Much More Powerful Pie Now!!!

[o00scorpion00o]

ok after taking it all apart AGAIN! I fixed 3 broken hall wires. Put it all together, this time I tightened the nuts  . Turned the throttle and it worked. Put it all together torque arms and all, turned the bike back over, connected the battery lifted wheel of ground turned throttle, wheel spun, great now time for a spin!

Got up on bike pulled throttle she took me about 15 feet stopped and 5 beeps LMAO I nearly cried with laughter, serious, with all the hassle taking the damn thing apart and putting it together so many times, you would actually think it would work!

OK now it's time to think what I want to do, I have a damn good motor that was handicapped by crap electronics. It could have been good if GM didn't fool people into actually thinking it was a 1000watt motor, For me max 750watts on 48-52 volts! and I'm very annoyed over that more than GM care to know!!! I should have sent it back and got a refund. To say that I wasted $33 on a usb cable that will not program the current as I wanted is an even bigger joke, that and the fact the cable grip of the usb end of the cable is totally useless and the wires pulled out!

I know it's my fault for opening it up and tinkering around with it, but serious, if it was doing what I wanted I would not have opened it at all!

Ok does anyone know if a crystalyte controller will work with the pie? analogue or digital? I love the way some of them have a key switch, the would be very useful! I love the (external) magic controllers ability to adjust the current and regen, but hiding cables is a pain and so is the way they connect to the controller, I prefer the way the cables exit on their old controllers! I think you can adjust the crystalyte controllers with the direct plug in cycle analyst?

I think I will go crazy trying to get GM to send me a new internal controller, and a new cable. I don't think I will bother anyway. So my  only way might be to go external. Wiring up phase and hall wires through the axle will be a pain for sure that and matching hall and phase wires! 

[Magneto81]

I think this should be a lesson to GM - the SW has GOT to work, and it should work in such a way that the user decides how many amps and what speed limits exist. Clearly, maximums must be set so that people can't overload the controller with too much power. It seems wrong to me that the voltage of your pack is what determines your maximum speed, it should be what you tell the controller you want the max speed to be. After all, 24V should be able to drive the thing very fast if you have no load (IE testing the bike with it lifted in the air.)

Given a set of specifications, I would seriously consider doing the programming for you Any chance I can get the API and what maximums have to be enforced?

[o00scorpion00o]

This Pie Aint going to beat me!!!

I'm going to order an external magic controller, new throttle. And while I'm waiting on them it will give me plenty of time to work on running the cables.

I will use cat 5 network cable, and I've got some 12g silicone wire, it's high temperature wire and I've used it for my ping battery before and it's good stuff and very flexible!

I don't want to loose this pie because it has MONSTER torque!!!

I'm going to show this pie who's boss!!! 

[Leslie]

Thats the spriit.  Dont give up...

May I ask some questions.

So did you replace all fets or just one.  Did you use the fet you picked or a different fet? Was it the fet you replaced that blew up or another fet and was it a on the same channel or a different channel.


Which channel did the first fet blow up and what fets are blown up out now.  Green blue yellow?

[o00scorpion00o]

Thats the spriit.  Dont give up...

May I ask some questions.

So did you replace all fets or just one.  Did you use the fet you picked or a different fet? Was it the fet you replaced that blew up or another fet and was it a on the same channel or a different channel.


Which channel did the first fet blow up and what fets are blown up out now.  Green blue yellow?

Hi 317537, I don't remember which fet blew the last time. And I haven't opened it up yet to see whats blown now. But it's the same symptoms. 5 beeps strong vibration in wheel!

I replaced all 6 fet's the last time with irfb3207zpbf.

[Leslie]

I think what went wrong with your fets is the internal diode fused.

The IRFB3207ZPBF body diode pulse max is 640 amps peak.

With your software amps set to max you may have exeeded this from a capacitor discharge that is after the shunt.  This is why I think using the lower farad tantalum caps are the go here for spike resistance.

