Mixing chemistries for most economic performance and range

[torrent99]

In the sketch above the discharge would be natural and the charge would be controlled. I can see now that it wasn't really what you were looking for.

Maybe putting the load in parallel with the high current battery would be better, yes I think so. S3 wouldn't be needed.

With that approach discharge would be "b) limit A" and charge would still be controlled. If I'm not mistaken that is all you have wanted.

Hmm, yes I see that meets the requirements I've put above. Unfortunately I forgot one!

* Don't use B unless it's needed!!!

(Why? well B is only small capacity, for those times when you need nitro like getting up a nasty hill. We want to save our nitro until it's really needed!)

One way I can think of is to use a current limiter on A (such as you suggest), and simply switch B on or off (either manually or when we detect that A's limiter has kicked in)...

 

[ZapPat]

All this reminds me of the idea of using ultra-caps for the booster pack. Check out this doc that's about exactly this:

View attachment Regenerative Braking for an Electric Vehicle Using Ultracapacitors and buck-boost PWM.pdf

It is similar to what bearing suggested, meaning it's a PWM circuit, but the difference is that both switches are used actively, which turns this circuit into a bi-directionnal buck/boost circuit that links both batteries. The power pack is on the low V side, while the output and range battery are on the higher V side.

I still think the best thing to try out first is just putting similar tech power and energy cells in parallel, which wouldn't cost much to try out. Have seperate shunts for each cell as to measure current in/out of each cell, plus another for the global current in/out if you don't want to do the math to figure this out (simple addition of the other two). Using this with any kind of variable load would give us a real good idea of how currents are circulating.

All I have here to do this with would be some NiMh cells and a couple extra LiPo cells I picked up... maybe I'll try since their voltages almost seem to match at a 3:1 ratio (3 NiMh's to 1 LiPo). These Lipo's are powerfull little suckers, this weeny 400mA cell should be able to give 8A if I can believe the 20C rating. My NiMH cells can maybe do 2C, and they are 4Ah cells, so maybe will be able to give 8A also, but for longer than the weeny LiPo of course. I haven't measured the LiPo's internal R yet, but I'm sure it'll not be far from these much larger NiMh cells (older tech).

 

[Aerowhatt]

All these ideas are just far more complicated than needed. Basically if you go whole hog on electronic controls you spend enough time (that you could work some overtime) and money to double the a123 packs capacity. That partially solves issues all by itself. The simplest electonic way to go would be to put a current limit circuit on the ping. Not like the one built into the BMS which drops it out, but one like a constant current charger uses which would keep it pumping out current but throttling the max output. Whenever the bike is drawing less, some goes to the a123's to bring them up from the last acceleration or hill.

Simple is better! Sometimes it requires altering the voltages of the component batteries. I have a LiMn eG0 cycle with a 30AH LiMn. The pack is 7 cells and about 27volts nominal. I needed more range on it so I paralled a 9 cell LifePO4 Ping 20ah. The fully charged Ping resting voltage doesn't exceed the max charge voltage of the LiMn (29.4volts). Simply connect them before taking off. The lower resistence of the LiMn does the heavy lifting and the ping keeps it topped up and contributes to the total load enough to increase top speed and acceleration performance. The Ping goes dead first on a very long (all she has) trip and is protected by it's BMS. The two are charged seperately, which is manual, but it could easily be automated with a couple of relays run by the key switch. No power losses to control electronics and it works quite well. Just like my SLA/LiFePO4 combination on the motorcycle which charges in parallel too, an added plus.

Think about exploring playing with the cell counts for a Ping/a123 hybrid battery pack to get it to work out.

Aerowhatt

 

[torrent99]

All I have here to do this with would be some NiMh cells and a couple extra LiPo cells I picked up... maybe I'll try since their voltages almost seem to match at a 3:1 ratio (3 NiMh's to 1 LiPo). These Lipo's are powerfull little suckers, this weeny 400mA cell should be able to give 8A if I can believe the 20C rating. My NiMH cells can maybe do 2C, and they are 4Ah cells, so maybe will be able to give 8A also, but for longer than the weeny LiPo of course. I haven't measured the LiPo's internal R yet, but I'm sure it'll not be far from these much larger NiMh cells (older tech).

