Evaluation of Built-In Lithium-Ion Batteries Used as Energy Source for Externally Powered Upper Limb Prostheses
Oyvind Stavdadahl Liselotte Hermanson Liselotte Hermanson K. Tomm Krristensen Kjell E. Malvig
New battery technologies bear the potential to greatly increase the benefits of powered upper-limb prostheses. The present study involved monitoring of the charge capacity of 35 Lithium-Ion batteries over an 18-month period, as well as a written survey among 100 amputees for quantification of important battery properties. The batteries displayed no capacity degradation during the test period, and performed superiorly to comparable NiCd batteries. The survey showed that for time-to-recharge, cosmetic appearance and the overall properties, the built-in Li-Ion batteries were rated as an improvement in comparison with conventional batteries.
The inability of available power sources to supply sufficient energy over a reasonable amount of time within reasonable weight and space limits, has been one of the major drawbacks of externally powered prostheses. As a number of new and promising battery technologies evolve, manufacturers of prosthetic equipment as well as the clinical society seem somewhat hesitant in taking these into use, and little or no information has yet been published regarding the pros and cons of these technologies and related fitting techniques in a prosthetics context.
Since 1993, more than 200 upper-limb amputees have been fitted with Lithium-Ion powered prostheses at Norwegian Technical Orthopaedic, Ottestad, Norway and the Orthopaedic-Technical Department, Orebro Medical Centre Hospital, Orebro, Sweden. Because of the fast recharge and the high energy content, all these amputees have had their batteries permanently built into the prosthesis socket, as shown in Figure 1. . This novel technique obviously influences the appearance of the prosthesis, but also potentially changes the use habits and the amputee's confidence in his prosthesis as empty batteries cannot be quickly replaced with charged ones.
The present study is an evaluation of the long-term performance and the user opinion on these batteries and techniques, and the evidence for their superior performance.
Materials and Methods
The first part of the study involved monitoring of the charge capacity of battery packs that were in normal use. The second part comprised a written survey among amputees who use Li-Ion powered upper-limb prostheses. All batteries involved were of the Easy Power or Easy Power Light (for children) types, comprising one or two Li-Ion cells, respectively, plus circuitry for voltage supervision and regulation. In most cases, Otto Bock electric hands were utilised for terminal devices. The batteries also powered other electrical components when present, such as powered elbow, wrist rotator and/or heating element integrated with the prosthetic socket, in which cases a double two-cell Easy Power pack was sometimes used. Each amputee had two identical prostheses at his/her disposal.
Charge capacity measurements
17 amputees repeatedly submitted their prostheses by mail for charge capacity measurements. The batteries were charged for three hours to make sure they contained maximum charge, and then connected to a test apparatus comprising an Otto Bock 8E38=1L7 1/4, reference hand and a combined cycle generator/counter (Figure 2. ). For convenience, a control circuit external to the prosthesis was applied between the charging socket and the cycle generator/counter. This circuit was identical to the one integrated in the battery pack, and the set-up was equivalent with connecting the battery pack's output terminals directly to the generator/ counter except for a negligible energy loss induced by the extra circuitry. The apparatus was set out to connect the control circuit's output terminal voltage of 7.2V (5V for the light version) to the motor terminals of the reference hand in order to make it open, disconnect when the stall current reached 1A or 300mA, respectively, wait for a few seconds and then repeat the whole action with the opposite voltage polarity to make the reference hand close. This cycle was reiterated until the battery pack was completely discharged. For every pack, the date of application, the dates of each measurement and the number of open/close cycles obtained were recorded. The whole trial involved a total of 35 battery packs and 72 individual measurements.
A questionnaire was designed to quantify subjective figures. Questions 1-5 were related to the test population: age and sex, prosthesis use-habits, recharge habits and -frequency; while in questions 6-11 the amputee was asked to compare several properties of Li-Ion batteries with those of conventional batteries: recharge frequency, maintenance and charging, weight, cosmetic appearance, problems related to the 'batteries not being interchangeable in the usual sense, and overall rating. For each question a set of alternative answers were given, and a "not applicable" option was provided where appropriate. Question 12 asked for further opinions or comments. Answers that were not uniquely defined, i.e. where a person had checked none or more than one option for the same question, were considered non-existent and removed before the data was analysed. Each amputee's questionnaire was marked with the amputation level prior to being sent out, but was otherwise anonymous. Questionnaires were distributed to 100 Norwegian amputees randomly picked among the clients of one of the orthopaedic workshops involved; of these, 69 were returned and form the basis for the results presented in this paper.
