ADC12DJ1600-Q1: Switching time

Piotr F.

Part Number: ADC12DJ1600-Q1
Other Parts Discussed in Thread: HD3SS3411-Q1, HD3SS3411

Team,

Customer is trying to validate the ADC12DJ1600-Q1 and in order to do that theyalso have to measure the switching time for the MUX.

Below is a screenshot of the oscilloscope where:

Yellow - SEL signal
Green - Port A
Blue - Port B
Purple - Port C

Two questions regarding this topic:

First will be at what percentage of the signal do we have to measure this? 80%?

And the second one is it: do we have to measure the time until the signal B becomes high? So the second cursor (the one with a dotted line) should be a little to the left? So the time would be below the limit of 1ns (right now it is at 1.64ns).

over 3 years ago

0 Neeraj Gill over 3 years ago

TI__Mastermind 20541 points

Hi,

Can you please elaborate on what you are trying to measure. Which MUX switching time are you trying to measure?

Regards,

Neeraj

0 Anca Carasel over 3 years ago

Prodigy 70 points

Hello,

I'm trying to measure the switching time (tSW) between the channels of the MUX. The maximum limit of switching time is 0.5ns. What is the correct procedure to measure this parameter? What should be measured precisely?

Note: MUX is HD3SS3411-Q1 One channel Differential 2:1 Mux/Demux.

Thank you!

0 Neeraj Gill over 3 years ago in reply to Anca Carasel

TI__Mastermind 20541 points

Hi Anca,

I will move your post to correct forum. So that you can get you question answered regarding HD3SS3411-Q1.

Regards,

Neeraj

0 Anca Carasel over 3 years ago

Prodigy 70 points

Hello,

Could you give me a link, maybe? Thank you!

0 Neeraj Gill over 3 years ago in reply to Anca Carasel

TI__Mastermind 20541 points

Hi,

I have moved your post to the correct forum. You should get a reply soon.

Regards,

Neeraj

0 Brian Zhou over 3 years ago in reply to Neeraj Gill

TI__Guru**** 163031 points

this is test setup

0 Piotr F. over 3 years ago in reply to Brian Zhou

TI__Expert 3020 points

Hi Brian,

More information is needed.

--> HD3SS3411-Q1 datasheet does not give TSW_OFF time - waht is value for it?

HD3SS3411-Q1 datasheet gives TSW_ON as 0.5ns_max.

--> Please help to understand meaning of TSW_ON

Is it propagation time between switching port A between Ports B and C? (so this time is 0.5ns_max?)

As you can see on the scopeshot in the 1st post here, customer captured VSEL switching time vs. PortA/B/C - and this is much longer than 0.5ns. So how this stand vs. 0.5ns_max datasheet value?

This is important to understnad, to account right method to system validation.

0 Stanton Weaver over 3 years ago in reply to Piotr F.

TI__Intellectual 2850 points

Hi Piotr,

To determine the switching time, please measure at the following conditions mentioned below:

SEL going low, SEL voltage = 0.8V and measure when Port B data is enabled and Port A is disabled.

SEL going high, SEL voltage = 2.0V and measure when Port A data is enabled and Port B is disabled.

Basically look at when Port A starts or stops driving in relation to Port B, when the SEL pin is changed.

Best,

Stanton

0 Stanton Weaver over 3 years ago in reply to Stanton Weaver

TI__Intellectual 2850 points

Any update?

Best,

Stanton

0 Anca Carasel over 3 years ago

Prodigy 70 points

Hello,

I've tried all the measurements and procedures proposed by everyone of you. The best timing I could obtain is roughly about 2 ns that's still greater than 0.5 ns.

It seems that the last measurement described give us a timing of approximately 12-13 ns, which is way more than the maximum limit of 0.5 ns specified in the datasheet. Is there any other suggestion on what to approach this timing measurement?

If not, are there any consequences if the limit is not respected? Should we worry about it?

Thanks!

0 Stanton Weaver over 3 years ago in reply to Anca Carasel

TI__Intellectual 2850 points

Hi Anca,

Sorry you have not been able to get the max limit of 0.5ns.

Can you please describe your process of measuring? I think this may still be contributing to the discrepancies.

Best,

Stanton

0 Anca Carasel over 3 years ago in reply to Stanton Weaver

Prodigy 70 points

Hello,

The setup that we used consists of:

Power Supply Konstanter SSP 240-40
Oscilloscope from Tektronix – DPO71604SX with 16GHz / 100Gs/s
Active differential probes from Tektronix – P7716 with P76STFLXA tips

Picture 1: SEL is asserted high -> Port A to Port C

In the middle of Picture 1 can be noticed the moment when port A(green) is switching from port B(blue) to port C(pink),

Port A oscilloscope trace stops following port B trace and starts following port C trace. The transition time from port B to port C was considered to be the switching time.

Port A amplitude starts to increase because port’s C amplitude is higher than port’s B amplitude, so there is a rising edge on port A.

On the rising edge of port A, cursors were placed at 20% and 80% amplitude and the time difference between them was considered to be the switching time(>1ns).

Picture 2: SEL is asserted low -> Port A to Port B

In the middle of Picture 2 can be noticed the moment when port A(green) is switching from port C(pink) to port B(blue).

Port A oscilloscope trace stops following port C trace and starts following port B trace. The transition time from port C to port B was considered to be the switching time.

Port A amplitude starts to decrease because port’s C amplitude is higher than port’s B amplitude, so there is a falling edge on port A.

On the falling edge of port A, cursors were placed at 80% and 20% amplitude and the time difference between them was considered to be the switching time(>1ns).

Note: If the quality of the picture above is too low, I've already sent them to Piotr to full quality.

0 Brian Zhou over 3 years ago in reply to Anca Carasel

TI__Guru**** 163031 points

Anca:

you need to measure from 50% of SEL to 10%-90% of A.

Best

Brian

0 Anca Carasel over 3 years ago in reply to Brian Zhou

Prodigy 70 points

Hello,

If we approach the procedure you mentioned (to measure from 50% of SEL to 10%-90% of A) we'd obtain a value ~10ns, which is way too much. As you can see in the picture, the cursors are now at 1.4ns.