Shimano steps - adding a second battery

Toiler

New Member
Region
Australia
Hi, New here. But others may be interested in an experiment I am running.

A while back (2018?) I bought two Avanti Inc. E bikes - Avanti is a New Zealand Brand common in Australia and the inc E (at that time) was equipped with Shimano Steps 6000 drive line, hydraulic disc brakes, Alfine 8 speed hub and Gates CDX belt drive. At the time was top spec and a good price.

So one bike has been sold (mine after 6,000km) and I kept the wife's (600km) for myself. But in this time both batteries basically would get to 25 to 30% State Of Charge (SOC) and then drop to nothing in a very short time/distance. Both batteries were replaced under warranty. The bike sold thus had effectively a new battery. The bike I have now - some time later - is showing the same symptoms but is well out of warranty now. The battery is a AUD900 thing to replace if it gets worse. Even worse - how long will they be available to buy.

So what to do about this?

So I thought - add a second battery. 36volt approx. ~10AH batteries are approx. 1/3 the price of the genuine battery. My thinking is that as long as the motor control electronics can communicate with the genuine battery (and the bar mount controller) then the "system" should still operate blissfully unaware of the second battery - even if the Shimano battery is basically dying - as long of the SOC is above the cutoff - the second battery can still be used to run the bike.

I will say upfront I am not trying to maintain the same SOC in both batteries. I have another motive here as hinted at above.

I used a common cathode Schottky diode array (MBR6045WT) summing both batteries into the motor controller (meaning batteries cannot discharge into each other) and the results are:
  1. It works. Putting a wattmeter in series with the new battery shows it IS contributing and from the bar controller display the rate of SOC reduction in the main battery is visibly slowed.
  2. Powering on the second battery will power on the eBike.
  3. The eBike will not power off via the bar mounted controller unless the second battery is switched off before doing this (I use a 30A circuit breaker for this).
Now for the more interesting bits.
  1. If the main battery is at say 50 to 60% SOC and the second battery is at 100% SOC, the SOC on the Shimano display does not change for a LONG time as the second battery is effectively running the bike. In my case this is the target - great for main battery longevity. The second battery is CHEAP - can be recharged often with no concern - thus reducing the re-charge cycles for the Shimano battery.
  2. The bike is more powerful - higher assist being provided. A watt meter in series with the second battery when the second battery has a higher SOC than the Shimano battery (supplying all the current) shows up to 400W peak power and up to 11amp peak being delivered to the motor. Average is lower eg 300 to 350Watt. This has implications I was not initially aware of:
  • The "250Watt" Shimano Steps 6000 motor seems to be more than 250W capable when not run using the stock battery. Hence the wattmeter is in play until I figure out what is going on.
  • The heatsink I designed for the Schottky diode array was sized as a compromise between 7amp continuous and it fitting inside the plastic motor casing - I may have to do some extra work here now higher currents thus higher heating is occuring.
  • There is now a risk of blowing the Shimano motor controller if these power levels are real. This is a WARNING for others that play in a similar manner. For me an annoyance only if it occurs - I have a spare (new) motor bought for USD100 some time back - I have already checked that it works on this bike.
So in summary, a successful experiment to date.

I intend to deliberately run the cheap second battery at a higher SOC (recharge often) than the Shimano battery (recharge much less often) thus this second battery carries most of the work load - thus lengthening the lifetime of the stock Shimano battery. If I want increased range, I will charge both up to 100% SOC.
This same "hack" is probably possible with other Shimano engine / battery combinations.

Some photos of the diode array fitment are attached
  • The heavy power cables fit between the rear of the motor and the Avanti frame quite neatly.
  • The heatsink is sitting on top of the light controller module. Yes it all fits - just!
  • The second battery is a bit ugly at present - the blue wattmeter will be removed once I am confident of what is going on and a smaller SOC meter will be mounted up front.
  • The circuit breaker sits below the second battery.
 

