Nanotech LiPo booster pack for my GEM e4...

[GGoodrum]

This experience also points out that a hybrid mixture of lead and lipo may be a good choice for other vehicular applications that are cost sensitive. Adding a small lipo pack to a lead powered bike would have similar results, most likely. Might be a market for a smaller version there.

Maybe, although with bikes, the weight factor is a much bigger issue. It would be better on a scooter.

Another "hybrid" combo I've been thinking about is a LiPo booster pack for a ThunderSky/Sky Energy setup. It could possibly make the TS/SE-type cells last a lot longer, since they wouldn't get worked so hard. I might look at this as a "replacement" option for my six 100Ah gels, at some point.

-- Gary

 

[auraslip]

I think, at least with lower voltages, a hybrid pack might be useful on ebikes. Unless I engage speed 120% I'm only pulling 20a at full speed(more like 30a with speed 120%). However I could easily do 60a amps at starts. It's a lot more fun like that, but I know it's hell on the 48v20ah ping I have! Given that 16s1p of a123 from cell-man only costs $150 shipped and 2p costs $250 shipped, this might be an easy way to push my hub motor and controller to the limits of what it could do at 48v. This would be fairly cheap, safe, and take advantage of lifepo4's extended cycle life.

The question is, would everything be dandy if all the two packs were in parallel together? What if the packs were in parallel, and each cell was in parallel via the sense wires so that they could share a bms? It'd be too much of a pain to wire each cell in parallel, but what if you did?

Or maybe something like the circuitry you made for the GEM could be used. It's fascinating to think about. People here have disregarded lifep4o because of it's performance woes, but it's become clear that lipo is no cheaper and safer. I imagine a 20ah pack made from 5ah of a123($250) and 15ah($300) of 1c lifepo4 could be cost effective and powerful if done safely.

 

[MitchJi]

Hi Gary,

Probably even a bigger market for roughly $2k-$2.5k plug and play 100ah-140ah lithium pack replacement. Just need to either find a good price or wait a year or two for prices to drop.

Using LiPo, that would be 20s20p, so 80 5s-5000 packs. At roughly $45 each that would be $3600, just for the packs. You'd still have to figure out a clever way to mount and connect them as well. The control electronics would be a lot simpler, though, because you wouldn't the special logic for controlling when the pack kicks in.

This thread killed HK Lipo as an option for me for large packs (although I'd love to hear your opinion):
http://endless-sphere.com/forums/viewtopic.php?f=14&t=27494

...This is (was) a 4s5ah 20C hard pack charged only to 16.2V (4.05v/cell) and set aside...never used. It only saw a few degrees C change in temp daily for the few weeks it sat there, and no direct sunlight, so I'm at a loss as to the cause. I'm just very lucky that it didn't spread to the other packs only inches away, and that the single pack didn't get hot enough to ignite the formica/particle board desk it sat on, or we would have surely lost our guest house that I use for ebike stuff and my office...

We are still a ways away from a $2500 Lithium-based replacement with a 100Ah capacity.

I hope that is incorrect, as I am hoping to buy a Gem that needs batteries and put in a lithium pack for $2k to $4k. Might be able to find a good price on 18650's (even if the prices below should be cell prices instead of pack prices they are pretty encouraging):
http://rumors.automobilemag.com/tesla-20000-model-sales-required-start-posting-profits-11763.html

Tesla’s 18560 cell battery pack, which is similar to our everyday laptop battery, has the benefits of preexisting R&D from major tech companies (Panasonic has invested $30 million in Tesla) and advanced economies of scale, not to mention enviable energy density. According to Martin Eberhard, Tesla’s co-founder who later left the company and has famously sparred with Tesla CEO Elon Musk since, the 18560-cell packs likely cost $200 per kilowatt-hour, which is 71- to 75-percent cheaper than large-form cell lithium-ion packs at current analyses. Additionally, the cells have already diverged onto a dedicated EV development route and are expected to see further year-over-year price drops from 6 to 8 percent.

Another option in the next couple of years should be used Leaf or Volt etc. packs from Wrecking yards:
http://www.dailytech.com/Tesla+CEO+...imitive+Boasts+About+Model+S/article19286.htm

Thus when Nissan claimed to have reached production costs of $375/kWh for its upcoming 2011 Nissan LEAF EV, it turned heads. After all, most auto companies were saying that they hoped to reach $400-$700/kWh with their upcoming models.

In any event, one way or the other, I think we'll be able to hit the $2.5k price within a couple of years.

A more practical way to do a lithium-based replacement, in my opinion, would be to use a 24s5p setup of a123 20Ah pouch cells, if we can ever get access to them. Maybe through Mavizen? Anyway, even if we could get them, at $50 each that would be $6k. That's almost as much as I paid for my whole GEM.

