Feature request: Travel charger

[OliverH]

As I sometimes went out for a weekend trip with the MSuper I need to carry all our chargers with me. So the back pack is nearly full (4 power chargers if I went on tour with the family and a power bar to connect all chargers to one power wall plug)  with stuff which could be smaller. In the forum here we saw photos from inside power supplies. It looks like that the power supply could be smaller by 25-50%. Not sure if this is true for Gotway also but it's worth to think about to make them smaller.

[dmethvin]

For all the chargers I've seen, it seems like there either needs to be a large heatsink or a smaller heatsink plus room for airflow and a fan. Either way they end up being about the same size. The big 5-amp charger I have uses an aluminum exterior which can act as a heatsink but you have to be more careful about the case being energized.

Thinking about it a completely different way, what if you used a big 5-amp or higher charger with 4 cables coming out of it? Wouldn't that work? You wouldn't need to carry the power strip and there would be a LOT fewer power cables.

[sbouju]

This low price 2A fanless charger is significantly smaller than the original 2A supplied with a MSuper2:

http://fr.aliexpress.com/item/2A-67-2V-charger-for-16S-lithium-battery-pack-67-2V-Electric-unicycle-charger-recharger-aviation/32475704873.html

I sometimes use two of them for charging my MSuper2 twice quicker...

 

[SlowMo]

46 minutes ago, sbouju said:

This low price 2A fanless charger is significantly smaller than the original 2A supplied with a MSuper2:

http://fr.aliexpress.com/item/2A-67-2V-charger-for-16S-lithium-battery-pack-67-2V-Electric-unicycle-charger-recharger-aviation/32475704873.html

I sometimes use two of them for charging my MSuper2 twice quicker...

 

I also have this one and haven't tried it yet. I am afraid it might not do it's job since it is much smaller and is like a baby brother of my gotway charger. 

[OliverH]

1 hour ago, sbouju said:

This low price 2A fanless charger is significantly smaller than the original 2A supplied with a MSuper2:

http://fr.aliexpress.com/item/2A-67-2V-charger-for-16S-lithium-battery-pack-67-2V-Electric-unicycle-charger-recharger-aviation/32475704873.html

I sometimes use two of them for charging my MSuper2 twice quicker...

 

Do you've a photo original Gotway and this beside? I like to see the difference in size.

1 hour ago, dmethvin said:

For all the chargers I've seen, it seems like there either needs to be a large heatsink or a smaller heatsink plus room for airflow and a fan. Either way they end up being about the same size. The big 5-amp charger I have uses an aluminum exterior which can act as a heatsink but you have to be more careful about the case being energized.

Thinking about it a completely different way, what if you used a big 5-amp or higher charger with 4 cables coming out of it? Wouldn't that work? You wouldn't need to carry the power strip and there would be a LOT fewer power cables.

Not sure how this should work. How can one charger detect which EUC has finished charging? Our EUC have different battery packs or we will end after a day with different battery levels on the MSuper and the KS16 because of weight differences.

But one multi charger would be nice.

18 minutes ago, SlowMo said:

I also have this one and haven't tried it yet. I am afraid it might not do it's job since it is much smaller and is like a baby brother of my gotway charger. 

That sounds promising. Need to order two 

[sbouju]

2 hours ago, OliverH said:

Do you've a photo original Gotway and this beside? I like to see the difference in size.

Sure!

The original is upper, on the first picture, and on the second picture, you can see the two smallers from aliexpress on the left, tied together and spaced with a short piece of an aluminium ladder foot, and that I use with a ChargeDoctor2 with two input connectors.

[dmethvin]

10 hours ago, OliverH said:

Not sure how this should work. How can one charger detect which EUC has finished charging? Our EUC have different battery packs or we will end after a day with different battery levels on the MSuper and the KS16 because of weight differences.

But one multi charger would be nice.

If the output to each EUC was protected by a diode so the power can't reverse between packs, I would think the charger should produce a constant 67.2v at 5a that gets consumed by all 4 packs. When one of the EUCs starts to reach full charge, the current will go down for that pack and the others will use more of the 5a until they are charged as well. Finally they will all be charged and the current will go down to zero. Or would it not work that way?

[esaj]

8 hours ago, dmethvin said:

If the output to each EUC was protected by a diode so the power can't reverse between packs, I would think the charger should produce a constant 67.2v at 5a that gets consumed by all 4 packs. When one of the EUCs starts to reach full charge, the current will go down for that pack and the others will use more of the 5a until they are charged as well. Finally they will all be charged and the current will go down to zero. Or would it not work that way?

