It looks to me that Xera's reflector is not really helping with the throw, but the wider spill.
Originally Posted by colleen c
night riding enthusiast
It seems the reflector is less efficient than the optic. It just restricts the output to a narrower beam without increasing the overall brightness significantly. However, it does make the foreground light less blinding.
Click the thumbs up button if you have a thumb...
A quick thank you for everyones advice. I am using the Gemini on my helmet and Olympia on the bars, I have nothing to compare to but, I can tell you this combination has greatly helped extend my biking season and allowed me to stay in condition. The winter season has been a blast.
just a Newbie
Originally Posted by GraXXoR
Yes, I too notice there is not too much difference in throw between the optic and reflector for the Xera. If anything, it is slightly more with the reflector if I point them straight at the tree instead of the bottle. The reflector does provide some side spill but not so much in the immediate foreground for the first 50ft or so.
Originally Posted by butasan
Effect wise, when using the xera/reflector along with the Olympia, it helps to cut down on the foreground glare while still helping with some extra throw. OTOH, Xera/reflector is not as useful when used alone. The optic is better. I run two Xera ony drop bar commuting bike. Two Xera/optic had too much spill and I cannot see the throw, but two Xera/reflector makes a nice combo light offering a balance of spill and throw.
Depending on the setu, the reflector does have some advantage and disadvantage.
my Magicshine 880 has Way more throw then my Olympia. MS880 for helmet and Olympia for bars is my best combo to date. Comparing the 2 the 880 is a v8 and Olympia is a turbo4
I think I'm going this route.
I like the Xera's light weight, but it seems to me just the weaker version of Olympia so I don't think I will get a good throw regardless of the use of the reflector.
Originally Posted by sevencycle
Originally Posted by sevencycle
Cool. Action LED will ship out the day you order. I have no connection with them other than a purchase.
I rode last night and took my standard t6 magicshine (spot) to compare.TheMS8?? helmet spot did work ok with the Olympia. But when I went back to the MS880 I didnt realize how much spill it added plus the size of the spot was times 2 and whiter.
I feel I have a good opinion on lights being, Old with bad eyesight, Kinda fast usually top 3 Cat1, Experienced, Raced the very first 24hr race in Slatey Fork WV... 1992?? and many after that.
just a Newbie
I just did a review on the Olympia at another forum. I will post the review here for all to see. There is some notes with looking at like the 6cell runtime and lux reading for the User Intuitive programing. Here is the Review I wrote:
Gemini manufacture introduced their line of bicycle lights with the introduction of the Titan P7, since that time, they have now introduced the Gemini Xera and Olympia in which both of these fine lights uses the Cree XML. I purchase both of these lights and was impress by the features and performance of these light. A little over a year ago, I did a review of the Titan P7 and I feel that the Gemini Olympia is worthy of an review. The Olympia light uses 3 XML led to provide a manufacture rated 1700 lumens using an optic that is 19-degree beam. The runtime from their 5200 mah pack is 2hrs as listed. They also offer an optional 6 cells 8400 mah pack, which allows for a 3.5 hrs of runtime. The listed weight of the Olympia light is listed at 78g which is consider pretty light enough to be mounted on the helmet. It has a user intuitive programming for the brightness in all of the three mode plus strobe mode. One can be overwhelm with these listed feature and is curious how well the Gemini Olympia will perform the claimed listing. I bought one of the Olympia light package which had all their accessories including the 6 cell battery, and after several weeks of using this light, I found the Olympia was worthy of a review.