Practical in terms of quality and ease of pack building but not in terms of price, although factoring in the cell life they might be worth it. Your current 110 ah pack (including the booster) cost about $2,400. If the A123 pack lasted 3x as long,which seems reasonable the long term cost is actually less.

I'm not sure how accurate these claims are (both the 300 and 700 figures sound low) but Lithium is a really an appealing option (I'm sure you are aware of the plusses ) :
http://www.allcelltech.com/component/content/article/35-products/62-rad-gem

Global Electrical Motorcars, a Chrysler company, has produced a pollution-free, all-electric vehicle with a speed of up to 25 mph and a 30 mile drive range. We replaced the original lead acid battery bank with our custom PCM protected lithium-ion system while maintaining the vehicle’s original range and performance. Our system is smaller, lighter, and longer lasting:

					  System comparison	
			 Weight		Volume		Projected life		Drive Range
			   (lb)		(cu ft)		(# of cycles)		(miles)
Original      396			3.3			300					30

AllCell		136			1.9			700					30

 

[redorblack]

18650 cells are not something you want to try to build into a car sized pack yourself...Tesla has each of those 8000+ cells in it's own steel can with full cooling/warming, and multiple circuit protections for each cell. You aren't going to just solder together thousands of bulk cells and have something that is safe... more than likely you wouldn't get it working unless you have the ability to oversee quality control of the cell manufacturers and have teams of skilled people doing the building and testing at each step. I'm very impressed with what Tesla has done with that style cell. in order to achieve that, they did things like put two fuses on each cell, then fuse each sheet, then fuse the pack... and on and on. That is just the beginning... check out the link for a bunch more.

http://webarchive.teslamotors.com/display_data/TeslaRoadsterBatterySystem.pdf

 

[MitchJi]

Hi,

Doc has a lot of Konions for sale. 20 sub-packs of 80p (1,600 cells) would make a 120ah pack (vs Gary's 100 ah pack). Total cell cost wold be about $2,800 (Doc only had about a third that many for sale) but it makes me think finding an affordable ~$3k Lithium solution should be possible.
http://endless-sphere.com/forums/viewtopic.php?f=9&t=26589

18650 cells are not something you want to try to build into a car sized pack yourself...Tesla has each of those 8000+ cells in it's own steel can with full cooling/warming, and multiple circuit protections for each cell. You aren't going to just solder together thousands of bulk cells and have something that is safe... more than likely you wouldn't get it working unless you have the ability to oversee quality control of the cell manufacturers and have teams of skilled people doing the building and testing at each step.

1. There's a huge difference between a 7kw pack, replacing six 12v golf cart/fork lift batteries in a golf cart, and a 50kw OEM quality pack with a 10 year warranty that has to work (for example) in Arizona in the summer and New Hampshire in the winter. Their pack design also has the flexibility to use any Lithium based 18650 Cells. Its such a huge difference in requirements I'm surprised that anyone would even suggest it as a comparison.

2. 1,600 is a lot less than "thousands".

3. Doc seemed to do fine soldering 432 18650 cells into a large pack:
http://www.evalbum.com/1947

432 Sony 18650V (known as Konion).used in Makita, Bionx, 3.75 Volt, Lithium-Ion
2.21 Kilowatt-hour LiMnO2

4. I wouldn't even consider soldering 100's of Cells. I'm sure I could build individual solderless 80p sub-packs. With proper setup building twenty wouldn't be that much more time consuming than building one. So I'm pretty sure (not absolutely positive) it could be done in a reasonable amount of time.

 

[GGoodrum]

You have a point, Mitch, this wouldn't have to costs "thousands", going the Konion route. The big advantage is that the BMS portion is much simpler, because you really don't need a full balancing type circuit. I would, however, still do parallel group-level LVC protection, to make sure you are protected if a parallel connection breaks somewhere. In that case, you could end up with a lot less capacity than the total pack would normally have. That's how I used to kill cells with my a123-based packs.

A solderless solution would definitely be most desirable, that's for sure. The less soldering required, the better.

The safety issue with these is still a big plus, in my mind. For my next LiPo-based booster pack, I'm going to put a 20s3p 15Ah setup of 12 5s-5000 packs in an extruded aluminum box I found on McMasters (it's called "build-a-box"...), which is 4.1" x 6.4" x 18'. Even with this "extra" protection, I'm not completely comfortable with level of safety this would provide, should a LiPo pack decide to puff/explode. After doing a little playing around, I see I can fit 20 5s2p Makita packs in this same box, with room to spare all around. These could be wired in a 20s10p configuration, which would be 15Ah. Hmmm... All you'd need to equal the gels, in terms of capacity, would be six of these 20s/15Ah packs wired in parallel. Actually, since you wouldn't have the Peukert effect to deal with, you'd actually get a ton more range with 90Ah of Konions than you'd get with six of the 100Ah gels. Four of these 15Ah packs, so 60Ah total, would still probably get you equal range. In terms of volume, in the space that four of the gels occupy, you could easily fit even as many as 12 of these 15Ah packs.

