DeWalt 36V Technology (A123 Systems)

DeWalt 36V Technology (A123 Systems) - revised 7/12/06, 8/18/06, 8/22/06

DISCLAIMER: These articles are the findings and opinions of one person. Long-term durability of these cells is not known for our application/construction techniques. Any cell technology that we use is potentially dangerous. Do not work with these cells if you do not have a healthy respect for them, or if you lack the requisite skills.

There has been a lot of hype over several months about a new cell technology being developed by A123 Systems and destined for use by DeWalt in their new, ultra-manly 36V tools. Well, they've finally arrived, and while not as good as the rumors implied, they are still a great deal for the money. We will refer to these new cells as "M1" cells per the manufacturer's designation.
Besides this introductory article, there are several other articles of interest:
Building Packs
Easy Packs
Charging Packs
Example Application
Anecdotal Evidence
Dapter
M1 cells

Although LiPo packs are becoming more and more prevalent, there are several shortcomings in the areas of

safety
cost
output
durability
form factor

The new M1 cells appear to be a significant improvement in all of these areas.

SAFETY - LiPo cells have caused significant property damage due primarily to their tendency to burst into flame if overcharged. They are also subject to catastrophic failure if over-discharged or called upon to provide high current, e.g. momentary shorts. These failure modes are exacerbated by any unbalance between cells in a pack. The M1 cells behave differently from other Lithium technologies. The specified end of charge voltage is 3.6 volts rather than 4.2 volts. If the charging source for a conventional Li cell is greater than 4.2V, the cell will continue to accept energy, sometimes with disasterous results. The M1 cell, on the other hand, accepts very little additional energy above 3.6V.
From the A123 site:
"A123 materials are designed to ensure all the Lithium is fully extracted from the cathode when the battery is fully charged. As a result safety issues relating to overcharge are eliminated because there is no Lithium available to plate on the anode in an overcharged state. This is in contrast to conventional Li Ion cells, which only extract half their Lithium content when they reach their upper cut-off voltage. Conventional Li Ion cells are easy to overcharge and once in this state they can continue to extract Lithium putting the cell in a dangerous mode and making it prone to fires and explosions."
The Figure below shows the cell voltage of an M1 cell charged with a Dapter and an Astro 112D NiCad charger. The M1 cell has a nominal 2.3 ampere-hour capacity. Normal Li technology dictates a 1C (2.3A) charge. The curve below was from a 6A charge. A123 Systems claims their cells can be charged in as little as 5 minutes without damage. Realistically, our chargers can't do better than a 15-30 minute charge. A123 Systems recommends a maximum of 10A charge. This is just short of 15 minutes. The Dapter cuts off charge at a bit less than 4.2V per cell. As can be seen on the graph, once the 3.6V point is reached, the cell voltage rises rapidly to cutoff (less than two minutes). After cutoff, the voltage quickly "sags" back below 3.6V. The purpose of this test was not to endorse the charging of M1 cells on LiPo chargers. This will cause long-term harm. On the other hand, their were no fires or explosions or self-destructive behavior. Based on A123 recommendations, the new LiPoDapter+ will cutoff at 3.7V and will handle up to 8A. This results in a full charge in less than twenty minutes. The Dapter will handle up to 10 M1 cells or 8 LiPo cells depending on the charger capacity. M1 overcharge

COST - LiPo packs are very expensive for larger planes. I believe that the ultimate acceptance of electrics by the "glow" community hinges on "40-size" and larger planes. For many, the cost of LiPos is prohibitively expensive for these planes. I consider the 20C cells from reputable brands such as Kokam and Polyquest (Hobby Lobby "Twenty") to be the measure of cost for larger planes. By this I mean that a "good" LiPo cell in the 2- to 3-AH sizes is in the $30 to $50 price range. I reject the claims of "Chinese copies" that appear to be much cheaper. Right now, if you buy a DeWalt 36V power pack for list price, you will pay $160 for 10 2.3AH cells. That's $16 per cell. You've also purchased a large electronic speed control module and a lot of plastic. Once DeWalt gets their supply line filled, other OEMs should eventually be offering cells and/or packs at less than $10 per cell. Considering that the M1 is equivalent to about two and a half NiMH cells, I think we've turned the corner on cost. The difference between the M1 and LiPo becomes a lot greater when you consider the ability to charge the M1 in less than half the time. For some, this means less packs to buy to fly as often as they would like.

