SSR-90 vs SST-90 on Cree star (Advice needed)

[Slartibartfast]

Hello CPF,
As a longtime lurker, I want to first thank everyone here for all the excellent information and advice that I have already received.

I am in the process of modifiying my 3D mag to run a SSR-90 or SST-90 in DD. I notice that 4sevens is selling the bare emitters (SST-90), and have read that some have soldered these onto cree stars.

Are the +ve/-ve connection pads and the central heatsink pad the same size on the cree star as on the luminus emitter? (I can't find bare luminus stars anywhere) If not, how much of a reduction of heatsinking capability am I looking at?

I have access to a machine shop and will be fabricating the largest heatsink I can fit in the mag body, but am concerned if I float the emitter onto a cree star I will create a point of thermal resistance.

Any help is appreciated


Build plan:
SSR-90/SST-90
largest possible heatsink
spring and switch low-resistance mods
3xNiMh D cells (largest mAh I can get that will support the high current)
DX Aspheric

 

[ahorton]

If you're going to use a MCPCB star anyway , I'd stick to the SSR-90.

The ones I've had have been very good as far as stars go.

 

[bigchelis]

Definitely get the SSR-90 because the STAR helps transfer the heat to the heatsink quicker/better.


Then use copper heatsink. There is a VARA2000 SSR-90 3D Mag I tested just check my sticky. ONe sample was copper and one was aluminum and the results are shocking.

Got to the link and use control find to find the VARA2000!
http://www.candlepowerforums.com/vb/showthread.php?t=260659


The heatsink used in the VARA2000's was only .9in long, but you should be able to fit 1.5~2in since you likely dont have that PWM knob switch.

bigC

 

[aurum]

@bigchelis ...

Your results doesn't say anything ... they are not comparable! :tired: @1sec there is a difference of almost 300lm otf. I doubt that such a huge difference is caused by the copper heatsink. There is just not enough heat generated in this period of time. We need to know the current of both emitters :shrug: You'll have to swap the emitters if you want to compare your results

 

[bigchelis]

@bigchelis ...

Your results doesn't say anything ... they are not comparable! :tired: @1sec there is a difference of almost 300lm otf. I doubt that such a huge difference is caused by the copper heatsink. There is just not enough heat generated in this period of time. We need to know the current of both emitters :shrug: You'll have to swap the emitters if you want to compare your results

I know I know. Im actually now working with PCC with my 3 heatsinks challenge.

He is making 3 identical heatsinks for me with the 3 famous metals used here, then I will swap the SSR-90 from heatsink to heatsink using the same hosts, cells, and with the same current we will know for sure how much better the copper really is.

PCC also mentioned I could underdrive the SSR-90 with the same 3 metals to see how perhaps at low 3~5A current in a Mag D size hosts copper is not needed.

For now my money is on copper as I supect from 4~13A it will prove to be an asset over brass and alum.

:touche:

 

[morelightnow]

I agree. I think copper will have a huge benefit over aluminum at high current, but only during the first few minutes until the host is heat saturated then it all depends on how fast it can shed the heat. I just held my heatsink in my hand and ran my sst-90. At 5 seconds I could feel the sink getting warm and at 45 seconds it was too hot to hold. That tells me the led is putting heat into the sink very fast and it needs maximum thermal transfer. I dont know what the amps are, but its 4 accupower c cells so I'm guessing around 7 amps.

 

[aurum]

At 5 seconds I could feel the sink getting warm and at 45 seconds it was too hot to hold.

I dont know what the amps are, but its 4 accupower c cells so I'm guessing around 7 amps.

I don't think so ... I think it is more than 7A. Please take a good clamp meter/multimeter and tell us the current.

 

[bigchelis]

I don't think so ... I think it is more than 7A. Please take a good clamp meter/multimeter and tell us the current.

There is another way** better then the clamp method that I will be using, but for now top secret. So, as to keep this thread on track.....


If you polish (mirror) the heatsink where the star will sit on it will transfer heat better too. Ideally you should try to polish the bottom of the star also.

This is what was done to the 3D Mag SSR-90 VARA2000 Copper I tested that did 2100ish real OTF lumens.

 

[65535]

There is another way** better then the clamp method that I will be using, but for now top secret. So, as to keep this thread on track.....


