Power PCB for Lebowski Controller to be combined with Bobc's processor board

3.1 l/100km, not bad indeed for a retro-technology I'm also still using (4 l/100km with BMW F800S) :wink: .
But back to the future... The choice of R1 on the Power PCB is much more sensitive to the battery voltage than I first thought.
Protect the processor-board from high voltage (how much is acceptable?) means high resistor values which means high losses (heat dissipation) at the 50mA needed by the processor board. Here's my conclusions based on lebowskis remarks:
R1_choice_lo600.jpg
To stay below ~60V as supply voltage for the processor board, 1 resistor with 0.5W is enough up to a 16S (67V) battery. For higher voltages several 0.5W-resistors are needed for heat dissipation.
10S-battery 1 x 100 Ohm or just a bridge for R1
13S-battery 1 x 100 Ohm 0.25 Watt losses
16S battery 1 x 150 Ohm 0.38 Watt losses
20S-battery 3 x 1.5kOhm in parallel, 1.25 Watt losses (or 3 x 150 Ohm in series) (or 1 x 470 Ohm 2W)
24S-battery 5 x 3.9kOhm in parallel, 1.95 Watt losses (or 5 x 150 Ohm in series) (or 1 x 820 Ohm min. 2W)
 
I'm a little confused. Why are we paralleling resistors instead of just buying single resistors in the power rating we need?


Also, I haven't seen an answer yet as to whether it is a bad idea to mount my brain board on the top, rather than the bottom, yet.
 
Sorry for the confusion, on course just one resistor with the right wattage would be the simplest (for 20S: 470 Ohm 2W (1.18W dissipation)).
I was just hoping to find standard 0.5W resistors I could combine. Stupid problem: If I order single resistors in an online shop, this costs a lot considering a min. ordering amount and shipment. And it takes several days until it arrives...
BUT: Maybe a reducton to 60V is not needed! I tried to identify the 15V power supply on the brain-board, it seems to be Texas Instruments LM5017. If this is true... the LM5017 accepts up to 100V, so we wouldn't need a heavy voltage reduction on the power-PCB.
With 20S battery and just one 150 Ohm resistor, the voltage is reduced by 7.5V (always below 77V to the brain) and dissipation is 0.38 Watt.
With 20S battery and whereswally's 220 Ohm resistor, the voltage is reduced by 11V (always below 73V to the brain) and dissipation is 0.55 Watt, so a 1W resistor would be needed.
Experts opinion?

My guess for the 'brain on top' question: I don't see a problem, it just would sit turned-over higher over the board with long connection pins.
 
You can change the location of the brain board, just make 100% sure all connections are correct and not flipped.

You can use a single high watt resistor to drop the voltage into the brain board. I just imagined a bunch of 100 Ohm 0.5W resistors would be the easiest to include in a kit because then everyone can connect as many in series as necessary....

Yes the LM5017 can take 100V, I just find that it gets too hot when it gets more than around 40V at the input. When the 5017 overheats it will shut down, stopping the controller.
 
Good info, thanks guys! I'm going to go ahead and put my brain on top. I want as thin of a package as I can get. I started cutting the second copper plate yesterday and will hopefully have it soldered this weekend. Should be rock and roll after that with the other components.
 
coleasterling said:
Good info, thanks guys! I'm going to go ahead and put my brain on top. I want as thin of a package as I can get. I started cutting the second copper plate yesterday and will hopefully have it soldered this weekend. Should be rock and roll after that with the other components.

Can you post a picture before powering up just so i can see how you connected the brain to the powerPCB ?
 
So the issue in my view cole is that when you attach the brain under the board as in lebowskis vids you can flip the whole construction brain side up and see the leds flash directly.

To place the brain pcb on the other side you will need some longer headers to avoid physical clashes with other discrete components on the power board and it will have to go brain chip facing into the power board so you will need the other headers for the halls/reset/prog to all face the other way. Another potential idea is to use a short ribbon cable but ill be honest I don't like it since I don't think those wires are designed to take 40 - 100v. you should be able to see all the led from the sides anyway so flipping the board to the other side is not too bad. just make sure non of the through hole parts are touching from brain to power board. :D
 
I need your opinion... the controllers I know have an "enable" function, which is a battery voltage input to the controller which can be switched on and off with any type of device (keyswitch, kill switch, rip cord, ...). The main battery line is not interrupted, just the brain is switched off which will switch off the Mosfets.
Would you recommend this for 'our' baby? Nobody was talking about this point up to now and I'm a bit unsure if I should do it... a possibility to put in a switch would be the pins over R1 (where the discussed power resistor sits) on the power PCB.
This would interrupt the BPOW voltage to the brainboard.
Or do you see a better way?
 
Yes that sounds ok to me, bas might chip in with a better informed opinion. Just don't turn off under power.
 
As a main on/off switch I would just completely disconnect the battery.