That sounds great I was actually thinking of NiMH for my own "booster pack"! Unfortunately due to a 20month baby I don't have the testing time or facilities...so at the moment this is just a paper exercise for me

My original borrowed design (from Tiberius), was to use a NiMH pack with slightly less voltage than the main LiPo pack. When the LiPo sags the NiMH should prop it up...or so the theory goes....

 

[dogman dan]

I just ordered a nicad 8 ah pack for my evgloblal. I already have 12 ah sla's in the bike. If I paralelled them, which one would be doing the heavy lifting? I was planning to just run them sequentialy, the nicad first, then the sla's. If the sla's charged the nicads, that could work ok. Later I do plan to just make it a 16 ah nicad bike.

 

[ZapPat]

All these ideas are just far more complicated than needed. Basically if you go whole hog on electronic controls you spend enough time (that you could work some overtime) and money to double the a123 packs capacity. That partially solves issues all by itself. The simplest electonic way to go would be to put a current limit circuit on the ping. Not like the one built into the BMS which drops it out, but one like a constant current charger uses which would keep it pumping out current but throttling the max output. Whenever the bike is drawing less, some goes to the a123's to bring them up from the last acceleration or hill.

Simple is better! Sometimes it requires altering the voltages of the component batteries. I have a LiMn eG0 cycle with a 30AH LiMn. The pack is 7 cells and about 27volts nominal. I needed more range on it so I paralled a 9 cell LifePO4 Ping 20ah. The fully charged Ping resting voltage doesn't exceed the max charge voltage of the LiMn (29.4volts). Simply connect them before taking off. The lower resistence of the LiMn does the heavy lifting and the ping keeps it topped up and contributes to the total load enough to increase top speed and acceleration performance. The Ping goes dead first on a very long (all she has) trip and is protected by it's BMS. The two are charged seperately, which is manual, but it could easily be automated with a couple of relays run by the key switch. No power losses to control electronics and it works quite well. Just like my SLA/LiFePO4 combination on the motorcycle which charges in parallel too, an added plus.

Think about exploring playing with the cell counts for a Ping/a123 hybrid battery pack to get it to work out.

Aerowhatt

Those setups sound groovy, Aero! This encourages me to try too.

It is true that the document I posted above refers to Ultra caps, and caps aren't like batteries in that their voltage goes down linearly as they discharge, thus the need of the bi-directionnal PWM circuit to "boost" the cap voltage up to and a bit over battery voltage. One thing though about this possible route is that it lets us have total control over current flow between banks, without regard to chemistry or even voltage difference between banks. It would take a little time to design, but could easily be copied by others (fairly simple circuit if an appropriate microcontroller is used). Your method's simplicity scores it good points though for many needs! Anyways, it's always good to try different routes, both simple and complex if need be.

Have you ever used shunts to observe current (in/out) through each battery seperatly?

I will be trying out the LiPo/NiMh possibility soon, and will report my observations of their current flows.

Ciao!
Pat

 

[torrent99]

I just ordered a nicad 8 ah pack for my evgloblal. I already have 12 ah sla's in the bike. If I paralelled them, which one would be doing the heavy lifting? I was planning to just run them sequentialy, the nicad first, then the sla's. If the sla's charged the nicads, that could work ok. Later I do plan to just make it a 16 ah nicad bike.

Well my plan is ping based. I have a 36V Bafang based bike (a Wisper 905SE) whose battery will come to the end of its life sometime this year.
The Wisper has an 18A controller and usually peaks around 12-15A. For the range I need a 12Ah Ping, but I'd like to keep the max C level as low as possible to lengthen the life of the Ping. But I need some good torque occasionally (I use throttle only and can hardly pedal due to dodgy knees), so I'd like to have something to take the extra load off the ping but only occasionally... hence my interest. (A 20AH or even a 16AH ping would be a bit bulky and heavy for my purposes)So my current plan is a 12AH Ping + some sort of (semi-disposable) booster... either NiMH/NiCD based or maybe a Bosch toolpack.....

