TPS92691EVM-001: Design Inquiry

Hello Christian,

The EVM limitations are likely somewhat arbitrary as far as the input/output voltages being dependent on the switch voltage. In a SEPIC the switch voltage is Vin+Vout and it is a 100V FET. That is easy to rectify for your voltage needs.

Your more difficult task would be driving a FET that can provide 100W. It can be done though with good layout, heatsinking, and external component choice.

Whether you use SEPIC or buck-boost really depends. The buck-boost is a more simple circuit requiring only 1 inductor, but the LEDs are referenced to Vin or Vout depending on where you sense the current. You can use the boost calculator and just enter Vin+Vout for the output voltage to get correct calculations of component values. The SEPIC requires 2 inductors (or a coupled inductor ideally) but it allows you to reference the LED string to ground which is desirable in a lot of applications. I will see if there is a SEPIC calculator available and let you know. But you would get close with the calculated values for the buck-boost if not.

Regards,

Clint

Hello Clint,

Thank you for your reply.

My next concern is the PWM dimming for my down light.

Here are the specs of the down light that I'm using:
Vf = 25.6V @ 250mA
LED Power = 6.5W

My PWM is set at 200Hz (3.3V) and I have set the Iadj to 350mV so I can set the output current to 250mA. When I set the Duty cylce @ 100%, I'm getting 245mA. But when I set the duty cycle @50% the down light will be completely turn off. Maybe my settings are wrong. Hope you can give me some advice.

Thank you,

Christian

Hello Christian,

It's hard to say, I am not aware of any issues with PWM dimming from my own experience, but I will check with the systems engineer that has the most experience. But nothing noted as far as I know. It may be possible that it has something to do with such a low IADJ voltage. That results in a 25mV current sense voltage and at that low of a setting offsets could come into play and/or the COMP voltage may not hold up as well. Have you tried using a higher IADJ voltage and different current sense resistor to see if there is any improvement?

If that isn't the issue I would be happy to look over your schematic and layout for issues if you want to provide them. FYI, I checked with the systems engineer regarding a SEPIC calculator and it should be available on the web in the next week or two at the latest.

Regards,

Clint

Hello Clint,

Thank you for your reply.

That's what I'm thinking too. Because the IADJ voltage is set to low when I set my duty cycle from 100% to 50% the current drops from 250mA to 0. I tried setting the IADJ voltage to 700mV @ 500mA, I can dim the 6.5W LED properly.

As of now, I still don't have a schematic and layout for this design. I'm using the TPS92691EVM-001 dev board. Maybe I need to change some components on the dev board to dim the 6.5W LED properly?

Also, I'm planning to use a PWM signal to adjust the IADJ value. What would be the required frequency range to drive the IADJ pin via a PWM signal. Yesterday I tried a 200Hz signal but I can't seem to make it work.

Thank you,

Christian

Hello Christian,

For best accuracy and range for either PWM or analog dimming you want to set your maximum current using the maximum IADJ voltage of 2.4V. This gives you a current sense voltage (CSP-CSN) of 2.4V/14 = 171mV. Then change the sense resistor value to R = 171mV/250mA = 0.684ohm (or the nearest standard value).

For PWM to analog dimming the frequency depends on how much you want to filter it. On the EVM there is already an RC-RC filter for this purpose using TP4, but it probably isn't enough filtering for 200Hz. You want a DC voltage on IADJ so you will need to filter the PWM signal so the frequency you want to use determines how big it is (or how small if you go to higher frequencies).

Regards,

Clint

Hello Clint,

That sounds good.

For my design, I'm planning to drive 3 different LED.
Here are the specs of the 3 LEDs:
1. 24W LED
I = 1.05A Vf = 22.8Vdc

2. 29W LED
I = 500mA Vf = 58Vdc

3. 6.5W LED
I = 250mA Vf = 26Vdc

So if I changed the Sense resistor value to 0.684Ohms, I just need to adjust the Viadj value to get the 500mA and 1.05A values.

Thank you,

Christian

Hello Christian,

It really depends on what you need for each. For example:

1. If you want (or need) the same circuit with the same values to do all three then you would use a sense resistor value to get 1.05A at a maximum IADJ voltage of 2.4V. Then you could adjust the IADJ voltage down to get 500mA and 250mA. However, this will provide maximum analog dimming range for 1.05A but a more limited dimming range for 500mA and 250mA.

2. If you want (or need) each of them to have maximum dimming range then you would select a different sense resistor value for each to get 1.05A, 500mA, and 250mA respectively at the maximum IADJ voltage of 2.4V. Then all three will have full analog dimming range. This is the case that you would select 0.684ohm for 250mA and approx. 0.341ohm for 500mA, etc...

Regards,

Clint

Hello Clint,

We are really aiming for option 1. We want the output current settable (100mA-1500mA) so we can drive any kind of LED. So the challenge that I'm going to face here is dimming the 250mA LED which has the limited dimming range. Unless I lower the maximum current that we need to 1050mA. Or maybe there is still some other way to maintain the 1500mA max and dim the 250mA properly.

Thank you,

Christian

Hello Christian,

That is a wide range, it would make it impossible for any regulator to have range on the low end.

Consider if you want to get up to 1.5A as you mention for a maximum current. That would result in a sense resistor of 0.114ohms. To get 250mA then you would have to reduce IADJ until the current sense voltage is only 28.5mV. So you need to consider two things:'

First, the offset voltage of the error amp could be around 5mV over all conditions, so the accuracy at 250mA would be very bad from circuit to circuit.

Second, you are already dimmed nearly as low as you can go for the 250mA circuit, you might get 4:1 at best and accuracy will drop the lower you go.

This is the reality with all LED drivers. All error amps will have some offsets so if you want different circuits with vastly different currents you will need to change the sense resistor value to get the best performance from each. I know that makes things a little more difficult with PCB variants and whatnot, but it's worth the tradeoff for the increased performance in most cases.

Regards,

Clint

Hello Clint,

Thank you for your reply.

Do you have an Altium file for the schematic and layout of the TPS92691EVM-001 dev board? I'm planning to add the circuit on my Altium project.

Thank you,

Christian

PWR735A (11-6-2015 7-08-08 AM).zipHello Christian,

No problem. They are attached.

Regards,

Clint

Hello Clint,

Thank you very much.

Christian