Jump to content

Decrease charging time! 5A High Current Charger Mod


Cranium

Recommended Posts

On 9.1.2016 at 1:39 AM, Cranium said:

 

Yeah, I'll be seeing if caps will do the trick.  I already played with some and went too large because voltage dropped.  I have A Capacitance Substitution box (Elenco CS-440) that I'll hook up later to try inline.  

Since you have that voltage divider there to give out that around 5V, you already know the R-value (limiting the current that goes to the cap while it's charging), and you also know the time during which the ripple goes up (and how long it takes to come down, in case you want to calculate how fast the voltage should drop and how much current the circuits that the regulator powers use). Using this information, I *think* you should be able to calculate a "perfect" capacitance value, but of course you'll have to go with the closest standard size, unless you order a specifically made or use multiple capacitors in series/parallel to get close enough to the "perfect" value:

t = 3RC → I = 5%, V = 95% - practically fully charged
t = 4RC → I = 2%, V = 98%
t = 5RC → I = 0.7%, V = 99.3%

C = t / (3*R)  => I = 5%, V = 95
C = t / (4*R) => I = 2%, V = 98%
C = t / (5*R) => I = 0.7%, V = 99.3%

At least I think? ;)  The 3*R -value is probably enough to lower the ripple already. Also, putting a diode with high enough amperage-rating between the output and the cap(s) is probably also useful, so the cap won't go discharging itself towards the regulator (through the lower side of the voltage divider). But then you'll probably have to adjust the resistors a bit to account for the diode forward voltage drop.

 

Quote

The 1054Z is a great little scope.  I have the 1052E which is a 50MHz scope that I applied a firmware hack to get 100MHz.  I had actually considered selling my scope to get a Mixed Signal Oscilloscope like the Rigol MSO1104Z-S (on the high end).  And then add on the Real Time Waveform Record and Replay Option and the Serial Bus Analysis Option.  This would allow me to also replace my signal generator and my Saleae logic analyzer.  

Good to hear, I might have to press the order-button soonish. ;) 400€ sounds like a lot of money to me, but having watched lots of review-videos, this looks like the scope to go with for hobbyist purposes. Not in that much in a rush, since the shop's in Germany and they deliver fast. Also, I'd have a lot of use for a bench power supply...

 

I just had to test this with some values I pulled from my hat:

Voltage divider by two 100 ohm resistors (to halve the voltage from 10.0...10.4 to 5.0...5.2), assuming 1µs risetime and 4µs falltime for the ripple. Calculating simply based on the risetime:

C = 10-6 s / (3 * 100ohm) = 3.333...nF

I then assumed the load resistance to be 10k ohm. But, this did not work out that good, the capacitor would discharge too quickly and the voltage ripple would be about as large as originally. Testing further, I changed the capacitor to 10nf and then to 100nF:

CSmAxAf.png

Now the ripple of the output (the green line in graph) is considerably smaller. Do note that it's also lower due to the voltage drop caused by the diode. So, using simply the rise time won't be enough, the discharge rate of the capacitor also affects a lot.

 

Something similar to this looks like it would work pretty nicely, if you can do some changes in the evaluation board (or just pull 10V out of it and do this "outside" the board:

SZ6YAKx.png

So, the highest voltage-graph (light blue, Vn003) in the above picture is the rippled voltage coming from the regulator. The darker blue line (Vn001) is after the resistor R3, and green line is after that (the voltage of the capacitor basically). The red line (Vn004) is the voltage from the divider. There's next to no ripple, the capacitor gets charged fast through R3 (limiting current from the regulator to around 500mA, probably could work with lower current too), and then the voltage divider formed by R1 and R2 is slightly modified to account for the resistance in R3 and the voltage drop in D1. I used E24-resistor values, and 4.7uF cap is very common too. Fairly low source impedance from the regulation, I've assumed the load to be 500 ohm here (10mA), probably you'll need more current though, depending all the things you're powering.

