H4 led

by burudis Fri 08 Jun 2018, 2:27 pm

https://es.aliexpress.com/store/product/Creadvent-Plug-and-play-alta-potencia-20-W-2000lm-H4-motocicleta-led-luz-bombilla-P43t-HS1/3190010_32865695168.html?spm=a219c.search0104.3.21.7cd63447dhutWA&ws_ab_test=searchweb0_0%2Csearchweb201602_4_10152_10151_10065_10344_10068_10547_10342_10343_10340_10548_10341_10696_10084_10083_10618_10307_10820_10821_10301_10303_10059_100031_10103_10624_10623_10622_10621_10620-10000802%2Csearchweb201603_55%2CppcSwitch_5&algo_expid=a62c5dee-7167-45f7-adcc-6abc11d9ea29-3&algo_pvid=a62c5dee-7167-45f7-adcc-6abc11d9ea29&transAbTest=ae803_1&priceBeautifyAB=0

for years.
much more light.
Not for tuv day.......

by **stormbringer** Sat 09 Jun 2018, 7:43 pm

I've tried fiddling with this subject - and it was a complete disaster! Thing is, yesteryear's reflector geometry/technology is not really compliant with LED light generation, leading to light splatter all over the place. Fitting one of these will get you no friends anywhere!

To see if things have improved, install a regular H4. Choose a dark night. Light up a wall. Does the dipped beam produce a reasonably sharp cutoff? Yes.

Exchange regular H4 with WonderWhizbang(TM) chinese elCheapo LED gizmo. Light up wall. Sharp cutoff?

If not, discard WonderWhizbang(TM) bulb. If yes, produce photo-documentation of both the regular H4 (reference) and WW(TM) light patterns here.

No. I won't hold my breath, waiting for the photos.

by Baddi Tue 12 Jun 2018, 8:42 pm

I was reading the post about H4 led bulbs not long ago and it got me thinking if it might be worth it to design a reflector myself with LED's along the front edge, shining backwards into the reflector.
Then if only the bottom LEDs are on, they would hit the top part of the reflector so it is thrown forwards and down and then for full light, you could have LEDs all around the edge to light up all of the reflector. By having the LED's spread across a bigger area and by having their backside along the outer edge of the reflector, they can easily cool off. Wouldn't be legal but if engineered properly, perhaps it could serve the purpose anyway?

by **stormbringer** Tue 12 Jun 2018, 9:44 pm

Baddi wrote:...to design a reflector - with LED's along the front edge, shining backwards into the reflector.
Then if only the bottom LEDs are on, they would hit the top part of the reflector so it is thrown forwards and down and then for full light, you could have LEDs all around the edge to light up all of the reflector.

Read up on 'parabola' and 'focal point' - which is the principle being used in headlights. Here' you will discover that to get a proper light pattern (light beams being sent out from the assembly in perfect parallel), the light source must be placed in the parabola focal point and have a physical extent of 0 (zero). Looking at a H4 bulb, it's easy to realise that either one or both filaments will be offset from focal point and that neither of the filaments have extent zero. In short; the H4 design is flawed by design. But given the constant extent of the H4's filaments, it's also trivial to see that a larger parabola will improve on the situation - making for a better and more precise cutoff.
H4 led Para210

Comparing one filament in a standard H4 with a corresponding light emitting component in a LED H4 tells you that the filament resembles a thin rod (length approx. 5mm, dia 1mm) whereas the LED is a COB (Chips On Board, a matrix of little individual LEDs on a plate) measuring, say, 5x8mm. To achieve the same proportional accuracy, the COB would have to be installed in a far larger parabola. It's not. Thus, light splatter!

Installing LEDs at the parabola rim would get you a nicely lit-up headlight allright, but no focus whatsoever. Bad idea. BMW 'angel eyes' are not for driving, but for looks.

The standard CB500 headlight may look oldskool and boring, but it's actually an optimal design! It's reasonably big and it's completely circular allowing for a mathematically perfect revolved parabola. Comparing it to WW2 searchlights, designed to wring out every little bit of efficiency possible, is not far off, apart from the scale.
H4 led 19236310

Again, Honda did it right!

by **stormbringer** Tue 12 Jun 2018, 10:04 pm

So, I went into my photo archive and dug out evidence.