The package with the 6 cells battery arrived a little later than I was hoping for. The DHL tracking showed the package was sitting at Hong Kong for several days and I figure there might have been a problem. Not long after that, I received a notice that Gemini has sent what seem like another package and sent me a notification. I was not home at the time of the delivery and Gemini sends me an email notifying me that the package did not get deliver. It was nice to have some assurance of customer support when there is problem with delivery and Gemini did just that. My package arrived safety and undamaged. The content in the box included the following:
Pro Head belt
The size of the Olympia is smaller than the Magicshine P7. I took a weight measurement of the light head with my scale and the scale showed 82 gram. The difference in the listed 79g vs. my 82g might be in the scale itself. Here are some pictures of the light head along with it in comparison with the Magicshine 808 P7
The weight of the 6 cell battery was weighted at 322 gram including the nylon pouch.. The 6 cells battery was shipped charged at 7.61v. I charged it but notice that the charger did not terminated at 8.4 volt. The charger had 8.7 volt written on it and I figure that was intentional. After emailing Gemini, they assure me that was Ok for the 6 cells battery. I charged the battery and the termination voltage read 8.76 volt. As soon as I unplug the charger from the battery and check the SOC of the battery, the battery volt read 8.42 volt. After charging, I plugged in the light to the battery and notice how bright this light was putting out. I have a lux meter and did a 5ft ceiling bounce test in my 10x8 room to see how many lux this little light head was putting out. I got a reading of 162.5 lux. Not bad for a little light.
One of the neat features of the Olympia light is the User Intuitive programming . It allows for adjustment for the brightness of the light in any mode. It has three modes and a strobe mode. You can access any of the three modes without having to go through the strobe mode. In each mode, you can program the level of brightness in any increment of 10 steps. This feature allow the user to be able to program each mode to anyway they want such that the unit can be program for sequence instead of being stuck in the OEM setting of H-M-L or L-M-H. This feature goes for the same in the strobe mode. What nice about that is that now you don’t have to put up with a strobe that is strobe at 100% brightness. The programming of the setting is very easy and can be set out in the field. I took a lux reading of each of the 10 increments step for the programming and here what I measured.
1- 6. lux
2- 11 lux
3- 24 lux
4- 40 lux
5- 57 lux
6- 74 lux
7- 92 lux
8- 111 lux
9- 130 lux
10- 162 lux
The performance from the 6 cells battery is promise as claim and more. I started with the battery fully charged with a SOC of 8.4v and the battery indicator showing green. Here is what I got:
After 2 ½ hours of total runtime: the indicator turned yellow and the battery read 7.09v.
After 3hrs 10 min of total runtime: the indicator was red and battery read 7.01v.
After 4hrs 5min of total runtime: the indicator was blinking red and battery read 6.64v.
Afterward the unit went into low mode and I stop the runtime test. I was please to find that the Olympia runtime was better than what the runtime was suppose to be. I took a Lux reading with a ceiling bounce test at the last ½ hours of runtime and the Lux meter read 150 lux. This tells me the regulation was pretty close to flat line.
The Olympia is very bright for use in the trail or road. The beam pattern has more flood than throw, however the throw is not really bad at all. It can provide enough throw to be used as a stand alone light, but adding an additional throw light such as the MS 808E xml or a thrower flashlight will spice thing up quite a bit. Gemini offers the Xera light with a throw reflector optic, which I am using currently and gosh does the combination with the Olympia, really lights everything up in front of me. Recently they are testing a different reflector for their Xera line for even a more throw pattern. Here is a picture of the Olympia beam. There is two 1 litter bottle and one 2 litter bottle on the grass at 100ft. The tree behind that is 150ft and the last tree is 200ft. The picture was taken at F4.0, 2sec, ISO200, daylight, and manual focus.
I have much more pictures with comparison of the Olympia beam shot along with many other lights I have in my collection like the MS P7, MS XML, Bikeray IV, Gloworm etc. This album also contain the control shot and daytime shot showing the setting. They can be found here:
Gemini Olympia and Xera beamshot comparison pictures by Colleenlc - Photobucket
In my helmet setup, I had mounted the Olympia along with the Xera on my helmet. I decide to use a different mount than what came with the Olympia. The Oring mount is OK for most but I prefer the mount offered by Cateye although the tilt adjustment is not as fine as using an Oring mount. On the other hand, the Cateye does offer easier removal and somewhat a lower profile. There some mod that has to be done to fit besides just removing the screw. The screw of the Olympia is a large head and the Cateye spacer had to be drill for clearance so that the screw sit flush and slide over the helmet mount with ease after filing the side of the mount. It is easy process and takes a few minute to adapt. The Cateye helmet mount is #534-1831 while the space is #533-8730 model H-27. Here are some picture of the Olympia at the top of my helmet and a Xera in front using Cateye mount.
