DS90UB964-Q1: DS90UB964-Q1: AEQ SP and EQ setting related issues

Hello Lucy,

Lucy Lee said:

. TI customers have no problems using the TI default method. Please provide the TI default procedure and parameters.

Not sure what default procedure and parameters are you referring to.

Lucy Lee said:

2. TI displays "lock complete" after testing a certain number of parameter values.

How many values ​​are needed?

Where is TI displaying that? and what parameters do you mean?

Lucy Lee said:

3. Does AEQ need to be re-run every power-up?

AEQ is active all the time. You do not need to do anything.

Every time you lose LOCK or power cycle your device it will start scanning for the best AEQ values.

Lucy Lee said:

4. What are the fastest and slowest lock times for AEQ?

That depends on your system. But there are typical and max numbers in the UB964 datasheet.

Lucy Lee said:

5. We want to prevent unlocking or screen flickering after power-on. If unlocking occurs, it will re-scan. Will this cause screen flickering?

If you lose lock that means your camera data is not anymore being sent to the display which results in flickering. The best way to prevent flickering at power up is to keep your display's back light turned off until you have stable LOCK.

Lucy Lee said:

6. Is it possible to achieve a flicker-free screen?

It depends on your system design and transmission link quality. If these are good enough and fulfills our channel specifications, then you should not see any flickering if you do what I recommended in last answer.

Lucy Lee said:

7. Our company scanned 20 systems, and the EQ values ​​are biased upwards compared to the TI default. Is this normal? How should we adjust it?

TI default: TI964 default
SP: Reg [0x41] = 0xA3 è 3 ~ 10
EQ: Reg [0xD5] = 0xF8 è 8 ~ 15 Scanned 20 systems using TI tool With the 4M/15M cable length shipped, the SP/EQ range is as follows:

(SP 5~8, EQ CAM1~6=5~7, CAM7~8=6~8)

Register 0x41 is reserved and should not be used on the UB964.

Register 0xD5 is a write register, so you are reading what you are writing.

The AEQ status can be found on reg 0xD3

Lucy Lee said:

8. After verification, our company found that the EQ range of two ICs (with the same cable and camera) is significantly different. What range difference is considered acceptable?

What register are you talking about? and what are the values?

Additional description Item 8:

We're using the MAP tool's margin analysis to verify the EQ status of our products. We found that the results for the first two channels of a DS90UB964-Q1 were poor. However, after replacing it with a new DS90UB964-Q1, the results improved.
We'd like to have the poor chip with the poor EQ results tested and analyzed to see if there are any abnormalities.

How many levels of difference between SP/EQ are considered acceptable?

Lucy Lee said:

How many levels of difference between SP/EQ are considered acceptable?

All the shown results are acceptable. Minimum acceptable is a 2x4 block of green boxes, similar to this example:

Hi Hamzeh,

item8,

How much difference in SP/EQ between two ICs is considered acceptable?

Hi Lucy,

There is no acceptable range differences between ports, because this may vary depending on the trace layout + connector + cable + camera of that specific port.

If you have a working result as shown above that should be okay. Important is not losing Lock or accumulating errors.

These two results are based on the same PCB/cable/camera; the only difference is the TI964 IC. Why would I want to understand the differences between the two?

Thanks for the clarifications. As I said, both results are okay to me. One IC has more margin at higher AEQ values though, increased AEQ from 8 to 10.

One question; why do you call this as Failing IC? what is it failing in your test exactly? 

It should be described as a TI964 IC with a smaller margin. Understanding that both are reasonable and acceptable results, what could be the reason for the significant difference in margin between the two?

As said earlier, anything on the link can lead to this difference also soldering or AC-coupling capacitor can make a difference. Is this a single device behavior or multiple devices?

Can you provide register dump from a good device and suspect device? Make sure to dump the registers of the relevant RX Port.

BTW, I do not see this as a significant difference in margin.