Attachments

  • StepsMotorDualBattery - 1.png
    StepsMotorDualBattery - 1.png
    538.9 KB · Views: 833
  • StepsMotorDualBattery - 2.png
    StepsMotorDualBattery - 2.png
    534 KB · Views: 731
  • StepsMotorDualBattery - 3.png
    StepsMotorDualBattery - 3.png
    559.6 KB · Views: 695
  • StepsMotorDualBattery - 4.png
    StepsMotorDualBattery - 4.png
    558.7 KB · Views: 775
  • StepsMotorDualBattery - 1 (1).png
    StepsMotorDualBattery - 1 (1).png
    653.7 KB · Views: 826
Hi, New here. But others may be interested in an experiment I am running.

A while back (2018?) I bought two Avanti Inc. E bikes - Avanti is a New Zealand Brand common in Australia and the inc E (at that time) was equipped with Shimano Steps 6000 drive line, hydraulic disc brakes, Alfine 8 speed hub and Gates CDX belt drive. At the time was top spec and a good price.

So one bike has been sold (mine after 6,000km) and I kept the wife's (600km) for myself. But in this time both batteries basically would get to 25 to 30% State Of Charge (SOC) and then drop to nothing in a very short time/distance. Both batteries were replaced under warranty. The bike sold thus had effectively a new battery. The bike I have now - some time later - is showing the same symptoms but is well out of warranty now. The battery is a AUD900 thing to replace if it gets worse. Even worse - how long will they be available to buy.

So what to do about this?

So I thought - add a second battery. 36volt approx. ~10AH batteries are approx. 1/3 the price of the genuine battery. My thinking is that as long as the motor control electronics can communicate with the genuine battery (and the bar mount controller) then the "system" should still operate blissfully unaware of the second battery - even if the Shimano battery is basically dying - as long of the SOC is above the cutoff - the second battery can still be used to run the bike.

I will say upfront I am not trying to maintain the same SOC in both batteries. I have another motive here as hinted at above.

I used a common cathode Schottky diode array (MBR6045WT) summing both batteries into the motor controller (meaning batteries cannot discharge into each other) and the results are:
  1. It works. Putting a wattmeter in series with the new battery shows it IS contributing and from the bar controller display the rate of SOC reduction in the main battery is visibly slowed.
  2. Powering on the second battery will power on the eBike.
  3. The eBike will not power off via the bar mounted controller unless the second battery is switched off before doing this (I use a 30A circuit breaker for this).
Now for the more interesting bits.
  1. If the main battery is at say 50 to 60% SOC and the second battery is at 100% SOC, the SOC on the Shimano display does not change for a LONG time as the second battery is effectively running the bike. In my case this is the target - great for main battery longevity. The second battery is CHEAP - can be recharged often with no concern - thus reducing the re-charge cycles for the Shimano battery.
  2. The bike is more powerful - higher assist being provided. A watt meter in series with the second battery when the second battery has a higher SOC than the Shimano battery (supplying all the current) shows up to 400W peak power and up to 11amp peak being delivered to the motor. Average is lower eg 300 to 350Watt. This has implications I was not initially aware of:
  • The "250Watt" Shimano Steps 6000 motor seems to be more than 250W capable when not run using the stock battery. Hence the wattmeter is in play until I figure out what is going on.
  • The heatsink I designed for the Schottky diode array was sized as a compromise between 7amp continuous and it fitting inside the plastic motor casing - I may have to do some extra work here now higher currents thus higher heating is occuring.
  • There is now a risk of blowing the Shimano motor controller if these power levels are real. This is a WARNING for others that play in a similar manner. For me an annoyance only if it occurs - I have a spare (new) motor bought for USD100 some time back - I have already checked that it works on this bike.
So in summary, a successful experiment to date.

I intend to deliberately run the cheap second battery at a higher SOC (recharge often) than the Shimano battery (recharge much less often) thus this second battery carries most of the work load - thus lengthening the lifetime of the stock Shimano battery. If I want increased range, I will charge both up to 100% SOC.
This same "hack" is probably possible with other Shimano engine / battery combinations.

