L6480 bootstrap charge pump stop after write of configuration register

[Phlip Cornelissen]

Posted on January 23, 2017 at 11:35

The L6480 controller charge pump frequency stops after a write to the configuration register. A pulse on the reset pin seems to do nothing at all. Is there a proper sequence to restart the charge pump frequency? How can I fix this?

Controller setup:

16Mhz internal clock with no output, OSCOUT, OSCIN disabled

PWM frequency devision factor = 2

PWM frequency multiplication factor = 1

Config register value: 0x2F80

[Enrico Poli]

Posted on January 24, 2017 at 17:13

The charge pump is disabled in the following conditions:

1. Device under reset
2. Device supply below UVLO
3. Wrong clock configuration (i.e. setting an external clock source with no actual clock available)

Cases 1 and 2 can be easily checked measuring the supplies and the STBY/RESET input values.

Case 3 should be recovered resetting the device (or through a power cycle), so check if your reset pulse is long enough to actually reset the internal registers.

[Phlip Cornelissen]

Posted on January 25, 2017 at 06:43

Cases 1 or 2 were not active. I then read the initial boot value to change one setting at a time to figure out which setting stops the charge pump, it was either PWM settings (multiplication or division factors), so a change in PWM frequency stops the charge pump. I tried a 30ms Reset pulse (datasheet recommends 500uS) but it did not sove the problem. Any recommendations?

[Enrico Poli · Accepted solution]

Posted on February 01, 2017 at 14:16

Ok... This is quite strange.

Lets proceed step by step:

1. Turn-on the device and check the charge pump: it should be operating.
2. Clear the STATUS register using GetStatus (UVLO failure is always asserted at power-up) and then read the STATUS register again... UVLO should be no more present.
3. Change the CONFIG register

   only

   . Is the charge pump still switching?
4. Clear the STATUS register (GetStatus). Probably a UVLO failure is reported cause the internal regulator takes some time to rise the VCC up to 12 V (above new VCCUVLO  threshold). R

   eading the STATUS register again the UVLO should be no more present.

5. Reset the device and read the CONFIG register: Is the configuration back to the default (2C88)? Is the charge pump switching?

[Phlip Cornelissen]

Posted on February 02, 2017 at 13:32

I did run through the whole sequence and found the problem, a double swop of the MSB and LSB values for the configuration register, instead of writing 2F80 the value written was 802F hex, problem now solved, Thanks!