Q8, PMS SEND TO THOSE WITH ISSUES BLF soda can light

TomE was so kind to set me up with parts for a Q8 v3 driver. The intent was to put the driver through its paces and to verify proper electrical operation.

The Q8 driver is basically a FET+1 driver (direct drive with the addition of one 7135 chip for regulated low modes). The basic design is the same as other BLF DD drivers. We did try to improve the design somewhat, notably by cleaning up the power feed to the MCU chip (proper decoupling layout) and switching the FET ‘softer’. This was to mitigate instability seen with this design in high-power applications, particularly when using the more temperamental tiny25/85 MCUs.

The driver did prove to be electrically stable in TomE’s tests, but what I saw on the oscilloscope was still less than ideal. The good news is there is a very simple solution and after this small addition I think we are ready for production. I will post more information on this solution in the attiny25/45/85 thread for those interested.

I used this setup for the driver tests:

The emitters are three XM-L2 U2s with a lone XP-L from a Manker A6. The cells are four 30Q’s (frequently recharged!). The wiring is mostly AWG 14. The setup gets 15.5 A on a set of fresh cells.

This animation gives a conclusion of all the testing (and double-checking in SPICE) that was done:

The traces show voltages in various parts of the circuit. The PWM duty-cycle is at around 30% here (so about 30% light output in theory). The PWM frequency is at 15.6 kHz, as set in the firmware of the MCU. The LEDs are powered when the power transistor is on. This is where the purple trace is at zero (0 V is at the first division from the bottom for all traces). Pay attention to the yellow trace. That is the voltage present at pin 8 of the MCU. The improvement is remarkable in the second image of the animation. These MCUs are rated for 6 V absolute maximum, and prior to the modification we are right there when the cells are fresh. The added resistor isolates the main capacitor of the driver from the power part of the driver, breaking the resonant ringing and resulting voltage boosting we see in the first image. This ringing was occurring every time the LEDs are turned off, at the end of each PWM cycle. In hindsight the solution was fairly obvious but I think it is valuable one, applicable to all DD drivers, brought along by the Q8 project for the benefit of BLF.

EDIT: Took out the scaling on the animation to clear up the text.