Golden Motor Build 48v and Beyond!

[icecube57]

OK so I left my bike on with the throttle lock on and I come back several hours later to move the bike one day it doesnt work. (Throttle lock that is) and this is what happens. If the controller comes on I can turn the throttle about 75% and the motor runs with the wheel off the ground. I can slow down and speed it up. If I go to 100% throttle the motor stops and the controller crashes and I have to unplug it several times to get it back going again. Sometimes I get lucky and it runs fine without issues and i can actually ride. But other times it just sits there and does nothing. I cracked it open and I found the afterbirth of some magic smoke. Pictures are below. I also used google and made a cheap camera mount for my digital camera.... Super sturdy too as for the motor controller... eh.... I called the guy up at Arizona Ebikes. Explain the situation. Sent the pictures of the fried components in the controller... He is out of my controller that I currently have but hes going to send me an old school black case GM Regen Controller. Not the new sensorless one but the old school larger one. He says its slightly larger than the one I have now. Its 1000w 48v 30A. I think its that one with the funky regen issues. But i was looking to get a regen controller anyway since my brakes are sub par. Hes upgrading me free of charge and should be shipping it out tommorrow or monday at the earliest so i should have a new controller by the end of next week.

[icecube57]

Two videos different routes. Real world conditions.

[icecube57]

Despite my bad controller Ive been riding my bike like its a rental. I did range test today on the Silver Comet Trail by my house.. I did 35 Miles total. I rode 17.5 miles and consumed 6.5 AH. I maintained a speed of 18-20mph. On the way back to my car I did WOT for the last 17.5 miles. I maintained atleast 30mph. I peaked up to 33-34mph. I consumed 9 AH. For a total of 15.5 AH. My bms cut out just as i was getting off the trail. It was only at 46v under load. I am going to bypass the bms on my next run to see if i can get anymore out of my pack. So on the first part.. if 48v is my nominal voltage i did 18wh per mile on the way back i did 25wh per mile for an average of 21wh per mile. This is flat terrain 1-2 percent grade at most in places.

[dnmun]

why not just fix the controller? it hasn't ruined your motor yet and is still fixable.

maybe you can get one of the infineon controller for 48V and use it to see if you get the same performance.

do you know for certain that the BMS on the battery was shutting down for LVC or could it be the controller LVC shutting down because that input power resistor is burned up and the voltage regulator is running so far out of control it cannot provide a good 5V source for the zener used by the the shunt voltage reference. which is what appears to be happening. imho.

[icecube57]

The controller is bad yes, Im getting a GM Regen Controller to replace this one RMA through the company i bought it from.

But Its the bms cutting out early on the battery because the last time i tried a full discharge it stopped at the same exact AH which was 15.5. But the controller was good on my last full discharge. When I tested the cells through the bms they were rather high at 2.7. some a little bit lower but not near the full discharge state of 2.2.The voltage was around 46 when it cut out.. The lvc on a 48v controller is normally 41 sometimes lower if you get one custom made.

The vendor says its normal that the bms is protectecting the pack. It doesnt lug or anything i still have good solid 30mph performance all the way to the 15.5AH mark and then it cuts out. But i think the lvc is set so high to prevent full discharge from damaging the cells and pack. Its well designed and i can see why the vendor sets it that way. But I ordered a 20AH pack and im getting 75% of that and verified by a volt meter that the cells arent fully discharged. Im going to charge with the BMS and discharge without it and either let the lvc of the controller kick in or manually back off when it hits 40 to 41 under load i bet i get better results.

[dnmun]

i see, this is a different thread from the one i responded to originally. i had only seen the new thread you started on this. anyway, i thought your FETs had burned up too. you just soldered up the trace on the side when the resistor shorted out on the case and melted that trace? that was what you asked about originally i thought.

anyway, i suspect your power resistor may be different in value from what it was originally, so you could check them with an ohmmeter and if they are far off in value from the color code value, then replace them, and you could check to see if your 5V reference bus is really at 5V before you take the resistors out. but if you can certify that, and voltage regulated on the 12V, i expect that controller would still work since it just shorted out on the case when you jumped a curb or hit something. you can cut a small square of polyethylene off the side of a milk jug and slide it under there to support the pcb to keep it from happening again.

they most likely have the regen feature somewhere available on that pcb too. you are getting a similar controller?

[icecube57]

The guy that i spoke with says its a old school 1st generation regen controller. The new regen controllerss are small and compact. The 1st generation are black and a little bit longer than the current 48v 1000w controller I have pictured on my bike. I looked at the board really closely after i cleaned it up a bit and it seem that the trace on the bottom edge of the board had something inline with it and then it linked up with the leg of the resistor on the left. the trace above it seems was fine. I took solder and made a connection between the burned trace and the bottom leg of the resistor it seems to have gotten rid of the erattic behavior... not totally but its better. It seems now if i plug my pack in a leave it a min i can ride but i cant plug it in and ride off immediately like its takeing time to charge something up or like you said regulate the voltage properly.

