It worked great! for 3 seconds... :(

Hello everybody!

I'm having a bit of trouble with my e-bike.
Controller: 12 FET 72-110V Mark II High Voltage Racing LYEN Edition Controller
Motor: Golden Motor Magic Pie 3

So i just finished up building my battery pack (LIPO 24s), was really excited so I connected the halls and phase wires, the throttle and the cycle analyst to the controller.
I followed the color codes from Lyen's website from his controller to the Magic Pie 3 motor.
Phase: Match all the colors.
Hall: Blue-Blue, Green-Yellow, Yellow-Green.

Everything's connected correct, right?
I hooked up the battery to the controller, the cycle analyst lit up and showed the correct voltage. Looks good so far.
I turned the throttle just a little bit, the motor started shaking and made a horrible whining noise. I figured it was the wrong wiring on the halls wires, and so it was. I switched out the hall wires to match, Blue-Blue, Green-Green and Yellow-Yellow.

I tried it again, turned the throttle (slowly) just a little bit for 2-3 seconds. It ran great! No weird noise, everything looked ok on the cycle analyst screen. I used the manual breaks to stop the wheel from spinning. When the wheel stopped, I wanted to try it again. So I turned the throttle (slowly), the Cycle Analyst blacked out and nothing more happened. No response when i turned the throttle.

My first thought was the battery. I opened up my pack, look at everything inside and received good readings from the voltmeter, inc the precharge resistor i built into the pack with a switch.

After checking the battery pack out, I connected it into the controller again, used the precharge resistor(1kohm 10w) to not get a spark, waited about 10 seconds and then connected the main red lead. Giant blue spark and I pissed my pants because i wasn't expecting one. No short or anything, just a standard (giant) spark when the electricity rush into the controller. Why didn't it work now when it worked earlier? I looked at it earlier in the battery pack and it looked like nothing was wrong with it. Anyways, the Cycle Analyst screen went completely green for almost 2 seconds, then blacked out. i tried to turn the throttle, no response.

So I started to use the voltmeter to get readings for the throttle to make sure it was working. Great results, nothing wrong with that. After the throttle I tested the hall sensors on the motor just to make sure nothing was wrong. And nothing was wrong there. Phase wires ok too.

I opened up the controller, I don't really know what to look for in that box, but nothing was burnt so I take that as a good sign...

But something happened, was it the controller? Did I do something wrong? I do not know what to do here. Please help me here...

PS. When the Magic Pie is not connected to anything, it spins really good and freely. As soon as I connect the motor to the controller, it doesn't spin as good, feels like there's more resistance and really lumpy.

Thank you, any help is really, really appreciated.

I'm not an expert on troubleshooting stuff but have you tried bypassing the CA? Need to narrow down the problem. Another controller you can test with?

How is the CA integrated?
Everything connecting to CA and single connection from CA to controller?

+1 on removing the CA from the equation for now. There might be settings that aren't configured for your throttle/motor/controller/battery.

Well, I've tried removing the CA, it's connected directly into the controller via a 6 pin plug.

I do not have another controller to test with.

The weird thing is, everything worked great, until i twisted the throttle the second time. And before the first test of the system, the CA ran just fine, no trouble at all. Now it wont even turn on when I connect the battery...

So I understand clearly, you bypassed the CA, and then twist throttle and nothing happened? Or it did work and then stopped?

drew12345 said:

So I understand clearly, you bypassed the CA, and then twist throttle and nothing happened? Or it did work and then stopped?

Click to expand...

Yea, I got no response at all. Nothing.

Also it seems like the controller doesn't want to be precharged, so there's a huge spark when i connect the battery.

While it's not the same kit as yours my ebikit stuttered on start and went dead a few times. It turns out while the hall sensors were soldered, they were soldered like crap. Soldered them better and the kit worked fine.

Could still be hall/phase wire combination. So you can still play with those. But you can test voltage of hall wires also. There are instructions at ebikes.ca in troubleshooting section. Also , make sure you have rear wheel off ground and watch cranks because they may spin.

PS. When the Magic Pie is not connected to anything, it spins really good and freely. As soon as I connect the motor to the controller, it doesn't spin as good, feels like there's more resistance and really lumpy.

Click to expand...

Sounds like shorted wiring or FETs inside controller. Wrong, yet kinda working combos can blow FETs.

Ykick said:

Sounds like shorted wiring or FETs inside controller. Wrong, yet kinda working combos can blow FETs.

Click to expand...

Ok. How can I check the FETs inside the controller to see if they are working or dead?

Nevermind, found out how to do it from ebikes.ca. Going to check the FETs out now.

Alright, so if anyone can clarify my results, that would be great!

So I measured the resistance between the power leads and the phase leads.

Red Lead - Blue 0.1 Ohm (i think there's a short here)
- Green 14.2k Ohm
- Yellow 3.45k Ohm

Black Lead - Blue 4.95k Ohm
- Green 9.85k Ohm
- Yellow 0.1 Ohm (also think there's a short here too)

Does this mean i have to replace the FETs between Red - Blue and Black - Yellow?
Why does the Red - Yellow connection has almost 11k Ohm lower resistance than the Red - Green measured?
Also a 5k Ohm difference between Blue and Green on the Black lead.

