Ideal Diodes for battery combining/switching

[Tiberius]

Hi All,

You may have seen this already. I was thinking about how to use FETs instead of Schottky diodes to combine batteries. I concluded it was workable and would save energy but was a bit complex.

Now I've just discovered Linear are making ICs to do just that.
http://www.linear.com/idealdiodes

I wonder if this could also be used as the master switch in a BMS.

Nick

 

[fechter]

That's cool.

I'll have to see how that works. Most enhancement mode devices look out though, since they have a body diode.

 

[Tiberius]

Yeah but if its connected the right way round, like a high side FET, then the body diode works as normal, and then the FET cuts in to eliminate the voltage drop.

Nick

 

[fechter]

Gotchya, that makes more sense. That would be ideal for battery combining.

 

[OpAmp]

Thanks for the outlook and the link. Linear technology has a very informative website.

 

[deVries]

Hi All,

You may have seen this already. I was thinking about how to use FETs instead of Schottky diodes to combine batteries. I concluded it was workable and would save energy but was a bit complex.

Now I've just discovered Linear are making ICs to do just that.
http://www.linear.com/idealdiodes

I wonder if this could also be used as the master switch in a BMS.
Nick

That's cool.

I'll have to see how that works. Most enhancement mode devices look out though, since they have a body diode.

Yeah but if its connected the right way round, like a high side FET, then the body diode works as normal, and then the FET cuts in to eliminate the voltage drop.

Nick

Hello Richard, I know you are busy with the BMS/LVC etc. but perhaps someone in the know can help out here...

I need to add two different voltage packs in series for long rides, so I want to be able to connect/disconnect the add-on batt-pack from the bike. Unfortunately, I'm not an electronics guru to understand how to try this using this new idea. I'm guessing the cost is not much different than the schottky? I would certainly like to try this if it's a good idea, so if anyone that's savvy wants to help out, please just post how to do this.

I'm certain I won't be the only one to benefit from this new idea.

TIA, deVries

 

[rkosiorek]

this is an interesting idea. for the all out money is no object design they may even be a good idea.

i'm not 100 percent sure but after a quick skim through the data sheet it looks like you could wire a switch in series with the GND pin. turn the switch off and the the device will be forced into a FAULT condition and effectively turn the fets off. need more study to actually confirm this.

the advantage to diodes like the 40CPQ60 or 40CPQ100 is the simplicity of using them. they are large parts with big fat leads that you could directly solder wires too. the disadvantage is that there is a voltage drop. the 60V device (40CPQ60) at 40A will drop about 0.7V the 100V part will drop about 1V. At 40A though these parts will heat up and need a heatsink. they are inexpensive at appox $6.00 each in single units. the 60V part can be used for 48V systems and the 100V parts for 72V nominal systems.

the LTC4355 is good to a maximum of 80V so the limit would be a 60V nominal battery system. other thinks will also complicate life. the parts are only available as either a 14pin or 16pin surface mount part. several resistors in addition to FETs are needed to make them work. being surface mount necessitates the use of a PCB.

the parts are much more expensive. added to the cost of making a PCB add $6.00 for the cost of the chip and about $8.00 for the FETS. then a buck or two for the passive components. the total cost would be close too triple the cost of the diodes.

another consideration for the ideal diodes is the limits of the individual FETS. never mind what the stated electrical current limit of the FET is. in a TO220 case the maximum physical dissipation of the FET would be limited to 15 or 20A or so per device. the device leads will act like fuses at 75A.to safely pass 40A per battery pack at least 2 FETS per pack would be needed. on the other hand you could just parallel a couple of the Schottky diodes. which would work out much cheaper.

so it comess down to a design decision. cheaper diodes would work well enough for every day use. for more demanding applications the ideal diode circuits may be a better choice.

rick

 

[Tiberius]

[

I need to add two different voltage packs in series for long rides, so I want to be able to connect/disconnect the add-on batt-pack from the bike.
TIA, deVries

Hi,

To add extra capacity for a long ride, the way to do it is to combine the packs in parallel. That's the classic situation that you use combining diodes for - because the batteries are not always charged and discharged together, or may be even different types or capacities.