I think this is definitely worth exploring further, so I may try and do one of these as a booster pack, and see how it compares.

-- Gary

 

[MitchJi]

Hi Gary,

Ride-4-fun has quite a few high performance options but nothing that addresses the problem you've solved with the Nano pack. I bet they'd be interested in what you've done, maybe even in purchasing some parts or getting some help developing a similar product.

I think I was incorrect. Ride-4-fun might not have this problem.

They recommend Trojan 150ah batteries, which I believe are Flooded Lead Acid:
1. These would require a lower C rating to handle heavy amp draws. For example 300a with your 100ah Gels is 3C vs 2C with the 150ah Trojans.

2. Your Gel cells probably have a lower C rating than Flooded Lead Acid batteries:
http://www.cameronsoftware.com/ev/EV_BatterySelection.html

The general evaluation of each battery chemistry is as follows:
* Flooded Lead Acid - inexpensive, heavy, requires maintenance
* AGM PbA - little more expensive, heavy, no maintenance, less capacity than FLA, non-spillable, high amp output
* Gel PbA - heavy, no maintenance, only suitable for lower amp draw systems (like high voltage AC systems)
* NiCd - very expensive, requires maintenance, great value in the long run
* NiMh - quite expensive, minor maintenance (memory effect), good energy density
* LiIon - expensive, no maintenance, great energy density, requires BMS

Update January 2005 After extensive research and consultation, it has been decided to go with 26 60Ah Deka Gel Batteries, model 8G34R. These gel batteries were selected because:
* Less expensive than AGMs
* If properly maintained will have over 1000 recharge cycles
* Are of a low profile, so will allow easy fit into the New Beetle
* Have been proven to be very successful in the Solectria Force vehicles (Now Azure Dynamics USA)
* Are less susceptible to developing charging imbalances than AGMs
* Over the long run are more cost effective than any other battery

These batteries do have one major drawback: The cannot put out as much current as other batteries (Wet, AGM). However, since they are going into a high voltage AC system (higher the voltage, reduces the current for the same amount of power), it is expected that the large current demands will be within the limits of the battery.

 

[redorblack]

Looking at lead acid amp hours in order to compare a lithium C rating is pretty useless. The Trojan battery you refer to at 56 amps (and I suspect that is a typo and is supposed to be at 50 amps) is rated for 102 minutes or 1.7 hours. 1.7 x 56 = 95ah at less than 1c. The 150ah is at a 20 hour rate... ie. 7.5 amps draw. That "150ah" battery probably delivers somewhere around 75ah in an EV and has a huge voltage drop when asked for 300 amps. I know my 72v 130ah battery pack drops about 6 volts when you nail it or are trying to climb a hill/rise and my controller limits it to less than 200 amps. 66volts x 200amps = 13200 watts. Eliminate that voltage drops and 72v x 200 amps equals 14400 watts... more juice available to the motor and I suspect the motor would normally run cooler. Now being that the Lipo solution can provide significantly more than 72volts because the charged voltage is what it'll stay closer to, rather than the '72v' rating. Charged lead acid is in theory 75.6 volts for a fully charged pack. A lipo booster is going to be over 80v when charged through with the same charger and will hang there a lot longer. Lets see if I've got this right... 80v x 300 amps =24000watts available. Assuming 66v on lead under load at 300 amps and you've only got 19800 watts available. Rounding everything to guesstimates, I suspect around a 15% gain in power. A significant gain in range should be seen because the lead isn't being subjected to the high amperage draws that suck it dry really really fast.

Someone correct me if I'm wrong... I'm sure I've over generalized a bunch of this and made some bad assumptions.

 

[MitchJi]

Hi,

Looking at lead acid amp hours in order to compare a lithium C rating is pretty useless. The Trojan battery you refer to...

Someone correct me if I'm wrong... I'm sure I've over generalized a bunch of this and made some bad assumptions.

You completely and totally misunderstood what I said, although I think the statements are pretty clear. I was not using "lead acid amp hours" to "compare a lithium C rating". I was comparing 150ah Trojan Flooded batteries used by ride-4-fun with the 100ah Gel's used by Gary.

 

[GGoodrum]

Hi Gary,

Ride-4-fun has quite a few high performance options but nothing that addresses the problem you've solved with the Nano pack. I bet they'd be interested in what you've done, maybe even in purchasing some parts or getting some help developing a similar product.

I think I was incorrect. Ride-4-fun might not have this problem.

They recommend Trojan 150ah batteries, which I believe are Flooded Lead Acid:
1. These would require a lower C rating to handle heavy amp draws. For example 300a with your 100ah Gels is 3C vs 2C with the 150ah Trojans.

