999zip999
100 TW
Alan B I do enjoy your input and your wisdom. I do enjoy the enlightenment your brain brings here 2 endless sphere. Thank you.
Why 50V-60V is considered the optimum battery voltage?
- Thread starter thunderstorm80
- Start date
Alan B
100 GW
999zip999 said:
Thanks for your kind words.
Near the end of my software, controls and electronics engineering work career at a National Laboratory I was retasked to Electrical Safety issues. I attended the Electrical Safety conference a few years ago in Louisville. I've been to a lot of conferences over the years, but none as moving. Some of the presentations were about situations where co-workers had perished, and what had been learned to improve future safety. We all need to appreciate the risks these folks take in their daily work of keeping our electrical grid functioning. I can hardly imagine the job they face getting the power lines repaired in the midst of a severe storm, in the wet and windy darkness.
I did experience 600 volts DC across my little finger when I was a teenager (from a commercial Ham Radio transmitter that had exposed terminals on the rear of the cabinet!), and the usual 120 VAC shocks, as well as the occasional TV high voltage zap. I try to avoid that anymore.
Good solid connections, well insulated, good and appropriate tools and personal safety equipment all goes together to make reliable and safe systems and workers - ebike or whatever.
One thing about lithium (li ions) cells that irks me is the large change in voltage between fully charged and empty. With my BBS 02 a 13S setup is nice when fully charged but not so good below half charge. Adding another cell for 14S would be good, but a 15S battery would be even better. At this point the BBS controller kicks in and says that the voltage is too high when the battery is fully charged.
So for my controller a supply voltage somewhere between 50 ~55 volts would be very good.
But where I am going with these thoughts is a buck controller chip like the LTC 3895. With this switch mode buck convertor I could run at "say" 16 or 18S and feed the BBS controller with "say" 55 volts. The chip is designed for battery operation and when the input voltage drops to 55 volts or below, the top switching power mosfet just stays on and passes the supply volts straight through with no further losses. While the battery volts are higher than the output to the controller the chip operates as a buck convertor. This means that the 20 amps (that's where my BBS 02 is set to currently) being taken by the controller , looks like some what less to the battery (since its being chopped to reduce the voltage). So perhaps the best of both worlds. High voltage battery pack with maybe less cells in parallel required, and a consistent supply voltage to the controller to keep full performance until the last.
Regards Bob
So for my controller a supply voltage somewhere between 50 ~55 volts would be very good.
But where I am going with these thoughts is a buck controller chip like the LTC 3895. With this switch mode buck convertor I could run at "say" 16 or 18S and feed the BBS controller with "say" 55 volts. The chip is designed for battery operation and when the input voltage drops to 55 volts or below, the top switching power mosfet just stays on and passes the supply volts straight through with no further losses. While the battery volts are higher than the output to the controller the chip operates as a buck convertor. This means that the 20 amps (that's where my BBS 02 is set to currently) being taken by the controller , looks like some what less to the battery (since its being chopped to reduce the voltage). So perhaps the best of both worlds. High voltage battery pack with maybe less cells in parallel required, and a consistent supply voltage to the controller to keep full performance until the last.
Regards Bob
liveforphysics
100 TW
The performance dropping with SOC% is most efficiently solved by running a motor with 1 less turn, and having a bit more phase current.
It would be equally easy to make a 6s bike have identical performance from 100% to 1% SOC as any higher voltage vehicle system. Either way it's a ratio of pack voltage to motor BEMF.
It would be equally easy to make a 6s bike have identical performance from 100% to 1% SOC as any higher voltage vehicle system. Either way it's a ratio of pack voltage to motor BEMF.
recumpence
1 GW
I am very glad this subject was brought up. I have stuck with 12S for numerous reasons. Manily for convenience. Being an RC drive guy, 12S is compatible with everything I already run. My chargers, all of my packs, my controllers, everything is setup to run 12S. But, as the years have gone by, I have begun pulling higher and higher amperage to get the performance I am looking for. A couple of my bikes peak at well over 600 amps (for 1/2 second) and 300 amps for 3 to 4 seconds. With such a short duration, these crazy high amp pulls are not difficult to manage. Pulling that much power continuously is another story. But, we are talking about a sub 100 pound vehicle here. Continuous power is rarely more than a few thousand watts for me unless I am running in very deep snow (which is alot of fun, by the way) 
. But, I have pulled 300 amps for 7 minutes through a water propulsion application recently and I learned a lot. In that application, my 2 gauge wire got hot. But, for our typical application, low voltage is safe and high amperage is easy to manage. As Luke said, we are only running a foot or two of wire. And, as I said, we are only pulling high current for very short durations.
For the forseeable future, I will stick with sub 50 volt setups.
Matt
. But, I have pulled 300 amps for 7 minutes through a water propulsion application recently and I learned a lot. In that application, my 2 gauge wire got hot. But, for our typical application, low voltage is safe and high amperage is easy to manage. As Luke said, we are only running a foot or two of wire. And, as I said, we are only pulling high current for very short durations. For the forseeable future, I will stick with sub 50 volt setups.