If you polish (mirror) the heatsink where the star will sit on it will transfer heat better too. Ideally you should try to polish the bottom of the star also.

This is what was done to the 3D Mag SSR-90 VARA2000 Copper I tested that did 2100ish real OTF lumens.

Shunts are good, for under 10Amps a Fluke is good, those Watts Up inline meters are cool.

Also polishing the surfaces will give you poor adhesion with thermal adhesive, so if you were so inclined you'd want to screw down the star and use thermal paste. When I was into computers people found they got the best heatsinking with a good 800 grit finish on a heatsink and cpu heat spreader. It was also proven the key wasn't the finish, but the part being as flat as possible, so if you want good adhesion and heat transfer sand the mating parts extremely flat.

 

[Slartibartfast]

Thanks for all the input.

-I will look into the copper heatsink. The problems may be 1. obtaining the billet stock, and 2.Machining the copper (contamination, as our machine shop is all cnc and all shavings are carefully recycled --> separate machines for Alu. and steel.)

-I was planning on lapping the top of the heatsink, and sanding the underside of the star, up to very high grit wet sand. Glad to hear I was on the right track with that.

-The star will be anchored with multiple screws (3 or 4) to evenly apply pressure to the star. Thermal compound will be used.

In regards to the Varapower link, it definitely seems something else is different between the Al and Cu versions. The initial lumen level is very different, and thermal effects related to heatsink efficiency shouldn't be evident in the first 1 second. As current flows through the LED, and heat is generated, then I would definitely expect the Cu heatsink version to start outperforming the Al. Hopefully, I can figure out a way to make it happen.

In regards to buying the SSR-90 right away versus buying a SST-90 and floating it on to a Cree star, I saw the good deals on the SST-90 at 4sevens (and free shipping!) and cree stars are cheap at DX/KD.

I am thinking now it will be better to just get the SSR. I have been looking on AVnet and photonfanatic, any other seller to consider?

Thanks,
Slartibartfast

 

[Nos]

led-tech.de has sst 90s and 50s now

 

[Mettee]

AVNET

 

[bigchelis]

AVNET

The whitest white SSR-90's are the 65k color ones and I cant find those on Avnet website

 

[Mettee]

I notice that if you search the Avnet site, use "SST-90" as your search topic and it will bring up both SST and SSR. Go to page 4 and 5 and the SSR's are listed there. Now they do not have stock right now, but they will sooner or later. I just check back once and a while for what I need.

They do have the star
http://avnetexpress.avnet.com/store...=500201&langId=-1&storeId=500201&listIndex=-1

And this bin 5700- SST-90-W57S-F11-GN201
http://avnetexpress.avnet.com/store...=500201&langId=-1&storeId=500201&listIndex=-1

 

[Slartibartfast]

Got the copper billet on order, at over $7 an inch it ain't cheap. Ordered >99.9% pure OFC in 1.5" OD to lathe down to the ~1.33" ID of the D mag.

Thanks for the AVNET links...probably going to order this one:
SSR-90-W57S-R11-GN201

Cheers,
Slartibartfast

 

[morelightnow]

I don't think so ... I think it is more than 7A. Please take a good clamp meter/multimeter and tell us the current.

I measured 5 amps with my snap-on dvom. I'm sure the wires suck a little voltage from the circuit.

 

[Der Wichtel]

I measured 5 amps with my snap-on dvom. I'm sure the wires suck a little voltage from the circuit.

It's not only the wires:
http://www.candlepowerforums.com/vb/showpost.php?p=3461550&postcount=29

 

[eatkabab]

I'm currently using XPG-R5's (driven at 850ma) in a lamp but I seem to need a little more efficiency (needs to be a bit brighter).

is the SSR-90 a bit brighter? Maybe the XP-G S3's? I think the actual chip these LED's are on are all similar sizes as well.

I'm a little new, sorry. Thank you very much for your help.

 

[morelightnow]

Just measured 8 amps with a clamp on ammeter. The part that sucks is these cells only last about 15-20 minutes at that level, then they drop off to about half the output.

 

[Slartibartfast]

I have ordered the LED's from Fred at Photonfanatic and decided to get them on star for ease of installation. Thanks to Fred for great customer service :thumbsup:, got to recommend him to anyone.