See between that cap after the shunt, fet and motor you will be lucky to get .7 ohm resistance.  Do some math here and 50 volts divided by .07 ohms =714 amps current.  Maybe explains why you failed to get out the driveway.  The software must of been allowing more current through to the fet from that capacitor in spikes in a manner that you can not see on your CA..

Its interesting out of the three fets we compared that the stb140nf75 has the lowest body diode rating out of the three and I am suprised it lasted longer than the IRFB3207ZPBF.


The Fet  IRFP2097 has a body diode that can take 800 amps pulse.  There was another reason why I picked the IRFP2097 over the other fets for an internal controller is because it is an automotive engineered spec.  It was designed to run under high temperatures, like in motors and such.  It has a higher dissipation max wattage than the IRFB3207ZPBF and its derating factor allows more reliable operations at 150 degrees...

You use the derating factor to accertain the max watts a package can dissipate.

The STB140nf75 has a max dissipation of 310 watts and derated factor (_dr) at 2 watts per C^o above 25C^o

Say your max temp gets to 150C^o

150C^o - 25C^o =125

125x2_DR = @150C^o, 250 watts max dissipation before failure.

The IRFB3207ZPBF has a max dissipation of 310 watts and a deraring factor of 2 watts  per C^o above 25C^o

125X2=  @150C^o, 250 watts max before failure

The  IRFP2097 has a max dissipation of 470 watts and derated at 3.1 watts per C^o above 25C^o

125X3.1=  @150C^o, 387 watts max before failure..

  The IRFP2097 is almost able to perform without any thermal resistances closer to its max ratings at 150 C^o than any of the other fets we have mentioned..

[Leslie]

Further more.

When you shunted to the noise cap with the first mod you at least were providing the noise cap a lot of current so it wasnt draining out.

When you shunted correctly from the main input caps and set the amps to max you were inadvertantly starving the after shunt cap and sucking everything that that 63v 470uf after shunt cap could provide.  Lots of noise and big ripples and more than likely exceeding the body diod of the fet with every dip in the DC rail volts drop over the after shunt cap.  

Seemingly the layout of the mod makes a difference.

Current can only move in one direction and little was coming from the gm shunt because of your first bypass shunt.

If the resistance of the PCB tracks and GM shunt is more than your first attempt to install the bypass, you were not using the input caps to there full ability.

Meaning you had 50v @ 120-15 amps of noise making switching being filtered by a single cap.  The software current limits in this situation may only better implemeted by a proper shunt soldered in the correct path of current.

How one would approach these difficulties would be as follows.

Beef up the tracks from you power pole where your current travels to the caps then to the shunts. Do not try to bypass any of the tracks with wires as you may create too many divisions of the current pathway.

Use a 30 amp rated shunt to the noise filter cap.  Set your max amps to 30 amps via the sofware.and continuous to 20~25 amps.

[Leslie]

I just did a simultion of a buz fet with a .2ohm load.



The 3.3 ohm resistor simulates the 15 amp shunt you had there.  And the 63v 470 cap is allowed to charge through the shunt and discharge into the motor.

The ocillocscope channels A is the voltage ripple allowed happen over the positive input of the motor and ground. And thh b channel is ground..

Its important to note in this simulation the current meter showing 13.8 amps exceeded the 15 amp shunt by 10 amps I saw it flickering at just over 25 amps peak.  For sure a fet with a low on resistance would do more peak from the cap than a buz fet.

The only thing I see doesnt work in the sim is the 3.3 ohm resistor compared to a shunt.  It still leaves a the of question to how the noise difference to the shunt and resistor would make over the 470uf cap.

As I decreased the duty cycle of the timer the current draw came closer to 14 amps and when I raised it the current draw was very spiked at 28 amps.

[Leslie]

If anything the sim shows how important it is to set you current max via the software to factory spec, or in Scorpion case his shunt was 15 amps.

