TPS922052: Dimming ratio questions

Hi Elan,

The minimum dimming ratio of hybrid dimming is 1/(2^8) = 1/256 = 0.4%. Please refer to Section 8.3.5.3 Hybrid Dimming from the datasheet. You can refer to this E2E thread.

The description - (2,000:1 at 20-kHz PWM, 10,000:1 at 4-kHz PWM, 1,000,000:1 at 120-Hz PWM) only corresponds to flexible dimming mode. Sorry for the confusion. 

May I understand what is your target dimming ratio and PWM frequency?

Best Regards,

Steven

Thanks for the useful thread link! I would be interested in obtaining the beta tool for component computation as well, if possible.

Ideally I'd like 10000:1 at 4kHz, but 2000:1 at 20kHz might work as well. Dim-to-dark is really important, so making it work at low levels is important, would love any pointers on how exactly to use Flexible mode for this.

Hi Elan,

For the beta calculation tool, I can share it with you. I have sent you a friendship request on this E2E forum. I will send you my e-mail address through private chat after you accept the friendship request. Please contact me through e-mail and I will share with you the beta tool.

Elan said:

Ideally I'd like 10000:1 at 4kHz, but 2000:1 at 20kHz might work as well. Dim-to-dark is really important, so making it work at low levels is important, would love any pointers on how exactly to use Flexible mode for this.

For your target dimming ratio, it is possible to achieve this with flexible dimming. The easiest way is to allocate 8-bit (256 steps) with analog dimming for part of your overall dimming ratio and then use PWM dimming to achieve the rest.

Let's take 10000:1 at 4kHz as an example. Excluding the 256 steps with analog dimming, you need to allocate 10000/256 : 1= 39 : 1 with 4kHz PWM dimming. This will need a minimum current pulse width of 1/4k/39 = 250u/39 = 6.4us. The rest of the work is to focus how to carefully select your inductor and output capacitor to meet this 6.4us minimum current pulse width requirement.

Best Regards,

Steven

OK, great, this gets to the heart of my confusion. I'm familiar with splitting ranges between PWM and analog dimming ("Hybrid Dimming") but this ends up with additive dimming ratio, not multiplicative.

The way you're doing the math makes it sound like you're combining 1:39 PWM * 1:256 to get 1:10000 total. I think that would imply some sort of mechanism whereby the analog and PWM are combined like least significant and most significant bits to get the total brightness, but I'm not really sure how that would work (hence my question).

Thanks so much for your help!

Hi Elan,

Elan said:

The way you're doing the math makes it sound like you're combining 1:39 PWM * 1:256 to get 1:10000 total.

You understanding of flexible dimming is correct. It is multiplicative.

Elan said:

I think that would imply some sort of mechanism whereby the analog and PWM are combined like least significant and most significant bits to get the total brightness, but I'm not really sure how that would work (hence my question).

The analog and PWM are not combined in the LSB / MSB way. They are multiplicative. For example, if you set ADIM/HD pin duty to 128/256 and set EN/PWM pin duty to 5/39, then you will get an overall output of 128/256 * 5/39 = 5/78 = 6.4% dimming ratio.

Best Regards,

Steven

Ah, thank you, that makes sense.

To be pedantic, is there a recommended strategy for how to combine the two? In your example the weighting was higher towards ADIM, but there are many ways one might do it, including equal weight, meaning 50% dimming ratio = 71% ADIM * 71% PWM = 50.4%, and 25% dimming ratio = 50% ADIM * 50% PWM = 25%.

Hi Elan,

How to combine the two ratios will depend on you and there is no recommended strategy. 

Best Regards,

Steven

I just wanted to thank you for the great support here, Steven.