Does it matter? Is it damaged or in good condition?

The phase wire colors should match the hall wire colors. B-B, G-Y, Y-G. At least that's what worked for me with my Hua Tong controller to a GM 901 and I'm pretty sure that was the same as was shown for Lyen's controller to a GM motor.
The reason you get a big spark is because the 1K resistor is too big. Try 150 ohm. wattage doesn't matter much but I use 5W wire wound. A 1K resistor won't let the voltage come up enough in the caps. I started with 550 ohm 10W and went down from there. It didn't work good until I had 3 in parallel for ~170 ohm. 150 ohm 5W works well with all controllers and voltages up to 100V tested, and probably a lot higher. Leave the CA disconnected until you get the thing working.
What's funny about this is I just changed the tire and freewheel on mine and when I put it back together, the motor just sat there and grind. I had got the halls swapped right but had forgot to swap the phase wires. Luckily it only took a few minutes to realize what I had done and there wasn't any damage.

wesnewell said:

The phase wire colors should match the hall wire colors. B-B, G-Y, Y-G. At least that's what worked for me with my Hua Tong controller to a GM 901 and I'm pretty sure that was the same as was shown for Lyen's controller to a GM motor.
The reason you get a big spark is because the 1K resistor is too big. Try 150 ohm. wattage doesn't matter much but I use 5W wire wound. A 1K resistor won't let the voltage come up enough in the caps. I started with 550 ohm 10W and went down from there. It didn't work good until I had 3 in parallel for ~170 ohm. 150 ohm 5W works well with all controllers and voltages up to 100V tested, and probably a lot higher. Leave the CA disconnected until you get the thing working.
What's funny about this is I just changed the tire and freewheel on mine and when I put it back together, the motor just sat there and grind. I had got the halls swapped right but had forgot to swap the phase wires. Luckily it only took a few minutes to realize what I had done and there wasn't any damage.

Click to expand...

Ok thanks I'll try a lower ohm rated resistor. I use the 1k ohm resistor because Lyen recommended me to use one in a precharge resistor to his controller. But I'll try another one.

And when I get the MOSFETs replaced, I'll try to switch the phase wires around. But the halls worked best as B-B, G-G, Y-Y.

By the way, is it easy to replace two FETs by myself?

Edit:
I'm probably going to send it back to Lyen so he can repair it. Just hoping it won't cost too much

I've replaced 8 fets in my 24fet lyen after my entire high and low side yellow exploded after I shorted the wires and ran 6kw through it. Worked like a charm afterword It's not that hard but you need to know how to solder well enough to make sure you don't kill/melt the PCB or any component near it.

My Lyen controllers have drain-down resistors, so I have to keep the precharge resistor connected while making the battery connection to avoid any spark, since the resistor drains the caps fairly quickly.

When working on wiring combos or even reconnecting a working motor, I use alligator clips with a small gauge wire to complete my battery connection until I've spun it up to prove proper wiring. The thin wire acts like a fuse, so there's no way for enough current to get through to blow anything. Before adopting this method I had 4 controllers blow before ever getting a ride, and none since. Alligator clip wires work great for using a precharge resistor, as well as inserting a multimeter to measure no-load current, the final step before riding to verify a valid wiring combo.

For a magic pie, wiring it up to lyen controller may or may not involve swapping the yellow and blue phase wires.

Also, the magic pie runs at a different phase angle than most hub motors. I believe the correct setting is 120 degrees for the 'pie. Or yo ucould just set it to 'auto'. I think that some controllers come pre-set to 60 degrees by default. On my early cell_man controllers, they were set that way anyhow, and it was necessary to change to auto in order to get the motor working.

Ugrd said:

I've replaced 8 fets in my 24fet lyen after my entire high and low side yellow exploded after I shorted the wires and ran 6kw through it. Worked like a charm afterword It's not that hard but you need to know how to solder well enough to make sure you don't kill/melt the PCB or any component near it.

Click to expand...

My skill with soldering is really thin, only have soldered wires together etc. I'll try to change the FETs myself. If it doesn't work out, Lyen will receive a half melted controller with solder everywhere.

John in CR said:

When working on wiring combos or even reconnecting a working motor, I use alligator clips with a small gauge wire to complete my battery connection until I've spun it up to prove proper wiring. The thin wire acts like a fuse, so there's no way for enough current to get through to blow anything.

Click to expand...

Great tip, John in CR! I'll try a thin wire with alligator clips and see which combination works best. Thank you.

neptronix said:

Also, the magic pie runs at a different phase angle than most hub motors. I believe the correct setting is 120 degrees for the 'pie. Or yo ucould just set it to 'auto'. I think that some controllers come pre-set to 60 degrees by default. On my early cell_man controllers, they were set that way anyhow, and it was necessary to change to auto in order to get the motor working.

Click to expand...

Okay, I honestly have no idea what the phase angle is... I'll have to read more about that. In the meantime, how do I set the controller to 120 degrees instead of 60? (Or set it to auto.)