I'll have to do some calculations & simulations to check this after work, but I suspect that the diodes might get "run for their money", if the voltages among the wheels are wildly different...

[sbouju]

Hobby16 has worked on a special ChargeDoctor2 with 2 input connectors, but with 2 output cables too... May be it is the same problematic, somewhere...?

https://www.espritroue.fr/topic/356-charge-doctor-v2/?do=findComment&comment=8647

Edit: the initial purpose of this special CD2 was only to have one usable on a Gotway *OR* a Ninebot!

And the question came soon (but only!) about its usability with the 2 chargers at the same time for only one wheel, just for an eventual faster charging.

[esaj]

17 hours ago, dmethvin said:

If the output to each EUC was protected by a diode so the power can't reverse between packs, I would think the charger should produce a constant 67.2v at 5a that gets consumed by all 4 packs. When one of the EUCs starts to reach full charge, the current will go down for that pack and the others will use more of the 5a until they are charged as well. Finally they will all be charged and the current will go down to zero. Or would it not work that way?

The usual disclaimer: I'm not a professional or expert in electricity or electronics yadayada, so this might be completely wrong...  But, I had a suspicion that differing voltages of the batteries could pose a problem and the diodes/packs might take disproportionate amounts of the current each. Here's the testing circuit:

I know I could have used current sources, but instead opted to try and see if I could create a "high-side" constant current source with a p-channel mosfet. The current measuring occurs at the low-side over a 1-ohm resistor between packs & ground (but could be on the high-side too, with measurements done with something like a differential- or instrumentation-amplifier circuit). The constant current source is set for 5A (5V voltage drop over the R_measure). Anyway, the "real beef" are of course the diodes (D1-D4) and batteries (BAT1-BAT4) connected in parallel with the charger output. I've deliberately set all the batteries to different voltage values here to see the effect:

The upper measurement pane shows the voltages from points Bat1_v - Bat4_v & V_before_diodes. The feedback-loop to the mosfet gate has caused the mosfet to drop the voltage before the diodes to around 60.65V to keep the current at 5A. The MBR20100CT -power-schottkys (which probably wouldn't be the best choice, because their reverse breakdown voltage is too low if the charger was connected in reverse polarity ) drop the voltage a bit further before the batteries (where they conduct). And where do they conduct in the above scenario? The lower measurement pane shows that the full current runs only through battery 1, with the rest sitting pretty tightly at 0A (actually, there's a few micro amps of leakage, but that doesn't matter much). If that was for example a single-pack battery of 132Wh (2200mAh cells), the charge rate would be above 2C, and if it's cheap chinese cells, that's probably not a good thing. Not to mention that the connectors & wires inside the wheel might not be rated for that high current.

If you look at the voltages above, you'll notice that for example Bat3_v sits at 72.2V. That's because the battery is set at 67.2V on the simulation, but its "ground" (negative side) sits against the current measuring resistor, which has a voltage drop of 5V, so the voltages sum up (67.2V + 5V = 72.2V). So just keep in mind that the voltages at the battery positive sides read 5V higher than the internal voltage.

So, to simulate a more "meaningful" scenario, where the battery voltages actually go up, let's do another run with time-based battery voltages:

This time I did use a current source (an ideal component that does not exist in reality ), marked as Current_charger, and set to provide constant 5A current. I also removed the measurement resistor, so now all the voltage readings are "correct" (directly against ground). I also later on added series resistors, representing battery pack internal resistance (assuming that they all have the same 0.5-ohm internal resistance, in reality, the resistance can and will vary between batteries) and tweaked the batteries to act as "arbitrary behavior voltage sources", a fancy way of saying that their voltages can be represented as equations instead of constant value, each starting from 50V, and then raising their voltage (like the batteries when they're being charged) with varying rates (BAT1 = 50V + 5V per time unit, BAT2 = 50V + 3V per time unit, BAT3 = 50V + 2V per time unit and BAT4 = 50V + 1V per time unit), limiting their maximum voltages to 67.2V (which they in reality wouldn't actually ever attain, due to voltage drops caused by the diodes... Well, it's just a simulation anyway ).