[dnmun]

yes, you can actually go check the voltage on your controller while it is plugged into your battery. with the hall sensors and phase wires still connected

you will see places on the pcb where they have +5V printed next to pins and holes in the pcb.

you can put the voltmeter on there, black lead to ground, and it should read 5V.

if not then you have a problem with the power resistor that got all discolored when it burned up. it burned up because somehow that resistor leg or maybe the edge of the pcb where that trace is, touched the case and shorted out the full battery voltage, so that resistor reached high enough temperature that the inside turned into plasma and the heat from that plasma discolored the inside of your box.

i suspect if you replace the resistor with the proper one, or even a little higher resistance, your voltage regulator will work again and your controller will behave again. make sense? other things coulda burned up, check the second resistor too, but it does not look dead. just close, so check it anyway. that is where you can read the 12Vbus too i think. across that 2nd resistor.

you really should make a plastic cap / tray to go around the end of the pcb there. like i said, a small polyethylene plastic square that includes the corner of the jug a little, around the end and underneath it for long term prevention. also cut off the long legs of the new resistor too.

before you put the pcb back inside, test the 12V and 5V before you put it back inside the box and if you get those back at normal, that is likely all it is. but i'm just guessing.

[icecube57]

I think i can vouch for the ebay vendor that sold me my battery pack. I felt kinda cheated about the capacity of my pack but the more i read the threads i began to understand that 15-16AH out of a 20AH pack with a BMS is conservative but respectable. Its taking it to the safe limit of 3,6 of 3.7 volts a cell to prevent damage to the pack. If you want to tap into that reserve you can but you better not take it below 2.1 or 2.2. Its good to know I more than likely have that 20% reserve left in my pack. I actually just got done wiring my pack with andersons. I can now charge and discharge with the bms. I can also now discharge without the bms if i get stuck out some where and im on that last leg home. I know my pack limit with bms is 15.5 AH so thats why i will try and limit myself to. I just also checked my pack and it pretty well balanced All my cells read 3.72 to 3.75 off the charger. I still have one cell thats a little low at 3.68 but its respectable its within 2% of my highest reading. I will give it one full true full dischage to the lvc of the controller which is 41.5. Right now the bms is cuttin out at 46v under load voltage.

[dnmun]

you should be able to discharge down to 35V without damaging your pack if no cell drops below 2.1V.
don't wanna replace that 27 cent resistor because your bike is a rental?
if this is how GM makes controllers, you will have a problem like this again i bet.
i don't really think much is wrong with that controller except the power resistor.
you could see if it is possible to change the reference voltage in the BMS for the low voltage shutoff, to discharge farther, but that way is dangerous if you drop it below 2.1V/cell
BOL,dm

[icecube57]

I honestly think there was something else in line with that resistor. Because the trace was lifted and there is a rectangular crater.I havent fooled around with it. The weather is bringing my spirits down. I cant ride. Georgia is the only place where you can experience 4 seasons in one week.

[dnmun]

i think that proximal end of the power resistor was connected to a trace which went diagonally over to the the input of the voltage regulator and that the second resistor connected across from the input to the output of the voltage regulator. but you have 3 voltage regulators in TO220 packages in a row, so not sure which is for 12V and which is for the 5V. it may be that when you filled in the void with solder, you ended up shorting out 2 traces which were close initially, but not connected. so when you filled the hole with solder, it ended up having a different circuit configuration, and the little bit of current leaking through the burned up power resistor is just enough to allow the logic and drivers to run, but gives you problems. it would be neat if there was someone who has a picture of that pcb and could post it up so we could compare, then have some idea of how to fix it.

if it is not supposed to be connected and you put a new power resistor in there, could be problems again. the proximal end where it burned the hole in the board doesn't actually have to go to the board, you could find a replacement resistor and connect it from the Vcc on the distal end all the way over directly to the input to the voltage regulator i suspect, then figure out what to do about the solder blob you put back.

you can go read the infineon transistor modification thread that geoff started following knuckles design for the 'any voltage' power resistor replacement. that was the only thing i had to go on until i looked in mine and saw how it worked. mine is the 9FET infineon, but is pretty much the same on all controllers i think.

maybe more closeups in there so we can see where the traces run underneath those 2 resistors and what the TO220 packages are would be a start.

if you can get the voltage regulators working again, and have a new controller to depend on, then you could use this one to hop up the shunt and swap out the FETS for more power. then if you blow this one up again , you would still have a working, if normal, controller.