Edited by esaj
  • Upvote 1
Link to comment
Share on other sites

While testing the monitor box circuits for different loads and voltages, I noticed that the current sensing appeared to be more erratic than when testing it using my bench power supply.  I am assuming that this is due to the high ripple output of the regulator.  I tried to use the formulas you suggested @esaj but I didn't get them to work for determining capacitance values with my circuit.  The voltage divider I have in place is only being used for sensing the voltage of the power supply and not for the 5V output.

I tried using several different capacitors but these did not seem to help and I’m guessing it is because I was using the capacitors on the breadboard whose circuits are powered from the regulator.  But the current sensor is in the regulator box right next to it. 

So rather than trying to attach capacitors to SMD components, I decided to move the output connector on the board to the low ripple mode.  A quick de-soldering and re-soldering few minutes later and I was powering up the board to test again.

I performed this test at the same 200mA load I did before and set up the Oscilloscope at the same.  Output voltage was down to 4.89V so I lowered the current to 100mA which is still a tad more than what my circuit will consume.  Voltage was back up to 4.99V.  Perfect!

Here is the 200mV ripple before I made the change:

NewFile2.png.eaec8b0474da673d3832339f6d7

Here is the 80mV ripple after I made the change:

Screenshot.png.ba96830dda73d2037c672279b

Very promising output.  I'll continue my testing tomorrow.

As a teaser, here is a brief capture from the SD card for a 3A, 50V test.  The Arduino captures to a CSV which is then opened in Excel for graphing:

Capture.thumb.PNG.f1ae162be3b75e48007d7a

Screenshot.bmp

 

 

Edited by Cranium
  • Upvote 1
Link to comment
Share on other sites

I received my heat shrink yesterday and opened up the E+ battery to look at the circuit board.

IMG_20160121_233648.thumb.jpg.12aa200715
The battery cells are LGDDBMG11865 which translates to LG 18650 2900mAH batteries (2850mAH nominal).  I was even able to put together a spec sheet from online (see attached PDF).  Normal charge current is 0.5C or 1.425mA.  Max charge current is 1.0C or 2.850mA.  Since the Ninebot has 2 cells in parallel, the normal charge rate is 2.85A and max charge rate is 5.7A.  So my desired 4A is well within the range.

IMG_20160121_233330.thumb.jpg.55d77d45f9

The power MOSFETs are IRFB3607 which are 75V Single N-Channel HEXFET Power MOSFETs.  (See Attached datasheet).  The first MOSFET on the left has the charging -Negative lead attached to the Source Pin with the Main -Negative lead attached to the Drain Pin.  This is the charge FET seen in the diagram below.

IMG_20160123_000436.thumb.jpg.5a52eae247

There a couple of other transistors at the +Positive lead for the charging cable that are completely unmarked.  This prevents current flowing out of the battery from the charging port.  The ground and source pins are both on the charging cable +Positive wire.

IMG_20160123_000226.thumb.jpg.f5a194dede

There is one UF2G ultrafast rectifier between the positive and negative leads to the control board.  Thanks @Chriull for finding the spec sheet for this.  See attached.

IMG_20160123_000532.thumb.jpg.939fcf748b

And then there are the main ICs that do all of the work.  This is the SH367004.  And it does a lot!  From the English translation of spec sheet: "SH367004 is a section 3/4/5 series of lithium battery protection IC provides precision overcharge voltage protection, over-discharge voltage discharge protection and over-current protection SH367004 integrated adjustable delay overcharge protection. When circuit, over-discharge protection delay circuit and discharge overcurrent delay circuit .SH367004 for wide temperature, wide voltage applications. And also features a low-voltage mode as lithium iron applications."  If you read Chinese, I've attached that as well.  