 

[dogman dan]

Which would have the lower internal resistance, sla or nicad? I suppose it doesn't matter that much, either way it goes should work ok. It's just a stopgap for awhile.

 

[swbluto]

I definitely advocate both directions to be pursued, the easy peasy simple and the sometimes needingly complex. If you're unlikely to encounter situations where you're going to run your run your booster dry from simple paralleling, then you may never cause the capacity pack's BMS to trip and the battery's charge currents to be too high. Also, if the range of currents you want to draw is only doubled or so, simple paralleling may work fine but I'd definitely try to stay on the "safe side" by testing if you decide to do so. If the amount you want to increase the current is much less, like say from 20 amps to 30 amps(1.5 times as much), you probably wouldn't have much to fear assuming the voltages of both packs are nearly the same.

But for those that want as much as you can get from both packs as possible, or even near the limits of both(assuming the boost pack can give much more than the capacity pack), I heavily suspect some kind of electronic control is the only safe, practical economic way to do so. This is more important for those that "want it all", i.e., those who want 80+ amps on tap for RC motor set-ups or high-powered crystalyte 530x motors while not breaking the bank nor the range in the process.

And, I'll reverb ZapPat, development time really isn't much of an issue. My time is pretty much free at this point(what little I have of it), and once enough time is invested in finding an economically solid working design, then it'd be readily copyable/modifiable according to whomever's needs.

 

[swbluto]

Which would have the lower internal resistance, sla or nicad? I suppose it doesn't matter that much, either way it goes should work ok. It's just a stopgap for awhile.

Well, the SLAs have an internal resistance that rises exponentially with current whereas nicad's don't increase nearly as much. This is what causes the "Peukert effect" with Lead acid, so for any higher currents, I can almost guarantee the SLAs will have a higher internal resistance than the NiCads. Of course, this depends what kind of configuration you're running with the SLAs and NiCads, but for any "normal ones", I don't think this is far off the mark.

In essence, it seems you should be ok using any high-C chemistry with the SLAs as the SLAs internal resistance increases so relatively quickly so its current output doesn't really change that much for higher currents. It'll keep increasing with higher output currents, however, so it might be wise to take some measurements.

 

[Aerowhatt]

Have you ever used shunts to observe current (in/out) through each battery seperatly?

I will be trying out the LiPo/NiMh possibility soon, and will report my observations of their current flows.

Ciao!
Pat

Yes checking the amperages in action is a necessity. Basically though you only need to look at your low output pack. (I believe that the high output pack should be such that it can handle peak loads alone) You just need to make sure your setup doesn't draw too many amps from the low output pack at wany time during normal operation. To check my hunches on the motorcycle I rode around for a few days with a 20Ah ping hooked in parallel with the SLA's and data loging setup on the ping. This gave me the data to optimize the size of the ping pack for the final setup.

Another simple strategy is to use a known load and seperately test each chemistries voltage sag with that load. It's pretty simple to do the math and predict (very acurately) the amperages contributed by both in use. the only equipment needed is a good volt meter and some kind of constant load in the ballpark of your average load on the bike.

I went down the electronic controls path a few years back when I had a 18650 based lithium cobalt pack paired with the SLA's on the motorcycle. In that case at least it was more trouble than it was worth.

Aerowhatt

 

[Aerowhatt]

Which would have the lower internal resistance, sla or nicad? I suppose it doesn't matter that much, either way it goes should work ok. It's just a stopgap for awhile.

It could be either way. SLA's of the same size but different brands and internal construction can have IR's up to three times of some others. Nicads usually don't have that broad of a range but it's not far behind.

The simple voltage sag measurement under the same load for each battery tested seperately will qualitatively get you a good Idea what to expect. Since the battery brands and types (in the same chemistry) vary so much you really have to test to find out.

Aerowhatt

 

[John in CR]

swbluto,

Forgive me if I'm taking a stupid step into what is mostly unknown to me, as most of the previous posts are over my head. Isn't there a way to limit current without unnecessary heat loss, and just have that current limiter on the high capacity pack (not a BMS cut, but something just limiting the flow), since it won't care if the current is going to recharge the high power side or going to the motor? Our controllers have some kind of high efficiency current limiting, so why can't a battery pack? Then you just parallel the two packs after that current limiter.

