Driver turn on/off delay

[Mattaus]

Hi all,

Bit left field here so I may struggle to get answers on the matter but it never hurts to ask!

Can anyone suggest to me what sort of time frame we are looking at for turn on delay for flashlight drivers? To explain it a little bit when we hit the switch we don't suddenly get maximum power running through the driver circuit - there is a delay while all the parts (caps, MCU etc) charge/discharge/concanflubulate* before we get the desired result. Obviously this 'delay' is so small we don't even notice it.

I will Google this more to see what the human eye is actually typically capable of noticing but it would be nice to see what the cheap Chinese drivers are like in this regards, or possibly even some of the high end gear from the likes of pcb-components.de, taskled and so on.

For arguments sake the human eye cannot see moving PWM above 15Khz, which works out to an on/of period of about 4mS. So theoretically anything faster than this would be basically impossible to differentiate a delay between press and on. I'd argue you could take even longer given pressing and 'seeing on' is a different ball game all together to seeing pulses of light (pure sight differentials would be easier to see than touch and sight differentials). Do any of my ramblings make any sense?

- Matt

* I made this word up in case you can't tell

 

[lwknight]

Well, its not a thrystered strobe so it takes a few milliseconds for all the circuits to charge up.
Even is strobe mode they really kinda fade in and fade out. Its just really fast so it appears to be a strobe.

 

[Mattaus]

Well, its not a thrystered strobe so it takes a few milliseconds for all the circuits to charge up.
Even is strobe mode they really kinda fade in and fade out. Its just really fast so it appears to be a strobe.

'thrystered' :thinking:

Can't seem to find what that means, but from what you've said even strobe has a turn on/off delay so my numbers are nothing to worry about

 

[lwknight]

I miss spelled it.
Its " thyristor "

Basically a thyristor can charge up like a capacitor and deliver huge voltages without discharging completely.

BTW: I really like your new word

concanflubulate

I am going to try to get it going at work. It should be fun.

 

[Mattaus]

I miss spelled it.
Its " thyristor "

Basically a thyristor can charge up like a capacitor and deliver huge voltages without discharging completely.

BTW: I really like your new word
I am going to try to get it going at work. It should be fun.[/COLOR]

That explains it then . My drivers turn on/off delay (from cold or on for an extended period) is 1.9mS and a turn on immediately after a turn off is 3.1mS.

Those numbers look good to me!

I like to use concanflubulate in a sentence immediately after someone has spoken to me using far too many acronym's or unnecessarily complicated words. It can be tricky to do but I usually get a 'wtf' look followed by a 'I feel stupid' look. I work in the Aerospace industry which is full of self important know-it-all genius' so the opportunity arises often

- Matt

 

[mvyrmnd]

I like to use concanflubulate in a sentence immediately after someone has spoken to me using far too many acronym's or unnecessarily complicated words. It can be tricky to do but I usually get a 'wtf' look followed by a 'I feel stupid' look. I work in the Aerospace industry which is full of self important know-it-all genius' so the opportunity arises often

- Matt

I work in IT and come across similar issues.

Quite frequently people have Quantum discombobulation failures in their flux capacitors.

 

[tolkaze]

at less than 4ms, the turn on (especially with a forward clicky) would have the light fully up to brightness before the first click, and definately before the second click of a clicky since the light turns on somewhere between fully out and first click. I think it would have to be measured at the driver with some pretty sensitive gear since you wouldn't be able to know when the juice starts flowing at once side of the driver, and when it comes out "conditioned" at the other side.

 

[Mattaus]

at less than 4ms, the turn on (especially with a forward clicky) would have the light fully up to brightness before the first click, and definately before the second click of a clicky since the light turns on somewhere between fully out and first click. I think it would have to be measured at the driver with some pretty sensitive gear since you wouldn't be able to know when the juice starts flowing at once side of the driver, and when it comes out "conditioned" at the other side.

Yeah that's a good point and something I didn't really think about. Makes it even less of an issue unless of course the person in question uses the momentary function a lot, but even then my times listed previously put me in good stead.

 

[georges80]

There's more to the turn on delay than circuitry powering up caps etc. OFTEN, a driver designer will intentionally utilise the Soft Start feature of the switcher IC to slow the ramp up time. This reduces what would otherwise be very large current spikes from the battery which can trip protection circuits etc. It also reduces initial currents going through the inductor and other circuitry.

So, a driver that powers up quickly may not be a good choice in some applications. Furthermore, a driver may also utilise PWM to control strobe features or low level dimming. In some of my drivers, I'll power down the entire switcher core to turn off the LED (say a boost driver) and other times I may use the PWM to temporarily turn off the LED.

i.e. no simple answer and in reality not something that really matters. Certainly in the case of my drivers, you won't be able to click them on/off fast enough for the LED to not momentarily light up - unless you have bionic enhanced reaction time

The original post has some errors, you don't need 15kHz to stop flicker effects and that is NOT a period of 4msecs. Above about 100Hz it becomes difficult to 'see' flicker and I tend to use at least 200Hz when I use PWM (often I run 400Hz). At 200Hz that's a 5msec period.

This shows the PWM response time of a H6CC driver. Takes about 54uSec to reach full regulation (3A in this case) from PWM being enabled and about 26uSec to reach 0A on PWM off. Current measurements made with a 50MHz current probe. Takes some fancy gear to 'see' how fast some of this stuff really is...

cheers,
george.

 

[Mattaus]

All good points George. Sorry for some reason when I did my maths earlier I used Hz as being per minute, not per second. Very stupid mistake

My driver does have soft start features for exactly the reasons you list. I'm still learning how it works so forgive the noobness!

As for my frequency numbers I based them from these sources. I couldn't find much that specifically related to to our scenario but the maths and explanations stood up to my admittedly uneducated (in this field at least) brain:

http://physics.stackexchange.com/qu...man-eye-flicker-perception#comment44989_19042

Of course I took worst case in my driver, as many point out in the above thread that 100Hz and above is fine. I should note my driver is constant current for brightness, but uses PWM for strobe, SOS and beacon etc modes.

All in all I'm still reasonably happy with the on/off time of the driver

- Matt