Matt
I see a controller like the phaserunner being the direction where controllers need to go. Fully potted for water-proofing, and also shock resistance. That last bit is more important than many realize, as I have had several electronic components arrive (from vendors I trust) and they had clearly been roughly mishandled by the drunken gorillas at FedEx/UPS. Found to be DOA, and I don't know where to start to diagnose. I may be forced to spend a couple months reading over old ES posts to learn about trouble-shooting and repair of controllers, but....it's safe to say that most people don't want the bother, and will pay extra for plug-and-play.
That being said, I would gladly pay an extra $100 to get a Phaserunner that is longer (to allow room for more FETs) so it can support more amps. The fact that it is completely sealed means that it can be hidden (along with the wiring clutter) without air-flow (not that it would be horrible to run ducting and computer fans inside a battery housing to a hidden controller) is a desirable benefit to many new ebike customers. Maybe they shouldn't care about stuff like that (priorities, right?), but they have money to spend, and they do care about that. I have several ideas to easily and cheaply mate the Phaserunner to an aluminum frame, so the frame is a heat-sink.
I recently read (from Luke and others) that high current requires good crimps and no solder. When it comes to the connectors and wire gauge, thicker than you think you need is always better. I am definitely starting to see some of the benefits of lower volts and higher amps, but...that path is still limited due to existing components. The 25R cell has been slowly dropping in price, so a 100A pack is getting affordable and fittable. 5P at 14S is larger than 5P at 10S...hell, 6P/10S is smaller than 5P/14S. (60 cells vs 70 cells).
Maybe Lyen will offer a "Luke" version of the 18-FET?
That being said, I would gladly pay an extra $100 to get a Phaserunner that is longer (to allow room for more FETs) so it can support more amps. The fact that it is completely sealed means that it can be hidden (along with the wiring clutter) without air-flow (not that it would be horrible to run ducting and computer fans inside a battery housing to a hidden controller) is a desirable benefit to many new ebike customers. Maybe they shouldn't care about stuff like that (priorities, right?), but they have money to spend, and they do care about that. I have several ideas to easily and cheaply mate the Phaserunner to an aluminum frame, so the frame is a heat-sink.
I recently read (from Luke and others) that high current requires good crimps and no solder. When it comes to the connectors and wire gauge, thicker than you think you need is always better. I am definitely starting to see some of the benefits of lower volts and higher amps, but...that path is still limited due to existing components. The 25R cell has been slowly dropping in price, so a 100A pack is getting affordable and fittable. 5P at 14S is larger than 5P at 10S...hell, 6P/10S is smaller than 5P/14S. (60 cells vs 70 cells).
Maybe Lyen will offer a "Luke" version of the 18-FET?
liveforphysics
100 TW
A Luke edition would be FOC based.
Shall we name it the "What the FOC"?
liveforphysics
100 TW
spinningmagnets said:
cell_man
100 kW
liveforphysics said:
I'd love to see or even put together a kit that follows along the lines of what you are suggesting. The primary difficulties I'd see, is making a suitably rated bms that could handle the extra current, getting a controller to run such a low turn motor well and be able to handle the extra current, without being huge.
I'm guessing that the lower voltage would offset the low winding resistance/inductance, to some extent, but still (not to mention the phase wires would need to be huge). I get that of course you don't need to have discharge fets on the BMS, that the controller can do the job.
I can fully accept the merits of this approach, but at the same time, I struggle to see how you wouldn't end up with a bigger/more expensive controller (and bms), than if you were to do the same job with a more typical 50V or so.
No offence, but I've not faced with big issues from running 14S packs and suitable controllers from them. At the same time, I'm also totally open to trying something new/different.
Would love to hear some specifics, of how the controller would be approached
The sort of components that would be required to pass this extra current and how it might the size/cost of the conplete system.My post is in no way intended to question your comments, but rather to dig a little deeper and get more details. I have the utmost respect for you Luke, and it's really cool that you still find the time to come onto ES and share your thoughts 8)
miro13car
100 kW
optimum for what?
ebike, escooter?
I carry 36V and dont see the need to add complexity and extra cells, weight if I can easly reach 47km/hour on my EPLUS , with pedalling 50km/hour.
ebike, escooter?
I carry 36V and dont see the need to add complexity and extra cells, weight if I can easly reach 47km/hour on my EPLUS , with pedalling 50km/hour.
999zip999
100 TW
Well who has a 1turn motor ? Sounds like a diy home brew.
A "1 turn motor" was an earlier response to suggest how far could this go. A more realistic example would be to switch from a 4T hubmotor at "X" volts, to a 3T motor at a lower voltage, but higher phase amps.
liveforphysics
100 TW
cell_man said:Would love to hear some specifics, of how the controller would be approached
Thank you for the kind words. For an ebike in the say 3-4kW and below range, I would use 6s battery with this MOSFET in a 6-fet layout:
http://www.infineon.com/dgdl/Infineon-IPT004N03L-DS-v02_00-EN.pdf?fileId=db3a30433e9d5d11013e9e0f382600c2
With ~24 of this capacitor:
http://www.digikey.com/product-detail/en/panasonic-electronic-components/35SEPF120M/P16342-ND/4204178
With the caps and FETs of the power stage mounted to custom formed busing like this:
And use these current sensors in each phase lead and one battery current lead:
http://www.digikey.com/product-detail/en/allegro-microsystems-llc/ACS770ECB-200B-PFF-T/620-1538-5-ND/4473977
Then run fast A2D's and a nex-generation FOC based motorcontroller brain, and/or lebowski's or whatever is capable of the fastest current control today.