March 2015 I fiddled with a LED H4 Motorcycle headlight bulb, and these are the results:
Looks nice, eh?
H4 led 20150310

This is 'dipped beam':
H4 led 20150311

Feel free to compare with an ordinary Osram Nightbreaker Unlimited H4:
H4 led 20150312

My point here is that LED headlight is not for the CB500.

by **stormbringer** Tue 12 Jun 2018, 10:12 pm

Arrgh! I did post something back then: full review.

I know. I'm a headlight tech nerd. Sorry.

by Baddi Tue 12 Jun 2018, 10:55 pm

First of all: who says that it has to be a parabola? A parabola is good if you only have one concentrated light source, but in reality, we could make it any shape as long as we remember than incoming angle=outgoing angle and place our light sources accordingly.
So the idea is as such: We make the upper and lower part such that a light shining from the lower edge bounces off of the upper part of the reflector and shines the light where we want it to go. Then for full blast, we have another light at the top which then lights up the parts that the light at the bottom doesn't. It requires some engineering to control the direction of the light but it is not unreasonable.

And it was actually that exact post you just linked that got me thinking of ways to get around the issue that LED's clustered to a bar to get rid of the heat isn't a point.
H4 led Headli10

by **stormbringer** Tue 12 Jun 2018, 11:29 pm

Baddi wrote:First of all: who says that it has to be a parabola?

No one says it has to be a parabola. And I do like your way of thinking outside the box. However, your drawing illustrates the issue perfectly. Look at what leaves the assembly; non-parallel rays. Not what we want. And this is only 2D. Revolve the 2D around a horizontal axis and what do you have?

Methinks a complete replacement headlight unit would be a practical possibility. I've looked at several prefab designs, featuring lenses and stuff. A sharp dipped-beam-cutoff is possible, but AFAIK any HID device is required by law to feature active direction management to be approved for road use in DK. And that's not so easy to achieve on a CB500.

And, regulations aside, I wouldn't want to be on the receiving end of a misaligned or wrongly installed HID headlight. Thus, I'll reserve the right to be very conservative in choosing 'possibly viable' solutions.

by Baddi Wed 13 Jun 2018, 7:22 am

stormbringer wrote:
Baddi wrote:First of all: who says that it has to be a parabola?
No one says it has to be a parabola. And I do like your way of thinking outside the box. However, your drawing illustrates the issue perfectly. Look at what leaves the assembly; non-parallel rays. Not what we want. And this is only 2D. Revolve the 2D around a horizontal axis and what do you have?

Methinks a complete replacement headlight unit would be a practical possibility. I've looked at several prefab designs, featuring lenses and stuff. A sharp dipped-beam-cutoff is possible, but AFAIK any HID device is required by law to feature active direction management to be approved for road use in DK. And that's not so easy to achieve on a CB500.

And, regulations aside, I wouldn't want to be on the receiving end of a misaligned or wrongly installed HID headlight. Thus, I'll reserve the right to be very conservative in choosing 'possibly viable' solutions.

The non-parallel lines was actually intentional to show that we can control the intensity of the light at the ground certain distances from the headlight by shining more light to the areas we want to illuminate if that is purposeful.

Revolving around a central axis: why is that a technical requrement??
I understand if you want it to look original from the outside but I have the full fairing so not the original headlight and that is kind of square when looking from the front and much wider than it is tall so it would be perfect for having that 2D drawn shape extruded to a width and then rounded at the ends to close it off and control sideways illumination .
Also, this would be LED and not HID (Xenon) so active direction management is not required by law.

by ZekeVal Wed 13 Jun 2018, 4:44 pm

Or...

...you could just install "Off the shelf" LED Dipped (here) and High Beam lights (here).

by Beresford Wed 13 Jun 2018, 5:38 pm

ZekeVal wrote:Or...

...you could just install "Off the shelf" LED Dipped (here) and High Beam lights (here).

They know how to charge, don't they ?

by **stormbringer** Thu 14 Jun 2018, 8:41 am

Baddi wrote:The non-parallel lines was actually intentional to show that we can control the intensity of the light at the ground certain distances from the headlight by shining more light to the areas we want to illuminate if that is purposeful.

