USB Step up voltage with LM3410

[Micro9900]

Hey guys, I could use some assistance. I am trying to get 40mA @ 20V using the LM3410 from USB.

However, by default, USB 2.0 can only deliver 5V @ 100mA, which is 500mW. And, 40mA @ 20V is 800mW. I know by using string descriptors I can get 800mW from USB 2.0. However, I need to understand a bit more about the LM3410 and I could use some help.

The reason I need 20V @ 40mA is to the backlight of an LED display. Here is a link to the LED TFT LCD screen:

https://www.newhavendisplay.com/nhd43480272efatxlt-p-5571.html

My Vin is going to be 3.3V @ 100mA and my Vout can be between 19.2 to 22V, so I am going with 20V @ 40mA.

Vin = 3.3V @ 100mA

Vout= 20V @ 40mA

Looking at the datasheet I calculated that the series resistance required is Rset= VFBmax/ ILED => 202mV / 40mA => 5.05 ohms.

To get the inductor value it is D= (Vout-Vin)/Vout => (20V-3.3V) / 20V =>.835

Ts => 1 / fsw => 1 / 525KHz => 1.9048 x 10^-6 seconds or 1.09048us

(525KHz used since this is the LM3410Y and not X)

L = (Vin/ (2*dela IL)) x DTs

delta IL = ??? it is unclear how to obtain this from the graphs and equations shown.

In sum, I am not sure how to calculate the inductor and capacitance values required as this part is a bit confusing to me. If anyone can assist me with trying to figure out how to properly calculate the required inductance and capacitance values it would be much appreciated. Thank you.

 

[kubeek]

First of all, this 100mA requirement is usually left dwelling in the specification of usb. All usb ports I have encountered have no provision for metering the actual current being used, only to trip somewhere over 500mA. So if you are not planning to sell this as a kit or something, I´d just go with it, or maybe get a self powered usb hub just to be sure.

 

[Micro9900]

Well, I need help with the LM3410 as far as the inductor and capacitors are concerned. I am trying to test 12V @ 40mA, which is 480mW and then scale it up when I have the proper current configured (or see if it will go without configuring the current). I am using 10 2W zener diodes with a VF of 1.2V (10*1.2=>12V) and each diode has a reverse voltage of 20V as a test circuit.

Right now I am only reading millivolts measuring across the first diode terminal to the last diode's terminal. This leads me to believe I either have it hooked up wrong or my capacitor, inductor, and possible resistor values are off.

I have the 3410Y and I am following the schematic with overvoltage protection on pg. 31/49:

https://www.electro-tech-online.com/custompdfs/2013/08/lm3410-1.pdf

I am using 22uH (was for 20V), but I am not sure if this is a viable value to use at 12V or 20V. I am aso using a 2.2uF (25Vmax) capacitor on Vout and a 10uF (6.3V max) capacitor on vin.

 

[ronv]

You get to choose ΔIL. I used 40% and came up with 28uH, but I didn't do exact calculations.
If you design it for 3.3 volts the 12 volts should be ok.
These can be real fussy about the PCB layout because of the high frequency so keep all the leads short.
In any case it probably should work so check it twice.

 

[Micro9900]

Thanks for the reply. I am testing this on a breadboard first. Let me post a picture and maybe someone can assist me as to what is wrong (maybe my leads are too long, etc). I'm going to be pulling directly from USB so,

Vin=5V, Vout = 12V (across zener diodes), Iout=40mA.

EDIT:

I've made a video instead (I apologize for the messy circuit, but due to the multitude of connections, it is sort of unavoidable):

https://www.youtube.com/watch?v=nXF-wmWdk8k&feature=youtu.be

Notes: First blue part is the 22uH inductor, the green part is the 5.05 ohm resistor, smaller blue part is a diode with .43 Vf, which is connected to the first of ten zener diodes

 

[ronv]

Kind of hard to tell. Maybe you could replace all the zeners with a 500 ohm 5 watt resistor to simplify it. It may be noise or maybe a wrong connection, but it is hard to tell from the video. Voltage readings at each pin might help, but I would see if I could get rid of all the diodes first.

 

[Micro9900]

Thank you for the reply. That seems like good idea. It is a bit of a mess, so replacing the bulk of the load with a resistor would simplify trouble shooting. And, 500 ohms at 40mA should mean that there is 20V across the resistor. I'll give it a shot, I just have to order one since I unfortunately do not have a 5W 500ohm resistor laying around. I'll also get two 2W 250 ohm resistors if I can for the thickness of the leads of the 5W resistor may be too much for the breadboard (I had a bit of a hard time getting the 5 ohm 3W resistor in there).

As far as voltage readings go, across all of the zeners and each individual zener diode I get in the tens of millivolts. The only place I see is across the schottky diode and it reads only ~100mV. In addition, I do not read 5V coming from USB. When I separate it from my test setup and just put the 5V pinout from USB to an empty column on the breadboard I then read 5V. I have powered microcontrollers in the past from USB and LEDs, so this has to be an issue with the test setup and not the port (also tried different ports to be sure).

 

[ronv]

Actually there is only 0.8 watts in the resistor so a couple of 1/2 watt ones would work if you don't leave it on to long once it's running.

 

[Micro9900]

Yeah, I would imagine they would get pretty hot. I could use two 1W flameproof 1K resistors which I could connect in parallel. I believe this should be a safe bet since this will be a current divider (each resistor sees 20mA which is 400mW per resistor).


EDIT:
I post a video and take some measurements and try to clean it up a bit

EDIT2:

Got it working! Okay the caps were in the wrong spot. I read ~17.5V (probably due to tolerances) across each resistor. I'll post a video and show measurements and you can double check me to ensure this is right. I'll have to retest with zener diodes as well and then I will go for the backlight of the display.

EDIT3:

May have burnt it out it now reads 4.5V across each resistor . Oh well, I have to order a few more... crap.