This also confirms that the Ninebot has over-temperature protection (top two pins on the right)

IMG_20160123_000316.thumb.jpg.e95f654f77

 

Typical circuit diagram for a BMS (only showing 1 IC for simplicity whereas the Ninebot's has 3 ICs).  Each IC can be configured for 3, 4 or 5 cell balancing and since the Ninebot has 15, it fully populates 3 ICs.

56a31f0c95cbb_Typical5cellcircuitdiagram

LG18650MG12900mAh.pdf

IRFB3607 Power MOSFET.pdf

UF2A-UF2K (UF2G) Rectifier.pdf

Edited by Cranium
Added some technical clarity on a couple pics.
  • Like 1
  • Upvote 1
Link to comment
Share on other sites

The results of all of this means I'm now very confident that I will be able to charge at 4A with no ill effects.  The charging wires can handle it, the BMS can handle it and the batteries can handle it.  Next steps are to try and use the high current charger limited to ~1.9A like the stock charger to make sure I am seeing the same behavior from the data captured.  Then I will ramp up the charging an amp for the next time and if it does good, ramp it up to 4A.

 

  • Upvote 3
Link to comment
Share on other sites

14 hours ago, Cranium said:

The results of all of this means I'm now very confident that I will be able to charge at 4A with no ill effects.  The charging wires can handle it, the BMS can handle it and the batteries can handle it.  Next steps are to try and use the high current charger limited to ~1.9A like the stock charger to make sure I am seeing the same behavior from the data captured.  Then I will ramp up the charging an amp for the next time and if it does good, ramp it up to 4A.

 

Regarding the amperage, someone was just asking about charging a wheel with a larger current charger, and I started to think about the limits... Do you know if the wiring and/or charge port of the Ninebot can take higher amperage? Are there specs for maximum amperage of the Lemo-plug somewhere?

Link to comment
Share on other sites

7 hours ago, esaj said:

Regarding the amperage, someone was just asking about charging a wheel with a larger current charger, and I started to think about the limits... Do you know if the wiring and/or charge port of the Ninebot can take higher amperage? Are there specs for maximum amperage of the Lemo-plug somewhere?

The Lemo plug on the Ninebot charger appears to be a PRG.M0.4GL.?J plug from the dimensions, material and appearance.  The catalog link is below.  The wiring is 22AWG stranded but I don't know how many strands it has to say the exact current rating (and I'm not cutting my cable to count) but 4-5A for the short run it has should be just fine.

From the catalog, on Page 19, it contains the rating for this 4 pin connector plug as 8A.

http://www.lemo.com/catalog/ROW/UK_English/P_series_catalog.pdf

Disclaimer....all of the LEMO plugs appear to be stamped on the side with REDEL.  The ones on the Ninebot don't have a stamp and neither do the ones I purchased.  I can't guarantee these are not different.

Edited by Cranium
Updated a link
  • Upvote 1
Link to comment
Share on other sites

I was asked about low voltage protection so figured I would add the robust feature list of the SH367004 IC that the BMS uses in the Ninebot and other EUs.

SH367004 features:

     ·        SEL0 / SEL1 pin switching section 3/4/5 series application

     ·        Precision voltage detection function: (for a single power-saving core)

          o   Overcharge protection threshold voltage: 3.3V - 4.5V (50mV  per  step)

               §  Threshold voltage accuracy: ± 25mV

          o   Overcharge release voltage protection 1: 3.2V - 4.5V

               §  Threshold voltage accuracy: ± 50mV

          o   Over-discharge protection threshold voltage: 1.8V - 3.0V (100mV  per  step)

               §  Threshold voltage accuracy: ± 50mV

          o   Over-discharge protection release voltage : 1.8V - 3.4V

               §  Threshold voltage accuracy: ± 100Mv

     ·        Two discharge overcurrent detection functions:

          o   Discharge overcurrent protective threshold voltage: 0.05V - 0.3V (50mV  per  step)

               §  Threshold voltage accuracy: ± 15Mv

          o   Discharge overcurrent protective threshold voltage: 0.2V - 1.0V (100mV  per  step)