That's the situation where an ideal diode circuit, rather than plain Schottky diodes, could save 2 to 3% of the energy. It is really aimed at lower voltage systems (5V or 3.3V) where the energy savings are much greater. But since it could also do the on/off switching and part of the BMS work, I think its well worth considering for electric vehicles.

Adding battery packs in series is normally done to get extra speed.

Nick

 

[deVries]

[

I need to add two different voltage packs in series for long rides, so I want to be able to connect/disconnect the add-on batt-pack from the bike.
TIA, deVries

To add extra capacity for a long ride, the way to do it is to combine the packs in parallel. That's the classic situation that you use combining diodes for - because the batteries are not always charged and discharged together, or may be even different types or capacities.

That's the situation where an ideal diode circuit, rather than plain Schottky diodes, could save 2 to 3% of the energy. It is really aimed at lower voltage systems (5V or 3.3V) where the energy savings are much greater. But since it could also do the on/off switching and part of the BMS work, I think its well worth considering for electric vehicles.

Adding battery packs in series is normally done to get extra speed.

Thanks! Nick for your ideas... I've read a lot of excellent one's from you in the few weeks I've been here. 8)

My booster pack long-ride pack will consist of 2 Milwaukee 28v packs that are in parallel adding about 5ah @28v, and my daily pack will top-off at 59v 8-9ah and start running nominal in the mid-50's, I'm guessing. So, adding that 2pack of parallel Milwaukee's in series to the main pack will cause these two Milwaukee's to run down first and go off-line with less AH. Initially, my volts will be way up there at about 85v, though I'm not going to use this to be a speed demon. This should allow my controller to still function correctly with the regular 50v pack, because the BMS in the Milwaukee's will cut these two 28v off first.

Let me know if I'm missing something, because I'm a rank rookie building my first e-bk. I need to avoid all the mistakes I can.

Thanks!

If I'm missing something, then please explain.

 

[Tiberius]

My booster pack long-ride pack will consist of 2 Milwaukee 28v packs that are in parallel adding about 5ah @28v, and my daily pack will top-off at 59v 8-9ah and start running nominal in the mid-50's, I'm guessing. So, adding that 2pack of parallel Milwaukee's in series to the main pack will cause these two Milwaukee's to run down first and go off-line with less AH. Initially, my volts will be way up there at about 85v, though I'm not going to use this to be a speed demon. This should allow my controller to still function correctly with the regular 50v pack, because the BMS in the Milwaukee's will cut these two 28v off first.

I think we need to check on the series/parallel arrangement here. Can you list the batteries you want to use, and then we'll try to work out the best configuration?

Nick

 

[deVries]

My booster pack long-ride pack will consist of 2 Milwaukee 28v packs that are in parallel adding about 5ah @28v, and my daily pack will top-off at 59v 8-9ah and start running nominal in the mid-50's, I'm guessing. So, adding that 2pack of parallel Milwaukee's in series to the main pack will cause these two Milwaukee's to run down first and go off-line with less AH. Initially, my volts will be way up there at about 85v, though I'm not going to use this to be a speed demon. This should allow my controller to still function correctly with the regular 50v pack, because the BMS in the Milwaukee's will cut these two 28v off first.

I think we need to check on the series/parallel arrangement here. Can you list the batteries you want to use, and then we'll try to work out the best configuration?

Nick

Thanks! for the offer Nick.

In order to keep this thread on-topic I posted the battery design ideas here:

http://endless-sphere.com/forums/viewtopic.php?f=14&t=5674

Topic: Help Design Booster Multi Dual Batt Pack -> 60v 88v Maybe ?

 

[Johnbear]

My diode prototype:

120v, 60 amp - for series joining!

Ran it today at 70v 26A, just barely got warm. Worked great. I will have to get some decent potting! View attachment IMG_2206.jpg

 

[Anonymous]

.

 

[Johnbear]

What voltage/ current are you running? Does the diode ever get warm/ hot? Are you using them on lithium? That is my intended use after I test them on my nicads.

Thanks!