New product: redundant battery system with 3S to 10S with ideal diodes (PDB)

[Auvidea]

Hi,

many have asked me how two flight batteries may be connected in parallel to increase power and to improve reliability. Simply connecting two batteries in parallel is a tricky issue and should be handled with extreme care only by an experienced user.

The batteries need to be of similar kind and should be equally charged. If not there will be a current flowing from one battery into the other. As the current is just limited by the cell internal resistance of the battery pack, this can be pretty high. This may be not healthy for the battery and the user. A good chance to take an impressive Youtube video.

It would be far better to take the classical approach. Just use two diodes to merge the packs and to avoid that one pack charges the other. However high current diodes are bulky and hard to find. But the biggest problem is the fairly high voltage drop across the diode. At high currents this means very high power dissipation. But please have a look at the simple schematic with the 2 diodes. It explains the concept.

So what we really need is an ideal diode. No voltage drop at high current. Now this is impossible. But we can come close by using a high current MOSFET with some supervising logic. It checks the direction of current flowing. If the current wants to enter the pack, then the logic quickly turns off the MOSFET and blocks the current. So with this we have a diode with a resistance (Rds) of 1.4 mOhm typically. If 100A are sourced in total and both packs equally provide 50A, then 50x50x0.0014 = 3.5 watts are dissipated in each MOSFET. Normally this is the peak current. At 100A a 10000mAh pack is discharged in only 6 minutes. Possibly cooling needs to be provided.

Specification:

Size: 36.5 x 44 mm (without XT60 connectors)

Hole spacing: 30.5 x 30.5 mm (M3)

Max. current: the MOSFETs are speced at up to 240A - but the system will be less (we will have to see what the real limit is)

Resistance: typically 1.4mOhm each

Power in: 3S  to 10S

Connectors: XT60 or wires soldered directly to board

Pricing:

EUR 49.90 net (for customers outside the EU)

EUR 59.38 incl. 19% German VAT

Status:

Design is finished now. It will go to PCB production now. First prototypes in November 2015

Attention: the 3D bodies in the 3D models may not show the components used. They just show the physical size of the component.

If you have some ideas or recommendations please let me know. I may still be able to incorporate them.

Regards,

Jurgen

[iskess]

Let me know what you think of this possible use.
I want to utilize Li-ion batteries in my Skywalker to take advantage of the higher energy density. The problem is that they don't provide an adequate power burst for a safe aborted landing when in the last 30% of the pack capacity.
The thought is to add a small high C Lipo to pick up the load. If I use a 4S li-ion and mix it with a 3S Lipo, it should work out nicely. The 3S Lipo is resting unused at 12.6v until the 4S Li-ion pack reaches 3.15v/cell. So they start sharing the load at about the last 20% of the Li-ion's capacity . The Li-ion can be safely discharged to 2.75v, but I think 3.0v is better for longevity. So on approach, the little Lipo will be at 12v with 80% of its capacity left. My Skywalker uses about 500w at max throttle, so that's around 40A if the Lipo took all the load. A little 1000mAh battery rated above 40C should work in theory, and not add much weight (100g) in comparison to the huge Li-ion pack savings.
At a climb rate of 5 m/s, I only need a 10s burst of power to make it to a safe altitude of 50m.

I purchased some high current diodes to experiment with, but haven't had the time so far. This MOSFET project could be perfect for such an application.

Thoughts?

[Jesus Alvarez]

Iskess has been reading my mind!

What we want/need is for taking off. We have GEB Liion cells and would like to use them in cruise while using a small "punch" lipo for take off and high AMPS maneuver.

That kind of lipo-liion mixing electronics would be just great.

[Auvidea]

Hi Iskess,

 you have a very interesting and creative idea. What you describe, make a lot of sense. I do not see anything wrong with it. Diodes will have the problem of a very significant voltage drop across them. My design with the ideal diodes should work in this application. You can connect 2 batteries which have a different cell count. And the lower cell count pack will only kick it, when the voltages match. So it should work.

At 40A max my design will give you a lot of headroom. The thermal power dissipation at one pack with 40A will be 40 x 40 x 0.0014 = 2.2 watts. For short uses this should be ok without specie cooing.

Regards,

Jurgen

[Philip Rowse]

This is a really good application of these Ideal diode units.  the mix of Li-Ion and Li-Po makes sense...

◤PHILIP ROWSE

    LEAD SYSTEMS ENGINEER

    3D Robotics Australia

    +61 (0)427326366

    website | facebook | instagram

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Federico Mari]

Hi Jurgen,

I think this is a great Idea! I did something similar few moths ago to use very low C rating battery for cruising. 

Make sure that your system has an interface that can be controlled form the AP. CAN bus will be the best. 

Without a proper communication with the AP this product will be useless... a failsafe device that doesn't tell you when you are in failsafe mode is useless. 