2. Your Gel cells probably have a lower C rating than Flooded Lead Acid batteries:
http://www.cameronsoftware.com/ev/EV_BatterySelection.html

I looked at the 150Ah flooded types but the problem is they just don't last very long in climates that can routinely get over 100F, which can happen here quite a bit. Instead of roughly four years of life, I'd only get about two years out of them before I'd have to replace them again.

 

[GGoodrum]

Looking at lead acid amp hours in order to compare a lithium C rating is pretty useless. The Trojan battery you refer to at 56 amps (and I suspect that is a typo and is supposed to be at 50 amps) is rated for 102 minutes or 1.7 hours. 1.7 x 56 = 95ah at less than 1c. The 150ah is at a 20 hour rate... ie. 7.5 amps draw. That "150ah" battery probably delivers somewhere around 75ah in an EV and has a huge voltage drop when asked for 300 amps. I know my 72v 130ah battery pack drops about 6 volts when you nail it or are trying to climb a hill/rise and my controller limits it to less than 200 amps. 66volts x 200amps = 13200 watts. Eliminate that voltage drops and 72v x 200 amps equals 14400 watts... more juice available to the motor and I suspect the motor would normally run cooler. Now being that the Lipo solution can provide significantly more than 72volts because the charged voltage is what it'll stay closer to, rather than the '72v' rating. Charged lead acid is in theory 75.6 volts for a fully charged pack. A lipo booster is going to be over 80v when charged through with the same charger and will hang there a lot longer. Lets see if I've got this right... 80v x 300 amps =24000watts available. Assuming 66v on lead under load at 300 amps and you've only got 19800 watts available. Rounding everything to guesstimates, I suspect around a 15% gain in power. A significant gain in range should be seen because the lead isn't being subjected to the high amperage draws that suck it dry really really fast.

Someone correct me if I'm wrong... I'm sure I've over generalized a bunch of this and made some bad assumptions.

I agree with most of this. My gels have a 20 hour rating of 108Ah, which drops to about 98Ah at C/5. My controller also limits the continuous discharge to just under 200A, but I've seen quick bursts, just starting out from a dead stop, of about 320A. Fresh off the charger, the gel total voltage is about 78-79V. After running for about 10-20 seconds, the voltage will settle in at about 75-76V. At the first hill, without the booster pack, the voltage will sag down to about 65-66V. About 3/4 through the capacity, it will sag down further, to about 61-62V from a resting voltage of about 73V. The booster pack sits at about 80V, after the initial surface charge is burnt off, and once the gel voltage dips below 70V, during that first hill climb, it kicks in and brings the total voltage back up to about 73V. Further on in the ride, the booster pack keeps the voltage up to about 69-70V, when it would otherwise dip down to 61-62V.

What I've learned is that although the booster pack does a great job helping out, I'm still getting some sag. I'm also seeing that my 10Ah booster pack is not lasting the full duration of the gels. To address both issues, I'm adding another 5Ah of capacity. I just received a big box of 5s-5000 Turnigy packs from HobbyKing. In addition to upping the capacity in my present pack, I have enough to make a second 20s3p 15Ah booster pack.

I found a new extruded aluminum box that is sized perfect to hold a 20s3p configuration, plus a custom board that has the balancing circuits and the control section on one board. I also got some smaller variants that would be perfect to hold four 5s-5Ah packs that could be configured as either 10s2p 37V/10Ah or 20s1p 74V/5Ah. The same custom booster pack BMS, or one of the regular v4 BMS PCBs would fit in this as well.

View attachment 3



I did a new custom single PCB for the entire booster pack BMS. It has a more compact set of cell circuits that have single shunt resistors per channel. The max balance current will be around 500mA with this setup, which I've found to be more than enough for these LiPos. This allows room for a lengthwise set of parallel adapters so that the balance plugs from all 12 5s-5000 can be plugged directly in. In addition to the custom end plate for the big box, I also did one for the smaller version, which I will probably use as an ebike pack for one of my folders.



The PCB in the middle is simply a surface mount version of the same BMS board. I will use this to verify the basic SM layout/parts operation. Because this test version will be built by hand, I purposely used larger versions of some parts, like most of the resistors. Once tested, I can scrunch things up a bit and convert this over to gerber files, with normal solder masks, etc., so that they can be machine-built. The basic booster pack BMS can also be used without the special auto-"kickin" logic enabled, in which case it will then simply be a no-frills BMS with full cell-level LVC and HVC protection and auto cell balancing. The goal is to have completely sealed 37V/10Ah ebike packs that basically mimic a 36V SLA setup. These self-contained 36V/10Ah batteries could be used in series, or in parallel, or both, and the whole lot bulk-charged with a simple charger/supply. THere's not even a separate set of charge wires that are required.

Anyway, that's the long-range plan. For now, I'm concentrating on the booster pack and its variants.