John

 

[swbluto]

swbluto,

Forgive me if I'm taking a stupid step into what is mostly unknown to me, as most of the previous posts are over my head. Isn't there a way to limit current without unnecessary heat loss, and just have that current limiter on the high capacity pack (not a BMS cut, but something just limiting the flow), since it won't care if the current is going to recharge the high power side or going to the motor? Our controllers have some kind of high efficiency current limiting, so why can't a battery pack? Then you just parallel the two packs after that current limiter.

John

Yeah, the controller does current limiting by changing the "duty cycle" of its switching. I suspect something similar would give what we'd desire but I don't know the circuit of the controller and I think it "pulses" its output to the motor by chopping it up into three separate phases signals to the motor, and I'm looking for a device that'll draw a steady DC current from the battery and supply a steady DC current to the controller. Looking for a switching current regulator that'll draw a steady DC current and supply a steady DC current is the challenge, and it probably already exists in the "real world", but it doesn't seem many of us are power EE's to readily understand what the solution looks like. But ZapPat's link to the super-capacitor circuit looks extremely intriguing though I suspect it pulses the input / i.e., the super-capacitors energy to the battery(I haven't analyzed it yet) but I'm still looking / investigating / "trying to understand". It seems like one of the guys on the electronics forum had a solid idea of what the circuit would be like (A "double bridge" something or other), but I don't.

However, pulsing the input/batteries might not be a bad thing, but I suspect it might not be good for the battery's BMS. Any EE's care to comment? If pulsing the current from the capacity battery is possible, then it seems there's some ready-to-apply buck-circuits out there already.

Anyways, i don't have a lot of time now, so I've unfortunately haven't done much to advance my understanding of some of the circuits. I probably shouldn't comment until I've made some marked advances in understanding.

 

[Aerowhatt]

I definitely advocate both directions to be pursued, the easy peasy simple and the sometimes needingly complex. If you're unlikely to encounter situations where you're going to run your run your booster dry from simple paralleling, then you may never cause the capacity pack's BMS to trip and the battery's charge currents to be too high. Also, if the range of currents you want to draw is only doubled or so, simple paralleling may work fine but I'd definitely try to stay on the "safe side" by testing if you decide to do so. If the amount you want to increase the current is much less, like say from 20 amps to 30 amps(1.5 times as much), you probably wouldn't have much to fear assuming the voltages of both packs are nearly the same.

But for those that want as much as you can get from both packs as possible, or even near the limits of both(assuming the boost pack can give much more than the capacity pack), I heavily suspect some kind of electronic control is the only safe, practical economic way to do so. This is more important for those that "want it all", i.e., those who want 80+ amps on tap for RC motor set-ups or high-powered crystalyte 530x motors while not breaking the bank nor the range in the process.

And, I'll reverb ZapPat, development time really isn't much of an issue. My time is pretty much free at this point(what little I have of it), and once enough time is invested in finding an economically solid working design, then it'd be readily copyable/modifiable according to whomever's needs.

There is something to be said for the exercise of design and mental experiemnts. Hey, alot of great things have come from this aproach. It all depends on what your time frame goals are. I also agree that when it comes to expensive batteries and protecting one's health and property, safety considerations first, make a lot of good sense.

I'll have to part company with you on the pushing a battery (mixed chem or not) near it's limits for any application. It just doesn't pay! There are limits that the specs say are doable that just shouldn't be approached. Thats' one thing you learn much more quickly doing high performance electric RC airplane flying, or electric RC car racing.

Aerowhatt

 

[Aerowhatt]

swbluto,

Forgive me if I'm taking a stupid step into what is mostly unknown to me, as most of the previous posts are over my head. Isn't there a way to limit current without unnecessary heat loss, and just have that current limiter on the high capacity pack (not a BMS cut, but something just limiting the flow), since it won't care if the current is going to recharge the high power side or going to the motor? Our controllers have some kind of high efficiency current limiting, so why can't a battery pack? Then you just parallel the two packs after that current limiter.

