Why Fast-Charging Rocks

[jbwheel]

2 hours ago, Keith said:

@jbwheel, you are misunderstanding me.

You are correct, if the cells IR is better then that the voltage drop will be less, but it will still be there.

Yes I did. So my guess is first try to find the reference of the cell used in your wheel then do the maths depending on the current you want to use (because 0,16V is not really significant but 0,4V is more problematic). What about measuring voltage during the charge around 65V, interrupt it then measure the pack voltage right after using the same voltmeter. Will the difference between the two measures be the number we are looking for ? then it would also allow to make an approximation of the internal resistance ?

[Keith]

29 minutes ago, jbwheel said:

What about measuring voltage during the charge around 65V, interrupt it then measure the pack voltage right after using the same voltmeter. Will the difference between the two measures be the number we are looking for ? then it would also allow to make an approximation of the internal resistance ?

Yes, good idea, that would work, Pack voltage would not need to be measured immediately, letting it settle for 10 minutes or so would be better to let the cells cool etc. I can't remember if Charge Doctor can switch off at a set voltage or only a set current but, if it is possible, it would be ideal setting it to (say) cut charge at (say) 66V and then measuring to see if the settled voltage has reached 65.6v (for 4.1V/cell) Adjusting the voltage and switching on again until the pack does reach the desired level. It would only need doing very occasionally as once the right cut off voltage has been identified Charge Doctor could be left set to that value for future charges. No need to consider internal resistance at all if done this way.

An occasional slow and full charge to ensure the cells a remain in balance would not do any harm.

[esaj]

Maybe just for determining the internal resistance, it could be done the other way around: measure pack voltage before starting to charge, then start charging, measuring accurately the charging voltage & current, and calculate the internal resistance from the difference of charging voltage and the known starting voltage of the pack, and the charging current? Of course it probably changes a bit as the cells heat up during charging... But if done at the start of the charging, the cells wouldn't have yet heated up much?

The thing about (or problem with) fast charging (ie. high current) seems to be that the power dissipation of the pack becomes very high with large currents (basically there's a voltage drop of the FULL charging voltage over the pack)? So for example 60V charging voltage at 5A current = 300W of power... But I don't know how much of that wattage turns into heat and how much is stored in the pack. Guess most of it gets stored and only part of it turns into heat? For the same reason, paralleling packs of different voltages is risky.

 

[jbwheel]

1 hour ago, Keith said:

Yes, good idea, that would work, Pack voltage would not need to be measured immediately, letting it settle for 10 minutes or so would be better to let the cells cool etc.

But then that would be resting voltage, which is lower anyway. Something is still not fitting : on Jason curve you can't see any voltage drop despite a sharp current drop from 5 to 0,5A. Following your theory that would be a .36V drop ? 

edit : didn't see that Jason's is theoretical ... but not the one below and still no voltage drop.

BTW, this battery capacity is close to 825Wh, just not completely empty at the beginning

 

[Jason McNeil]

Got some charging data today with the 5A charger & the 840Wh battery pack—I'll be using this as a baseline when the new dual-voltage (64.8v/67.2v) chargers arrive in a couple weeks for comparative testing.

I guess the big question is that when partial charging to 4.05v/cell (or whatever), is the current tapering off more of a drop than a curve, as in a conventional charge cycle? Is the CV trickle curve a function of the cell's chemistry on attaining the 4.2v? My assumption was [is] that a cell can reach the cut-off current (100mA) more quickly at the 4.05v, because more of the Lithium ions are available from the anode to accept free electrons—this remains to be put to the test!

This page has some interesting info about the standard charge cycle  http://www.mpoweruk.com/chargers.htm
"In order to maintain the specified constant current charging rate, the charging voltage must increase in unison with the cell voltage to overcome the back EMF of the cell as it charges up. This occurs quite rapidly during the constant current mode until the cell upper voltage limit of the cell is reached, after which point the charging voltage is maintained at that level, known as the float level, during the constant voltage mode. During this constant voltage period, the current decreases to a trickle charge as the charge approaches completion. Cut off occurs when a predetermined minimum current point, which indicates a full charge, has been reached"

On 4/20/2016 at 10:59 AM, Keith said:

What I have said (and put some numbers against it to demonstrate) is that if you wish to charge quickly AND to a lower voltage to protect the cells then you need to take internal resistance into account. If you stop a (say) 5Amp charge when the voltage reaches 4.1V per cell (65.6v) then, actually you are nowhere near 4.1V per cell or 86% charge at that point.

