Need help on temperature of voltage regulator board

[Surya P]

Hello Guys,

Need help please.

My wive dive shop is retailing Brightstar 12W LED
http://www.brightstar-hid.com/product_info_e.php?UID=860

A friend of mine bought one and within two dives, the torch died.
So I took a look and found out that the voltage regulator board is out of action.

I am not so into electronics but I can understand a bit here and there:duh2:.
Anyhow I suspect of overheating as the culprit on this voltage regulator board failure.

I can't yet post picture, and I been digging in CPF and someone has photos of the voltage reg I speak of.
http://www.candlepowerforums.com/vb...htstar-Darkbuster-LED-12&highlight=brightstar

The 1st photo is exactly the torch with that white heat shield. Inside that white heat shield is where they hide the voltage reg board. That voltage reg board is then stashed at the side of the battery and near touching the battery, hence they use that white heat shield to probably protect heat from touching the battery.

The 2nd last photo in that thread showed the voltage reg size with the white heat shield remove, but it does not show that on the opposite side of this voltage reg board, it has a flat heat radiator. Just like a coin but so thin. Maybe only 0.2mm.

Anyhow, I have a thin wire type thermocouple ( digital temperature gauge ) and placed that thermocouple wire on the flat heat radiator of the voltage reg board.
I use a small clipboard type clip like this to........:
http://www.rockler.com/product.cfm?...edium=CSE&utm_campaign=V9124&utm_term=Rockler

......to fasten the thin wire thermocouple to the flat surface of the voltage reg heat radiator. I also use heat shield between the thermocouple thin wire to that clip so that there is no bimetal interference and also the clip can't be a heat radiator.

Testing is open air, ambient temp 31C.

In 30 seconds, the voltage reg board temp at the thin heat radiator side is 49.6C
In 60 second it is 60.8C
In 3 minutes 3 seconds, it is 77.6C
In 4 minutes 1 second, it is 81.4C
Slow rise after this and in 7:12 , it is 84.6C
In 8:06 it is 86.7
In 9:28 it is 89.5C
In 10:07 it is 90.3C
Slow rise after this
In 12:57 it is 90.8C
In 13:30 it is 91.8 and maitaining 90is until 17:24 it is 92.5 and later dropped to 90.4C at 18:20.

I stopped the test.

Start with full battery voltage, with light on voltage reads 12.6V at regulator input and its output at approx 12.8V

After 18+ minutes test, with light on voltage reads 12.2V at regulator input and its output at approx 12.7V . Battery use is 4,400Mah 11.1V on nameplate/sticker.

This voltage reg board uses LM5171 by HTC.
http://www.actron.de/de/data/HTC/LM5171.pdf

I looked at its maximum junction temp, it is 125C


My questions are :

01. Am I correct to assume that if I am reading 92.5C at heat radiator of the VR board, that is not LM5171 actual junction temperature since I am reading some centimeters away from the LM5171 ?

02. Is it fair to assume that in a voltage reg design, if it is that small like what Brighstar uses and with so thin heat radiator, the 92.5C temperature I am reading may also be the temperature of other supporting semiconductors on that VR board ?
WIll be be fair to assume that no matter how good the LM5171 temperature tolerance is, in a board like what Brightstar uses, the failure can be from 1 component on that VR board which has a lower temperature rating than the LM5171 ? I never seen heat radiator so thin like that. I thought semiconductor heat sink are supposed to be with lots of fins or at least the steel frame of say a radio ?

The MC-E Cree used on this 12W torch did get warm in open air but not hot. Approx 45C after 18 minutes.

03. Can I assume that in a 45 minutes dive, the MC-E will warm up the so-small air space within the torch internal, heat from the VR board too and the battery itself being discharging all in all will bring the ambient temperature inside the torch to be 45C. If assuming 45C is the ambient temp, can I add 14C to that 92.5C I am reading as estimated heat of the voltage reg board if those 18 minutes test were actually of the torch being used in real life* application. ( *my test is in open air with every component of the torch laid out on a table, so it gets great natural ambient cooling ).

I am kind of sad looking at this torch compared to Underwater Kinetic AquaSun where it even has a special cooling radiator that touches sea water outside of the torchlight.
Look at the middle of the lens assy, that is the heat radiator. UK lights are plastic body, so it can't benefit cooling from its own body submerged in water.

http://www.amazon.com/gp/product/B001P4DLYW/?tag=cpf0b6-20

UK AquaSun has all the electronics on that torch head and has aluminum frame for cooling, unlike Brighstar which hides this hot VR board tucked near the lith-ion battery.

I am kind of afraid when something so small like this VR board and so hot , is actually placed next to a lithium ion battery ( cramped space in there )and with only that white heat shield plastic/rubber like material. Can the battery get hot spot and overheat too ?

UK light makes lots of lights for explosive environment, fire fighter and so on. I am sure they have lots of safety consideration into their design.

Brightstar is HID manufacturer for automotive and for some reason they do dive lights recently. There is not even hydrogen adsorber pellets on their torch lights with regular disposable batteries. They don't even understand that it is important.

What is a good working temperature of VR board or its components for longevity ?
I am behind time with semiconductors. In the 1980s 55C is the max temperature for ICs , that's all I could remember.

Many many thanks for those who will help answer my questions.

Regards,
SP

 

[MikeAusC]

Any consumer electronics where parts rise above 50 deg C is a poor design - people will be injured and the electronics will have a short life. Putting parts reaching 90 deg near a LiIon battery in a pressure tight container should be classed a criminal act.

I have a 10watt Halogen bulb torch in a similar size plastic housing and have had no problems - even plastic provides reasonable thermal conductivity when immersed.

I even made a 50 watt halogen bulb dive torch in a housing that was somewhat larger, but I had an internal fan to distribute the heat internally.

 

[Surya P]

Agree Mike, agree.

Thank you

 

[SemiMan]

Perhaps as designed it was never supposed to get that hot?

I would check all the solder connections to start and go from there.

Semiman

 

[Surya P]

I dont know, I dont really understand electronics all that much.
However since the VR board can increase voltage, probably the heat comes from that switching job.

To raise DC voltage in the old fashion way was to convert to AC and use transformer and again to DC . I dont know how does the LM5171 and its supporting circuits does the voltage rise. However, based on the LM5171 spec, such temperature rise from 30C to 92C does not produce very stable output both in voltage and current. So I guess this is one of those bad design.

The designer may know electronics but as part of a dive torch equipment, they should do more R&D for real life application for this VR board.
If only they can increase the torch dimension by say 3-4 cm longer, stacked this VB board below the LED module, increase cooling fin size, use non heat conducting screws to prevent heat transfer from the LED module to this VR Board, this could be a better design. Better add a battery temperature sensor on the battery to turn off the light at X celcius battery temperature.

If they are more creative, they can make a torchlight head using aluminum and keep the plastic main body. That torch head at that size when exposed to sea water will be so cool and they share the VB board cooling fin ( no need to isolate ) with the LED module if such hybrid design is the way.

Expansion rate of aluminum head vs plastic body can be sorted out with suitable o-ring to accomodate expansion of the alu head ( if any ). They then don't waste money new mould for body and perhaps US$10 extra cost can make such an alu head.

I had doubts when I first saw ALL plastic body torch at that near 700 lumens. I know the heat of the LED module will be plenty but I never thought the VR board is another problem for Brightstar since I assumed they will place such hot VB board to the same LED module cooling fin.

Thanks

 

[Surya P]

All soldering on VR board is OK.