               §  Threshold voltage accuracy: ± 100mV

     ·        Two charge overcurrent detection functions:

          o   Charging overcurrent protective threshold voltage: 0.05V - 0.3V (50mV  per  step)

               §  Threshold voltage accuracy: ± 15mV

          o   Charging overcurrent protective threshold voltage: 0.1V - 0.5V (100mV  per  step)

               §  Threshold voltage accuracy: ± 40mV

     ·        Discharge temperature protection:

          o   Charging temperature protection threshold temperature: -20 ° C, -10 ° C, 0 ° C

               §  Threshold temperature accuracy: ± 2 ° C (typ)

     ·        Discharge temperature protection threshold temperature: 50 ° C, 60 ° C, 70 ° C

          o   Threshold temperature accuracy: ± 2 ° C (typ)

     ·        Balancing:

          o   Balanced turn-on threshold voltage: 3.1V - 4.4V (50mV  per  step)

               §  Threshold voltage accuracy: ± 25mV

     ·        Disconnection detection function

     ·        External capacitor can be set to delay the overcharge protection, over-discharge protection delay, delay discharge overcurrent protection delay and charge overcurrent protection delay.

     ·        Charge / discharge overcurrent protection delay 2 and delay the internal temperature protection Fixed

     ·        CTLC / CTLD pin priority control CHG / DSG pin output

     ·        Wide operating voltage range: 3V - 26V

     ·        Wide operating temperature range: -40 ° C ~ 85 ° C

     ·        Can be cascaded

     ·        Low power consumption:

          o   Normal operating current consumption: 25μA (typ)

          o   Low power mode current consumption: 4uA (typ)

     ·        Package: 24-pin TSSOP

 

Edited by Cranium
  • Upvote 1
Link to comment
Share on other sites

I recovered my battery with shrink wrap.  I bought both clear and blue shrink wrap but decided to just use the clear.  The battery has two layers of wrap.  The second layer is put on at 90° from the first to provide 100% coverage.  It came out [almost] perfect.  I just need to apply some caulk around the seams of the second layer to make it as water resistant as the original.

@SlowMo 200mm was the perfect size for the heat shrink.  I cut the length to 7" but could have used a little less.

I cut the labels off the original heat shrink to reuse.  

56a479447c0be_Batteryre-shrinkwrapped1.t

And there is something about being able to see the circuit board that I really like!

56a4794668bc1_Batteryre-shrinkwrapped.th

 

Edited by Cranium
  • Upvote 3
Link to comment
Share on other sites

24 minutes ago, Cranium said:

recovered my battery with shrink wrap.  I bought both clear and blue shrink wrap but decided to just use the clear.  The battery has two layers of wrap.  The second layer is put on at 90° from the first to provide 100% coverage.  It came out [almost] perfect.  I just need to apply some caulk around the seams of the second layer to make it as water resistant as the original.

@SlowMo 200mm was the perfect size for the heat shrink.  I cut the length to 7" but could have used a little less.

I cut the labels off the original heat shrink to reuse.  

Thank you so much! Its really nice to have a clear wrapping to see the internals.

Link to comment
Share on other sites

Back to circuits, I received some ICL7660S voltage converters today.  The purpose I would use them is to create a negative voltage rail.  So if I have +5V, it will create a -5V rail for me so I would have 10V to work with on the OpAmp to maximize my range output.  I will have to play with this a bit to see if it will really help.  For some reason, I ordered 10 DIP packages and 10 SOIC SMD packages.  It must have been right after a rum and coke.  lol

Link to comment
Share on other sites

51 minutes ago, Cranium said:

Back to circuits, I received some ICL7660S voltage converters today.  The purpose I would use them is to create a negative voltage rail.  So if I have +5V, it will create a -5V rail for me so I would have 10V to work with on the OpAmp to maximize my range output.  I will have to play with this a bit to see if it will really help.  For some reason, I ordered 10 DIP packages and 10 SOIC SMD packages.  It must have been right after a rum and coke.  lol

Humm, but if the analog-inputs of Arduino are limited to 0...5V, what is the negative rail good for? :huh:  I'm probably missing something here :P

Edit: Maybe you meant that you won't hit the rails so easily, and can then level-shift the signal to 0...5V for the Arduino?