On 4/20/2016 at 3:07 PM, esaj said:

The thing about (or problem with) fast charging (ie. high current) seems to be that the power dissipation of the pack becomes very high with large currents (basically there's a voltage drop of the FULL charging voltage over the pack)?

But are we aren't we missing sight of the fact that it's only the charge/cell that's important with fast-charging? Even with a 5A charger on a 16s4p pack, the Amps that are delivered to an individual cell is only 1.25A, much less than the 2A charging that is found in 95% of Electric Unicycles, because they are predominately 16s1p. Newish cells are typically rated to 4A charging, which is  +3x faster than what we're doing here! This being the case, I think the internal resistance is going to be of negligible account for bulk battery fast charging. 

 

[Keith]

@Jason McNeil, as ever, I may be wrong but, a battery isn't fully charged at 4.2V; it would accept, albeit with progressive damage, up to 4.35V so I'm not at all convinced the CV curve will be any different at 4.00V than it is at 4.2V.

More to the point, we have established that a voltage, some value above the current EMF of the pack, is required to "push" a constant current of whatever value is required into the pack. Obviously the higher that current, the higher the potential difference required in order to 'push' that increased current.

I would maintain that that potential difference, and indeed the current curve once the charger hits the CV stage are a direct result of the internal resistance of the pack. Indeed, if the internal resistance of the pack was zero I do not believe there would be a CV curve, the current would drop to zero as soon as the charge voltage is reached.

[Jason McNeil]

34 minutes ago, Keith said:

would maintain that that potential difference, and indeed the current curve once the charger hits the CV stage are a direct result of the internal resistance of the pack.

This is good, because I can test it experimentally using two different types of cells, like the HG2s (18mOhms vs. something with at least double these). 

 

[jbwheel]

 

14 hours ago, Jason McNeil said:

But are we aren't we missing sight of the fact that it's only the charge/cell that's important with fast-charging? Even with a 5A charger on a 16s4p pack, the Amps that are delivered to an individual cell is only 1.25A, much less than the 2A charging that is found in 95% of Electric Unicycles, because they are predominately 16s1p. Newish cells are typically rated to 4A charging, which is  +3x faster than what we're doing here! This being the case, I think the internal resistance is going to be of negligible account for bulk battery fast charging.

For Panasonic NCR18650PF used by GW, max charging current is 0,5C (1.4A) for a daily use. Wheels manufacturers (KS, GW) use even lower rate (half of it for GW 340Wh but ridiculously low for a 850Wh battery). There is also no easy way to dissipate the heat from a big pack which is anageing and risk factor, especially if you charge just after a wheel run...

Below is charging of NCR18650PF at 1A (similar to 5A on a MSuper 850). Cutting at 4,05V (CC phase) will give you 80% capacity. It looks a bit different from your curve... because your measure start from 60,5V which is not an empty battery.

@keith :   Higher charge voltages boost capacity but lowers cycle life and compromises safety. Source: Choi et al. (2002)

 

 

 

[Jason McNeil]

On 4/24/2016 at 5:31 AM, jbwheel said:

Wheels manufacturers (KS, GW) use even lower rate (half of it for GW 340Wh but ridiculously low for a 850Wh battery).

This is not on account of any technical consideration, but the fact that a 2A charge can be bought in volume for less than $20, whereas a decent quality fast-charger is at least three times more than this.

On 4/24/2016 at 5:31 AM, jbwheel said:

There is also no easy way to dissipate the heat from a big pack which is anageing and risk factor, especially if you charge just after a wheel run...