Also I will suggest to make this PCB potted in heat conductive encapsulation epoxy. Will help to dissipate heat, insulate the PCB and make more robust the XT60 connectors that otherwise will be the weak point of this product. 

Regards,

Federico 

[jdennings]

> Max. current: the MOSFETs are speced at up to 240A - but the system will be less (we will have to see what the real limit is)

With  same XT60 connectors on input and output you are limited at somewhere around safe max 60amps. For many planes this could be fine but for multirotors this is quite low,  the  two batteries capacities would be limited in discharging only 30 amps max, for 20c rated packs that's 1500mah each for a total of 3000mah, not much. Maybe an option for EC5 connector rated at  safe 120amps on the output?

At 240amps max seems like the MOSFETs are going to be plenty current wise, but the size of your board trace on the output is going to be a big factor in reliability, would  they be able to withstand twice the  60 amp rated input, for instance,  and for how long?

Also, any voltage limitation?

[Auvidea]

Thank you for your feedback and your ideas.

I actually have a second design in the works, which addresses your concerns. 

1. higher current: here there are 4 holes each for input and output. They are 3.4mm in diameter. You can either solder in a wire or solder in a 3.5 mm gold connector. These are rated at 80 or 90 A each. The current sensors are rated at 150A or 200A each. The MOSFET is the same with 240A.

2. each battery input has its ideal diode and its voltage and current monitoring. So there are 2 six pin connectors (Pixhawk style).

3. there are 2 power converters with 5.3V out (up to 3.5A) plus one 12V power converter

4. last there is a CAN interface to monitor all these parameters (including temperature)

The design is optimized for high current. Positive voltages on top and GND on the bottom. However things could be improved further by applying more copper. 35um is standard. Up to 210um seems to be possible.

The design is 80% done. If you have any ideas for improvement, please let me know.

Regards,

Jurgen

[Auvidea]

I would like to add some more details on the 38174 dual PDB:

Key features:

- 2 battery inputs

- each with current (150A) and voltage monitoring

- ideal diode to OR batteries together - no reverse current into batteries

- very low resistance in „diode“ (typ. 1.4mOhm)

- two 5.3V 3.5A power converters

- one 12V power converter

- CAN bus interface to monitor currents and voltages

- 2 legacy ports (analog out) for current, voltage and 5V)

Flexible power connection with multiple 3.5mm holes:

1. power cable may be directly soldered in

2. solder in 3.5mm gold connectors to connect to PWD board (mounted in top or bottom)

3. hex standoff (fixed with screws) to connect to PWD board - this needs to be tested first

Specification:

- size: 65 x 65 mm

- mount hole spacing: 45 x 45 mm

- current measurement with low resistance (0.1mOhm) Hall sensor

- 2 ideal diodes (up to 240A)

[john...@gmail.com]

Great little project. The principle is well known and works. You will find something similar in most balanced charger designs etc.

My only concern is the PCB layout. With numbers like 240A mentioned, I think you will need some bus bars and high current vias.

[Akshat Deshpande]

From what I understand , this product will help us while using two lipos with similar voltages in parallel and if the voltage of one pack is more than the other it would avoid the charging of the low voltage pack by the high voltage pack but at the same time most of the load current will be supplied by the higher voltage pack and the low voltage pack would just sit there not contributing much to the load till the time that the voltage of the other battery does not become equal or lower.

When using two lipos of the same voltage, capacity and discharge the load will be shared equally till such a time that the voltage difference between the two packs does not increase by much, if the voltage difference increases beyond a point then it would suggest a faulty pack, but using two lipos of significantly different voltages ex:(3S v/s 4S) this would not be of much use as the 4s pack will continually try to charge the 3s pack.  While this device will stop that from happening but still most of the load will be shared by the 4s pack  while the 3s pack would continue to sit there till the time that the voltage of the 4s pack does not become lower than the 3s pack at which point your 4s pack will either be on fire or will become a balloon which is ready to pop its nasty contents out.

The same thing applies to using two packs having similar voltages but different capacities and discharge rates(lipo v/s Lion), the battery with a higher discharge capacity will maintain its voltage better than the one with a lower discharge rate and both the packs will keep exchanging the load till the lower capacity pack runs out of juice and the other pack can continue powering the load.

I don't see anyway this proposed product could be used in a scenario other than using two packs having the same voltages , capacity and discharge rates.
  
 

How can we solve this problem            

[Akshat Deshpande]

This problem can be solved by having calibrated voltage and current sensors connected to both the battery channels and having a small micro which can switch power between the two battery channels smoothly when certain conditions are met, we can have this micro communicate with the AP via MAVLINK or serial (MAVLINK preferred) using which we can configure the voltage capacity and current for each of the two channels.

So when using  packs of same voltage range but different capacities and discharge rates, we can ask the micro to switch to the battery with a higher discharge rate when we are consuming more than X-amps from the primary high energy density pack this will ensure reliable operation. Also when the primary pack runs out of juice we could have the other pack kick in so that we continue to power the load.