And perhaps for a sporty ebike or scooter, say 5-50kW, I would go 20s and use x~48-72 of this capacitor:
http://www.infineon.com/dgdl/irfp4568pbf.pdf?fileId=5546d462533600a40153562c7c32201b
And this MOSFET:
http://www.infineon.com/dgdl/Infineon-IPT015N10N5-DS-v02_01-EN.pdf?fileId=5546d4624a75e5f1014ac94680661aff
Then I would package all the parts that get hot against one side, and monolithically encapsulate the whole controller in a thermal potting compound pour, and have only the power busing terminal ends setup for bolting on ring terminals, and the signal level connector poking out of the potting.
999zip999 said:
Bus bar windings should be far less human work than winding, and enable higher efficiency motor performance with higher continuous torque levels.
Honda figured out how to do a 4t bus bar winding!
Doing a bus-bar 0.5t, or 1t or 1.5t etc as a DIY'er should be reasonable to do.
We can make ebike tech go to whatever level we want through creating it.
Alan B
100 GW
Beautiful Components.
999zip999
100 TW
I want the honda one.
liveforphysics
100 TW
Miles's uber motor had enough teeth on it that a series winding could likely be done as a single bus bar strip per phase taking a series path through the coils.
Kinda like Hondas winding in this early tranverse flux prototype that failed due to soft-magnetic-core material limitations. If the core of that same transverse flux honda motor were made from an amorphous glass-metal magnetic material it would be a game-changing motor topology with respect to efficiency, native torque density, and power density. If you want to know why this winding rocks, recognize that the copper in the slot is what's the critical performance metric, and all the copper outside the slot (end-turns, end-turn-losses) is just added resistance and heat to your motor that isn't making torque for you.
The objective in a winding is to complete the magnetic circuit with as great of percentage of the copper in the slot itself, with as little as possible outside the slot.
That's why this is such a sexy winding, and how to do it right:
Even more-so than the beautiful bus-bar origami from that accord hybrid 4t bus bar winding, or this new Toyota winding.
Kinda like Hondas winding in this early tranverse flux prototype that failed due to soft-magnetic-core material limitations. If the core of that same transverse flux honda motor were made from an amorphous glass-metal magnetic material it would be a game-changing motor topology with respect to efficiency, native torque density, and power density. If you want to know why this winding rocks, recognize that the copper in the slot is what's the critical performance metric, and all the copper outside the slot (end-turns, end-turn-losses) is just added resistance and heat to your motor that isn't making torque for you.
The objective in a winding is to complete the magnetic circuit with as great of percentage of the copper in the slot itself, with as little as possible outside the slot.
That's why this is such a sexy winding, and how to do it right:
Even more-so than the beautiful bus-bar origami from that accord hybrid 4t bus bar winding, or this new Toyota winding.
liveforphysics
100 TW
999zip999 said:
Beautiful sexy origami end turns, but still end turns.
Absurdly high fill factor, but still a decent chunk of end-turns.
That is like a 30% lower copper loss winding that amazingly achieves the same copper slot fill %
We need Miles/Biff/Farfle etc to make some pattern segment that can be wound by stitching through some strip of coated copper bar-stock.
It seems relatively DIY friendly to make a jig that smashes the barstock into some special bend.
999zip999
100 TW
Such beautiful porn. I have to go to my room for a while. Excuse me.
cell_man
100 kW
Thanks Luke for coming back with all that cool stuff.
Those are some amazing looking motors 8)
The bus bar in some of those motors is enamelled? I've only ever seen regular round section enamelled wire. I'd imagine the insulation could easily be damaged on an enamelled wire with square cross-section and it must be fairly hard to work with? Or maybe they insulate the windings in a different way? I guess if you only have 1 turn, then you don't need to worry so much about shorts, you just wrap it in an insulating layer and you are done. That solid slot design is very smart.
Are there are any commercially available controllers that would work well for such a setup (6S, several KW) for such a low turn motor?
Apologies for all the questions
Those are some amazing looking motors 8)
The bus bar in some of those motors is enamelled? I've only ever seen regular round section enamelled wire. I'd imagine the insulation could easily be damaged on an enamelled wire with square cross-section and it must be fairly hard to work with? Or maybe they insulate the windings in a different way? I guess if you only have 1 turn, then you don't need to worry so much about shorts, you just wrap it in an insulating layer and you are done. That solid slot design is very smart.
Are there are any commercially available controllers that would work well for such a setup (6S, several KW) for such a low turn motor?
Apologies for all the questions
Similar threads