Now running a 20 "time unit" (seconds in LTSpice, but here it doesn't matter as there's no frequencies etc. involved), we see the voltages rising and the currents across each battery:

So now what happens? The slowest charging battery (BAT4) will draw the largest portion of the current. Of course, this doesn't 100% match what would happen in real-life, as the other battery voltages raise beyond that of the charger output, but (probably) in a real situation, the other batteries would reach the charger output voltage (minus the diode forward voltage drop), and draw (almost) zero current, leaving the "brunt" of the current running through a single pack. If you look at the current & voltage graphs, you can see that in the end, the voltage before diodes actually rises ABOVE the charger output (67.2V) and the current steadies to 1.25A per pack (1/4th of total current). That's just an "artifact" of the simulation, because there's a constant current source, that "forces" 5A current even after the total voltage has been reached, and wouldn't happen in real life (instead, the current would drop to zero for the BAT4, all others alredy drawing 0A).

Finally, I still ran one last simulation with the battery voltages raising slower:

This shows better how the current would divide unequally between packs with different voltages (again the voltages raise above the charger-voltage, as they're still time-based in the simulation, probably should figure out how to simulate this more properly ).

Simplified, for a current of 1.25A (1/4th of the total current) to run through all the battery packs, the voltages would have to stay equal (assuming equal internal resistances). If any one of them will charge slower than the others, a higher proportion of the current will start running through that pack.

During constant current phase, to keep a constant current of I (5A in this example), and with a constant internal resistance of the packs (0.5ohm per pack, 0.125ohm with 4 packs in parallel), you get:

U = 0.125ohm * 5A= 0.625V    (Do note that I always use U to denote voltage, as that's how it was taught to me back in the day, U = voltage, I = current, R = resistance, X = reactance, Z = impedance)

So the charger voltage should be 0.625V higher than the packs. Actually due to the diodes, the charger output voltage should be lowest battery pack voltage + 0.625V + diode forward voltage. It also probably doesn't help that the diode forward voltages are likely not all equal, plus they change according to the current and heat (also the battery internal resistances will change with temperature, and they may heat up during charging). The thing is, if the battery voltages aren't equal, that drop will happen over the lowest voltage battery, and the current would be something like

I = (charger output voltage - pack voltage - diode forward voltage) / internal resistance of the pack

AFAIK, the charger constant current circuitry should keep the voltage correct to allow constant current of 5A (or whatever it's for that particular charger) to keep running as long as the maximum voltage isn't reached, but it won't "know" how the current is being divided over each pack. I don't know how the actual chargers measure it, but at least to my knowledge, the only "control" over the current that the charger has is by adjusting the output voltage (current sources like the one used in above simulations do not exist in real life). If that voltage for any reason goes "haywire" and starts to rise faster than the lowest battery voltage, the current will start to shoot up and bad things will happen...

Also, like I (think I) mentioned above, the simulations are not that exact: probably if and when any of the pack(s) would "fall behind" in voltage of the others (ie. charge more slower), the current for that/those packs would rise higher, and drop for the others, until the voltages would "catch up" each other again. So it could be possible that the system sort of "auto-adjusts", and stays pretty much at "equilibrium". So likely the most "critical" part is at the start of charging, when the pack voltages are at "most different" from each other.

I guess this is a very long way of saying that I'm not 100% sure what would happen if you used a larger current (for example 5A) charger to charge multiple packs in parallel, as there are many factors at play. Do note that this situation is DIFFERENT than when charging multiple packs in one wheel: in a single wheel, the packs are connected in parallel through the discharge wires WITHOUT a diode in-between, so the voltages "equal out automatically" in such case. Here, the idea of using diodes between the packs is to isolate them from each other (so they cannot cause large current to flow between them due to voltage differences), which is a good idea if placing packs with potentially different voltages in parallel, but at the same time, can cause the entire charging current to only go through a single pack. Also, diodes exhibit something called "diffusion charge/capacitance", which can cause a (short-lived) backwards current to flow when the power shuts out, but I believe it's (usually) fairly minimal (milliamperes at best?), so probably not an issue here?

If someone decides to try it out, I'd suggest at least using large power-diodes per pack (preferably with 200V or more reverse breakdown voltage, as mistake in polarity can cause the voltage to become charger voltage + pack voltage, which would mean something like >130V for fully charged pack and 67.2V charger output) , and using a charger with low enough maximum current that it can't destroy a single pack (keep charge rate below 1C) should the current start flowing through one pack only... And to monitor it closely in case of trouble. Also, the charger output could be adjusted to overcome the diode forward voltage drops for it to be able to charge the packs to full 67.2V (but then again, there's the slight problem of the forward voltage changing with current... probably should use the value at low current, ie towards the end of charge).

But now, what if you placed bleed-power resistors between the packs..?