Edited by esaj
  • Upvote 1
Link to comment
Share on other sites

  • 2 weeks later...
On 23.1.2016 at 7:51 AM, Cranium said:

...

There a couple of other transistors at the +Positive lead for the charging cable that are completely unmarked.  This prevents current flowing out of the battery from the charging port.  The ground and source pins are both on the charging cable +Positive wire.

...

This should be the voltage inversion protection diodes do prevent harm, if a charger would be connected with wrong polarity. With Gate and source connected to C+ you have the body diode (invers diode) conducting to Drain.

Or they used a schottky diode (which i just saw are also available in the TO263 housings http://www.produktinfo.conrad.com/datenblaetter/575000-599999/596637-da-01-en-DIODE_SCHOTT_VB30100S_E3_4W_TO_263AB_VIS.pdf) - But the pinout would point towards the invers diode of a MOSFET? But maybe the pinouts for diodes in this housing are not as standardized as for Mosfets and they also exist the other way round, too?

Quote

There is one Zener main diode between the terminals but I couldn't find specs for it.

...

 

 

This UF2G is an ulta fast rectifier (http://pdf1.alldatasheet.com/datasheet-pdf/view/14849/PANJIT/UF2G.html). With between terminals you mean between the discharge wires?

On 23.1.2016 at 8:01 AM, Cranium said:

The results of all of this means I'm now very confident that I will be able to charge at 4A with no ill effects.  The charging wires can handle it, the BMS can handle it and the batteries can handle it.  Next steps are to try and use the high current charger limited to ~1.9A like the stock charger to make sure I am seeing the same behavior from the data captured.  Then I will ramp up the charging an amp for the next time and if it does good, ramp it up to 4A.

 

Just the two mosfets (if they are...) working as voltage inversion protection could theoretically cause problems. A pity they have no number and so its not possible to get the data sheet. But as normal invers diodes from Mosfets they should have a forward voltage from 0,7-1,5V. So with the double charge current the have double dissipation power. Which should rise in this assumption from about 1 Watt per Mosfet to 2 Watt per Mosfet - so hopefully still well within the limits... 

 

Edited by Chriull
Strike out repetition of cited text
  • Upvote 2
Link to comment
Share on other sites

7 hours ago, Chriull said:

Or they used a schottky diode (which i just saw are also available in the TO263 housings http://www.produktinfo.conrad.com/datenblaetter/575000-599999/596637-da-01-en-DIODE_SCHOTT_VB30100S_E3_4W_TO_263AB_VIS.pdf) - But the pinout would point towards the invers diode of a MOSFET? But maybe the pinouts for diodes in this housing are not as standardized as for Mosfets and they also exist the other way round, too?

This UF2G is an ulta fast rectifier (http://pdf1.alldatasheet.com/datasheet-pdf/view/14849/PANJIT/UF2G.html). With between terminals you mean between the discharge wires?

Just the two mosfets (if they are...) working as voltage inversion protection could theoretically cause problems. A pity they have no number and so its not possible to get the data sheet. But as normal invers diodes from Mosfets they should have a forward voltage from 0,7-1,5V. So with the double charge current the have double dissipation power. Which should rise in this assumption from about 1 Watt per Mosfet to 2 Watt per Mosfet - so hopefully still well within the limits... 

 

Good find on the schottky diode.  These could indeed be diodes rather than a MOSFET and would make more sense to use than a MOSFET for the intended purpose.  I will be checking temperatures of these and other components when charging with higher current to see if there is any notable increase in temperatures.  I also haven't checked the temps of the components during regular charging yet so still need to do this in order to get baselines @ 2Amps before increasing to high currents.