OUTPUT - The main reason why LiPo packs can fly our planes for so long is that they are severely limited in how much current they can safely deliver. The paralleling of cells was necessary to supply the current levels we required. Thus the early 4P 8AH packs flew our planes for way longer than we could stand up and weighed almost as much as the NiCad packs they replaced. Newer cells boasted 15C and 20C continuous current and a premium price. A123 Systems specifies 30C (70A) continuous and over 50C (120A) for 10 seconds! Implied in these specifications is that these cells can handle heat better than LiPo cells.

DURABILITY - The aluminum and plastic bag construction of LiPo dictates careful handling just to avoid permanent damage. A drop on the ground or similar minor incident could destroy a LiPo. A plane crash could mean loss of a big dollar investment or even possibly a fire. The M1 cells are more like NiCad/NiMH sub-C cells in their durability.

FORM FACTOR - For many, the conversion from Ni to Li was difficult due to the the flat, rectangular shape of the LiPo cells and the inability to build packs in shapes to fit difficult installations. The M1 cells, though larger than sub-C cells, allow "conventional" pack construction of side by side, end to end, and "bricks" of stacked cell rows. I have successfully soldered cells end to end and so far have measured no performance degradation. The M1 occupies slightly less volume than a PQ2100 cell and slightly less than twice the volume of a GP3300 NiMH cell.

PERFORMANCE - As mentioned earlier, there was some disappointment felt when these cells finally became available. Early rumors had capacity closer to 4AH and case size somewhat shorter. I think these will be good for 40-size planes for those of us who fly less than ten minutes. For larger planes, higher voltages (geared motors) or paralleled cells will be necessary. The M1 cell, being definitely different from other Lithium cells, has a nominal output voltage of 3.15V, versus the 3.75V of LiPo. This implies that you would need 7 M1 cells to replace 6 LiPo cells. This is not unlike the early NiMH versus NiCad use except the extra NiMH was due to higher internal resistance, the voltage being essentially the same. M1 internal resistance is somewhat lower than LiPo cells of comparable MAH capacity. The striking thing about M1 performance is that the voltage stays "flatter" than LiPo and NiMH over the discharge curve, as seen in the Figure below. discharge comparison

M1 (left) and E-Moli
The chart below compares some parameters of different cells. The basic numbers were taken from ElectriCalc data. I am sure you can get a better price than I have listed on "X" cells. This is just a reference chart, not a buyers' guide. The first cell is NiMH. The next two are good 20C LiPo cells. The E-Moli is the Lithium cell used in the Milwaukee V28 power tools. It comes in a cylindrical steel case, but it performs like LiPo. The photo on the right shows the M1 [DeWalt 36V] (left) and the E-Moli [Milwaukee V28].
cell comparisons
The voltage used by ecalc is the average "open-circuit" voltage over the middle 80% of the discharge curve. The resistance is in milliohms and is the "DC" resistance averaged over the same 80%. The first derived comparison number is R/V, which is the internal cell resistance divided by the nominal voltage. For this value, lower is better. This is a way of comparing apples to apples,`because for a pack of X volts, the total resistance would be X volts times this number. The M1 is bested only by the PQ3300. Larger cells of similar construction and chemistry generally have lower internal resistance. The next comparison is milliwatt-hours, obtained by multiplying milliamp-hours by cell voltage. M1 is next to last here, but about 70% better than the GP3300. The next column is also MWH but figured at a discharge rate of 10C. The previous column is really an upper bound. The next column is kind of an energy density measure. The dimensions are watt-minutes per ounce. For this value, bigger is better. The two LiPos win here, but they do not have rugged cases. The E-Moli and M1 still offer much better energy density than the NiMH. The last column is energy per buck, and clearly, more is better. NiMh still wins, but the E-Moli and M1 stomp the LiPo. Again, the M1 can be charged much faster than LiPo without special equipment. NiMH charge really should be limited to 1C to avoid shortened life.