Some photos of the diode array fitment are attached
  • The heavy power cables fit between the rear of the motor and the Avanti frame quite neatly.
  • The heatsink is sitting on top of the light controller module. Yes it all fits - just!
  • The second battery is a bit ugly at present - the blue wattmeter will be removed once I am confident of what is going on and a smaller SOC meter will be mounted up front.
  • The circuit breaker sits below the second battery.
I have gone a bit of a different route. I bought a 12 volt charger for Shimano from Powerbutler.de Turned out it can charge the bike while it is being used. It can charge at about 100 watt which means in ECO mode the state of charge is maintained.
This means I can carry a 12 volt battery with me and charge the battery that is in the bike.
The powerbutler does need to be vented so I am still looking for a good way to attach it to my bike together with the battery. Problably something like the Ortlieb E-trunk should work.

When using my Burley coho trailer I can have the 12 volt battery and the powerbutler in there and either charge the Shimano battery in the bike or choose to charge my second Shimano battery in there and change the bike's battery every now and then.
I can also have solar on top of the trailer.
 
My setup is still running ok. The cheap second battery is still able to run the bike with little to no discharge from the main (Shimano) battery until the cheap battery gets down to 40 to 50% SOC. Then the Shimano battery starts to contribute to the motor power. The cheap battery is however reaching 40 to 50% SOC faster now. ie it is wearing out a bit. But that was the point for me - have a cheap battery wear out and not the $$$ battery.
 
I'm having range anxiety
I cannot really say as I have never pushed it until both batteries are flat. The second (cheap) battery is 11AH as is the stock battery so in theory double the range is possible. The next cheap battery will be a 15AH triangle pack - will fit the frame better. But I could also carry a third battery if required - unplug the second - plug in the third - again leaving the stock battery alone.
 
Thanks so much for the update.

I use my bike for pizza deliveries and get around 35ks on average using the middle assist (norm) setting. I really need to double that in a cost effective way.

I think you are a genius, it’s like everything, it’s so simple but needs someone clever to think outside the box.

I am going to use your setup but I have very little knowledge on electrical. I have already ordered the diode and I am studying your photos to see what connects to what etc. would it be possible to attach a quick sketch of how I should wire everything?

Thanks once again
 
the basic premise that that The positive supply from the stock battery flows to the motor via a diode (anode 2) as does the second battery positive supply via anode 1.
The common point of both diode outputs (the common cathode) then goes to the motor.

DualDiode schematic.png


Then:
  1. Either battery can power the motor.
  2. No battery can discharge into the other battery.
  3. The battery with the highest state of charge (higher voltage) will deliver the most current.
  4. The battery delivering the current will droop in voltage. If it droops enough the other battery will start to deliver current.

Some annotated pictures attached following:

The second battery cables are run in the gap to the rear motor housing. Run the heaviest gauge of wire that will jam in here.
Dual diode second battery cables.png

Cut that alloy angle to fit in a snug manner on top of the lighting control box. I left some vertical parts but had to cut most off to allow the cable to reach the dual diode. Make the heatsink as large as you can. I designed for approx 7 amp to the motor but I found that the second battery will delivery way over this (400Watt peak) to the motor so ideally more heatsink than I have is better.
Fit the diode to the heatsink using a silicon heat sink pad. The heatsink must be isolated from the common cathode of the diode. The nature of the diode case should ensure this but the heatsink pad is an extra step in this regard plus needed for heat transfer.
The outer sheath that covers the cable from the stock battery to the motor will need to be removed.
This exposes a heavy red and black wire and some smaller gauge signalling wires (yellow and blue).
Dual diode +ve preparation.png

The large red wire in the factory loom needs to be cut and each cut end stripped ready for soldering.
Ditto for the second battery red wire.
Be neat - there is no room for loose cable :) oh and do not forget to place the heat shrink BEFORE soldering - then shrink
Dual diode +ve connections.png

The negative connection of the second battery needs to be spliced into the large negative wire that runs from the stock battery to the motor. There is no need to cut this wire.
Dual diode -ve preparation.png

Dual diode -ve connection.png

Add some insulation to protect the signal wires from the alloy heatsink
Dual diode - protect signal cables.png


The finished result should look similar to the above.
Note the cable routing.
There is NOT a lot of space.
Beware the motor cover plate screws. They are fine thread stainless threading into alloy - they WILL strip if you are not careful. I suggest laying the bike on it's side for the cover plate re-fitting to allow more accurate lining up the holes. Be very gentle with the screw driver. (yes you guessed it - I have stripped all mine - replaced with self tappers).
My space is reduced from stock even further as I have a speed hack fitted - the module is jammed between the motor and crank housing - but it all fits.