I’m please with this light so far. The price is higher than some lights but yet lower than some others. I think it is a good upgrade for those who are consider a single brighter light for their lighting system. The small size and User Intuitive programming features will helps with those who seeks a light that is user selectable which is one of the best feature I found with this light.
Nice review... try one more thing if you get a chance.
Take the light indoors and turn it on the highest power output in steady burn mode and just let it run and see what happens...
If you have access to a "laser" thermometer, take some readings at various places around the case and note when/if thermal cutback occurs. If the temperature of the case gets above 140F before any thermal cutback occurs, probably best to abort and shut it off.
Inquiring minds want to know...
just a Newbie
Done! (I work in an industry shop and just happen to have a thermo gun)
Originally Posted by pethelman
I took some reading. The hot spot is dead center top near the bezel. It was only about 2 degree difference elsewhere on the body. I am not sure what the Emissivity should be set at on the thermo gun. I know it is different on Aluminium and not really sure what it is on silver anodized AL. I just left it at the turn on default of 0.93
The unit did not cut back at 140 F.
Ambient = 63
1 min = 83
2 min = 96
3 min = 113
4 min = 123
5 min = 136
5 1/2 min = 140
6 min = 143
Last edited by colleen c; 03-12-2012 at 10:35 AM.
Reason: Corrected ambient and added 1 min reading.
The conclusion from this is that you should probably avoid running the light at full power under conditions of little to no convective cooling, not only to prevent burning yourself but also damaging the LEDs.
EDIT (3/15): Pure conjecture on my part with limited data... Gemini states below in a later post that they have tested repeatedly that cutback occurs at a 63-ish degC case temperature (hopefully under these same circumstances), which is roughly 145 degF. So hopefully just a few seconds more, and colleen would have seen a reduction in power. Please take any analysis below as a "what if" scenario, and realize that turning on ANY high power LED light, STARTING from ambient temeratures with no air-flow is an absolute worst case scenario, and should normally not occur.
So lets assume a 14 watt power draw just in the LEDs. We want to estimate the LED junction temperature based on the measured external case temperature. We know, based on CREE's datasheet that the XML die itself has a thermal resistance of 2.5 degC/watt. So the problem now becomes, "what is the total remaining thermal resistance to the outside world that would be required to keep the junction temp below 150C when the case temperature is at 143F (62C)?" (Obviously the temperature rise was not going to stop at 143, but just for sake of argument). This thermal path includes:
Die-to-substrate junction (usually low),
Substrate-to-mount junction (probably low but dependent on quality of thermal compound),
Mount-to-housing junction (dependent on method of assembly and/or thermal compound),
Housing-to-ambient (dependent on the design of the case, total exposed surface area, and presence/lack of forced convection).
Or simply stated (at 6 minutes):
150C = 14watts * (2.5 degC/W + ? degC/W) + 62C
Solving for ? = 3.78 degC per Watt
This would be VERY difficult under static ambient conditions. As a designer with a conservate bent, I would have liked to have seen the thermal protection kick in around minute 3. Just my preference, NOT a necessity.
It'd also be interesting to see similar test results over a longer time period with simulated air-flow, comparable to climbing speeds.
Last edited by pethelman; 03-15-2012 at 09:14 AM.
Colleen, just out of curiosity is there any way you can do one of those thermal scans on the Gloworm X2....pretty please!
I would think by now that most people would know this about "any" LED lamp. Seeing that there are more and more lamps now using multiple LED's it is real important to power down when stopped or slow climbing for periods longer than ten minutes.