Revolving around a central axis: why is that a technical requrement??
I understand if you want it to look original from the outside but I have the full fairing so not the original headlight and that is kind of square when looking from the front and much wider than it is tall so it would be perfect for having that 2D drawn shape extruded to a width and then rounded at the ends to close it off and control sideways illumination .
Also, this would be LED and not HID (Xenon) so active direction management is not required by law.

Ah. Point taken. However, if it is so easy to do as you suggest, everyone would be doing it. Distributing N LEDs over a large area reduces not only the requirements for cooling them, you would get some nice spatial degrees of freedom too.

With auto manufacturers spending lots of money in development, the absence of designs adhering to your outline hints to one of two things; either the combined expertise of all these R&D departments have failed (because you've come up with a novel idea) or your idea is somehow flawed.
Hmmm. How about firing up eg. Blender or some other raytracing-capable tool and draw your idea. Then raytrace it. This would tell you if there's something to be found.

by Baddi Thu 14 Jun 2018, 3:18 pm

stormbringer wrote:
Baddi wrote:The non-parallel lines was actually intentional to show that we can control the intensity of the light at the ground certain distances from the headlight by shining more light to the areas we want to illuminate if that is purposeful.

Revolving around a central axis: why is that a technical requrement??
I understand if you want it to look original from the outside but I have the full fairing so not the original headlight and that is kind of square when looking from the front and much wider than it is tall so it would be perfect for having that 2D drawn shape extruded to a width and then rounded at the ends to close it off and control sideways illumination .
Also, this would be LED and not HID (Xenon) so active direction management is not required by law.
Ah. Point taken. However, if it is so easy to do as you suggest, everyone would be doing it. Distributing N LEDs over a large area reduces not only the requirements for cooling them, you would get some nice spatial degrees of freedom too.

With auto manufacturers spending lots of money in development, the absence of designs adhering to your outline hints to one of two things; either the combined expertise of all these R&D departments have failed (because you've come up with a novel idea) or your idea is somehow flawed.
Hmmm. How about firing up eg. Blender or some other raytracing-capable tool and draw your idea. Then raytrace it. This would tell you if there's something to be found.

I'm fairly certain that the reason to why it is not seen otherplaces comes from a combination of the complexity in design, in manufacturing, in price and in servicability along with the fact that the automotive industry is terribly slow and conservative when it comes to applying new ideas and technology to their products. Such a thing as the headlights are usually determined by the designers rather than the engineers and it is always a weighting of costs against reliability when new things is to be introduced.
I'll try and draw up the idea in the summer vacation, when I'm done with my examns and this seasons Formula Student events but for now, I'll see how many flaws of this idea I can find that would make a car or a large-scale motorcycle manufacturer discard the idea:
- Computer and man-time spent simulating costs money so the simpler the design of the headlight can be made, the cheaper the product can be and thus more money can be spent on the interior, radio and heated everything which the costumers appreciate when choosing their new vehicle.
- A single light bulb of standardised design is MUCH cheaper than an assembly of LED's that is different for every single headlight design. Also, being a single component, it increases servicability if it is easy to replace (often overlooked point).
- A conventional parabola can be injection-molded in one-piece by a two-piece (or very few-piece) mold as the inside of the reflector will have only negative angles. If the reflector is actually a parabola revolved aroudn the central axis, machining of the mold can be done on a lathe which further decreases costs. This proposed design has to have an edge that shields the low-beam LED strip from shining onto the lower half of the reflector and this will likely result in a shape that is difficult to injection mold as it requires even more seperate mold pieces or it requires the headlight to be made in multiple pieces that click together (snap-fit plastic things often breaks). I don't consider this a problem though, as I own a 3D printer, which is in itself a very costly way of mass-manufacturing, but OK if you only need less than a couple of hundred of something this size.

So all in all, if this design was proposed to an R&D department of a major car or motorcycle manufacturer, they would say: Why should we spend money on that when we have something that works and is cheap?
For these motorcycles however you have made it clear that the original reflector shape isn't good for the LED H4 'bulbs' because it is designed for a point-shaped light source and that the LED's can't do that so the choice is really if you want to keep on using a convenctional bulb or redesigning something completely, and then to me, this seems like it is easier to develop and manufacture than the automatically adjustable headlights that seems to be the way car manufacturers tackle it. I believe this idea to be the easier way of achieving LED illumination but are quite aware that it may not be the better option in large-scale production.

by Sponsored content

H4 led

H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led

Re: H4 led