BALANCE - Keeping LiPo packs balanced has become almost an obsession among fliers. NiCads are pretty much self-balancing because the peak detection algorithm actually results in one or more cells being overcharged if the pack is not balanced. NiCads are rugged enough to handle this and it allows the other cells to eventually "catch up" to those that are overcharged. When I observed the charge behavior of the M1 cells, I wondered if this behavior would provide a similar balancing effect. On a 3-cell pack, I forced a 130-millivolt imbalance by charging one cell. This is a HUGE imbalance due to the small change in voltage with discharge (per the graph discussed previously). Charging was done using a Dapter at a 6A charge rate. A higher charge rate was chosen because higher charge rates cause normal LiPo cells to become unbalanced. The charge regimen could also have used an Astro 109 Lithium charger. The pack was discharged at a constant 20A to 2.7V. Each successive cycle decreased the imbalance. After several cycles there was still some imbalance but only a fraction of the original. It appears that unbalanced packs should not be as much of an issue with the new M1 cells as with LiPo. My first packs will not sport balancing connectors. My LiPo packs, which were never discharged more than 80%, showed no imbalance after dozens of flights, so I don't foresee a problem with the M1 packs. If these cells do somehow get out of balance, there is a problem. As the charging curve above showed, voltage rises dramatically. In a large pack, if one cell has a much higher voltage, it will reach that point before the others. Using a LiPo charger would mean a damaging voltage could be placed on this cell. Always make sure all cells are within 10 millivlots when assembling packs.

SELF-DISCHARGE - One of the frustrations of Nickel cells is the high rate of self-discharge. If you charged them Friday night, by Saturday morning they had a noticeable loss of capacity. The first flight of the day was always the shortest. If they sat for several days, they had to be cycled to regain capacity. LiPo cells changed all that. You could charge Saturday night after flying, and the next Saturday they would still be charged. Good thing, because it takes so long to charge them. So how do the M1 cells compare? Tests so far show that over the course of a couple of weeks, the loss of capacity is close to zero. That's a good thing!

OBSERVATIONS/CONCLUSIONS

The new M1 cells are not as light as LiPo for comparable MAH capacity. Offsetting this is a much higher "C" rating, fast charge capability, ruggedness, much lower price, and, perhaps most importantly, safety.
Additionally, A123 Systems claims much longer life:
"A123Systems outstanding chemical stability and a flat plateau at a lower oxidation potential than other Li Ion chemistries provides our products with record life."
"Our batteries are uniquely engineered (pat. pending) so their internal resistance (impedance) will decrease with their use. This is the opposite effect to most Li Ion cells which experience a growth in their internal resistance as they are cycled at high rates or temperatures. This is a significant benefit in applications requiring long calendar life such as hybrid electric vehicles . ."
This remains to be seen, but the only way to find out is to try them. I am!
Although the charge method for M1 cells is incompatible with current LiPo chargers, it appears that the M1 chemistry is forgiving of overcharging. A123 Systems claims that less than one hour of overcharge will not affect cell life. That remains to be seen. Dapters shipped after July 7, 2006 have both LiPo and M1 capability.