Originally Posted by pethelman
I'm not so sure I'd go that far and say "most" in general, but definitely most of the folks that frequent the boards here have a keen sense of how to handle the high power lights. Hopefully this conversation will increase awareness to some degree.
Originally Posted by Cat-man-do
Last edited by pethelman; 03-15-2012 at 09:03 AM.
just a Newbie
No problem, I was a little curious myself to see how the Gloworm X2 compare. The number was almost identical in term of temperature and time. The mode was at 1200 lumens. Here they are:
Originally Posted by Cat-man-do
Ambient temp = 67.7 F
1 min = 86 F
2 min = 100 F
3 min = 114 F
4 min =125 F
5 min = 135 F
5 min 44 sec = 140 F
6 min = 142 F
6 min 30 sec = 147 F
No cutback even at 147 F
This was the 1200 lumen mode. I can only imagine it will rise faster in the 1300 lumen mode. The temperature rises for the X2 at 1200 lumens mode using two LED is almost the same as the Olympia at the 1700 lumens mode using three LED. Interesting huh?
Originally Posted by colleen c
Just to shed a little light on the situation (couldn't resist) - the thermal cutback at the board is set at 70deg C (158F). We will be reducing this to 60deg (140F) for V2.
I can only suggest that the Watts produced by the X2 are the same or similar to that of the Olympia hence the temp increase similarities. The thermal/surface area characteristics of the cases also seem to be similar. We're looking at 14W max for the X2 at 1200 lumens.
Hope this stimulates or helps the conversation.
OK, I just couldn't resist... per Jim's comment in post #51, bold emphasis mine:
Originally Posted by colleen c
It should be pointed out that the model equation that I proposed earlier is a very simple "steady-state" equation and does not take into account the thermal mass and temperature gradients that exist in the system as the unit is heating up. By virtue of the fact that the shape of the temperature ramp was still fairly linear when colleen pulled the plug, it tells us that the system still had a good ways to go before hitting "steady-state." In other words, it's "possible" that the LED temperature was already actually "higher" than the 150C limit that the model equation predicts at 6 min, but again this is just a rough prediction during a worst case scenario.
Originally Posted by Action LED Lights
The trip point temperature that Jim pointed out (78C) refers to the location of the temperature monitoring sensor, NOT the LEDs themselves. This temperature monitoring point, as he states, is ON the controller board itself. So the issue here is very likely that the thermal coupling (resistance) between the LEDs themselves and the controller circuit is not sufficiently low to enable adequately fast (for my liking) thermal cutback during the transient "heating up" phase of operation as demonstrated here. There are three ways to address this problem:
1. Dramatically lower the thermal resistance between the LEDs and the temperature sensor.
2. Significantly lower the temperature trip point (down from 78C).
3. Make it very clear to the user that this scenario really should be avoided.
#1 can be expensive to implement with a remote temperature sensor. #2 is feasible, but due to the overall low surface area of the case, it's likely that they had to choose this "high" value to sustain operation at steady state in the presence of moderate to low air-flow without the like cutting back pre-maturely. Since addressing #1 or #2 requires extra cost and/or design change, #3 is the only option, and to Gemini's credit, they have put some warning in the literature. Could it be a little more strongly worded? Perhaps, but that's just a matter of opinion. I'm totally OK with a design that is, by nature, capable of operating outside of safe limits, both to the electronics and to the user, as long as adequate warning is given.
Begin Edit (3/15)
My sincere hope for this design and overall preference would be that it could truly protect itself, even from any incremental thermal damage under any circumstance. And it may indeed do that. I can't say definitively one way or the other. Clearly there has been enough testing from the manufacturer to determine that the light will not self-destruct or become an immenent burning hazzard, even under a worst case scenario.