[icecube57]

Looking at the board this is how i interpreted the board. The retangle is whatever blew up and the green lines are supposed to be the traces. The trace on the right was fine the trace on the left was exposed going to the rectangle and then going to the leg of the top resistor.

[icecube57]

I was checking to see how much reserve was in the pack after the bms cuts out. So i crimped on new anderson connectors and bypassed the bms for discharge. I performed a discharge test. I turned my bike upside down and let the motor spin. It was drawing 1.5A and near the end of the discharge 15 hours i was drawing about 1.2A. I hit the lvc of my controller at 40.5v. I got 15.5AH out of the pack again. WTF!!! I checked the cells strings on the pack through the bms to check their voltage and they were way across the board. everything from 1.7... I tested some around 2.5v i even had some sitting at 3v. Im like wtf. I hurried and put it on the charger to get the low cells back to a safe range. Two hours later i took it off the charger and they were. sitting around 2.7v. I was kinda mad. I sent the vendor a harsh email. Then i proceeded to take the pack apart. After 45 minutes. I ripped through 1 layer of duct tape. One lay of laminated construction paper .... a crap load of hot glue and another layer of ductape i finally came down the the cells. I had 1 cell that had a crush top... tested fine though... voltage wise. I cell fell out the pack completely and i had 4 loose tabs. Those cells with the loose tabs were sitting looooooow. but not to low. I had one that was a dud.. barely registered on the meter. On the low cells... im broke so i took a cell phone charger and stripped the wires. 4.2v dc at 1 A taped it to the terminal for so many minutes depending on the SOC of the cell. And it brought the cells back up to about the same voltage as the rest in the string. I sanded the terminals. Applied a thin layer of solder on the tab and on the battery terminal and put them back in the strings. There were 4 loose tabs in different strings across the pack. Plus the one that fell out of the pack despite the glue and crap. GRRR. After i checked all I could physically and visually i carefully assembled the pack up. With Ductape. One layer of self adhesive neoprene foam on all sides and another layer of tape. I charged the pack. After it got done charging all were sitting at 2.72 to 2.76. One string was at 2.60. I let it sit another 8 hours on the charger while I went to work. It raised up to 2.64. In time i think it would have came up to the rest of them but i could wait i had to ride. I rode and drained my whole pack. I gained a little over 1AH and finished up around 16.7 AH. Its an improvement. Maybe if the cells balance a little more i can a little bit more but its a step forward in the right direction.

[icecube57]

By the way the pack is 16s 18p by the way... Yeah i know it sucks... lol

[icecube57]

A friend of mine tracked the information on the cells on my battery. Its not exactly the same cells. The tops seem different but the logo of the company and ebay vendor is uncanny.
What do you think?

http://cgi.ebay.com/LiFePO4-48v-20ah-BATTERY-FOR-ELECTRIC-SCOTTER-E-BIKE_W0QQitemZ170318686849QQihZ007QQcategoryZ11332QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
http://www.chinatoppower.com/battery/News_View.asp?NewsID=20

Model TP-LiFePO4 18650 1500mAh 3.2V
Nominal Voltage 3.2V
Battery Capacity 1500mAh
Cut-off Discharging Voltage ≥2.0V
Maximum Charging Voltage ≤4V
Maximum Charging Current 1500mAh
Maximum discharging Current 6000mAh
Charging Method CC/CV
Overall Dimensions Diameter：18±0.3mm
Height： 65±0.3mm

Weight ≤50g
Operating Temperature Charging 0℃---65℃
Discharging -20℃--65℃

If the specs are right this could discharge up to 4c with my current pack configuration thats 108A the bms is supposedly peaks up to 60A. Something is under rated or over rated. But 108A discharge is in a perfect world to.

[The Stig]

Hmmm it's probably hard to positively identify those cells in your pack...

I just got my 48V build running for a test run. Dual suspension, full size AND folding.
For me the only thing that worked perfectly was the ping battery. Other than that I had to machine the axle to fit, change hall sensors TWICE, took 3 times to get a working controller with it. All because of a bad vendor Brett... It's running beautifully now and I'm going to finish it up nice.

What rack bag are you using?

[biohazardman]

The controller is bad yes, Im getting a GM Regen Controller to replace this one RMA through the company i bought it from.

But Its the bms cutting out early on the battery because the last time i tried a full discharge it stopped at the same exact AH which was 15.5. But the controller was good on my last full discharge. When I tested the cells through the bms they were rather high at 2.7. some a little bit lower but not near the full discharge state of 2.2.The voltage was around 46 when it cut out.. The lvc on a 48v controller is normally 41 sometimes lower if you get one custom made.

