OMAPL138, McASP clock divider

Other Parts Discussed in Thread: OMAPL138, ADS1274

Hello,

I have a customer using an OMAPL138 and ADS1274 and they are running into synchronization issues with the McASP bus. The data to the OMAP is either over-running or under-running. They have set up their clock source of the A/D to be 15.737 MHz to get 61kS/s on 4 input channels. The customer is using the time-division multiplexing to have all 4 data outputs on one signal.

The customer says that the 15.737 MHz SCLK line needs a 256 divider on the OMAP side, but the McASP interface's divisor goes from 0 to 32 only.

Please advise whether the McASP bus can be used for the 61kS/s output of the ADS1274. If so, how can the two parts synchronize?

Thanks

over 12 years ago

0 Brad Griffis over 12 years ago

TI__Guru*** 125430 points

Can you please show a diagram of how the two devices are connected and which device is driving which signal?

Faraz.F said:
They have set up their clock source of the A/D to be 15.737 MHz to get 61kS/s on 4 input channels. The customer is using the time-division multiplexing to have all 4 data outputs on one signal.

For example in this case it's not clear WHAT is the data source. Is it the OMAPL138? Is it an external clock? Does that external clock then connect to both devices? The diagram requested above would eliminate all confusion here and then hopefully we can make some suggestions.

0 FF over 12 years ago in reply to Brad Griffis

Prodigy 245 points

Thanks Brad,

Left side is ADC the Right side is OMAPL138

0 Brad Griffis over 12 years ago in reply to FF

TI__Guru*** 125430 points

Your image didn't come through. Instructions for posting images on e2e can be found here:

http://e2e.ti.com/group/helpcentral/w/e2e/149.aspx

0 FF over 12 years ago in reply to Brad Griffis

Prodigy 245 points

here it is,

Thanks

0 FF over 12 years ago in reply to FF

Prodigy 245 points

Hi Brad , Did you get the attached file? Thanks

0 Brad Griffis over 12 years ago in reply to FF

TI__Guru*** 125430 points

No. Is the picture showing up when you look at the thread? How about when viewing the thread from another computer?

0 FF over 12 years ago in reply to Brad Griffis

Prodigy 245 points

5488.omap_L138.pdf

the left side is ADC and the right side is the OMAP

thanks

faraz

0 FF over 12 years ago in reply to FF

Prodigy 245 points

HI Brad,

could you open the file?

thanks

0 Brad Griffis over 12 years ago

TI__Guru*** 125430 points

Yes, I can now see the connection. However, there are many questions unanswered by the schematic:

How are the various MODE/FORMAT pins being driven, i.e. can't tell since it's simply connected to GPIO?
What is driving the CLK pin of the ADS1274, and what is the frequency of CLK?
I assume the SCLK pin is being driven by the OMAPL138 at 15.737 MHz, right?

Why is SCLK so fast? My calculations for the desired SCLK frequency are:

61000 samples/sec/channel x 24 bits/sample x 4 channels = 5.856 Mbps

So in other words I believe your ACLKX should be 5.856 MHz rather than 15.737 MHz. Then you can set XMOD = 4 and everything works.

Faraz.F said:
The data to the OMAP is either over-running or under-running.

Are they using CLK and SCLK signals derived from different sources? If so, you're going to have clock skew issues that will cause this exact behavior. Both CLK and SCLK must be derived from the same parent clock in order to maintain synchronization.

Processors

Processors forum

OMAPL138, McASP clock divider