Ed Scherer

Months later, I finally got back to this. Even though I see a couple of you guys have made progress on your own designs, I just wanted to see my stuff through to completion (didn't want the effort, materials, and tooling to go to waste).

See the latest details in the other thread I started on this (Building a replacement cap for sealing headlamps after converting to HID or LED).

traxtar928

Great Job, Ed! I'm still looking to get mine made from a 3D printer or CNC from aluminum and replicate the original shape of the cap (roughly). Anyone with machining experience care to quote what they think it would cost for those to be cut from billet aluminum? Might be crazy expensive and not worth the cost.

Ed Scherer

Thanks!

Some quick relative price info (using McMaster pricing; might be able to do better) for 12" by 24" by 3/8" sheets (that's what I'm using for now for 10 caps; might be able to squeeze more caps out of that size, but I'm not sure).

• ABS: $36
• Acetal: $61
• Delrin: $71
• Cast nylon: $81
• 6061 aluminum (just picked that since it's common; I haven't researched about what aluminum might be best for this application): $120
• "highly machinable" MIC6 aluminum: $107

OK, that's just one consideration.

But what's the best material for this application, and why?

I think I'd actually prefer some kind of plastic (as did Porsche). This is just a cap. The only reason I can think of maybe wanting aluminum would be to add additional heat sinking. But then you'd need to transfer heat across the seal for the bulb barrel. It would require a different design than the simple O-ring design I've used.

traxtar928

As for material... when comparing aluminum vs ABS (or similar plastic), the material shouldn't have a big impact on performance so long as it can withstand the heat of the bulb.

Those heatsinks get very hot (way too hot to handle without gloves) just during some brightness testing I did. I'd be afraid of melting the plastic housing if it was straight ABS. Since you mention Porsche, they did not use a generic machinable plastic... the cap is made of an insulating bakelite-type material or FRP, not ABS. It's very brittle and can handle high heat without melting.

It DOES have a melting point, since that's how the metal tabs are secured to the inside of the cap to connect the H4 bulb... but I think that's where the similarities to ABS end. Regardless, I'm watching your build and eager to see the results of real-world use, like a long night drive.

Ed Scherer

I'll be taking a close look at temperatures before I decide on final materials. I wanted a prototype first where I could examine temperatures in an installed environment, where the bulb is sealed in the actual headlight it will be used in, etc.

FWIW, in my current cap design (prototype #2), the bulb barrel doesn't touch the ABS, although it is ~1 mm away from it, separated by a high-temperature silicone O-ring rated to 400 °F. Likewise, the heat sink doesn't come any closer than about 2 or 3 mm.

You're right, though: the heat issue is very important.

FredR

Where exactly is all this "heat " coming from?

It may be a completely different thing but my new home is fully kitted with LED lighting and what we noticed is the very low level of heat being generated by LED bulbs thus why the amount of light emitted by an incandescent 100w bulb is achieved by a LED equivalent rated at something 10 watts. Thus for the same amount of power you can get vastly more light or achieve the same lumens for a greatly reduced power consumption- the difference being heat.

Apologies if I am missing something.

Rgds

Fred

jcorenman

LED's are much more efficient than traditional halogen lamps, but they still generate heat. Your typical 60w H4 haogen is around 15% efficient, which means around 50w heat and 10w of light. LEDs are much more efficient, but there is still some heat.

The problem is that the LED chips themselves don't like high temperatures-- I think the max is 50-60C where halogen lamps are happy as long as the glass doesn't melt. So LEDs come with fancy heatsinks to get rid of the heat and keep the temerpature down, but the heatsinks can still reach 50C-- too hot to touch for very long.

ABS is rated for 60C, a bit marginal. Delrin (acetal) is better (80C) and machines more nicely in my experience, but costs a bit more. Nylon is often used for molded under-hood intakes etc and can approach 100C but doesn't machine well-- and the LED chip would be thoroughly melted at that point.

P.S. Good info here: https://www.mcmaster.com/#about-plastic/=17hry96
And then there is PEEK, good for 250C and "only" $600 for a 12 x 24 x 3/8" sheet...

Ed Scherer

Yeah... once I'm done with prototypes, Delrin was at the top of my list for production pieces (assuming I get that far). Machinability is important for what I'm doing, too, which is why Delrin was at the top of my list. Once I'm done with prototypes, I'm going to be retooling a bit, too, to get machining time down (bigger cutters for some pockets to reduce machining time, and possibly re-optimizing the layout of parts so I can get more out of one sheet of plastic). Once I'm doing 10 (or more) at a time, (manual, for my setup) cutter changes are worth the effort.

I'm going to fire up a couple of LED H4s and get some preliminary temperature readings, just so we have a few real numbers to talk about from the actual devices we're looking at.