If you just want to shut down the low voltage part, I would interrupt both the high voltage going into the 15V DCDC and also the line where it measures the battery voltage (else it will supply a few mA via this resistor to the 5V).

On the bottom of the PCB, on the left (so with the temperature sensor on the right), there is a line going from the battery voltage to the two resistors that feed the battery voltage to the processor board. I would put a switch in this line...

( Well actually I would not, I would put a proper switch that completely disconnects the battery. What if for some reason you on/off switch fails and puts the battery voltage on the metal of the switch ? I don't really like the idea of a $0.50 cheap switch in the 80 V line...)
 
A good disconnection including a spark prevention would really be nice, I tried to find what is around in the Sphere and found Fechters and Methods designs. My proposal based on their knowledge is:
BatterySwitch_FechterVedder.JPG
It includes a connection to Justins Cycle Analyst. Hope that this will work... any objections?
 
I was soooo close, but now there's dispair: The power PCB is ready and I connected the processor board. On 35V from the lab power supply the 15V and 5V LEDs were first flickering. I installed 2x270ohm resistors for R1 (540ohm) to lower the voltage to the brainboard, so I had only about 18V as supply to the brainboard (PBOW). After bridging R1 everything was fine with 15V and 5.1V at the connector pins 1 and 2.
IMG_20180709_154801_lo600.jpg

Then... I wanted to communicate with the processor for the setup and something must have happened. While bridging again R1 on the power PCB, I heard a sparkling sound and now the 15V LED is a little bit flickering, showing ~5V at pin 1 and 0.1V at pin 2.
My suspicion:
- ev. I touched the wrong side of R2 (after the 39k resistor) which delivered V_battery (35V) via pin 14 to analog input 23 of the processor (battery voltage measurement).
- or for some reason the 15V DC/DC converter lm5017 (the one I wanted to protect with R1=540ohm) died.

I need your advice. What could be damaged? What could I measure to find it out?
ThanX for your assistance!
 
Hey Emmgee just found this post, surprised you haven't PM'd me

I can see whether I can use my hot air gun and repair the LM5017. not the easiest job due to the cramped conditions on the boards.

I really want to see these 3 prototypes out in the wilds working. Havent seen any on a bike yet.

Anyway Pm me if you want me to have a go at fixing it. I cant promise anything though.
 
I'm just tooling away slowly on mine. It'll probably be another month before I'm finished, but I am making progress! Nothing exciting to share yet, though.
 
So has anyone got their controller running yet?

I wonder what would be best way to power the brain up in order to set initial parameters. ~50V with lab PSU to pin 14 on main connector? 5VDC through halls pin? I would hate to blow something up.
 
Well, just as I was afraid of - blew a component: SOT23 between throttle and main connector, next to R4.


Applied 45VDC to pin 14 (BAT+) and 17 (GND) on main connector.

Now two questions - is it repairable (or did something else gave away its magic smoke?) and if yes then what's that component?
 
I will probably have a spare somewhere. Pm me your address and i will send you one. And by spare i mean the sot23 part not the board
 
If I got it right then it's D4 in schematic - BAT54S, a schottky diode, rated for 30V (!).
I was already able to get it locally, but thank you for the offer.

I'd appreciate to hear, what did I do wrong and some advise, how to proceed once I have replaced it.
 
Id imagine that part of the circuit is only 15v or 5v so 30v should be fine. No idea what went wrong maybe a short somewhere?
 
Posted this on the Facebook group. Finally got some time to work on the board again. I think I finished beefing up the high voltage path, but am open to input. Plates are around 1mm thick. Maybe more beef on the top side closer to the caps?
 

Attachments

  • bottom.jpg
    bottom.jpg
    76.4 KB · Views: 2,314
  • topq.jpg
    topq.jpg
    61.7 KB · Views: 2,314
  • top.jpg
    top.jpg
    83.4 KB · Views: 2,314
More progress! Got current sensors and the first three fets on. Took a lot of heat, hope I didn't cook something. Also wondering if I should have bent the fet leg to sit on the plate shared by the current sensor instead of passing it through the board. Too late now, anyway.
 

Attachments

  • IMG_20190423_182538.jpg
    IMG_20190423_182538.jpg
    117.1 KB · Views: 2,274
  • IMG_20190423_182556.jpg
    IMG_20190423_182556.jpg
    15.9 KB · Views: 2,274
More progress. Got all the components that require high heat done. Still worried I cooked something, but I guess we won't find out until I run the dang thing. I'm hoping it will go a lot more quickly now that the tough stuff is done.
 

Attachments

  • IMG_20190424_113155.jpg
    IMG_20190424_113155.jpg
    75 KB · Views: 2,262
More progress. Down to the last resistors and then pin headers.
 

Attachments

  • IMG_20190425_104925.jpg
    IMG_20190425_104925.jpg
    116.1 KB · Views: 2,244
Back
Top