bitPimps.lixlink.com

On the last SE I FET modded, I noticed that the basic FET mod gave the usual boost in performance, but the gate jumper mod gave no noticable increase in performance. So I took a closer look, and I notice that they've started putting in a new transistor that essentially performs the function of the jumper mod.

I've labeled the new transistor as QD5. It allows the base of QD1 to be pulled down through its own path rather than through QD4, as in the 1st gen PCB. That provides a better pull down to get QD1 turned on.



The gate jumper mod normally works by connecting the MOSFET gates of QD1 and QD3 directly, which achieves basicly the same thing.

I've tested an SE with FETs and no gate jumper (using just this new transistor) and with the gate jumper, and there isn't any noticable performance difference, so it looks like the gate jumper mod is no longer necessary, or at least doesn't provide a much of a boost as it used to. This may help explain some of the improved SE performance.

That's cool to hear, thanks for sharing the info, but one question.

After I do that jumper mod, the car has a surge of power to the motor everytime I turn on the car. Does that happen with the stock ZZs now? Does it still happen when you do the jumper mod?

I know it's probably not a big deal, but that always bugged me.

crazydave wrote: After I do that jumper mod, the car has a surge of power to the motor everytime I turn on the car. Does that happen with the stock ZZs now? Does it still happen when you do the jumper mod?

Nope, it doesn't happen with the new setup. If you take a look at the schematics I've posted on my FET mod page, which I updated just last night, you can see why.

From my updated FET Mod page:

After performing the gate jumper mod you will probably notice that the motor will give a little surge of power when the car is first turned on. There is a small delay between power-up and the activation of the car's CPU, during this time the RDP1 pullup is holding QD1 off and QD3 on, which in turn turns on QD2, allowing current flow to the motor. By removing RDP1 this surge can be eliminated.

I've also added a bunch of additional schematics to show exactly what the different steps do to the h-bridgh circuit.

great work mate, your prowess in mapping pcb track layouts to a definable schematic is 2nd to none. I get lost after the first solder through hole!

code, as you have stated before , the duty cycle appears to be better in the latter versions of the SE. Perhaps (it's a true stab in the dark) these circuit changes are governed from changes to possible only 1 or 2 passive components. ie: resistor, capacitor....

if it wasn't for all the black shit over the pcb and the fact that 99% of smt caps are unmarked these days we might of had a way to possibly track changes in other parts of the circuit.

my theory is based on the fact that alot of this asic based stuff (custom rx/tx chipsets, etc, etc....) usually only requires a limited amount of external components.

like in the bcg where it was possible to change the IF frequency of the tx/rx combo. you would change the same resistor on the tx pcb and rx pcb and end up having a new seperate channel....it worked, but when tested against the same freq with the normal IF, the normal tx killed it's use.

cheers,

ph2t.