As has been pointed out below, and again, purerly just my opinion, but I do like to take a conservative approach to heat management and stay as far away from the limits as possible without hampering the functionality of the light. Can a more aggressive approach work? Absolutely, and as the manufacturer has pointed out, they have at least included some necessary warning to the user.
Upon further reflection, it was completely unfair for me to make any claims of fact on the safety features of these lights without firsthand empirical data. So I'll leave the technical interpretation of colleen's test as an interesting aside, but no hard conclusions can be drawn from it. END EDIT
Last edited by pethelman; 03-15-2012 at 09:26 AM.
Realizing that this should really be it's own thread... total cross-pollination here.
Originally Posted by Gloworm Manufacture
60C is definitely more appropriate, but you still better have a darn good thermal path to the temperature sensor at 14W. Considerations for heat-sinking the controller board involve MUCH more than just using a good layer of thermal grease. JFWIW
Originally Posted by pethelman
I suspect that most of the drivers are using either a thermister on the board as an input to the MCU or an embedded thermister in the MCU. PCB material will present a poor thermal path between the the housing and thermister. Yes, eventually it will lower the drive current, but only after the emitters have overshot the temperature setting. A low thermal mass housing exacerbates this scenario.
The thermal resistance between the LED and temperature is very low. We have the LED Aluminum MCPCB mounted onto the same piece of aluminum as the Driver PCB. The PCB is multi-layered with heatsinking and has many thermal vias transferring this heat to the outside retainer. The thermal vias have lots of soldering contact with the aluminum retainer. That's as close as you can get without building a light engine on one PCB.
As for thermal protection, it is indeed built in. The threshold is set a little higher than 140F. As Cat has mentioned, it's quite common knowledge that LEDs get hot, especially with such high output these days. Yet, the trend is to fit more power into lighter and smaller products.
We have safety warnings in our user manual. They are located on the very first page under "Safety Precautions". We do inform our users about the extreme brightness (potential eye damage when in direct vision) and heat.
Here it is:
1. When turned on, do not look directly into the light. The OLYMPIA is extremely bright and can damage the eye.
2. The OLYMPIA is intended for outdoor use. The LED emitter can generate a lot of heat is used without airflow to the aluminum casing. Please be careful.
3. Use only the supplied Gemini charger to charge the battery pack. Do not use any other chargers, as it may cause damage to the battery.
With that said, we do expect some people to go without reading through the manual which is why we implement our overheat protection. The protection feature kicks way before any chance of damaging the LEDs.
You are entitled to your opinion, however what you have just said is entirely not true.
Originally Posted by pethelman
EDIT (3/15) You know, I don't mind admitting when I'm wrong, and I must say, I have to agree with you here. It was completely unfair of me to make such an emphatic static as if it were fact, based on so many assumptions, so please do accept my apologies for making that insinuation. End EDIT
Originally Posted by Gemini Lights
We have Jim's (Action LED) comment originally stating a "threshold" of 78C (or thereabouts). This is equivalent to 172.4F. So when you say that the "threshold is set a little higher than 140F", what exactly does this mean? Did the design change at some point from Jim's original comment, or was he just wrong to begin with? It's just confusing.
(Nitpicking here for sure) To the extent that we now have empirical data showing the case temperature going above 140F with no cutback, and this is the generally accepted temperature at which skin begins to burn, I don't believe it's a "completely" false statement to say that the light will become hot enough to burn the skin. Granted, I doubt anyone would actually touch the light long enough at this temperature to get burned, but still, there's at least some rationale. I'm sure if colleen were to continue to let the light run, we'd eventually see the thermal protection kick in, and hopefully, under these conditions, it would not get so hot that it would "instantly" burn when touched. Would colleen be willing to run the test again to the point where we see thermal cutback occur? Hmmm.
It's not my intent to spread misinformation, so I agree with you, I cannot state with certainty that damage WILL occur to the LED under the test conditions that colleen performed. We have limited test data, so all we have to fill in the gaps is best engineering judgement. I've made a couple of "critical" assumptions in the process that could certainly stand to be corrected if need be.
