Softstart- The correct choice?

[Overclocked]

Device being used: Max 1703 Boost Converter

Load=5v @ 400mA

I am using this IC for my flashlight, according to the datasheet:

(pg10) Do Not Apply Full Load until the output reaches 2.3V (max) in PWM Mode

(pg13) To Implement soft-start, set mode select low on power up, this forces the IC into PFM operation and reduced output current to 150mA. Once the circuit is in regulation, mode select can be set high for full power operation

The IC Has 2 Modes, Low Power mode (PFM) and high power mode (PWM). I assume that its Ok to start the IC up in Low power Mode and then switch over to high power PWM mode. I was going to use a switch to do it, But is there a low power circuit (Since this is a battery application, power and effciency are of concern) that will allow the chip to start in low power mode and then switch to High power mode?

To go to High power mode, a high level is to be placed on the mode select pin, to select low power mode, it has to be at a low. I assume a RC combination can be used, but I dont know how long the chip takes to start up (I assume its in the range of milliseconds).

Err wait, The time it takes to start up depends on the output capacitance doesnt it? I have for my design 220uf Caps in parallel (440uF in total). So Id just have to do a little math to see how long it would take to reach full charge and then, based on that time, I could implement a RC circuit with a longer time constant to insure the IC can start up.

 

[justDIY]

tie the power mode switch high using a capacitor (and resistor possibly), so you have an RC type delay ... this is a typical soft start circuit ... the chip will start in low power mode until the capacitor charges up past the threshold at which the pin is considered "high".

 

[Hero999]

Well the PFM mode limits the current to 150mA so the voltage rise time can be calculated from the following formula:
[latex]\frac{dv}{dt} = \frac{C}{I}[/latex]

The value in the RC circuit on the low power mode pin will also depend on the turn on threshold voltage.

 

[Overclocked]

tie the power mode switch high using a capacitor (and resistor possibly), so you have an RC type delay ... this is a typical soft start circuit ... the chip will start in low power mode until the capacitor charges up past the threshold at which the pin is considered "high".

The Max Low Level seen by the mode select pin is .2*Vout, So in mycase a Low would be 0v to 1V, and a the min voltage for a high would seen as .8*Vout which would be 4v to 5V. In highschool, we drew diagrams to illustrate the voltage inputs of CMOS and TTL Technology.

------ 5V
------4V
###
### Anywhere in between
###
------ 1V
------ 0V

Which means once it reaches 4V, The chip would be in PWM mode. Anywhere in between would be, Uh, Undefined. But what happens if you stay in that undefined region?

 

[justDIY]

i've got no idea. switchmode controllers are pretty robust items, if you want to worry about starting up in low power mode and then switching to high power, a simple RC circuit seems to be the most often recommended way to do it. Alternately if your flashlight contains a microcontroller, you could have it switch the pin for you after a few millisecond delay.

 

[Overclocked]

i've got no idea. switchmode controllers are pretty robust items, if you want to worry about starting up in low power mode and then switching to high power, a simple RC circuit seems to be the most often recommended way to do it. Alternately if your flashlight contains a microcontroller, you could have it switch the pin for you after a few millisecond delay.

Nope, no MCU. Just a SMPS. I would just have to find the right RC time constant to switch modes, then again, without doing any math, I could use a combo with a time constant of .5 seconds, which should be more than enough.

 

[Hero999]

I expect that 0.5s will be more than enough, I would use a 1M resistor and a 470nF capacitor.

 

[Overclocked]

I expect that 0.5s will be more than enough, I would use a 1M resistor and a 470nF capacitor.

I completely disregarded the fact that a RC circuit takes 5 time constants to charge up to 99%. I need to know at what time will the Voltage be 80% (since 80% of 5V is 4V, which is when the chip see's a high). Using a 2.2M ohm resistor and .1uF Cap will give me around 350mS of time at 80% and overall time will be around 450 mS, which should be enough.

I studied the data sheet a bit more, and noticed theres a power on delay of 7mS. Transients take about 4mS to settle (atleast looking at the graphs on the data sheet) which brings the total time to start up the chip to 11mS. Theres only one way to know for sure, Use a data logger

Luckly, I have a Data Q data logger. I can build the circuit, Have a switch to switch between modes and see how long it takes to start up in both modes. I could also incorporate a 100mA Load to see if it has any effect on start up time.

In terms of light, can a delay of .35 seconds be seen by the eye?

 

[justDIY]

yes, 350ms will be quite noticeable

are you sure the datasheet is talking ms = milliseconds not us = microseconds? 350 milliseconds seems an awful long startup delay...

 

[Overclocked]

Yes its in ms. 350 ms will be my delay from RC circuit, 7 ms is the start up delay (Well, Its actually 5 ms, but it seems the inductor settles after 2 ms) for the IC. I can probably reduce it to a 30 ms delay, but I want a bit of a margin of error just in case. OK, maybe 350ms is overkill....

The eye can see a Max of 30hz, So I would need a delay less than 33 ms. So A delay ranging from 20 to 30ms would suffice.

And I found out the exact technology this is based on; BiCMOS.

 

[justDIY]

I don't know about your eye, but my eye can see well beyond 30hz ... and peripherally I can see 'flicker' in lights and crt monitors up to about 72hz

but that's for things that are strobing / animation. anyones eye will notice the light starts out dim and switches to bright ... just sell it as a feature, not a bug - "soft start circuitry prolongs blah blah..."