[hobby16]

On 23/5/2016 at 3:57 AM, dmethvin said:

If the output to each EUC was protected by a diode so the power can't reverse between packs, I would think the charger should produce a constant 67.2v at 5a that gets consumed by all 4 packs. When one of the EUCs starts to reach full charge, the current will go down for that pack and the others will use more of the 5a until they are charged as well. Finally they will all be charged and the current will go down to zero. Or would it not work that way?

On Gotway wheels (and King Song and some generics too), there is no voltage inversion-protection diode on their BMS. The wheel's charging port has the same battery voltage as the motor. Some see it as a nice thing because it can be used as a power port, for charging smart phones or various electronics using a step-down converter. Other may consider it a dangerous thing.

Anyway, no, you can't connect two wheels on a same charger, especially Gotways.

You can connect two chargers to one wheel because each charger has its own diode, but not the other way round.

[hobby16]

On 23/5/2016 at 9:16 PM, esaj said:

If someone decides to try it out, I'd suggest at least using large power-diodes per pack (preferably with 200V or more reverse breakdown voltage, as mistake in polarity can cause the voltage to become charger voltage + pack voltage, which would mean something like >130V for fully charged pack and 67.2V charger output) , and using a charger with low enough maximum current that it can't destroy a single pack (keep charge rate below 1C) should the current start flowing through one pack only... And to monitor it closely in case of trouble. Also, the charger output could be adjusted to overcome the diode forward voltage drops for it to be able to charge the packs to full 67.2V (but then again, there's the slight problem of the forward voltage changing with current... probably should use the value at low current, ie towards the end of charge).

Inversion is not really a problem because the connector prevent you from it. Using a 200V diode is inconvenient because it means you must use a PN junction diode with a forward voltage of 0,6V, causing more power loss by heat and so a big heat dissipator is needed (0,6V x 5A = 3W to dissipate). It will work of course but not necessarily the best design. It's better to use a schottky diode (metal-oxide junction diode) with lower voltage (eg 80V) but less forward voltage, around 0,2V instead of 0,6V.

Anyway, your setup with external diodes should work (with copious heat dissipators probably), but it's not really for the light-hearted.

 

[esaj]

1 minute ago, hobby16 said:

Inversion is not really a problem because the connector prevent you from it. Using a 200V diode is inconvenient because it means you must use a PN junction diode with a forward voltage of 0,6V, causing more power loss by heat and so a big heat dissipator is needed (0,6V x 5A = 3W to dissipate). It will work of course but not necessarily the best design. It's better to use a schottky diode (metal-oxide junction diode) with lower voltage (eg 80V) but less forward voltage, around 0,2V instead of 0,6V.

Anyway, your setup with external diodes should work (with copious heat dissipators probably), but it's not really for the light-hearted.

Haha, I'm not going to try it in real life, just some simulations & general thoughts on dmethvin's suggestion   I've been trying to get my head around electronics for the past 6 months or so, and thought it would be a good "exercise". Schottky's should be fine as long as it's (more or less) impossible to reverse the connectors or wiring...

Good for you that you don't visit the forums so often anymore, I would have bombed your forum-inbox full with all sorts of stupid questions

[hobby16]

Just now, esaj said:

Good for you that you don't visit the forums so often anymore, I would have bombed your forum-inbox full with all sorts of stupid questions

Never seen a stupid question from you. So far... lol.

[hobby16]

On 23/5/2016 at 1:43 PM, sbouju said:

Hobby16 has worked on a special ChargeDoctor2 with 2 input connectors, but with 2 output cables too... May be it is the same problematic, somewhere...?

https://www.espritroue.fr/topic/356-charge-doctor-v2/?do=findComment&comment=8647

Edit: the initial purpose of this special CD2 was only to have one usable on a Gotway *OR* a Ninebot!

And the question came soon (but only!) about its usability with the 2 chargers at the same time for only one wheel, just for an eventual faster charging.

I know at least one user who has modded a standard 67V charger to a Ninebot 63V charger to connect it in parallel with the original Ninebot charger through the above Charge Doctor to fast-charge his Ninebot at 4A. And it works nice.

[esaj]

Split the electronics question-stuff here:

Sorry for derailing yet another topic

[Mono]

I recently bought one from Banggood for $16.55, but can't re-find it right now on the website. With 375g it is slightly lighter than the others I own and it is considerably less noisy. In particular it seems that the fan operates depending on the temperature.

EDIT: looks like I ordered this http://www.banggood.com/Electric-Unicycle-Charger-AC-to-DC-Power-Supply-p-966060.html one, though the model I got looks definitely different from the one on the pictures.