ex. Vol. 19, #3, pp. 285-293. In two subjects the angles were measured as
total movement:

Inversion-Eversion Plantar-Dorsi Flexion
Subject 1 15.5! 21!
Subject 2 14.5! 31!

Other sources measure inversion-eversion to have a range of 120!
and rearfoot motion a range of 110! (Nigg, Biomechanics of Running Shoes,
p. 43 and Cavanaugh, Biomechanics of Distance Running, p. 149,
respectively). These sources did not note the corresponding range of
plantar-dorsi flexion.
If we assume the Penny & Giles allowable values to be linear, with a
Plantar-Dorsi Flexion of 30! we would observe 1.5! of crosstalk in the other
channel. This is 10% of the 15! inversion-eversion angle. This value is not
acceptable for our measurements.

Conclusion: Given the best case scenario of the Penny & Giles allowable levels
these goniometers are not acceptable for our measurement needs. This
calculation of error is almost one-third of what we recorded in our
measurements. For our use these electrogoniometers need to have no crosstalk,
be more precise and more accurate.

I hope this information is helpful for you. I also played
trombone for ten years, so I am familiar with the motion of the right arm
of the trombonist. If the errors are suitable for you, the amplifier or
data-logger should be able to record the channels you desire at the same
time. Please let me know how it goes.

Jon Fewster
(617)297-4915 (W)