How to perform external Triggering in TSW1400 and provide same external reference clock of 10 MHz to TSW30SH84 EVM and TSW1266 EVM

Girish,

Are you trying to trigger a TSW1400+TSW30SH84 and a TSW1400+TSW1266 together?

In general the triggering modes are used for a single EVM in an event based mode (trigger in or trigger out) or synchronizing multiple DAC or ADC.

I do not believe we have documented how to perform a DAC master and ADC slave mode.  Is this what you are trying to achieve?

Ken

Hi Ken,

Sorry for the confusion. I discussed with my supervisor and I am re-framing the problem statement here:

1. As per your document, when we are putting external trigger ON in GUI, no trigger (i guess it should should look like strobe/impulse) appears in any of the 'sync port'. However, when we toggle the tab "external trigger" to OFF in the GUI, surprisingly a trigger is coming in sync-4. However other sync ports have still no trigger signal. I guess, it should be opposite as well as when trigger tab in put -ON, one should see trigger signal at all four sync ports (SMA connectors) in the board. Is this a GUI issue or hardware issue in your opinion?

2. We have two EVM's: EVM#1: TSW1400+TSW30SH84 and a EVM#2: TSW1400+TSW1266. Our purpose is to use the trigger signal generated by DAC in EVM#1 as marker (by connecting to external trigger in EVM#2)  for the ADC in EVM#2. 

3. The overall aim is to synchronize clocks of two EVMs using trigger signal for Digital predistortion application. 

Please advice.

Hi,

I have not used the 10MHz syncing input with the TSW1266, but the clocking on that EVM is shared with the TSW1265 and the 10MHz syncing was used with that EVM.  There are two configuration files that come with the TSW1266 EVM GUI, one for external clocking that passes an external clock through the LMK clock chip and the other that lets the LMK generate the clock internally and that configuration should already be set up to accept the 10MHz reference.   Since the external clocking setup uses the same SMA input for the external clock as is used for the 10M reference for internally generated clocking, the transformer coupling for the SMA input needs to handle a wide range of frequencies.  I think the bandwidth of the transformer used for the transformer coupling is not wide enough to pass the 10MHz well.  I note from the datasheet for the transformer that the transformer is not specified for under 50MHz so the transformer rolloff might be too great at 10MHz.

You could try increasing the amplitude into the SMA, but more likely you may need to swap out that transformer for one with a lower bandwidth that would pass 10MHz.

Regards,

Richard P.

Girish,

I have completed verifying that the trigger mode works as expected.  Based on your information it seems you want to use the DAC+TSW1400 setup and the ADC+TSW1400 setup to enable a trigger to drive the DAC pattern and ADC capture at the same time.

This can be done by enabling the Trigger modes in the DAC and ADC setups of HSDC Pro.  You can have either of the DAC or ADC be slave or master.

Both Slave:

In this mode you can just enable the Trigger On DAC and ADC setups within HSDC Pro.  Now on the DAC, you click WRITE DDR, this will load the TSW1400 buffer wit the pattern and arm the trigger on the TSW1400.  On the ADC by default it self arms and waits for a trigger after each capture.  To analyze the data you will need to click Read DDR memory.  Or if you set it to the 2nd ADC trigger mode, you can click Arm On Next Capture Button Press.  This would require you to click the Capture button which arms the ADC system and now it waits for the trigger to come in on J11 Trigger in.  As soon as a rising edge trigger comes to J11 on both TSW1400 you will trigger both the DAC and ADC.

EIther one as a Master:

In either the DAC or ADC setup Trigger options you can select Software Trigger Enable.  This means that it will generate Trigger pulse when the capture button is pressed on the ADC setup or the Generate Trigger Button is pressed on the DAC setup.  The ADC setup will issue pulses on SYNC1..4.  You will need to connect this trigger pulse to the Trig In on the same TSW1400 (trigger itself as it is still in trigger mode, just the pulse is generated on board).  The other 3 triggers can be connected to other TSW1400 slaves or to the DAC slave to trigger it.

If the Software Trigger is enabled on the DAC, it will only issue trigger pulses on SYNC1 and SYNC2 when the Generate Trigger button is pressed.  One of the SYNC must be used to self trigger the TSW1400 and the other can be used to trigger another DAC or ADC TSW1400 setup.

Only one of the TSW1400 can be setup as a master.

I have attached a ppt which I did which had the DAC setup as the master - it will generate a pulse on SYNC1 and SYNC2 when the Generate Trigger Button is pressed, one of which is used to trigger itself, and the other is used to trigger the ADC setup.  The ADC is setup as a slave (Software Trigger disabled) and you can either have it set for self arming trigger or trigger on clicking Capture.

See attached document.

Ken.

Trigger on TSW1400.pdf

Hi,

Checking in the LMK0480x datasheet, I see that the minimum amplitude for the 10MHz clock into the LMK is listed at 0.25V swing on each leg of the differential, or 0.5V peak to peak differential.  What amplitude do you see at the pins of the LMK? 

I'm told that if the amplitude is too low and if Holdover mode is set that the LMK0480x will essentially 'freeze' the frequency it is at instead of locking to the external 10M.  But we checked the configuration files that come with the SPI GUI and I believe the Holdover mode is disabled by our normal config file.   You are using the config file named LMK04808_614p4M_Dual_Loop_10M_Ref.txt?  This config file should set the Holdover mode to be disabled. 

If the 10M amplitude at the LMK is sufficient and you are using this config file, then we may need to consult our experts in the Clocking forum.

Regards,

Richard P.

Hello Ken,

I am also trying to achieve synchronization between transmitter and receiver for DPD application.

Previously, I was able to get good setup using TSW1266EVM and TSW30SH84EVM using altera FPGA in ohio state university. But I am not able to do the same while using TSW1400EVM.

I am attaching the file provided by Ti for previous setup, which worked nicely. 

May be you can suggest equivalent connections of this in TSW1400 EVM.

Thanks,

Meenakshi1205.Altera-TI Demo.docx

Meenakshi,

From the document it looks like you used the TSW30SH84 clock output to drive the TSW1266 clock.  This would work to allow synchronized sampling on the ADC and DAC.  The LO source is free running, I'm assuming that since the up and down conversion is with the same LO that any frequency or phase issues would cancel out.

What is the problem that you are seeing with the TSW1400?  The connections for the ADC and DAC boards should be similar.  Obviously the ARRIAV RF setup will be different from using 2 TSW1400.  However if the ADC and DAC boards are phase locked then you should be able to get a good capture.

Ken

Meenakshi,

You cannot use 307.2MHz when using PLL mode. If you are using clock distribution mode, this should not be a problem.

Regards,

Jim

Girish,

I was able to make the TSW30SH84 drive an external clock to the TSW1266.  Both the DAC and the ADC were running a data rate of 737.28Msps.  To prove that the system is in sync we would source a coherent buffer in the DAC side, and capture the same length on the ADC side.  If they are timing synchronized then the capture signal would also be coherent with that buffer size.

I was able to verify this worked ok by switching the TSW1266 external clock to another bench source and the captured signal became non-coherent.  Switching back to the clock from the TSW30SH84 showed coherency again.  This was done with no windowing in place.  Please have a look at the attached file which has the ppt which outlines the setup, as well as a TX buffer from HSDC Pro TSW1400+TSW30SH84 and a captured coherent buffer from  TSW1400+TSW1266.  Let me know if this cross correlation works better.  You will need to normalize as I did do anything to the captured data and it is somewhat attenuated due to the RF path.

Ken

temp.zip

www.dropbox.com/.../file_15.txt