John

This is a good and simple solution that I mentioned in passing earlier. Basically it's a motor controller with the right amp limit set at full throttle between the lower power output pack and the rest of the system. There is some loss in doing this and considerable expense in high power Fets etc. The losses are less than 5% for a decent design. Originally when setting up these hybrid chemisty systems I worried about cross charge rates between the batteries. What I found was that if they were sized (both voltage and capacity) to get along more or less OK during discharge they will easily remain within specs with one recharging the other at a red light. That only leaves the discharge side to be concerned about.

Aerowhatt

Aerowhatt

 

[John in CR]

Aerowhatt,

I'd think that the losses should even be a lot lower, since we're not talking about any conversion (eg DC to PWM'd AC), and we just want a restriction on the current. Isn't that what a shunt does?

John

 

[swbluto]

Aerowhatt,

I'd think that the losses should even be a lot lower, since we're not talking about any conversion (eg DC to PWM'd AC), and we just want a restriction on the current. Isn't that what a shunt does?

John

Actually, you are talking about a conversion. To "limit the current", you have to drop the battery's voltage - This is exactly what the current limiter in the controller does - it decreases the duty cycle which drops the voltage fed to the motor so that the motor "wants less current" and so the amount the battery supplies decreases(the conservation of power also comes into play with the switching(PWM), but let's go with that for now.). Going with ohm's law, V = IR, you see that I = V/R and so to decrease the current, you must decrease the voltage.

The other way to decrease the current is to increase resistance but that's know as linear regulation which basically means we're stilling dropping the voltage the same, but we're dropping it by burning it off instead of "switching it". Burning it off is very inefficient.



So, in essence, to limit the current, you have to convert the voltage - whether by burning it off or by more efficiently switching it.

Right now, the difficulty is finding a switching circuit that does what we want. Basically, it should draw a constant current from the battery and supplies a constant current to the controller, and many common switching step-down voltage converters pulses it - brief periods where it's drawing current and then not drawing current. It's basically switching the current "on" and "off" really fast, and it's usually the battery current that's being switched on and off really fast.

The "shunt" doesn't limit the current - the controller measures the current through the shunt-resistor by measuring the voltage across it, so it's a way for the controller to tell what the current is. The controller uses the current it calculates to decide what to do next. I.e., if the current is above the current limit, reduce the voltage going to the motor by decreasing the duty cycle(The percentage of time the switch is on and compared to being off.).

 

[swbluto]

swbluto,

Forgive me if I'm taking a stupid step into what is mostly unknown to me, as most of the previous posts are over my head. Isn't there a way to limit current without unnecessary heat loss, and just have that current limiter on the high capacity pack (not a BMS cut, but something just limiting the flow), since it won't care if the current is going to recharge the high power side or going to the motor? Our controllers have some kind of high efficiency current limiting, so why can't a battery pack? Then you just parallel the two packs after that current limiter.

John

This is a good and simple solution that I mentioned in passing earlier. Basically it's a motor controller with the right amp limit set at full throttle between the lower power output pack and the rest of the system. There is some loss in doing this and considerable expense in high power Fets etc. The losses are less than 5% for a decent design. Originally when setting up these hybrid chemisty systems I worried about cross charge rates between the batteries. What I found was that if they were sized (both voltage and capacity) to get along more or less OK during discharge they will easily remain within specs with one recharging the other at a red light. That only leaves the discharge side to be concerned about.

Aerowhatt

Aerowhatt

What kind of motor controller? A DC brushed motor controller? Does a DC brushed motor controller get fed a constant current from the battery and then outputs a constant DC current? I was under the impression a DC motor controller limits the current by engaging in switching, using PWM to limit the amount of time current is being sent to the motor, so I'm not sure how appropriate that'd be. Or, maybe it'd work... I don't know much about brushed motor controllers, honestly(Even though I used to have one, darn it.).

Please correct me if I'm wrong, but you are suggesting this "motor controller" would just limit the battery current being sent to the main controller that actually drives the motor, right?

 

[torrent99]

What kind of motor controller? A DC brushed motor controller? Does a DC brushed motor controller get fed a constant current from the battery and then outputs a constant DC current? I was under the impression a DC motor controller limits the current by engaging in switching, using PWM to limit the amount of time current is being sent to the motor, so I'm not sure how appropriate that'd be. Or, maybe it'd work... I don't know much about brushed motor controllers, honestly(Even though I used to have one, darn it.).

