The quetion of TPS92661 OPEN and SHORT Detection

Hi Jarod,

When you break the connection from the IC pin to the LEDs, the IC now has no reference point for an open. All of the switches are floating relative to the stack voltage on their source connection. If you remove the connection, you have eliminated the reference point for the open detection circuitry and the Vds of the switch is now undefined in the switch open condition.

And just to be clear, your test presently shows a triple system fault (2 broken connections to the LEDs and a shorted LED. Was this the intent of your test?

A question for you: Are you utilizing phase shifting in the system presently?

The open detection can be tested by removing an LED from the string. You need to maintain the connection from the IC to the LED anode/cathode.

I would also suggest, based on your schematic, that you shift your 8 LEDs to be located on the lower 8 switches or the upper 8 switches. This will make your firmware much easier to code and improve your bandwidth based on the register mapping.

Best Regards,
Mike

Dear Mike,

Thanks for your reply.

I answered your quetions and there were still somethings to be confirmed with you, pls read the attachment.

Best Regards

JarodThe issue description of OPEN and SHORT Detection 2.pdf

Hi Jarod,

Your schematic is better now and will make your firmware engineer much happier. ;-)

To answer your questions:
1. You are correct. The decisions are made based on voltage comparators.
2. This is correct. If you were to sever both connections from the IC, the switch couldn't do anything to fix the open anyway.
3. When you say you are simulating, are you using actual hardware or a circuit simulation?
4. If all switches are trying to turn on at the same time, you have a possible race condition - based on system parasitics and operating point - as all switches are trying to dynamically change simultaneously. Depending upon your system, you can get false faults in this case because everything is moving at the same time. By phase shifting them, you remove much of the dynamics in the system and allow more accurate detection of faults.
5. It doesn't particularly matter. If you want the FET to be on so you don't get false faults, you can program ENOFF = 1 and ENON = 0. You can also just mask out the fault bits in your firmware and ignore those bits completely.

Best Regards,
Mike

Dear Mike,

Thanks for your answer and it is helpful for us.

And I have two questions to ask. please read the attachemnt.

Questions of TPS92661.pdf

Thanks!

Jarod

Hi Jarod,

Yes, shorting across D3 should allow you to detect a fault. If it is not registering, my first guess would be that you have not programmed the device for PWM mode (LEDON != LEDOFF and ENON = ENOFF = 1). The device does not report faults if it's not actively switching. Note that the open LED protection is always engaged, but the fault reporting only occurs while in PWM mode. The IC comes out of POR will all switches closed and therefor has no way to detect a fault. How have you programmed the D3 LED registers?

With regard to phase shift... In the phase shifted case, there is no decrease in light output. Each LED is independently programmed, and the datasheet trace is showing the interleaved stack voltage. In both oscilloscope traces, the light output (duty cycle) is identical, 128/1024, for each LED. The difference is that they aren't all turning off and on at the exact same moment due to the phase shifting. So, LED1 is turning on at TCNT = 0 and off at 128. LED2 is turning on at TCNT = 85 and off at (85 + 128) 213. LED3 on at 171 and off at 299. LED4 on at 256 and off at 384. If you were to take the same scope measurement across each individual LED, you would see that they all have the same duty cycle. The trace shown is the cumulative effect of each LED on the entire string. The thing to note is that with phase shifting applied, the most you ever change the stack voltage is 1 LED Vf at any TCNT increment.

Best Regards,
Mike

Dear Mike,

I will confirm with software engineer about the D3 LED registers programming today later and try what you mentioned.

About phase shifting, I would like to learn more.

I upload the attachment, please read it.

Quetions of TPS92661 Phase Shift.pdf

Thanks!

Jarod

Hi Mike,

I have three quetions to confirm with you. Pls read the attachment.

Questions of TPS92661 Step 8th.pdf

Could you give some suggestions for us.

Looking forward to your reply.

Thanks!

Jarod