LTC6952 controller with two LTC6953 followers in passthrough sync mode - synchronisation issues if all channels are active

Hello,

I've been testing the following configuration of a LTC6952 controller with two LTC6953s followers in passthrough sync mode as per the following schematics which is a modified version of the LTC6952 datasheet page 27 Figure 12 to suit my purpose:

LTC6952 modifications prior to the test:

      - Capacitor C83 installed for external REF

      - Channels 0, 1, 4 and 5' capacitors replaced with 0R links

LTC6953 modifications prior to the test (on both boards):

      - Capacitors C59 and C60 replaced with 0R links

      - Removed R24, R27, R26 and R30

      - Installed R25 (100Ω) and R37 (160Ω)

The image shows how LTC6952 is configured with the ACE GUI and how the two LTC6953s are configured using the same parameters for both devices.

When the devices are configured in the way shown from the image, when synchronisation is run on the LTC6952 device all four outputs are successfully synchronised within the devices specifications and their datasheets. The screenshot below shows that:

However, the problem comes when the powered down channels are being activated on any of the three devices. After some or all of the channels are activated and a synchronisation is run on the LTC6952 device the result is unpredictable and most of the times the LTC6953 outputs are not synchronised. The screenshot below shows that (all channels activated on all three devices):

Another observation is that if only one or two of the powered down (unused) channels gets activated then we still see synchronisation issues but the skew is not that high as if we activate more channels, in other words the more number of unused channels we activate the worse it gets. The screenshot below is with only two unused channels activated:

For reference I have been looking at this topic LTC6953 vs LTC6952 multi-stage sync - Q&A - Clock and Timing - EngineerZone and this topic LTC6952 controller with LTC6953 followers EZSync multi-chip sync issues - Q&A - Clock and Timing - EngineerZone

The last test that was performed was to terminate channels 8 and 9 on the LTC6952 with four 50Ω caps and activate only them but the problem was still there and similar result as the last screenshot shown above.

I think everything about the way the devices are configured, connected and measured is clearly described and shown but if I missed something I'll be happy to add it if you let me know. It'll be very much appreciated if anyone can suggest a reason for this issue, is there something I miss as a step in configuring the devices and if yes which device etc.

Thank you very much for the time.

Regards,

Todor