-- Gary

 

[redorblack]

Being that you are going to seal it up inside an extruded aluminum casing, are you planning on fitting a vent of any sort to prevent pressure from building up inside the case in the event of a catastrophic cell failure?

 

[MitchJi]

Hi Gary,

Ride-4-fun has quite a few high performance options but nothing that addresses the problem you've solved with the Nano pack. I bet they'd be interested in what you've done, maybe even in purchasing some parts or getting some help developing a similar product.

I think I was incorrect. Ride-4-fun might not have this problem.

They recommend Trojan 150ah batteries, which I believe are Flooded Lead Acid:
1. These would require a lower C rating to handle heavy amp draws. For example 300a with your 100ah Gels is 3C vs 2C with the 150ah Trojans.

2. Your Gel cells probably have a lower C rating than Flooded Lead Acid batteries:
http://www.cameronsoftware.com/ev/EV_BatterySelection.html

I looked at the 150Ah flooded types but the problem is they just don't last very long in climates that can routinely get over 100F, which can happen here quite a bit. Instead of roughly four years of life, I'd only get about two years out of them before I'd have to replace them again.

Good information but I wasn't suggesting the Gels are a bad choice for you. I was merely stating that my previous comment that Ride-4-fun would probably be interested in booster kits might not be correct as it addresses a problem they might not have.

Question:

At that point, the equalization phase kicks in, where the current is kept at 2A and the voltage is allowed to rise up to 98-99V.

Does this imply that the system can handle more voltage (could you run 8 Gels?) to get more capacity (about 15%) and better performance? I realize this would require modification of the charging system.

 

[GGoodrum]

Being that you are going to seal it up inside an extruded aluminum casing, are you planning on fitting a vent of any sort to prevent pressure from building up inside the case in the event of a catastrophic cell failure?

Yes, the end plates have a rubber seal that serves to make the box "water-resistant". This can also act as a pressure seal/vent.

 

[GGoodrum]

Hi Gary,
Good information but I wasn't suggesting the Gels are a bad choice for you. I was merely stating that my previous comment that Ride-4-fun would probably be interested in booster kits might not be correct as it addresses a problem they might not have.

I didn't take it that way. Anyway, I do agree that Ride-4-Fun may or may not be a potential market for these. I have heard that the floodeds also sag quite a bit, so I'm guessing these booster packs would still be useful for those living in hilly areas.

Question:

At that point, the equalization phase kicks in, where the current is kept at 2A and the voltage is allowed to rise up to 98-99V.

Does this imply that the system can handle more voltage (could you run 8 Gels?) to get more capacity (about 15%) and better performance? I realize this would require modification of the charging system.

I'm sure the motor and controller can handle more voltage, but I'm not sure about the instrument panel/computer as the "gas gauge" is calibrated for the six gels. Also, the controller would have to be reprogrammed or you will still be limited to 25 mph. The charger would have to be replaced, or at least have its profile reprogrammed, if that's possible. Finally, there just isn't room to add two more 100Ah gels. They have another setup with nine 8V gels, so maybe you'd have room to add one more somewhere.

In the future, I might look to go all LiPo, and use maybe six of these 74V/15Ah packs in parallel. That would give me 90A and there's be plenty of room for more, if I need more range.

-- Gary

 

[MitchJi]

Hi Gary,

A couple of 2005 GEM's have come up locally at reasonable prices, both with bad batteries.

The problems with the 2005's are:

• The controller is lame, fixed in mid 2006.
  The charger is lame (Delta-Q V1), fixed at the beginning of 2006.

In your opinion if I purchase a 2005 with the intention of putting in a Lithium pack how big an issue is a lame OEM charger? Maybe the shortcomings wouldn't make any difference for a Lithium pack, which you might not know.

But would you even want to use the OEM charger with Lithiums?

How much would a good alternative cost (worst case is a $600 ride-4-fun "Super" unit)? I guess you'd recommend a Zephyr and some meanwell's?

Also in 2005 GEM improved the steering and handling, and upgraded the brakes to disc's. GEM improved the steering geometry again 2007. Have you driven a 2005 or 2006 and if so how does the steering compare with your 2007?

 

[texaspyro]

Being that you are going to seal it up inside an extruded aluminum casing, are you planning on fitting a vent of any sort to prevent pressure from building up inside the case in the event of a catastrophic cell failure?

Yes, the end plates have a rubber seal that serves to make the box "water-resistant". This can also act as a pressure seal/vent.

I'm not sure that I'd trust that to vent properly since the end plates are screwed down... those seal rings can be surprising effective at preventing pressure release. In similar situations, I have done things like drill a sizeable holeor two in the case and plug it with a plastic hole plug or rubber stopper. Those plastic hole plugs can be very water tight. The washer bottle in my 560SL has one that seals a hole used by the optional headlight washer pump.