In a large (64 cell) pack, even maintaining the maximum cruising speed of 30kph, will only demand 5A across the entire pack, on an individual cell basis this is only 1.25A. As can be seen in the below graph (for a different cell but it's roughly equivalent with something like the MJ1), this is predicted to produce only a couple °C above the ambient temperature, but you're probably correct that the battery best practice should be allow your wheel 10-20 minutes after a run to cool down a bit before charging.

[jbwheel]

5 hours ago, Jason McNeil said:

As can be seen in the below graph (for a different cell but it's roughly equivalent with something like the MJ1), this is predicted to produce only a couple °C above the ambient temperature, but you're probably correct that the battery best practice should be allow your wheel 10-20 minutes after a run to cool down a bit before charging.

These measures might be skin temperature of one cell on a bench. You need to take it in the middle of a pack and taking into account that it is in an enclosed space where electronics, wires and maybe even the motor produce heat.

[Jason McNeil]

It's a valid point, but since the cells are in direct contact with each other & pressed next to the side-panel, this heat should dissipate into your legs while riding. Anyone have a thermal camera that can be used to capture this? 

[esaj]

14 hours ago, Jason McNeil said:

It's a valid point, but since the cells are in direct contact with each other & pressed next to the side-panel, this heat should dissipate into your legs while riding. Anyone have a thermal camera that can be used to capture this? 

Eryk88 posted some images in the BMS-modding thread:

You could ask if he has anymore images or information about the battery temperatures. A decent FLIR-camera ("Forward-looking Infrared" or something like that) is probably something like a thousand or two, so I doubt many people have them   Maybe someone works somewhere where (s)he could borrow one...

Edit: A quick Google-search revealed that FLIR-cameras (apparently it's also a brand name?) cost between 1000€ (+VAT) for a basic handheld with something like 120 x 90 pixel resolution up to over 20000€ for high-resolution professional cameras   So yeah, I doubt anyone owns one "just for fun"

[dmethvin]

17 hours ago, Jason McNeil said:

It's a valid point, but since the cells are in direct contact with each other & pressed next to the side-panel, this heat should dissipate into your legs while riding. Anyone have a thermal camera that can be used to capture this? 

I had an "energy audit" done by a local heating and air conditioning company a few years back, and they used a thermal camera. It might be possible to contact a local company and see if you could use their camera. That's assuming it is at their office and not rolling around on a truck every day. I doubt they would let it out of their sight but they would probably be interested enough in the EUC to let you borrow it for some quick photos.

BTW, spring and fall are a good time to ask because they rarely do energy audits this time of year. They want a large difference between inside and outside temperature because it makes leaks and lack of insulation easier to see in the thermal image.

[jbwheel]

20 hours ago, Jason McNeil said:

It's a valid point, but since the cells are in direct contact with each other & pressed next to the side-panel, this heat should dissipate into your legs while riding. Anyone have a thermal camera that can be used to capture this? 

 

2 hours ago, dmethvin said:

BTW, spring and fall are a good time to ask because they rarely do energy audits this time of year. They want a large difference between inside and outside temperature because it makes leaks and lack of insulation easier to see in the thermal image.

A thermal camera wont see trough your wheel case. In my MCM4 680Wh, the mainboard and the battery share the same volume of air, there are 2 layers of batteries in one pack. The battery is hold with a plastic holder and then there is the case... on which I put some self adhesive foam.

What about a 10$ usb temperature logger ? You would still need to chose the right place to put the probe.

[Hunka Hunka Burning Love]

Here's some options:

http://www.flir.ca/flirone/display/?id=69324

http://www.diy-thermocam.net/store/

 

 

[jbwheel]

On 04/25/2016 at 9:02 PM, Jason McNeil said:

It's a valid point, but since the cells are in direct contact with each other & pressed next to the side-panel, this heat should dissipate into your legs while riding. Anyone have a thermal camera that can be used to capture this? 

This morning after I charged my MCM4 680 on 1,75A (did not ride it before), I noticed that the temp given by the GW app was  28°C after 1h30 of charging  (ambiant around 19°C). I suppose this is not battery temp but mainboard...