Assumption 1: Max power draw on high = 14 watts
Assumption 2: LED case-to-outside world thermal resistance = 3.78 degC per Watt
I made assumption no. 2 by giving the light the benefit of the doubt at 14 watts, and assuming that the steady state output had been reached at the recommended LED die limit of 150C or 302F. My experience says that this number for thermal resistance for this link in the chain is at least VERY conservative.
It's reasonable to assume (based on the good design technique that you highlighted) that the thermal resistance to the on-board temperature sensor is indeed lower than the composite value to the outside world. This means that the temp sensor should be higher than the value measured at the outside of the case during the 6 minute test.
Lets assume a LED-case-to-temp-sensor thermal resistance of 2.5 degC per watt (for a total of 5 degC per watt, when including the LED die itself). In this case, using either assumption (140F or 172F) for the temperature threshold for cutback, we should have seen the light cut back during the test. If the thermal resistance was lower than this, then we should have seen the light cut back even sooner.
Can you explain why the light didn't cut back during colleen's 6 minute test? And perhaps more importantly where I've made any gross mistakes in my assumptions? The chart attached here is reflective of those assumptions. Again, I have to re-iterate that these are estimates based on assumptions, so my apologies for making it sound like confirmed facts. Would love to see some actual test data confirming your statement that the "protection feature kicks *way* before any chance of damaging the LEDs" under the conditions of starting from ambient and turning the light on full power with no air-flow.
Last edited by pethelman; 03-15-2012 at 09:41 AM.
The cutoff point normally occurs at 63°C ± 1°C (temperature of the outside case). We have tested and retested our lights to ensure the safe handling for our customers. Whilst we appreciate your concern for safety, you can be rest assured that it is perfectly safe to use our light. We don't appreciate your spreading of false claims, saying that the light will in fact burn people and permanently damage the LEDs. We would not release a product if this was the case.
As for Action LED Lights statement of 78°C, I feel that it was perhaps a miscommunication and the 78°C was referring to the cutoff point of the temperature sensor on board.
Now put this all aside, there is never the case where the use of a high powered LED light (i.e. Olympia) must strictly be used on the highest mode, indoors and without ventilation. We say this because low mode is all that is ever required to see things indoors and without moving. We will leave the other reasoning up to you because we don't feel it is necessary to state such obvious facts. Normal common sense tells you that if you feel the light heating up, reduce the power. If by chance, you can't do this for whatever reason, the overheat protection kicks in.
Much ado about nothing?
Guys, I'm feeling conflicted about this discussion. On the one hand, I'm running 2 Xeras and and Olympia and I love the way they light up the road for me, especially at the price point. On the other hand, I have a Designshine DS500 on order and I know it will be a kick-a** rear light. You both have slightly different philosophies - Gemini may be a little on the edge of pushing the specs, Designshine a bit more conservative. I believe there is room for both points of view in the marketplace. I understand that my Gemini lights may overheat if I am not attentive. I hope other users also understand this. What it comes down to is buyers doing adequate research and making informed decisions that fit their riding style. While your back-and-forth is interesting and helps keep us informed of the technical issues accompanying every design decision, at some point, methinks you protest too much. While I fully expect to enjoy my Gemini lights for the next couple of years, I realize that the technology is moving so fast that I will probably lust after some new hottie before my Geminis give up the ghost. The DS500 should serve forever, because I don't think more than 500 lumens will ever be needed off the back. So let's not argue the fine points too aggressively- you both are great companies with a proven record of listening to the customer and providing superior customer service.
Thank you jgmarcotte. I very much agree.
We do try to push the performance to the very edge to give you the maximum you can get out of your equipment. We have never put down other products, in fact we feel the DesignShine makes a great set of tail lights. We strongly promote healthy competition, where it only drives product development further than it is today. In the end, with more competition around, only the consumers win.
By Timo in forum Endurance XC Racing
Last Post: 04-29-2012, 06:37 PM