Forward Operating Base In-a-Can

[Another Ahab]

I use a lot of the LED strips for any number of lighting tasks. One bit of advice I would offer is to make sure they are attached to something that can act as a heatsink. This is especially true if you are driving them near their current limits.

Hope this info helps.

I thought the LED generated light without heat (and that was the whole special feature about them).

Is there a "yes-and-no" explanation to that?

 

[DieselAddict]

There is and answer. I hope this isn't too long winded. I have been known to over explain things.

LEDs generate much less heat relative to their light output than conventional lamps but still generate heat. The key difference is how that heat is carried away from the light. With a normal filament lamp the heat is carried away in the form of Infrared radiation. This is because the filament operates at thousands of degrees (about 2700 deg K or about 4400 deg F). The excess heat radiates away to keep things in balance.

LEDs in contrast don't generate their light by heating a filament to thousands of degrees so they don't have this very high temperature difference between the die and the environment to encourage heat to leave via IR. They instead require their heat to be removed via conduction. Excess heat is the enemy of LEDs since they are actually semiconductors and more like computer CPUs than filament lamps. What happens if you let your computer CPU overheat? It will have a short life. The same is true for LEDs.

To put this in perspective - When we look at overall system efficiency of a filament lamp we see that for each watt of electrical energy you get about 0.02 watts of usable light. Meaning that 98% of the energy went into something other than usable light. More advanced lamps such as halogen lamps have brought that up to about 4% efficient (or 96% wasted energy if you like).

LEDs are much more efficient relative to filament lamps but are far from being totally efficient in converting electricity into light. For each watt of energy you put in a reasonably efficient white LED you get about 0.3w of usable light. This means they waste "only" 70% of the electricity (mostly as heat that has to be removed from the die).

This should give you an idea about how much heat you are trying to conduct away. If you put up a LED strip that and you are going to input 50w of electricity you will get about 15w of light and have to get rid of about 35w of waste heat.

I hope this helps explain why mounting your LED strips to a thermally conductive surface is important for them to have a long and productive life.

 

[tim292stro]

...One bit of advice I would offer is to make sure they are attached to something that can act as a heatsink...

Indeed, I'm killing two birds with one stone here - since PWM is the generation of an AC waveform and I'm driving enough current to illuminate power LEDs, a long strip of AC waveform makes an antenna... If not shielded.

The 1/2" wide strips come with a 3M VHB clear acrylic tape, this will be used during assembly to temporarily attach the strip to the center of a 4"-wide 100-mesh copper screen backing. Over the face of the LED strip (the emitting side), a 3/4" wide strip of frosted vinyl will be heat laminated to the LED strip and the edge of the copper mesh backing. Over the top of that will be a 4-3/4" wide strip of 22-mesh copper screen, with a half round relief rolled in for the strip - this will be sewn on along the heat sealed edge of the frosted facing to bind the front screen and back screen with tinned 38AWG copper wire (like a thread). Every 1/8" along the sewn border and at the edge of the two screens a solder dot will electrically and thermally bind the two screen halves - this will mean that the heat dissipating surface is increased by a factor of 8. The back will then get a heat/light reflective faced vinyl sheet, and the front a clear fiber reinforced vinyl face - then will be heat sealed to each other, which will melt the vinyl through the copper screen making it a solid/monolithic structure, and will water seal the assembly. Finally the white back of the light panel will have two strips of 2" Velcro sewn on with nylon thread. The two strips of Velcro (instead of one wide one) will create a small but effective air gap between the light strip and the inside of the tent liner. Two wired connector cables will come out the center of the long axis of the strip perpendicular to the length of the strip - I'm designing these so that they can be individually replaced and moved around, and the light controller (in the rack) will learn the new layout automatically allowing for quick simple setup. I intend to publish the design files here in case anyone want to try to replicate it.

Finally, the strip is not the most powerful one I could have obtained, and as you can see above by the algorithm, only half of the LED dies in each package (cool or warm white) can be 100% active at a time. Note that this tent would be VERY bright if they were turned up to 100% - a 305 Base-X has about 450 square feet of floor, one roll of this LED strip is 3175lumens with the warm white illuminated (less output that the cool white), 5-rolls (one roll per 5' section of tent) at 100% warm white would give me just shy of 16,000 lumens - that's ~35.25 lumens per square foot of floor (about 380lux). That's brighter than most classrooms/offices during the day, and approaching laboratory lighting levels. Cool white will give me "Showroom" brightness levels.

The central controller is planned to have a max level control for white and green (separately). The knob will set limits and the battery level cause the controller to taper of the output as the battery level falls, to eek out as much run-time as can be managed. In the end I should have full control from NVIS-compatible 1lux, all the way up to an "illuminating" 450lux (couldn't help myself ).

The whole thing wrapped in copper mesh acts as a Faraday Cage, and with shielded cables, power filters on the supply and differential signalling for the control (extra shield and twisted pair), it should be electrically/RF quiet too.

 

[tim292stro]

I thought the LED generated light without heat...