 

[GGoodrum]

Hi Gary,

A couple of 2005 GEM's have come up locally at reasonable prices, both with bad batteries.

The problems with the 2005's are:

• The controller is lame, fixed in mid 2006.
  The charger is lame (Delta-Q V1), fixed at the beginning of 2006.

In your opinion if I purchase a 2005 with the intention of putting in a Lithium pack how big an issue is a lame OEM charger? Maybe the shortcomings wouldn't make any difference for a Lithium pack, which you might not know.

But would you even want to use the OEM charger with Lithiums?

How much would a good alternative cost (worst case is a $600 ride-4-fun "Super" unit)? I guess you'd recommend a Zephyr and some meanwell's?

Actually, I think the charger might be just fine. The "lameness" might come from not being as clever in how they do the charge profiles, like the weird equalization phase mine has, that holds the current to 2A and lets the voltage rise to about 98-99V. I think this is unique to the gels, but in any case, all that is needed for a Lithium-based setup is a just a single standard CC/CV profile. I definitely think being able to make use of the existing OEM charger would be desirable. The BMS can be tweaked in order to match any of its weird profile features.

Also in 2005 GEM improved the steering and handling, and upgraded the brakes to disc's. GEM improved the steering geometry again 2007. Have you driven a 2005 or 2006 and if so how does the steering compare with your 2007?

I haven't driven anything but the 2007 model I have, so I don't have anything to compare to, but I will say having disc brakes is a good thing. I did get a very noticeable handling improvement, however, switching to 14" wheels and tires. It feels more like a small car now, rather than a golf cart that is about to be tipped over. This is really apparent at 40+ mph, but I do notice a difference at slower speeds as well.

I don't know about the controllers either. From what I understand, the older ones might be easier to program, but I haven't tried this just yet on mine. I'd love to be able to get in and at least fix the speedometer function, so it reads right, and maybe bump up the current limit a bit. In any case, Ride-4-Fun can reprogram the ones earlier than mine, if you don't want to buy the kit from that link you provided before.

Which motor do these 2005 models have? There is the stock 4.5hp version, and then an upgraded version that is supposedly rated for 7hp. Ride-4-Fun claims the 7hp version can't really deliver this much power as it gets too hot, but frankly, I don't know. It might just be hype on their part, in order to keep selling their 7.5hp version, which is beefier, and has cooling fins, etc. I had the 7hp GEM version in mine, and I did replace it with the 7.5hp from Ride-4-Fun, but I never tested the stock 7hp version. Part of the reason was because I needed the motor in order to get the speed increase, because it came with a modified speed sensor. Knowing now that there is a programming option for my controller that might let me get the speed increase without changing the motor, and also getting that increase without screwing up the speed readout, I might not make the same decision about swapping the 7hp motor for the 7.5hp version from Ride-4-Fun.

I'm just finishing up building the BMS board for my new booster pack, which go in the box shown above. As I said, this one will be configured as 20s3p, or 15Ah. I hope to finish this up, and get the new pack tested and installed this weekend. After that, I will reconfigure my current pack into a 20s3p setup, and build a second one of these completed units. I might then add this reconfigured new one in parallel with the new booster pack, and see what kind of range increase I might get. Eventually, I could see adding four more of these 72V/15Ah packs, for a total of six, and replace the gels completely.

Anyway, I'll post some more progress pics soon.

-- Gary

 

[GGoodrum]

Being that you are going to seal it up inside an extruded aluminum casing, are you planning on fitting a vent of any sort to prevent pressure from building up inside the case in the event of a catastrophic cell failure?

Yes, the end plates have a rubber seal that serves to make the box "water-resistant". This can also act as a pressure seal/vent.

I'm not sure that I'd trust that to vent properly since the end plates are screwed down... those seal rings can be surprising effective at preventing pressure release. In similar situations, I have done things like drill a sizeable holeor two in the case and plug it with a plastic hole plug or rubber stopper. Those plastic hole plugs can be very water tight. The washer bottle in my 560SL has one that seals a hole used by the optional headlight washer pump.

That's a good idea. I'll definitely look into doing this, thanks.

-- Gary

 

[GGoodrum]

More progress has been made, albeit slower than I'd like. I finished the BMS assembly, and got it fully tested. The next step is to plug the packs into the BMS, and button everything up in the box. This I will do later today. Anyway, here's a few shots of where I'm at right now:



The two right-hand sections each have 10 cell circuits. As can be seen, I'm using a single power resistor per channel, instead of the normal two. This limits the balance current to about 450mA, but I've found that in this application, that's more than enough. It takes far longer to recharge the gels in the GEM, than it does for this pack, so it doesn't matter if it takes a bit longer to balance. Besides, with healthy cells, only a little balancing is usually required. There's no LED for each channel, for obvious reasons, but other than that, these cell circuits are identical to those found in the Zepher BMS. The control section, on the left, has 8 back-to-back 4110 FETs that are used to control charging and discharging.