Usually for this setup to make sense keeping weight into consideration you would have a high capacity low discharge rate  primary pack and a low capacity high discharge rate secondary pack both with the same voltage range of operation and when the primary pack runs out we can power the setup for a very short amount of time using the secondary pack given the lower capacity and trigger a battery failsafe  

but if flying aggressively causes the secondary battery to run out before the primary battery we would have to trigger battery failsafe when the primary battery reaches a particular voltage level.
overall configuring two packs having different capacities and discharge rates for exceptions and failsafe will be very tricky.

 
    

On Monday, October 5, 2015 at 9:45:47 PM UTC+5:30, Auvidea wrote:

[Iam Bouret]

I think the point you are missing is that the Li-ion pack is safe down to 2.5v/cell, it will not catch fire as you say. For its health I would not plan on bringing it below about 3v/cell unless I needed to perform an unexpected aborted landing. So in normal ops, both packs would begin the landing approach at 12v, the 4S would be at 3v/cell and the 3S would be at 4v/cell. If I perform a missed approach, the total voltage would be planned to be down to 11v. That's 2.75v/cell on the 4s Li-ion and 3.66v/cell on the 3s Lipo. Everyone is happy, nobody catches fire. 

I think you misunderstand how the characteristics of a Li-ion differ from a Lipo. You mentioned that the Li-ion would "balloon", which tells me you must be thinking of a Lipo. 

Iam

--
You received this message because you are subscribed to a topic in the Google Groups "drones-discuss" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/drones-discuss/K-VmlLjn7YQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to drones-discus...@googlegroups.com.

[Akshat Deshpande]

I just used that as a case when two LIPO packs having different voltage levels would be used (3s Vs 4s) no one suggested this but I still went ahead and considered this case ,I was thinking out loud ☺.
but for your use case I have elaborated that in the last paragraph, for packs having similar voltage ranges but different capacities and discharge rates. Here I did not specify the chemistry because they could be anything,

in your case it is Lipo-Lion but in some other case it might be something else.  

[Jesus Alvarez]

For the application that Iskess and I are thinking the thing is to prioritize based on total AMPS demanded.

So there should be some kind of supervisor logic running and measuring both packs voltage and AMPS. So if AMPS_DEMANDED < THRESHOLD_LI_ION then LI_ION is active and LI_PO is disconnected. Otherwise LI_PO Is ON and LI_ION is OFF.

It needs rethinking but this would allow to have the best of both worlds and to create safe battery packs with higher total energy density.

Hope AUVIDEA can do it!

[pritam....@gmail.com]

Iam,

I just wanted to clarify one thing as I plan to use LiIon soon for a pet project of mine. As you have already used it and this may be of interest to others as well. When you say safe till 2.5V or good practice to 3V. Do you mean on load or off load. Because batteries jump in voltage as soon as you remove load. I just bench tested a NCR18650B. I switched off the load as soon as it reached 2.51V under 1C load. It immediately jumped to 3.1V. So what will be safe voltage under load. 2.5v or 3v.

On Thu, Oct 8, 2015 at 10:49 PM, Iam Bouret <iamb...@gmail.com> wrote:

--

You received this message because you are subscribed to the Google Groups "drones-discuss" group.

To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Regards,

pritam

[Iam Bouret]

I will try to never discharge my Li-ion batteries below 3v to lengthen their life. It doesn't matter how you get there. If you put a very large load on the batteries, they will hit 3v sooner, then rebound. The bigger the load, the bigger the rebound. This means that as I approach 3v, I will need to reduce the current load to not fall below that number. This is difficult because I actually need a higher current to offset the low voltage to maintain the same power. This is a troublesome place to be, and the reason I am seeking a hybrid chemistry solution. 

So to sum up my answer to you question, yes the low voltage limits are under load. However, as the amp draw is decreased and the battery starts to rebound, you can continue to use more battery capacity at a lower rate. 

Iam

[john...@gmail.com]

As said, low voltage limits are always 'under load'. Or to put it in other words. The battery cells should never ever under any circumstance be below the critical limit (3v safe, 2.7v critical). The since the cells will be at their lowest during hard usage, then that is what counts. This is also why pretty much all speed regulators have a built in low voltage protection system, to stop it from doing damage to the cells.

[Andy Little]

This all sounds great... but I'm struggling to figure what it has to do with this newsgroup ?

regards

Andy 

[iskess]

This thread stopped cold after the last post pointing out that it doesn't belong on DronesDiscuss. It is however an important topic for many of us, can you please create a new thread on DIYDrones or RCGROUPS so we can continue this?

Jurgen, I would love to test mixed chemistry batteries with your board if you want to send me one.

[Fourfingers]

Just in case anyone is interested, Jurgen opened a thead at http://www.rcgroups.com/forums/showthread.php?t=2516107#post33130101