The terminals for the UF2G I was referring to were the Positive and Negative terminals going to the control board.  Any idea what the purpose of this would be?

  • Upvote 1
Link to comment
Share on other sites

13 minutes ago, Cranium said:

The terminals for the UF2G I was referring to were the Positive and Negative terminals going to the control board.  Any idea what the purpose of this would be?

inverting polarity protection?

Link to comment
Share on other sites

44 minutes ago, Cranium said:

...

The terminals for the UF2G I was referring to were the Positive and Negative terminals going to the control board.  Any idea what the purpose of this would be?

To save the electronics/liion cells from negative spikes? Caused by wire inductance from the load changes. And they had no place for fat filtering capacitors on the BMS - or the schottky diode is more effective for this? But that's just wild guessing... don't really know why.

29 minutes ago, Rotator said:

inverting polarity protection?

imho not - then it should be in the current path and not between the terminals. And in the discharge path there is too much current for such a measure with a diode...

Edited by Chriull
  • Upvote 2
Link to comment
Share on other sites

2 hours ago, Chriull said:

imho not - then it should be in the current path and not between the terminals. And in the discharge path there is too much current for such a measure with a diode...

Oh yes. It is not the charge side, but the discharge one. Sometimes devices put a inverted diode in parallel to put any wrong connected power supply in shortcut. But this is not the case. My fault.

A negative spikes absorber (like driving relays) is a better theory.

  • Upvote 1
Link to comment
Share on other sites

I took my Ninebot One P 340 Whr battery cover off to see what was underneath.  To my surprise, it is very different from the battery of my E+. 

I have pictures but for some reason the site says I can only upload 0.16MB!  :angry: @esaj, can you help with this?

So a brief synopsis of what I found...

Batteries

3000mAH very good quality LG batteries.

1500mA standard charge

4000mA fast charge

Max discharge current: 20A

BMS

The control IC is different but there are still only 3.  I haven't been able to identify the manufacturer. It has a M logo that looks like that of Microchip Technology but I couldn't find a product they manufactured that would fit the bill.  The numbers are 502 beside the logo and 474G3 at the bottom. 

The MOSFETs are the same IRFB3607 80A 140W power Mosfets.

There are 2 Schottky diodes protecting the charging cable to discharging (like on the E+) but these are marked and I found the datasheet.  They are each rated at 10A and 170V with a typical voltage drop of 0.69V @ 5A. 

The same UF2G 2A diode is being used between the positive and negative leads going to the control board.

Link to comment
Share on other sites

1 hour ago, Cranium said:

I took my Ninebot One P 340 Whr battery cover off to see what was underneath.  To my surprise, it is very different from the battery of my E+. 

I have pictures but for some reason the site says I can only upload 0.16MB!  :angry: @esaj, can you help with this?

Every user has a limited amount of storage for attached images etc. Use a image-hosting site (like http://imgur.com/ , if memory serves, you can actually upload without having an account there, but then finding those images later on could be tricky unless you have the url written down ;)) and link here, that's what I do.

 

 

  • Upvote 2
Link to comment
Share on other sites

Battery Pack before cutting open

d3mfXbO.jpg

LG HG2 218650 batteries

Online review of these batteries: http://batterybro.com/blogs/18650-wholesale-battery-reviews/57179459-lg-hg2-review-20a-3000mah

CcHQML0.jpg

Battery cell specs:

GmRV2U8.png

Ninebot One P BMS

6fZeQmV.jpg

Power MOSFETs

TAxbbYP.jpg

170CB Schottky Rectifier

jscIS7j.jpg

Mystery IC - The numbers are 502 beside the logo and 474G3 at the bottom. 

dg9BE4s.jpg

 

 

Edited by Cranium
Added battery review link
  • Upvote 3
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...