There has been a lot of hype over several months about a new cell technology being developed by A123 Systems and destined for use by DeWalt in their new, ultra-manly 36V tools. Well, they've finally arrived, and while not as good as the rumors implied, they are still a great deal for the money. We will refer to these new cells as "M1" cells per the manufacturer's designation. Besides this introductory article, there are several other articles of interest: Building Packs Easy Packs Charging Packs Example Application Anecdotal Evidence Dapter
Although LiPo packs are becoming more and more prevalent, there are several shortcomings in the areas of safety cost output durability form factor
The new M1 cells appear to be a significant improvement in all of these areas.
SAFETY - LiPo cells have caused significant property damage due primarily to their tendency to burst into flame if overcharged. They are also subject to catastrophic failure if over-discharged or called upon to provide high current, e.g. momentary shorts. These failure modes are exacerbated by any unbalance between cells in a pack. The M1 cells behave differently from other Lithium technologies. The specified end of charge voltage is 3.6 volts rather than 4.2 volts. If the charging source for a conventional Li cell is greater than 4.2V, the cell will continue to accept energy, sometimes with disasterous results. The M1 cell, on the other hand, accepts very little additional energy above 3.6V. From the A123 site: "A123 materials are designed to ensure all the Lithium is fully extracted from the cathode when the battery is fully charged. As a result safety issues relating to overcharge are eliminated because there is no Lithium available to plate on the anode in an overcharged state. This is in contrast to conventional Li Ion cells, which only extract half their Lithium content when they reach their upper cut-off voltage. Conventional Li Ion cells are easy to overcharge and once in this state they can continue to extract Lithium putting the cell in a dangerous mode and making it prone to fires and explosions." The Figure below shows the cell voltage of an M1 cell charged with a Dapter and an Astro 112D NiCad charger. The M1 cell has a nominal 2.3 ampere-hour capacity. Normal Li technology dictates a 1C (2.3A) charge. The curve below was from a 6A charge. A123 Systems claims their cells can be charged in as little as 5 minutes without damage. Realistically, our chargers can't do better than a 15-30 minute charge. A123 Systems recommends a maximum of 10A charge. This is just short of 15 minutes. The Dapter cuts off charge at a bit less than 4.2V per cell. As can be seen on the graph, once the 3.6V point is reached, the cell voltage rises rapidly to cutoff (less than two minutes). After cutoff, the voltage quickly "sags" back below 3.6V. The purpose of this test was not to endorse the charging of M1 cells on LiPo chargers. This will cause long-term harm. On the other hand, their were no fires or explosions or self-destructive behavior. Based on A123 recommendations, the new LiPoDapter+ will cutoff at 3.7V and will handle up to 8A. This results in a full charge in less than twenty minutes. The Dapter will handle up to 10 M1 cells or 8 LiPo cells depending on the charger capacity.
COST - LiPo packs are very expensive for larger planes. I believe that the ultimate acceptance of electrics by the "glow" community hinges on "40-size" and larger planes. For many, the cost of LiPos is prohibitively expensive for these planes. I consider the 20C cells from reputable brands such as Kokam and Polyquest (Hobby Lobby "Twenty") to be the measure of cost for larger planes. By this I mean that a "good" LiPo cell in the 2- to 3-AH sizes is in the $30 to $50 price range. I reject the claims of "Chinese copies" that appear to be much cheaper. Right now, if you buy a DeWalt 36V power pack for list price, you will pay $160 for 10 2.3AH cells. That's $16 per cell. You've also purchased a large electronic speed control module and a lot of plastic. Once DeWalt gets their supply line filled, other OEMs should eventually be offering cells and/or packs at less than $10 per cell. Considering that the M1 is equivalent to about two and a half NiMH cells, I think we've turned the corner on cost. The difference between the M1 and LiPo becomes a lot greater when you consider the ability to charge the M1 in less than half the time. For some, this means less packs to buy to fly as often as they would like.
OUTPUT - The main reason why LiPo packs can fly our planes for so long is that they are severely limited in how much current they can safely deliver. The paralleling of cells was necessary to supply the current levels we required. Thus the early 4P 8AH packs flew our planes for way longer than we could stand up and weighed almost as much as the NiCad packs they replaced. Newer cells boasted 15C and 20C continuous current and a premium price. A123 Systems specifies 30C (70A) continuous and over 50C (120A) for 10 seconds! Implied in these specifications is that these cells can handle heat better than LiPo cells.
DURABILITY - The aluminum and plastic bag construction of LiPo dictates careful handling just to avoid permanent damage. A drop on the ground or similar minor incident could destroy a LiPo. A plane crash could mean loss of a big dollar investment or even possibly a fire. The M1 cells are more like NiCad/NiMH sub-C cells in their durability.
FORM FACTOR - For many, the conversion from Ni to Li was difficult due to the the flat, rectangular shape of the LiPo cells and the inability to build packs in shapes to fit difficult installations. The M1 cells, though larger than sub-C cells, allow "conventional" pack construction of side by side, end to end, and "bricks" of stacked cell rows. I have successfully soldered cells end to end and so far have measured no performance degradation. The M1 occupies slightly less volume than a PQ2100 cell and slightly less than twice the volume of a GP3300 NiMH cell.

PERFORMANCE - As mentioned earlier, there was some disappointment felt when these cells finally became available. Early rumors had capacity closer to 4AH and case size somewhat shorter. I think these will be good for 40-size planes for those of us who fly less than ten minutes. For larger planes, higher voltages (geared motors) or paralleled cells will be necessary. The M1 cell, being definitely different from other Lithium cells, has a nominal output voltage of 3.15V, versus the 3.75V of LiPo. This implies that you would need 7 M1 cells to replace 6 LiPo cells. This is not unlike the early NiMH versus NiCad use except the extra NiMH was due to higher internal resistance, the voltage being essentially the same. M1 internal resistance is somewhat lower than LiPo cells of comparable MAH capacity. The striking thing about M1 performance is that the voltage stays "flatter" than LiPo and NiMH over the discharge curve, as seen in the Figure below.