Please correct me if I'm wrong, but you are suggesting this "motor controller" would just limit the battery current being sent to the main controller that actually drives the motor, right?

A brushed controller? Sounds like exactly the right thing for a ready made current limiter. Probably over complex, but not that expensive! It can be made variable by changing the throttle input.

BTW Wouldn't ALL these methods pulse the battery? Only a linear (wasteful) regulator wouldn't...
This is apparently NOT good for battery life (see http://www.mpoweruk.com/life.htm "Load Sharing"), but since ALL controllers by definition do it anyway....

Also wouldn't we need to change the LVC on the primary controller?

 

[dogman dan]

Thanks guys. Now my idiot level electronics can understand how to proceed, and how to tell if it's a problem if I paralell these batteries. Theoreticaly they will be very similar capacities, and I don't need more amps or anything like that. I just want a bit more range at times and have two 12 ah sla's in the bike, and an ebikes ca 24v 8ah nicad pack on the way. And I'll get more cycles out of the sla's if I paralell them and ride a shallower discharge. I'll get another nicad pack soonish, but right now paralelling them, and then riding a less than full discharge may work ok. When I ride my sla bikes, I invariably slam em with too deep a discharge. Lithium would be nice, but I've learned the hard way not to ride brushed hubs in my climate as far as a Ping would take em.

 

[ZapPat]

A note about the use of PWM (or pulsing) to do the current regulation talked about here:

The input capacitors used on a PWM circuit are the ones providing the current pulses, not the battery at the input (so will have no effect on cycle life of the battery). These caps are chosen, just as in a controller, to smooth out any ripple the PWM circuit would otherwise subject the input to. For the buck switcher's output, an inductance + capacitor is needed to also smooth out the output waveform. A brushed controller would work as a voltage step-down (buck) regulator, but a fairly large inductor would be needed at the output since the switching frequency is usually quite low. I would tend towards making a custom PWM circuit instead of this though. Bucks circuits are not very complex, specially if you limit yourself to a uni-directionnal converter (non-synchronous, or only switching one FET, using a diode as a passive switch).

Personnally, I'm trying out AeroWhatt's simple path with a direct parallel connection of one LiPo to three NiMh cells (actually I'm using two to six - same thing). I charged them all up yesterday and the voltages match exactly, at least while fully charged.

Aero's suggestion of testing both cell/battery strings seperatly to see their discharge curves with varying loads is a very good idea if we want to figure out how the combo's going to react. I'll do this after I try them together though, since I'm impatient!

Maybe we should have some way of posting good combo results on line here somewhere, but we would have to decide on what figures would be the important ones to post. Like maybe cell count, chemistry, capacity, internal R, resulting current sharing ratio...?

Pat

 

[swbluto]

What kind of motor controller? A DC brushed motor controller? Does a DC brushed motor controller get fed a constant current from the battery and then outputs a constant DC current? I was under the impression a DC motor controller limits the current by engaging in switching, using PWM to limit the amount of time current is being sent to the motor, so I'm not sure how appropriate that'd be. Or, maybe it'd work... I don't know much about brushed motor controllers, honestly(Even though I used to have one, darn it.).

Please correct me if I'm wrong, but you are suggesting this "motor controller" would just limit the battery current being sent to the main controller that actually drives the motor, right?

A brushed controller? Sounds like exactly the right thing for a ready made current limiter. Probably over complex, but not that expensive! It can be made variable by changing the throttle input.

BTW Wouldn't ALL these methods pulse the battery? Only a linear (wasteful) regulator wouldn't...
This is apparently NOT good for battery life (see http://www.mpoweruk.com/life.htm "Load Sharing"), but since ALL controllers by definition do it anyway....

Also wouldn't we need to change the LVC on the primary controller?

For pulsed applications the peak load on the battery can be reduced by placing a large value capacitor in parallel with the battery. Energy for large instantaneous loads is supplied by the capacitor effectively reducing the duty cycle and stress on the battery. The capacitor recharges during the quiescent periods. Claims of a sixty percent increase in cycle life are made for this technique.