I used XT60 connectors for the second battery connections.
I would recommend using a 12volt automotive circuit breaker to connect / disconnect the second battery. I have found it difficult to get 48volt DC high rated current switches. (note battery pack is a nominal 36volt pack but when fully charged is more than 40 volt).

Per my original post. With the bike completely off, turning on the second battery will start up the bike including the Shimano systems. Turning off the second battery will NOT turn off the Shimano systems - the normal Shimano turn off method is used to do this.
 
Some more photos of the finished ensemble - the second battery is being charged hence the strange wire hanging out the front - the stock Shimano battery is stable at 69% SOC at this point = nice for longevity
The power switch is not ideal. It goes "pop" sometimes when I turn it on = not good.

DualBatteryEBike - 1.pngDualBatteryEBike - 2.pngDualBatteryEBike - 3.png
 
Absolutely great explanation! I can follow that quite easily. Do you have a link where I can buy you a beer?
 
Hi really interested in these setups, how is going after a while? Cheers
Are you refering to the Powerbutler setup I have?
It works great, however, charging under load can mean that a the bike thinks the internal battery's SOC is still good and then suddenly the % drop. Not really something to worry about.
Also, it is hard to get 100 watt or real power from the panels.
 
Are you refering to the Powerbutler setup I have?
It works great, however, charging under load can mean that a the bike thinks the internal battery's SOC is still good and then suddenly the % drop. Not really something to worry about.
Also, it is hard to get 100 watt or real power from the panels.
I did look at the power butler and found your post interesting but found it too expensive so I was looking at toilers experiment
 
It's true the powerbutler is not cheap. But it is a great piece of hardware that solves the problem of DC charging for shimano. Without wanting to sound arrogant, I think the time and parts needed for a solution like toiler proposes is close to the same costwise .
I can directly plug in my 18 volt solarpanel as long as it does >25 volt and run with it. Normally I connect it to the load port of an mppt controller and sometimes use a small buffer battery....
 
Last edited:
If I wanna add a second battery , couldn't I just put connectors on the wiring positive and negative wire outside the motor , and unplug the main battery , and plug in the extra battery in place of the main battery , So when my main battery runs low , I just unplug it and plug in the extra battery , Because it appears with your design that you've just made it more automatic so you don't have to plug in and unplug one battery , so can I just do that put new plugs connectors on the main wires and unplug one battery and plug in the other battery ,Yeah outside the motor case where the batteries run into the motor before they go into the motor or up by the battery and again just unplug the main battery and plug in the extra battery once the main battery runs low ,Yeah so I can put it cheaper battery and this wired in to plugs and unplug main battery put the cheaper battery plugged into the system select without having to modify it question , this is the bike I have 2016 felt versa
00j0j_axoty9WoGjpz_0CI0t2_600x450.jpg
 
I am of the belief that all parts of the Shimano system - motor/battery/handlbar controller - talk to each other over the eBus. Thus, I always assumed that removing the official battery from the system will not work - the system will shutdown. This is why in my solution the factory battery is always present as I did not want to take the risk that the system would end up in an irrecoverable and non functional state or at best a state requiring re-programming - the programming tools I do not have to do this.

From my testing above, the factory battery is doing more than dumbly supplying power to the motor. There is certainly current limiting occuring to keep the motor inside the 250watt power limit. The second battery (as I have fitted it) allows up to a measured 400W and this can be noticed while riding when the second battery is at a higher state of charge and thus running the motor in preference to the factory battery.
 
Yes I began to think of this after I posted , That unless the battery I'm using the extra battery , that I'm using has the right programming and yeah that the E bike will no what's going on, anyway buoght an extra battery for this bike but yeah your way is really great and thank you for your reply ,
 
Note that the switch to turn on/off the extra battery is goingn to be diffficult to find. ie it needs to be at least 48volt DC and 10AMP rated. Higher current is better. A 240AC (mains) switch is NOT going to be suitable - the DC current will cause internal arcing and will destroy it. In fact I used an automotive 30A circuit breaker rated to 48 volts.
 
Back