The fets can create a lot of noise if the max current in the software is set to max and the load is allowed to over demand and out source the pysical GM shunt installed.

If you plan to mod the shunt in the controller like so many have done before set your max amps to match your installed shunt specification.

[o00scorpion00o]

Hi 317537, I have ordered a new usb cable, and this time also the external magic controller, just in case things don't go right. I can give the internal controller one more shot.

Do you think I can leave the shunts as I have them now if I can turn down the amps in the software?

I have 2 spare fet's and maybe the same one is blown again, if so I will replace it.

I think at the end of this experiment we might have learned a lot about this pie!

I got one word of caution to anyone modding their pie, My axel threads are wearing because I've taking the nuts off and put them back on so many times, the nuts that come with the pie are not made of strong metal at all. Fine for opening and closing maybe 5 times, but they are NOT made for tightening so much on an E-bike and the threads are gone long ago and I have being using proper steel nuts. Maybe GM should start shipping proper nuts!

[Leslie]

Hi 317537, I have ordered a new usb cable, and this time also the external magic controller, just in case things don't go right. I can give the internal controller one more shot.

Do you think I can leave the shunts as I have them now if I can turn down the amps in the software?

I have 2 spare fet's and maybe the same one is blown again, if so I will replace it.

I think at the end of this experiment we might have learned a lot about this pie!

I got one word of caution to anyone modding their pie, My axel threads are wearing because I've taking the nuts off and put them back on so many times, the nuts that come with the pie are not made of strong metal at all. Fine for opening and closing maybe 5 times, but they are NOT made for tightening so much on an E-bike and the threads are gone long ago and I have being using proper steel nuts. Maybe GM should start shipping proper nuts!

I think you should sit the shunt in place of the old shunt and solder up the tracks all round the power pole and caps there, You want a single pathway to your shunt.

Make it shorter than in your pics..

Set the controller to 20 amps cont and 23 amps max and see if you can pull 20 amps reliably, once you achieved this mile stone bump the max up to 30 amps.

I want to know whether its your new fet or old fets.  The GM ones may have been damaged in the first failure event.  Or the newer fet has damaged the older fets when it failed or the older fets damaged the newer fets when they failed.  this is why I would of replaced all fets.

No harm in trying.  Honestly you have some hard skin on you and 100 points for trying.

Good luck there matey,

[Leslie]

Here.



You can remove the PCB varnish above the brown cap and lay some thickish soldered tinned wire down on the PCB from the power pole to the cap and bend it to around meet the shunt and beef it up with solder so it attactched the new wire onto the PCB.

Watch out for the positive terminal of the brown cap there.

Install a proper rated 30 amp shunt and you got it right.

[Leslie]

The original shunt could of been fine to do more than 15 amps certainly looks better then 15 amps, its hard to tell core width from pics..  But it looks like it wasnt soldered properly in the first place.

Remove the goo glue and investigate.




The bottom right arrow I can see the PBC fibreglass below the tracks.  The track could be lose from its board.  Maybe!

[o00scorpion00o]

I don't have or know where I could get shunts of any kind in this country! lol honestly it's the hardest place to get anything! The fet's had to be ordered from the U.K!

317537
what about leaving the resistor wires as shunts? and trying the software to limit the amps or remove one of the three?

Also I did replace all 6 fet's. I just replaced the same one that blew last time.

I just want to say that I'm disgusted with the lack of support from GM. I e-mailed Tom, fair enough he replied, but is not interested in sending spare parts by the looks of it. yes I messed up my pie's cable, but at my own expense and I'm not asking for spare parts for free!!!
There should be spares available for the pie on the website!

[Leslie]

Did you install 3 15 amp shunts.

Here?

The bigger picture tells more.

Your idea leading the fets to caps via the fets is a good one.

Make the shunt as close to the same length.

I can still send you something rated and tested if you want.