What DieselAddict said. Think of it this way, a filament light bulb has a long filament to dissipate heat with, a typical LED die (the actual part that emits the light) is about a millimeter squared. A 1Watt LED thus has a heating effect of 450Watts/square inch. You'd need only 3 square inches of LED die to equal a 1500Watt room heater! LEDs rely on the package design and ultimate installation to deal with the heat generated. If you've read the XM1027 build thread you'd know that the 500Watt COB LEDs I'm putting in the spot lights are generating enough heat in a small area to need water cooling - 500W * 0.7 = 350Watts heat, or ~1200btu/hr of heat... It should melt ice from the lens with just what is coming out the front.

 

[DieselAddict]

Thats interesting to go to that length to shield the LEDs. What frequency is the controller operating at?

Am I understanding it correctly that the entire strip will be fully enclosed in a vinyl cover? You may want to test that to make sure you have enough thermal dissipation.

 

[tim292stro]

...What frequency is the controller operating at?...

It free runs internally at 8MHz, data-rate I'm using is 2Mbps (max is 25Mbps), and PWM clock is around 2.5Khz - no spread spectrum on anything. The proximity to RF devices is a key reason to shield this much - CB and HAM bands (160Meter-6Meter) could pick up interference if it's not handled well. Also with the direct attachment to the UPS that powers the ToughBook (sources the tent's analog audio), DC-powered network switch (phones and WiFi AP attached), and video display controller - extra RFI/EMI suppression and filtering is the name of the game. Radios can "hear" a lot better when the area around them is quiet: quiet RF = longer receive range and better signal quality.

I have thermally modeled this in AutoCad already, the first build will help me validate/tune the design before I start building them all (so testing is part of the plan) - when heat laminated, the total assembly thickness on the "wings" sidewards of the LED strip is about 3mm. The other intent of the vinyl is that it will also help with some acoustic noise suppression, as the PWM frequency is right smack dab in the "annoying whistle" frequency range.

Here's some of the rest of the math since you seem intrested:

LED Tape width: 1/2"
LED run length: 50"
Mesh backing width: 4"
Mesh backing run length: 54"
LED strip max power: 12Watts (~66lumens/watt, warm white)
LED strip area: 25"-squared
Mesh area: 216"-squared

To oversimplify you can get a rough average temperature if you assume half the area of the heat dissipating surface, 108"-sq. Assuming the 3M 8805 thermally conductive VHB tape spec holds out, I should get an energy density of roughly 100mWatts/sq-in, and an LED die heat rise of less than 20°F above ambient. Assuming a 90°F day with no air conditioning (just circulating air from outside), I have 150° margin to the long-life rating of the LED strip, and over 240° margin to the melting point of the vinyl (both LED assembly and tent). The controller is including a linear power de-rating of 30%/50% if the ambient temperature is that high (starts at 80° inside ambient) - both to protect the LEDs, and the battery pack which would have trouble cooling itself passively at that ambient temp. 50% is if it's only battery powered, 30% if there is a charging source powering the UPS.

Not too complicated to understand for everyone else I hope

 

[DieselAddict]

Good info. I'll be curious to see how it looks when you have it put together.

I have the LED strips all over the house, the back porch, and the outbuilding. So far I've not had issues with the PWM dimmers being a problem on HF. The controllers I'm using run at 40khz. The CFL lights in the house are HORRIBLE. The LEDs are nice and quiet.

I had looked at using that same tape since the tape that comes on the strips is really crappy. Is so bad that I've wondered if it was a counterfeit. I had some contact cement left over from a counter top and it worked so well and was so much cheaper I've stuck (pun) with it.

I've enjoyed reading along with this project since its an awesome field day setup waiting to happen.

 

[Another Ahab]

Meaning that 98% of the energy went into something other than usable light. More advanced lamps such as halogen lamps have brought that up to about 4% efficient (or 96% wasted energy if you like).

LEDs are much more efficient relative to filament lamps but are far from being totally efficient in converting electricity into light. For each watt of energy you put in a reasonably efficient white LED you get about 0.3w of usable light. This means they waste "only" 70% of the electricity (mostly as heat that has to be removed from the die).

And we like to believe that we're living in an age of high-tech?!

There is still just so much that we don't know....

 

[srodocker]

So now you got me looking at toughbooks...

 

[rustystud]

So now you got me looking at toughbooks...

Tim already got me to buy one, and it was a great idea ! For the total cost of $210.00 I have a really nice Toughbook with a 1TB hard drive ! Later on when more money becomes available I'll be adding 8GB of RAM memory to it.
Thanks Tim for all your help ! Your a great guy !

 

[Another Ahab]

Tim already got me to buy one, and it was a great idea ! For the total cost of $210.00 I have a really nice Toughbook with a 1TB hard drive ! Later on when more money becomes available I'll be adding 8GB of RAM memory to it.
Thanks Tim for all your help ! Your a great guy !

Yee-Ow. That's a lot, right?!

Wonder how much total computing power Mission Control had to land Apollo 11 on the Moon?

 

[tim292stro]

For the record, RustyStud was already interested in buying, I just explained how accessible it was. I did the same for srodocker last night via PM. NO collusion (only a sorted eBay search was shared), just my method for getting a cheap ToughBook which included a general call to pace one's self when thinking about buying (no panic buying!), a few things to look for in listings (models, pictures of machines booting, etc), and some resources to use for learning about them in general.

