Bycicle LED POV reliability
kvas_off
assembly are much more reliable (in mechanical way, at least) than THT
ones (which are suggested to use in future hardware version 1.1). And
it's very important in such kind of applications like transported with
serious jerks and accelerations devices (even in simle remote control
units IR-LED fall off very often - this is a primary malfunction
reason).
There are two ways: low price or reliability. There should be a
compromise decision.
Casainho
Don't know If you read at wiki page of V1.1 hardware: "5mm LEDs are
larger, so the gaps between LEDs will be nearly non-existent. " --
looks like some large LEDs gives like a better light glow...
I apreciate your words of knowledge.
kvassoff, would you like to participate in the development? What do
you like/have knowledge, hardware and/or software?
Casainho
From: kvas_off
Date: Tue, Sep 16, 2008 at 3:06 PM
Subject: Re: [Bicycle LED POV - 198] Bycicle LED POV reliability
To: Casainho
>Don't know If you read at wiki page of V1.1 hardware: "5mm LEDs are
>larger, so the gaps between LEDs will be nearly non-existent. " --
>looks like some large LEDs gives like a better light glow...
Blue ones give the best glow, by the way.
Yes, I've read it. And that's why I posted this topic.
First of all, there still will be gaps anyway. Just look at SpokePOV
project: solid spaces of picture are stripy.
You can avoid it only if you'll use some optic diffusors or if you
take rectangular LEDs and place them like this:
//////////////////////////////// But rectangular LED is used just for
indication purposes; it hasn't built-in lens in its package and will
be very poor in luminous efficacy.
What I said about compromise?.. :) You can use SMD 5mm LEDs or usual
5mm LEDs, mounting them like SMD ones (with previously re-formed
outlets). Just an idea, don't kick me.
And I think, that batteries should be placed as near to the spinning
axis as possible (because of centrifugal forces and moments).
AND you could use widespread cell phone battery (nokias or smth.) It's
compact, flat - good :) But, again, little more expensive then 3 good
accumulating AA or AAA batteries...
Is there some way to use more powerful (but still cheap) current
drivers with one output for 2 parallel LED-resistor groups (one on
each PCB side)? No, I suppose. What do you think? Well, forget about
it :)
P.S. My words:
"...even in SIMLE remote control units/..." You call this "words of
knowledge"?! It's more like words of mistyping :)
BTW, if you interested: DIY ambilight for PC display -
http://alex.piter-furry.ru/index.php?page=w_ambilight
(in russian, but understandable :))
I can offer only my poor knowledge. :)
And I am not very reliable person in speedy and/or robust
developments... :(
DZiems
> First of all, there still will be gaps anyway. Just look at SpokePOV
> project: solid spaces of picture are stripy.
the cost.
> Is there some way to use more powerful (but still cheap) current
> drivers with one output for 2 parallel LED-resistor groups (one on
> each PCB side)? No, I suppose. What do you think? Well, forget about
> it :)
the LEDs, and use the current chips for that. It wouldn't work for
blue/white, but red, yellow, and green would. The reason we chose not
to is because we wanted to have each side display an independent image.
Casainho
> the LEDs, and use the current chips for that. It wouldn't work for
> blue/white, but red, yellow, and green would. The reason we chose not
> to is because we wanted to have each side display an independent image.
The reason for choosing that drivers were also because we wanted to
use just 2 batteries, and so we didn't use the Mikey Sklar Super POV
approach because would waste to much energy, 2*1.2v ~ voltage of LEDs.
kvas_off
> the LEDs, and use the current chips for that. It wouldn't work for
> blue/white, but red, yellow, and green would. The reason we chose not
> to is because we wanted to have each side display an independent image.
Of course, there is no need to mess with additional current chips,
numerous microcontroller ports, etc. Shift registers themselves are
good enough for that purposes.
How about cell phone battery (3.6V, 1000 mAh, for example)? At least 5
(10) hours of working in the average according to my calculations
(100% (50%) load-due to 100%(50%) filling image; 3mA to each LED from
64)... When I write "cell phone battery" I mean such form-factor and
third-party.
How about on-board charging via USB connector (from usual 5V PWM power
supply-charger for PDAs, GPSs, etc.)?
Jorge Pinto
> How about cell phone battery (3.6V, 1000 mAh, for example)? At least 5
> (10) hours of working in the average according to my calculations
> (100% (50%) load-due to 100%(50%) filling image; 3mA to each LED from
> 64)... When I write "cell phone battery" I mean such form-factor and
> third-party.
> How about on-board charging via USB connector (from usual 5V PWM power
> supply-charger for PDAs, GPSs, etc.)?
Well, if using the 3,6 volts 1000mAh battery means complicate the actual
schematic with one more IC (the charger), no, IMO.
Also 3,6 volts is very high voltage to source the LED with actual IC
drivers and resisters, would mean a lot of loss of energy, large
resistors :-(
Well, but If you have one idea, why not put it on schematic and show to
everyone? explaning the advantages, etc?? -- I think just talking we may
be missing something.
Anyway, we are in the midle of making the V1.1, If you want to
participate...? -- let's continue discussing but please make a practical
schematic of your circuit idea.
Can I add you as a member on code.google page?
kvas_off
> schematic with one more IC (the charger), no, IMO.
> Also 3,6 volts is very high voltage to source the LED with actual IC
> drivers and resisters, would mean a lot of loss of energy, large
> resistors :-(
Nobody cancelled voltage regulators (I know, p-n transitions, power
loss, additional components, etc.)
And, BTW, this is you who want to use 3 AAAs (for blue LEDs). Using
one-time alkaline ones is waste of money, so it'll be accumulators. I
guess you want to use not only 3*1.2V but also 2*1.2V. Anyway, AAAs
are bulky a little bit.
> Well, but If you have one idea ... -- I think just talking we may
> be missing something.
My idea was about handy charging and ergonomic widespread power source
that fit for design (that wouldn't jumping out during a ride :)).
> Anyway, we are in the middle of making the V1.1, If you want to
> schematic of your circuit idea.
> Can I add you as a member on code.google page?
kvas_off
"Due to the large number of mechanical connections present when using
spring-loaded battery holders, you may experience reliability problems
due to vibration, dust, moisture, etc".
How about refuse bulky spring-loaded battery contacts at all?
DZiems
Using li-ion/polymer is definitely the way to go, but the extra
circuitry needed to charge and prevent over-discharge is a bit
cumbersome and expensive, though very much required. I think perhaps
the way to do it, is to completely remove the battery from the boards,
and then create something that wraps around the hub. Batteries go on
one side of the hub, protection circuit goes 180 degrees to the other
side of the hub, and then wires are run from that to the boards. One
central battery that close to the hubs would minimize weight and
rotational inertia more than any other single improvement I think.
PS-I was working on the 1.1 schematic, but school has consumed my time
for the next 2 weeks, and I can just about promise I will not have
time to work on it until sometime next Friday at the soonest.