The vendor says its normal that the bms is protectecting the pack. It doesnt lug or anything i still have good solid 30mph performance all the way to the 15.5AH mark and then it cuts out. But i think the lvc is set so high to prevent full discharge from damaging the cells and pack. Its well designed and i can see why the vendor sets it that way. But I ordered a 20AH pack and im getting 75% of that and verified by a volt meter that the cells arent fully discharged. Im going to charge with the BMS and discharge without it and either let the lvc of the controller kick in or manually back off when it hits 40 to 41 under load i bet i get better results.

A packs rated AH may only be reached at a very low draw. Sometimes there is even a rest period half way through the test. Makes them look good and is a practice often used. It's about getting the most from their batteries. Getting 75% of rated AH in a real world ride is not all that bad at all.

[dogman]

You're kidding right? There's runt cells in that there pack. A good lifepo4 pack will discharge at least 95% of the rated ah at a 1c discharge.

[The Stig]

I get 17AH... 18AH+ out of my 16AH ping pack when I test discharge it at 0.5C and lower!
Be the pack leader!

[icecube57]

I went out and bought a rc charger used on rc planes. Takes 11-15 DCV input. It can output 4.2, 8.4, and 12.6 volts. I used my regular charger to bulk charge it. I let it try to balance itself. I then disconnect the BMS/Balance connector. Inserted the charging pins into the appropriate connector holes for the given string i want to charge. It charges them up to 4.2v. The bulk charger doesnt let the cells get that high only up to 3.75 which is the bare minimum for charging lipo.
. Its taking a while to charge... about 1hr or more a string at 1 amp so I know they werent balanced. It shuts off and alarms when its done. I can charge at 250ma, 500ma, and 1 amp. This should help a bit with the capacity. After my repairs they were all discharging evenly voltage wise after i pulled 10AH out of them.... that was a rare sight. I knew there were runt cells when i repaired the pack but i had no way to charge them properly. And with the bulk charge it always cut off before the weak cells got full. It took several days for the voltages to come up to the rest of the pack.

[icecube57]

Speed 15-18mph. Head wind. 14.55Ah. 751wh. 52miles. 14.44 wh/per mile

[Russell]

I went out and bought a rc charger used on rc planes. Takes 11-15 DCV input. It can output 4.2, 8.4, and 12.6 volts. I used my regular charger to bulk charge it. I let it try to balance itself. I then disconnect the BMS/Balance connector. Inserted the charging pins into the appropriate connector holes for the given string i want to charge. It charges them up to 4.2v. The bulk charger doesnt let the cells get that high only up to 3.75 which is the bare minimum for charging lipo.
. Its taking a while to charge... about 1hr or more a string at 1 amp so I know they werent balanced. It shuts off and alarms when its done. I can charge at 250ma, 500ma, and 1 amp. This should help a bit with the capacity. After my repairs they were all discharging evenly voltage wise after i pulled 10AH out of them.... that was a rare sight. I knew there were runt cells when i repaired the pack but i had no way to charge them properly. And with the bulk charge it always cut off before the weak cells got full. It took several days for the voltages to come up to the rest of the pack.

I'm confused; is your battery pack LiPo (3.7 nominal/4.2 peak) or LiPoFe4 (3.2V nominal/3.65V peak)?


-R

[icecube57]

Ive discarded the charger because of that issue.It was a lipo charger but i was assured it worked on Lifepo4.It dont.

[dnmun]

ice cube, i finally got one of the golden motor controllers to work on, it was the 48V1000W with regen, pcb revision 4/08. withthe black relay in the corner for the LVC.

this is in another thread, for the controller problem, it burns up the zeners they added for the regen function problems:

viewtopic.php?f=1&t=9746

but he gave me another goldenmotor controller, for 36V with the shenzen type controller board. with the LVC resistor bridge on the outside of the board.

you can see the trace from the input battery voltage running across the edge of the pcb to the LVC bridge in this picture:

36V goldenmotor controller with the NEC 3435 40V FETs.

DSC00243.JPG (154.06 KiB) Viewed 914 times

can you take some pictures of your board and paste them on the regen controller problem thread?

you were correct that the trace you rebuilt went across the edge, but there was never anything but a trace there.

make sure the resistor on the right doesn't make contact with the trace, that resistor is the input to the first voltage regulator and is at a different voltage level.

at the time it shorted, it may have damaged the resistors to the right in the resistor bridge, which you can check with a voltmeter and a magnifying glass. possibly the op amp which takes the input from that bridge is damaged, which may account for your intermittent problems.
>>edit added, after thinking about this i doubt if it could have damaged it if the power trace shorted at that spot, since the bridge is already sitting on the battery voltage.<<

now that i got a controller to look at, i suspect we can fix yours too.

does yours have the zeners on the bottom too? would appreciate pictures over there if possible, thanks, dm