The jist is the good deals happen if you an buy one that is a little rough but perfectly functional and then spend time buying the repair parts and installing them yourself - the "trick" for coming out ahead is finding one that is in good enough shape at the price point to be cheaper than a better one once you've repaired it. Not unlike the MV hobby if you take a step back and think about it . Admins may feel free to review my PM box if needed to verify for CYA.

 

[tim292stro]

Yee-Ow. That's a lot, right?!...

1TB:

• about 250,000 MP3s (assume 5MB each)
• about 1,000,000 JPG images (assume 1.25MB each)

8GB is enough to run 64-bit Win7/Win8.1 well, and more than enough to run any current Linux flavor - if you're running 32-bit, it's twice as much as the system can actually use.

IIRC the LEM guidance computer had 2K of RAM (0.000002048GB), and 32K of ROM storage (0.000000032768TB), and used almost 60Watts running at 2MHz with a multi-line 7-segment display (the CF-30 should use round 50W running at 1.66GHz, with a full color LCD screen). I think any smart phone with a touch screen now has more than 100x the compute power of NASA in the 60's, and only uses 1Watt...

From IBM's web-page: http://www-03.ibm.com/ibm/history/exhibits/space/space_gemini.html

"...During a Gemini flight, five IBM computers in Houston, Texas, performed 25 billion calculations every 24 hours to provide NASA flight controllers with almost instantaneous reports on the moment-by-moment progress of the mission..."

Thousand = Kilohertz
Million = Megahertz
Billion = Gigahertz

So over-simplifying here: any current/future smart phone with a multi-core processor running at >=1GHz will do about roughly as much "work" in 10-20 seconds as NASA's compute farm could do in a day during the Gemini program... put another way, your smart phone is about 4000 times faster than NASA's Gemini-era data-center! What have you done with your phone lately?

Drifting...

 

[Another Ahab]

IIRC the LEM guidance computer had 2K of RAM (0.00000248GB), and 32K of ROM storage (0.000000032768TB), and used almost 60Watts running at 2MHz with a multi-line 7-segment display (the CF-30 should use round 50W running at 1.66GHz, with a full color LCD screen). I think any smart phone with a touch screen now has more than 100x the compute power of NASA in the 60's, and only uses 1Watt...
/QUOTE]

 

[tim292stro]

I purchased the external video card adapter and the rack shelf today - I received the second docking station Monday. This will let me plug everything into the docking station (nearly hardwired) and drop the laptop in and lock it, the only other cable I need to insert for now is the express card to the external video card.

I have 9U to work with in the cabinet (1U "unit" = 1.75" high). The switch is 1U, the laptop with docking station looks to be about 3U, I expect the UPS will be about 3U, that leaves ~2U for a locking drawer to put the AP and a few other accessories into for travel.

 

[Another Ahab]

 

[srodocker]

Tim already got me to buy one, and it was a great idea ! For the total cost of $210.00 I have a really nice Toughbook with a 1TB hard drive ! Later on when more money becomes available I'll be adding 8GB of RAM memory to it.
Thanks Tim for all your help ! Your a great guy !

Yea Got a pm. Going to plunk the brain and hopefully build a similar setup to yours!

 

[Another Ahab]

I think any smart phone with a touch screen now has more than 100x the compute power of NASA in the 60's, and only uses 1Watt...

Drifting...

And I remember seeing those Mission Control guys pulling out slide rules once when they had some problem creeping up...and we not only put men on the Moon that way, but brought them back, too. How cool was that?!

Don't understand all the CPU talk here, but loving every chance to learn something new. Thanks for the thread.

 

[rustystud]

For the record, RustyStud was already interested in buying, I just explained how accessible it was. I did the same for srodocker last night via PM. NO collusion (only a sorted eBay search was shared), just my method for getting a cheap ToughBook which included a general call to pace one's self when thinking about buying (no panic buying!), a few things to look for in listings (models, pictures of machines booting, etc), and some resources to use for learning about them in general.

The jist is the good deals happen if you an buy one that is a little rough but perfectly functional and then spend time buying the repair parts and installing them yourself - the "trick" for coming out ahead is finding one that is in good enough shape at the price point to be cheaper than a better one once you've repaired it. Not unlike the MV hobby if you take a step back and think about it . Admins may feel free to review my PM box if needed to verify for CYA.

Sorry if I implied that Tim "strong armed" me into buying a Toughbook, but he was "tremendously" helpful in getting a good unit, and even helped supply me with some parts I needed !
I'm all for getting "expert" help when needed , and Tim is an expert in this field !
I've already installed over 100 albums of music and All my military manuals on the Toughbook and still have so much more space to use ! I never thought I could afford a lap top which sells for thousands of dollars !
Thanks again Tim for all your help !

 

[tim292stro]

Rack is coming tomorrow, I ordered the last of the missing parts for the ToughBooks (cosmetic at this point), and a Hardigg medical case to transport the computers and phones.