RE: MSP432 and TI-RTOS: PID Library Part 2 - Real World Example

DAB — Thu, 26 Jan 2017 20:25:41 GMT

Great post Jan.

Nice review of filter theory too.

DAB

RE: MSP432 and TI-RTOS: PID Library Part 2 - Real World Example

jc2048 — Thu, 26 Jan 2017 11:25:16 GMT

Would it be alright if I make an observation here? You'd do much better with a second-order filter (it only adds two more components).

Here's your filter circuit

[View:/resized-image/__size/396x164/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-ae92ef07-f7ca-4a53-ab97-e6d94ec61bef/0172.contentimage_5F00_183771.jpg:396:164]

this is the frequency response

[View:/resized-image/__size/620x408/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-ae92ef07-f7ca-4a53-ab97-e6d94ec61bef/4314.contentimage_5F00_183772.jpg:620:408]

you've had to position the -3db point at 10Hz to get a response that's down by 60db at the PWM frequency in order to make the ripple reasonable - here's the ripple

[View:/resized-image/__size/620x398/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-ae92ef07-f7ca-4a53-ab97-e6d94ec61bef/3718.contentimage_5F00_183773.jpg:620:398]

but with the -3db point so low the transient response is poor

[View:/resized-image/__size/620x352/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-ae92ef07-f7ca-4a53-ab97-e6d94ec61bef/7041.contentimage_5F00_183774.jpg:620:352]

Here's the circuit of a second-order filter. I've chosen the values so that the -60dB point is roughly similiar, so we should get about the same ripple

[View:/resized-image/__size/532x164/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-ae92ef07-f7ca-4a53-ab97-e6d94ec61bef/6445.contentimage_5F00_183775.jpg:532:164]

here's the frequency response

[View:/resized-image/__size/620x412/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-ae92ef07-f7ca-4a53-ab97-e6d94ec61bef/3302.contentimage_5F00_183776.jpg:620:412]

the ripple is similar

[View:/resized-image/__size/620x398/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-ae92ef07-f7ca-4a53-ab97-e6d94ec61bef/1258.contentimage_5F00_183777.jpg:620:398]

but now look at the transient performance - it's about ten times better. Much more agile and responsive - though it does mean your loop tuning will be more challenging to control it.

[View:/resized-image/__size/620x352/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-ae92ef07-f7ca-4a53-ab97-e6d94ec61bef/0753.contentimage_5F00_183778.jpg:620:352]

RE: MSP432 and TI-RTOS: PID Library Part 2 - Real World Example

jc2048 — Wed, 25 Jan 2017 22:46:27 GMT

"The 10K resistor in parallel with the capacitor is there to have a discharge path."

I didn't understand that bit. Why do you need a discharge path (other than the 5K resistor)?