The custom end plate has the main battery terminals, plus a single 24-pin VAL-U-LOK connector that can be used as a test plug. Connections to each cell are brought out, as well as the LVC/HVC opto signal. This end plate is connected to the BMS via a strap piece that connects the main negative terminal to the output FET connection point on the BMS. The positive connection goes to a special 3-1 4mm bullet plug adapter that I made, via a heavy gauge wire. It is not so easy to see in the pics, but there is a second one of these 3-1 adapters that is used to join the negative pack leads. This one is bolted to the pack side of the FETs.




The "12-pack" is now ready to be connected. There are four sets of three balance plugs that are fitted into matching connectors on the back side of the BMS board. For the main pack leads what i'm doing is running three separate strings of four packs in series. They are only connected in parallel at the pack level via the 3-1 adapters.

Next up is to plug everything in and make sure there's no magic smoke. I'll then slide everything into the case/box, and seal it up. After that it gets installed in the GEM, in place of the test pack I've been using, and then I'll start doing some more tests. I'll post some more pics then, when I get to that point.

-- Gary

 

[MitchJi]

Hi Gary,

Actually, I think the charger might be just fine. The "lameness" might come from not being as clever in how they do the charge profiles, like the weird equalization phase mine has, that holds the current to 2A and lets the voltage rise to about 98-99V. I think this is unique to the gels, but in any case, all that is needed for a Lithium-based setup is a just a single standard CC/CV profile. I definitely think being able to make use of the existing OEM charger would be desirable. The BMS can be tweaked in order to match any of its weird profile features.

Thanks! I think you are probably correct, and that the standard charger can probably be used to put out some kind of a single standard CC/CV profile.

I haven't driven anything but the 2007 model I have, so I don't have anything to compare to, but I will say having disc brakes is a good thing. I did get a very noticeable handling improvement, however, switching to 14" wheels and tires. It feels more like a small car now, rather than a golf cart that is about to be tipped over. This is really apparent at 40+ mph, but I do notice a difference at slower speeds as well.

There are a lot of factory and special wheels that fit the Gem. I'm pretty sure I can get something very good for a lot less than ~$1,000. Something like $25-$50 per wheel plus some good radials for ~$100 each should do nicely.

I don't know about the controllers either. From what I understand, the older ones might be easier to program, but I haven't tried this just yet on mine. I'd love to be able to get in and at least fix the speedometer function, so it reads right, and maybe bump up the current limit a bit. In any case, Ride-4-Fun can reprogram the ones earlier than mine, if you don't want to buy the kit from that link you provided before.

Does the kit I linked to work with the newer controllers (I don't remember )?

It sounds like ride-4-fun can reprogram your controller:
http://www.ride-4-fun.com/gem-car-motors-reprogramming.html

NOW AVAILABLE FOR ALL GEM CARS INCLUDING All 1999-2010 Models

Ride-4-Fun is proud to be the exclusive worldwide distributor of the only high performance Gem car motor. The performance of your 7.5HP Ride-4-Fun powered GEM Car is further enhanced with expert Re-Programming. We can Re-Program ANY GEM Car and are the only programming source covering all GEM cars, to achieve peak 7.5HP torque and top-end performance....

Re-Programming Services: To get the most from your 7.5HP motor, you can take advantage of our Re-Programming services. Ride-4-Fun is the only company setup to Re-Program ALL GEM car makes and models - including 05'-08' GEM's! We've done the research and testing to increase performance and reliability.

and:
http://www.ride-4-fun.com/frequently-asked-questions.html#Can I reprogram my 05' Gem?

Can I reprogram my 2005 or newer Gem?

Yes, Ride-4-Fun has been testing it's 2005 reprogramming systems for months and we have begun 2005 reprogramming services for our customers. The 2005 Gems use a different controller interface thus making the old reprogramming systems obsolete on these new Gems. Ride-4-Fun has made the investments in developing and testing new reprogramming equipment and we are ready to serve our 2005 Gem Car customers.

Which motor do these 2005 models have? There is the stock 4.5hp version, and then an upgraded version that is supposedly rated for 7hp. Ride-4-Fun claims the 7hp version can't really deliver this much power as it gets too hot, but frankly, I don't know. It might just be hype on their part, in order to keep selling their 7.5hp version, which is beefier, and has cooling fins, etc. I had the 7hp GEM version in mine, and I did replace it with the 7.5hp from Ride-4-Fun, but I never tested the stock 7hp version. Part of the reason was because I needed the motor in order to get the speed increase, because it came with a modified speed sensor. Knowing now that there is a programming option for my controller that might let me get the speed increase without changing the motor, and also getting that increase without screwing up the speed readout, I might not make the same decision about swapping the 7hp motor for the 7.5hp version from Ride-4-Fun.

http://www.ride-4-fun.com/gem-car-accessories.html

Performance Controllers - $599

The final stage in building out your high performance GEM Car is a High Output Ride-4-Fun Controller. At Ride-4-Fun we take your stock controller and hand build it to put out 45% more power, with increased reliability for all GEM car applications.

