TVL4 Development Thread, mains-powered dimmable multi-MC-E

[TigerhawkT3]

TVL1 is four Q2s in series, AAed to a block of aluminum with shallow fins, powered by a single 12V SLA.

TVL2 is nine Q2s in series, AAed to a block of aluminum with dense, deep fins, driven by an LM317T and powered by three 12V SLAs.

TVL3 is a single Q2, AAed to a piece of steel, superglued to a wood block, with 2AA alks on top for DD.

All these lights can be focused by those $1 optics from KD or DX (I forget which), and all have 1/4-20 mounting holes for use on standard tripods.

Now I'd like to begin development of the newest iteration, TVL4. It'll be smaller, more powerful, and have more features than TVL2. Here's what I have planned so far:

-Old CPU heatsink with room for nine stars, fan on the back
-Multiple white MC-Es in series
-Multiple red XR-Es in series
-White MC-Es will be mains powered, through the use of a variac, bridge rectifier, high-voltage capacitor, and bleeding resistors, with an LM138 driver
-Red XR-Es will be battery-powered, with an LM138 or LM317 driver

There are several things that I don't know, however:

-How much red will be filled in by the red XR-Es? What's a proper white/red emitter ratio, given equal drive current? I plan on being able to pot between not enough red and too much red (roughly ), so I need to know the "just right" value for that.
-What kind of connectors should I use? I was thinking of PowerPoles, but how about those connectors for automotive bulbs? I want something durable, low-resistance, DIYable, and maybe weather-resistant.
-I don't think there'll be enough room for optics, so I'm probably just going to use bare LEDs. Maybe I'll make some sort of white paper barn door assembly to fit over it. However, I would very much like to protect the bare LEDs. Maybe some sort of polycarb square bolted a few millimeters in front of the LEDs? If that, I was thinking machine screws that screw through the polycarb and into the heatsink, with nuts holding the polycarb in place.

That's pretty much all I can think of for now. Comments more than welcome.

 

[mudman cj]

Re: TVL4 Development Thread

Great project! Multi-emitter arrays offer a lot of potential for lighting, and I think you are on the right track with your incorporation of extra red in the spectrum. If I may make a suggestion, the addition of cyan emitters would also help to fill in another gap in the spectrum. This spectrum is for a 5A XR-E emitter, but the same bin is available in MC-E; though, you will have to be patient for them to become available again - as will I.

As for how many MC-E vs. colored LEDs to use, I would say two each of cyan and red since you can dial them down and because 5 MC-E LEDs could be driven at up to 50 Watts. You may get a nice result with only one of the cyan or red emitters, but I think we would just be guessing at this point. And more MC-E LEDs quickly adds power.

That is a lot of heat to dissipate if you want to keep the LEDs running at efficient die temperatures. I recommend a heat sink capable of 1C/W or lower if you can manage it. That means large heat sinks or noisy fans, but it looks like you are expecting a large heat sink already.

 

[TigerhawkT3]

Re: TVL4 Development Thread

Cyan... interesting. Where would I get cyan emitters, hopefully in the XR-E series?

I've been thinking about the red/cyan portion, and I realized that there might be a problem: if white and red/cyan were individually controlled, then decreasing the output of the white by itself would shift the color balance to red/cyan. Even if a user (me) attempted to decrease all colors at the same rate, the color balance would probably shift to some degree, and the camera would have to WB again.

Do red and cyan have the same current/output curves as white (blue>phosphor)? The XR-E datasheet (page 8) only has graphs for blue/blue>phosphor and green (and they're slightly divergent at higher currents). If they do match, I could just add one or two red and cyan emitters in series with the MC-Es to fill the spectrum, albeit to a fixed degree.

I'll be using a CPU heatsink and fan rated for a CPU of about 70W (Athlon XP 2000+).

 

[mudman cj]

Re: TVL4 Development Thread

Unfortunately, the XR-E series does not include cyan. The Lux III series does though. And there wouldn't be much difference in the output anyway, mostly just the distribution of light output as a function of angle.

I see the problem with dimming. Cyan follows the white/blue Lux III curves but red deviates significantly. So in this case the red would be more of a problem. But, instead of reducing output by reducing voltage, why not use high frequency PWM?

That heat sink sounds like it will work well.

Edit: The XR-E series does not include red either, that is the XR-C series, and it is quite limited in output. The best option for high output red or cyan is still the Lux III AFAIK. Someone please correct me if I am wrong.

 

[LukeA]

Re: TVL4 Development Thread

can be focused by those $1 optics from KD or DX (I forget which)

KD, and now they're $1.10/piece

 

[TigerhawkT3]

Re: TVL4 Development Thread

Hmmm... Maybe my best option would be to stick with just white for the TVL4 and leave the dimmable cyan and red for a separate light.

 

[TigerhawkT3]

Re: TVL4 Development Thread

Any recommendations for a tap/screw size for attaching the stars to the heatsink?

What value should I use for the bleed resistors? 50K? 100K?

How do I properly ground what needs to be grounded and keep everything as safe as possible? (<--important)

Any recommendations for connectors?

Any recommendations for enclosures?

Here's what I'm thinking of getting as far as power:
Variac
Rectifier
Capacitor
Bleed resistors

 

[TigerhawkT3]

Re: TVL4 Development Thread, lots of questions

I just bought a small heatsink/fan for the LM138. It's aluminum with a copper surface, so I'll be able to solder to it (I'm planning on the TO-3 package LM138).

I'm probably going to go down to OSH with a star and check out a few screw sizes.

I think I'll use PowerPole connectors, unless there's something better for this application.

As far as grounding, is that just connecting the heatsink's base to the ground prong (round pin)? If not, how would I go about that? What would advantages/disadvantages be?

 

[TigerhawkT3]

Re: TVL4 Development Thread, lots of questions

I ordered the linked parts yesterday. They should arrive within a few days. ($18 shipping... ouch.)

I went to OSH while WoW was down for scheduled maintenance, and got some 1/4-20 1/2" screws, as well as 1/4" washers. The screw size I chose for the stars was 4-40, with a head that's round on top but flat on its underside. I also got a 4-40 tap and appropriate drill bits (of course). I forgot to get the screws and nuts/spacers for the plastic shield, but I can get those later once I've finalized the design a bit more.

 

[TigerhawkT3]

Got the order in today. Man, the variac is beautiful (in a technological/industrial sort of way).

The next step is building the assembly containing the rectifier, capacitors, and resistors.

 

[TigerhawkT3]

:bump:

Any comments?

 

[TigerhawkT3]

I wonder what kind of enclosure I should use for the rectifier assembly.

 

[shomie911]

Not sure what kind of enclosure you should use but I'd like to say that a mains powered dimmable MC-E is a very cool project.

Can't wait to see how it turns out. :twothumbs