Good link! That makes a LOT of sense. I still like to err on the side of "safety" so I'll still simulate/test it, but attaching an appropriate capacitor in parallel with the battery sounds like it'd draw a relatively steady current from the battery while outputting a steady current to the controller while reducing the voltage at the same time using something like at http://en.wikipedia.org/wiki/Buck_converter to actively control the output voltage to stay under the current limit. If this is successful, I'll be sure to try different types of buck converters to see which one works "best"(highest efficiency, lowest cost, etc.).

Anyways, all switching converters have either a pulsed input or output BUT it is possible to put something that has a pulsed output behind something that has a pulsed input, like a boost and then buck converter(And so the first input is steady current and the output current is steady) or to use input/output filtering when needed(Effectively, the above paragraph is just adding an input filter to what would otherwise be a pulsed input buck converter). But I think just placing a capacitor on the battery before the "buck converter" sounds reasonably simple, effective and economic. The right values have to be chosen, of course. :wink:

And the diode can be replaced by a tightly controlled mosfet/switch to increase efficiency, as bearing has suggested.

 

[Aerowhatt]

Personnally, I'm trying out AeroWhatt's simple path with a direct parallel connection of one LiPo to three NiMh cells (actually I'm using two to six - same thing). I charged them all up yesterday and the voltages match exactly, at least while fully charged.

Aero's suggestion of testing both cell/battery strings seperatly to see their discharge curves with varying loads is a very good idea if we want to figure out how the combo's going to react. I'll do this after I try them together though, since I'm impatient!

Maybe we should have some way of posting good combo results on line here somewhere, but we would have to decide on what figures would be the important ones to post. Like maybe cell count, chemistry, capacity, internal R, resulting current sharing ratio...?

Pat

Looking forward to your results. One combination I haven't fooled with is NiMh and LiPo. From using both in airplanes before my guess is the LiPo will do more of the work at higher loads.

I think it's a great idea to publish results for everyone to use as guidelines. The Informations usfulness would be limited because of the wide variations in voltage sag between different battery brands at the same amp load. But still it would be a starting place.

As was stated earlier DC motor controllers dampen the ripple the battery will see pretty well. If you have ever looked inside one they are loaded with good size caps for just that purpose. For someone wanting amperage control on the low output battery. Someone who doesn't have the electronics expertise to build a simpler current throttle it makes a good alternative.

Aerowhatt

 

[Aerowhatt]

Just a note about resistance and paralleling two non matched batteries. If they are both conected to the load then having resistance in between one of them and the load isn't as bad as it seems. Intuatively adding resistance produces alot of heat and wastes power. This is very true in a single source system. But with dual sources the system just askes more of the second source of power rather than forcing the power through the resistance causing a large loss to heat.

Similar to the difference between adding resistors in series and adding them in parallel. The results are very different! For example on my motorcycle which has some pretty large peak amp demands. A direct paralleling of the LiFePO4 with the SLA's using the same guage and length wires would yield higher than desired peak loads on the LiFePO4. A simple solution is to use smaller gauge cable for the LiFePO4. So that at the controller (or junction of the two batteries) it's voltage drop is more than it otherwise would be. Intutition jumps to the conclusion that the wires would heat significantly and waste power. Why, because if you add resistance and have to pull the same amount of power you get a big voltage drop and an increase in amps resulting in hot cables. What intuition forgets is that there is another source of power (similar to parallel connected resistors)! So rather than FORCE the higher resistance line it pulls from the other one. Basically it doesn't take much additional voltage drop on the LiFePO4 to get the SLA's to step up a bit more and give a more ideal balance of current delivery during peaks loads. So if you have a dual battery system that is naturally pretty close to each performing where you want them too. Changing the harness resistance of either, or both batteries to fine tune it results in a very low loss.

In the case of the motorcycle going from a 6ga cable to 10ga cable on the LiFPO4 fine tuned the balance of load for each battery with only a 3 degree F increase in the harness for the LiFePO4. Arguably this is less loss than the losses associated in throttling that battery with electronics would be. With the added advantages of no additional cost or complexity.

Aerowhatt