Now available for all GEM Cars equipped with a Ride-4-Fun 7.5HP motor. Please note that a working core and a waiver form is required at time of purchase.

I trust their claims on their motor because I called and asked them a few questions about their motor and controllers and Jim said something like: "You don't need our Performance Controller, its only for the crazies." I said I might be a little crazy and he continued to explain why he thought it was a bad idea. He said it would give a higher top speed but in his opinions Gem's are not safe at the higher speeds. I said isn't the improved acceleration worthwhile? He said that the acceleration with the reprogramed factory controller is more than sufficient and that the more powerful controller (550A vs 350A) goes through batteries too quickly (less applicable with your booster packs or with a lithium pack) to be useful in any case. I don't believe they would try to convince me that their controller is unnecessary and hype their motor dishonestly at the same time. It also sounds like they could bump up your controller to use 350A more frequently, unless you want the 550A, for crazies model .

After that, I will reconfigure my current pack into a 20s3p setup, and build a second one of these completed units. I might then add this reconfigured new one in parallel with the new booster pack, and see what kind of range increase I might get. Eventually, I could see adding four more of these 72V/15Ah packs, for a total of six, and replace the gels completely.

Sounds great, I think it will be excellent.

But, personally, particularly after John's problems, I'm a little hesitant to trust HK lipo's in packs this size:
http://endless-sphere.com/forums/viewtopic.php?f=14&t=27494&hilit=+lipo

Anyone thinking Lipo is safe needs to think again

Sorry LFP et al, but this stuff simply isn't ready for prime time. This is (was) a 4s5ah 20C hard pack charged only to 16.2V (4.05v/cell) and set aside...never used. It only saw a few degrees C change in temp daily for the few weeks it sat there, and no direct sunlight, so I'm at a loss as to the cause. I'm just very lucky that it didn't spread to the other packs only inches away, and that the single pack didn't get hot enough to ignite the formica/particle board desk it sat on, or we would have surely lost our guest house that I use for ebike stuff and my office.

 

[GGoodrum]

Good to know, re: the new programming options at Ride-4-Fun. I might just go ahead and get one of their "crazy" version controllers. I don't want to take mine out, and send it to them. I use my GEM almost everyday, so it would be a pain to be without it for two weeks. I might also still look into just getting the reprogramming kit, and trying it myself.

As for John's issue, those were those special "hard packs", that usually are used with RC cars, I think. In any case, I think the "regular" Turnigy's we've all been using will be fine, especially if they in an enclosed metal box. As suggested, I'm going to add a pressure vent plug, that will vent any overpressure down away from anything that could burn. I'm also going to be monitoring the charge and discharge cycles for awhile, to get a good idea of how well the cells are behaving. Later I'll just check them once in awhile, to see how they are doing. It's pretty easy to spot cells starting to go hinky, if you monitor a cycle once in awhile.

I've now got the BMS mated to the pack, and no magic smoke. Next I will slide everything into the box, snap some pics, and then install it in the GEM. I just finished making an adapter cable for my triple CellLog monitoring unit, so that I can use it with the new pack which has a single 24-pin connector on the front, instead of two 18-pin versions.

-- Gary

 

[GGoodrum]

The shots below are the buttoned up booster pack, side-by-side with the test unit I've been using. It is only a little longer and a bit wider, but has 50% more capacity.





The pack is charging right now, and then I'll get it installed in the GEM. The packs all came about half full, at 3.88-3.89V, so I just connected everything together. I'm using a lab supply to charge it, just so I can monitor the process the first time. Also, the gels in the GEM are full right now, so better to start with the booster pack full as well.

-- Gary

 

[mikebikerad]

Nice bms!

I have a gem EL that i jus converted to lithium. I used 24x calb 70ah cells and signalab bms. I have 2x45volt 12 amp chargers. So far i have cycled the cells three times, and i am amazed with the preformance after gettin the lead out. I havnt done a range test yet but will post pictures of my car here when i get a chance. These are great little towncars!!! Good to know some others on here are rockin them. I might be interested in doing a booster pack also to save the calb cells a bit.

Mike

 

[MitchJi]

Hi Mike,

I have a gem EL that i jus converted to lithium. I used 24x calb 70ah cells and signalab bms. I have 2x45volt 12 amp chargers. So far i have cycled the cells three times, and i am amazed with the preformance after gettin the lead out. I havnt done a range test yet but will post pictures of my car here when i get a chance...

Mike

I'm really interested in your range results or wh per mile.

I'm also really interested what Gary thinks his range would be with a 60 AH pack, now that he has more experience.

Thanks!