Hi All
Does anyone know the KL28's DMAMUX/TRGMUX operation well enough to work out its details with respect to periodic DMA triggering based on LPIT channel(s)?
First of all I have tried to represent the operation in a simplified form as follows:
The periodic trigger functionality is available on the first 4 DMA channels (this is the case generally for Kinetis parts) and the single "always-on" slot is slot 63, called DMA MUX.
In other Kinetis parts the always on slots are connected to PIT sources but the KL28 seems to route via the TRGMUX to the LPIT, as is shown in the diagram (note that all other peripheral slots look to be connected directly to their corresponding peripheral trigger source).
The TRGMUX_DMAMUX0 is thus presumably dedicated to the DMA_MUX slot in the KL28. Also, this can connect to 4 trigger sources (SEL0, SEL1, SEL2 and/or SEL3), which can be either channels from the LPIT or various other peripheral triggers.
This all makes sense apart from the fact that I doubt that more than one trigger source would be used at a time (?)
The only reference that I can find to back this up is this extract from the user's guide:
This looks to confirm that the LPIT trigger really passes through the TRGMUX but it is confusing due to the following reasons:
1. It states that LPIT0 channel's 0 and 1 could be used, whereby the TRGMUX allows any of the 4 channels to be routed.
2. It limits to DMA channels 0 and 1 whereby periodic triggering should be possible on channel's 0 to 3.
3. It fixes LPT0 channel 0 with DMA channel 0 and LPT0 channel 1 with DMA channel 1, whereby there is only one (always on) DMA MUX slot, which should be able to connect to any DMA channel.
Therefore I am also wondering whether this extract is really correct or whether the table is a cut and paste from a different Kinetis part that has a fixed slot/channel/PIT source arrangement.
Any one with more details, explanations and/or experience with this?
Regards
Mark
Hi All
I have results from some practical tests to add:
- LPIT0 is sent to run at 10kHz
(channel 0, 1, 2, or 3 are tested)
- The corresponding PIT0 channel trigger source is selected in TRGMUX_DMAMUX0 and the DMA Mux set to use this in period triggered mode.
- DMA channel 0 is set up to copy samples in a sine wave table to the DAC output each time it is triggered.
The result is that any LPIT channel (0, 1, 2, or 3) can be used to do this, verified by the sine wave appearing at the expected frequency on the DAC output.
Repeating the test on other DMA channels (1, 2 or 3) which "may" support periodic triggered mode according to the user's manual shows that none are successful - the DMA trigger never occurs.
Comparing again to the following:
it suggests that it is misleading:
LPIT0 Channel 0, LPIT Channel 1, LPIT Channel 2 and LPIT0 Channel 3 could be used on DMA0 Channel 0.
None could be used on DMA0 Channel 1 (nor other DMA channels).
What is the official stance on the accuracy of the user's manual in this case and whether the results of these tests reflect the expected behavior?
Regards
Mark