Ed Scherer

I got a few numbers, given:

• 928 H4 headlight assembly;
• ABS cap (prototype #2) on;
• H4 LED bulb (the type I showed in post #48 in this thread, which I believe is a re-branded version of the one shown in post #1 of this thread) installed;
• Headlight cover not on, car in garage;
• Low beam (I also tried high beam briefly, but didn't take any separate measurements, as I noted that both low and high beam drew the same amount of current — 1.65 A — so I decided to omit separate readings for now).

Temperature measurements taken with an IR thermometer (Fluke 62 MAX+) after temperatures appeared to reach steady state (approximately 30 minutes, I think; I didn't time it):

• ambient air temperature: 22 °C (72 °F)
• LED barrel and heatsink area maximum temperature: 85 °C (185 °F)
• LED driver module maximum temperature: 56 °C (133 °F)
• ABS cap maximum temperature: ~60 °C (~140 °F)

That last one is very approximate, as it's difficult to isolate the hottest area of the cap with an IR thermometer. I need to put a hole in the next prototype near the center O-ring in the cap, track down my Fluke 80PK-1 bead probe thermocouple, and get some temperature readings of the ABS where it will be hottest. THAT will be the best indication of all of what we're up against.

The ballpark measurements above, though, clearly indicate that we're dealing with temperatures at the high end of the recommended temperature range for reasonably priced machinable plastics. And the temperatures above are conservative; they'll only go up once the headlight cover is on and the ambient temperature is at "hot summer day" levels with direct sun on a black headlight cover.

I'm glad there's not direct contact between the hottest element (the LED bulb's barrel) and the cap.

traxtar928

Got the covers finished and printed! I'm waiting on the o-ring to arrive from McMaster Carr that will seal the inner hole to the outside of the LED bulb. The design I went with emulates the original cover design, both in appearance and in venting method. You can see the small channel that goes from the inside of the cap, around the circumference, and through a pinhole. The stock cap has a filter to prevent debris from clogging the hole, but I don't think it should be an issue. I'll test these caps for a few weeks and see how the material holds up to the head of the bulb and lens assembly.

Let me know your thoughts and any suggestions for improvement. I printed 3 sets, and can print more if there's an interest in purchasing them. Not sure what I should price these at... thinking $40 a set is fair if they hold up to testing without issue.

Alan

I noted this earlier on this thread:

When I said there was no headlight bulb out warning - I wasn't guessing - there really is no warning on any year 928.

There are 2 dash warning bulbs related to front lighting:

1) The Green Marker Lights Only warning. This shows you don't have headlights on - It is controlled by the headlight switch only - it has no relation to the actual status of the headlights, only the position of the HL switch. Its function is unaffected by the presence, or not, of headlamp bulbs.

2) The Blue High Beam Active indicator - this is driven by the high beam output of the relay - so a wiring issue could cause this to illuminate when it should not. It certainly does not indicate bulb failure though - in a stock configuration car it would illuminate just fine regardless of the presence, or not, of headlamp bulbs.

Question - Do we have any better headlamp pattern photographs for these Phillips/Lumileds bulbs? I'm interested in both Pattern & Brightness/Range. I'd like to go LED if they are brighter than 80/100W halogens, and if they have a good cutoff pattern with no compromises. What I have seen so far seems to be a little inconclusive. I would also want a well sealing cover as discussed here. Not sure if we are there yet? - though they certainly look nice!

Alan

Michael Benno

Nice work Alan, yes I am interested in one of the sets. PM me when you are ready. Thanks again!

traxtar928

I'll be taking beam pattern pics and luminosity readings of the new LED bulbs compared to the H4 halogens they replace.

The covers should go on in the next week or so and begin a month-long test to see how they hold up. Want to make sure the material is robust enough for the environment and conditions. Usually 3D printed parts don't last forever... Especially in extreme conditions like high heat and moisture.

These covers use the factory seal/gasket as you can see, with the alignment tabs in the plastic and holes for the stainless steel retaining clips. They are also the same thickness as the original covers and use the same swivel clamp that the stock covers use, so it should be any different. We shall see.

Alan

What Halogen bulbs do you have for the comparison: Stock wattage? or 80/100W?

I'd be much more interested in the 80/100W comparison - Since I have been running this way for a long time, I think many others have too. I believe you have the local relay implementation also and that is also a desirable comparison point for me. No sense in going functionally backwards here.

Alan

Ed Scherer

Those caps look pretty good; it's interesting to see the texture of the 3D printed material. What is the material? If it's ABS, you can smooth it out using an "acetone vapor" process: look it up (lots of "how tos" on that topic). For this application, the texture might not matter as long as it seals OK.

Funny that you got yours done the same day I finally got a batch made using a completely different approach (subtractive process rather than additive process).