The results of the charge measurements for the two-cell battery packs are shown in Figure 3. , where the cycle counts recorded are plotted as a function for time-since application (i.e. the amount of time from the application of the battery pack to the measurement in question; the batteries' "age" at the time of the measurement). The circles indicate the arithmetic mean value for all measurements within the same one-month age interval, suggesting a slight capacity increase over the first 18 months of a battery's life. To find out if this surprising result was caused by drift in test apparatus parameters or settings, four unused battery packs were tested four times each to provide 16 new cycle counts. The resulting measurements are included in the data presented in Figure 3. , where the left-most bar is generated from these plus eight previous measurements; they are consistent with the rest of the data and suggest no drift in the test equipment. The results for one- and four-cell cases agreed qualitatively with those presented. For comparison, five unused Easy Power Li-Ion packs and five unused Otto Bock 757B8 NiCd batteries were put through the same test as described above; the average cycle counts obtained were 5.640 for the Li-Ion for NiCd batteries, a ratio of 4:1.
The results of the survey are presented as histograms showing the distribution of answers for each question, with "Not applicable" answers represented by dotted bars. The age and sex distribution of the test population is shown in Figure 4. , while the upper part of Figure 5. shows the distribution amputation levels according to the following codes: transcarpal (R), wrist disarticulation (WD), long below elbow (LBE), short below-elbow (VSBE). Elbow disarticulation (ED), standard above-elbow(SAE), short above-elbow (SAE), humeral beck (HN), shoulder disarticulation (SD) and intrascapulothoracic (S). The lower part of Figure 5. shows the use habits; in fact, only one person stated that(s) he uses his/her prosthesis.
When asked about recharging habits, more than 60% reported that they do not recharge until the batteries are completely drained. Approximately 25% recharged routinely (for instance every night), while 10% recharged randomly. Virtually all reported to charge their batteries 100% every time, though the charger used indicated when 90% charge is reached, and the latter takes only about one hour (half of what it takes to reach full charge). Figure 6. represents some of the most important finding of this study, namely related to time-to-recharge. A large majority of the test population claimed that they can use their prosthesis for two days or more before needing recharge, with nearly 40% exceeding three days. Of the four persons reporting to need recharging more than once a day, all were below 10 years of age and thus presumably use the one-cell battery pack. As shown in the lower part of the figure, the resulting recharge rate is perceived to be lower or much lower than that of conventional batteries.
The results concerning ease of maintenance and charging also were positive, with more than 60% claiming the Li-Ion to be easier or much easier to handle. This is assumed to be partly because of the Li-Ion batteries' displaying no memory effect and their short recharge time, as well as the user not needing to keep or carry spare batteries. Less than 5% of the test population gave negative reports, most of which were related to the mechanical design of the charging plug.
As shown in Figure 7. , the weight of the prosthesis was perceived to be almost as with conventional batteries (top section), while the cosmetic appearance was claimed to be better or much better by more than 50% of the population(bottom).
The majority did not report any problems related to the batteries being built-in, but close to one third did (Figure 8. , top section). A commonly reported problem was the inability to tell how much charge was left in the device. This is probably partly because the EasyPower Li-Ion units yield a constant output voltage throughout the discharge cycle, as opposed to NiCd cells. Also reported was the need for a charged spare prosthesis. The bottom section of Figure 8. shows the overall rating of Li-Ion batteries to conventional ones, where a definite majority claim the Li-Ion batteries represent a large to moderate improvement when comparing their overall properties with those of conventional batteries. In Figure 9. and Figure 10. we have split the data related to cosmesis and overall rating, into separate histograms for each amputation level represented. The results show a tendency towards more satisfaction among high amputation levels, but even the transcarpal and wrist disarticulation amputees report a preference for the Li-Ion batteries over conventional ones.
The Li-Ion batteries studied did not display charge capacity degradation, but rather a slightimprovement, over the 18-months test period. Part of the variance of the test results may be attributed to uncertainties in the test equipment. However, the apparent time dependency of this variance as well as the consistency between the initial measurements and the final test measurements suggest that it be caused by other factors, such as individual differences between the battery packs and the different packs being exposed to different patterns. The test population involved the Li-Ion batteries as being a moderate to large improvement over conventional batteries for a range of properties. The only exception was their weight which was perceived to be virtually identical to that of conventional batteries, with a negligible tendency towards rating the Li-Ion as lighter. This is an interesting result, because an Easy-Power Li-Ion pack is in fact 62% heavier than the 757B8 NiCd battery.
The study indicated that built-in Li-Ion batteries are a superior alternative to more conventional batteries used as energy source for powered upper-limb prostheses, regardless of amputation level.
The authors wish to thank Jostein Mathisen at Labtech Industrial Electronics for designing and supplying the test apparatus used for the laboratory measurements. We also thank the staff at Norwegian Technical Orthopaedics and the Orthopaedic-Technical Department. Ore-bro Medical Centre Hospital, for assistance in connection with measurements and distribution of questionnaires. Parts of this study was sponsored by the Research Council of Norway under grant 109533/320.
Please address correspondence to:
Department of Engineering Cybernetics
Norwegian University of Science and Technology