Treat D2/V3 as a 3.6V zener...the lowest zener available was 6.8
Trace shows Q1 collector voltage
Thanks DUG. Waiting for new PCBs with independent coils.
Think the chip is somehow affected by the different tests. Right now I have independent coils on on channel and no coils on the other, and it still starts up as shown in the video, so I have to disconnect the speaker at startup ....
Asked TI if it could be run with common coils, and they said yes ..... maybe the slight difference I know was on the last coil between the 2 windings is enough to make it not work well ....
Think the chip is somehow affected by the different tests. Right now I have independent coils on on channel and no coils on the other, and it still starts up as shown in the video, so I have to disconnect the speaker at startup ....
Asked TI if it could be run with common coils, and they said yes ..... maybe the slight difference I know was on the last coil between the 2 windings is enough to make it not work well ....
You can try with one of the common coils reversed input and output. Probably can work like this but no benefit from this.
Update 🙂
Made a new version (8x9cm). Firstly it had the exact same problem!!!
Wondered about whether it is because it tries to start up too fast before all is properly settled.
I had put in an extra transistor for external control of the reset pin (e.g. for a DSP), used this with a 100k + 22 uF in series from 12V to the base, to give quite some settling time, and it works perfectly.
See updated schematics for the reset (on this version use either R11 or R12 .. not both 😉 )
Used 7G17D coils from aliexpress ... seems to work very well.
Have also tied a new dual coil DLM1623C which I will use for an even smaller version (7x7cm PCB)
Have mounted a alu plate for cooling using a 1 cm thick thermal pad ... also seems to work well .... bit counter intuitive to use something this thick ....
Made a new version (8x9cm). Firstly it had the exact same problem!!!
Wondered about whether it is because it tries to start up too fast before all is properly settled.
I had put in an extra transistor for external control of the reset pin (e.g. for a DSP), used this with a 100k + 22 uF in series from 12V to the base, to give quite some settling time, and it works perfectly.
See updated schematics for the reset (on this version use either R11 or R12 .. not both 😉 )
Used 7G17D coils from aliexpress ... seems to work very well.
Have also tied a new dual coil DLM1623C which I will use for an even smaller version (7x7cm PCB)
Have mounted a alu plate for cooling using a 1 cm thick thermal pad ... also seems to work well .... bit counter intuitive to use something this thick ....
Attachments
I personally would modify the schematic on the reset circuit:
1: 47K...100K from the reset pin to GND,
2: R5=22K,
3: Capacitor of 22µF...100µF from the base of Q1 to GND,
4: R11, R12, C27, Q3 deleted from the schematic.
The current through D4 may be too low for stable operation over time and temperature.
1: 47K...100K from the reset pin to GND,
2: R5=22K,
3: Capacitor of 22µF...100µF from the base of Q1 to GND,
4: R11, R12, C27, Q3 deleted from the schematic.
The current through D4 may be too low for stable operation over time and temperature.
Q3 is there to ensure ca 5 seconds delay on start up. This way it settles fully, and you have no pop at all.
Only down side is turning on off on rapidly will not ensure the large cap is discharged and the new start up delay will be shorter, and a pop will be heard.
Only down side is turning on off on rapidly will not ensure the large cap is discharged and the new start up delay will be shorter, and a pop will be heard.
The modifications I suggested do the same thing, slowing down the ON switching of the transistor that ensures the start. The capacitor at the base of Q1 delays the ON switching of the transistor and thus the start-up delay through its value.
I think it will slow down the startup, but also the turn off .... you want the first not the later