The chart below compares some parameters of different cells. The basic numbers were taken from ElectriCalc data. I am sure you can get a better price than I have listed on "X" cells. This is just a reference chart, not a buyers' guide. The first cell is NiMH. The next two are good 20C LiPo cells. The E-Moli is the Lithium cell used in the Milwaukee V28 power tools. It comes in a cylindrical steel case, but it performs like LiPo. The photo on the right shows the M1 [DeWalt 36V] (left) and the E-Moli [Milwaukee V28]. The voltage used by ecalc is the average "open-circuit" voltage over the middle 80% of the discharge curve. The resistance is in milliohms and is the "DC" resistance averaged over the same 80%. The first derived comparison number is R/V, which is the internal cell resistance divided by the nominal voltage. For this value, lower is better. This is a way of comparing apples to apples,`because for a pack of X volts, the total resistance would be X volts times this number. The M1 is bested only by the PQ3300. Larger cells of similar construction and chemistry generally have lower internal resistance. The next comparison is milliwatt-hours, obtained by multiplying milliamp-hours by cell voltage. M1 is next to last here, but about 70% better than the GP3300. The next column is also MWH but figured at a discharge rate of 10C. The previous column is really an upper bound. The next column is kind of an energy density measure. The dimensions are watt-minutes per ounce. For this value, bigger is better. The two LiPos win here, but they do not have rugged cases. The E-Moli and M1 still offer much better energy density than the NiMH. The last column is energy per buck, and clearly, more is better. NiMh still wins, but the E-Moli and M1 stomp the LiPo. Again, the M1 can be charged much faster than LiPo without special equipment. NiMH charge really should be limited to 1C to avoid shortened life.
BALANCE - Keeping LiPo packs balanced has become almost an obsession among fliers. NiCads are pretty much self-balancing because the peak detection algorithm actually results in one or more cells being overcharged if the pack is not balanced. NiCads are rugged enough to handle this and it allows the other cells to eventually "catch up" to those that are overcharged. When I observed the charge behavior of the M1 cells, I wondered if this behavior would provide a similar balancing effect. On a 3-cell pack, I forced a 130-millivolt imbalance by charging one cell. This is a HUGE imbalance due to the small change in voltage with discharge (per the graph discussed previously). Charging was done using a Dapter at a 6A charge rate. A higher charge rate was chosen because higher charge rates cause normal LiPo cells to become unbalanced. The charge regimen could also have used an Astro 109 Lithium charger. The pack was discharged at a constant 20A to 2.7V. Each successive cycle decreased the imbalance. After several cycles there was still some imbalance but only a fraction of the original. It appears that unbalanced packs should not be as much of an issue with the new M1 cells as with LiPo. My first packs will not sport balancing connectors. My LiPo packs, which were never discharged more than 80%, showed no imbalance after dozens of flights, so I don't foresee a problem with the M1 packs. If these cells do somehow get out of balance, there is a problem. As the charging curve above showed, voltage rises dramatically. In a large pack, if one cell has a much higher voltage, it will reach that point before the others. Using a LiPo charger would mean a damaging voltage could be placed on this cell. Always make sure all cells are within 10 millivlots when assembling packs.
SELF-DISCHARGE - One of the frustrations of Nickel cells is the high rate of self-discharge. If you charged them Friday night, by Saturday morning they had a noticeable loss of capacity. The first flight of the day was always the shortest. If they sat for several days, they had to be cycled to regain capacity. LiPo cells changed all that. You could charge Saturday night after flying, and the next Saturday they would still be charged. Good thing, because it takes so long to charge them. So how do the M1 cells compare? Tests so far show that over the course of a couple of weeks, the loss of capacity is close to zero. That's a good thing!
OBSERVATIONS/CONCLUSIONS The new M1 cells are not as light as LiPo for comparable MAH capacity. Offsetting this is a much higher "C" rating, fast charge capability, ruggedness, much lower price, and, perhaps most importantly, *safety. Additionally, A123 Systems claims much longer life: "A123Systems outstanding chemical stability and a flat plateau at a lower oxidation potential than other Li Ion chemistries provides our products with record life." "Our batteries are uniquely engineered (pat. pending) so their internal resistance (impedance) will decrease with their use. This is the opposite effect to most Li Ion cells which experience a growth in their internal resistance as they are cycled at high rates or temperatures. This is a significant benefit in applications requiring long calendar life such as hybrid electric vehicles . ." This remains to be seen, but the only way to find out is to try them. I am! Although the charge method for M1 cells is incompatible with current LiPo chargers, it appears that the M1 chemistry is forgiving of overcharging. A123 Systems claims that less than one hour of overcharge will not affect cell life. That remains to be seen. Dapters* shipped after July 7, 2006 have both LiPo and M1 capability.