Ideas to hack ELENCO soldering station lamp

[Back2Volts]

I have one of those lamps with with an illuminated magnifying lens, with the third hand, etc...    Using it for soldering, I can see down to 0603 with some imagination  .   The problem with it is that the LEDs in the lamp are too weak for my tired eyes and I have to use my halogen desk lamp in the max power setting.   This is a bit awkward because the lens itself will project a ring shaped shadow which I have to somehow avoid.   

I am wondering if it would be possible to hack the lamp, replacing the LEDs for more powerful ones and possibly introducing some way to have adjustable intensity.   Today I opened the ring to take some pictures.   The lamp is powered by 4 batteries in the base, which I could replace with an external brick or wall-wart. 

The LED size seem to be 3mm by a bit less than 3mm, which I guess could be 3028 or may be 3528.   The lamp is rated as 6V/.5A or 4xAA batteries.   

Any ideas would be appreciated.

The lamp is this one

https://www.amazon.com/s/ref=nb_sb_noss_1?url=search-alias%3Daps&field-keywords=elenco+soldering+station

[Back2Volts]

I feel overwhelmed to say the least !   

I can think of some possibilities, all of which will require change in power supply / driving in some degree.   In order of my preference:   

• Over driving the existing LEDs.
• Replacing the LEDs in the existing ring with higher luminosity ones
• Replacing the whole ring assembly with an off the shelf 110mm 12V ring (eBay "angel ring")
• Designing a new ring system -- I rather not.

 

I am a little puzzled about the actual configuration of the LED strings.   It operates out of a 4xAA batteries, 6V.   A search on Mouser for similar package LEDs comes up with parts with nominal voltage of 3V or 6V (must have two inside).  I suspect the 6V ones are more exotic parts, so it probably uses 3V ones.   Would they be in 2x strings with barely no resistor in series or single LED with limiting resistor.   The first option would be affected more by battery decay, while the second have a significant waste of battery energy.

The only way know if to do continue with the tear down and take some measurements.     


           

I think I need to figure out how the LEDs are configured in the lamp.   I was looking at available LEDs in similar packages and seems that most white ones have voltage ratings around 3V or 6V.   

[jitter]

First of all you need to find out how the LEDs are arranged. From your photo, I'm guessing parallelled strings of each 2 LEDs in series, so most likely 3 V LEDs.

I don't see any current limiting resistors, does it have any?
Could you measure the current drawn from the batteries?

Unless they're driven really low (how hot do they get after some time?) I don't think overdriving would be good for the longevity of the LEDs, it doesn't look like the copper traces are big enough, nor is it a metal cored PCB.

To get an idea what the copper pad surface must be like to get enough cooling from a non metal cored pcb I attached a picture. Never mind the inside of the driver, look at the pcb, you can see the pad size shine through. This is a 14 W LED-luminaire with 28 LEDs, so 0.5 W each.

[Buriedcode]

Those are standard PLCC2 LED's - they likely don't have much in the way of heat sinking as they aren't power LED's, so the pads won't be that big, and I very much doubt the boards are backs with aluminium.
You can however try and find power LED's (I believe Cree have some) that will fit on that landing, it would just a be a case of wiring them up and either changing any series resistors, or replacing them with 0-ohm jumpers and using a CC power supply (ebay has many modules!).

It wasn't designed for high power, 1-3W LED's, so I doubt it'll have the capability, but it doesn't hurt to try.  With a CC supply you can keep ramping up the current to the LED's maximum whilst checking their temperature.  If you find its bright enough, and the board remains warm - you'll be golden.

Just noticed their two LED in series.  Which is odd because 4 alkaline cells = 6V, so thats 3V per LED.  It doesn't look expensive enough to have a built in boost converter, but you never know!  Still, that means an off-the-shelf buck converter, constant current controlled only has to kick out ~7V, so a standard 12V PSU will do.

[Back2Volts]

I opened the base of the lamp just in case there are any electronics.   Nothing there other than connection to battery and optional external PS (which I have never tried - I just noticed the jack yesterday).

I removed the ring from the lamp and found that the four segments are interconnected with two pin headers, making very easy removing one for examination.    There are no resistor anywhere.   Looking at the board with back light, it is clear that there are two LED strings. 

I measured the current consumed by the whole ring is about 20mA, which I comes down to 2.5mA per string.   I think that even for not being power LEDs, this is a very, very low current.   The voltage is 5.4V.   Battery pack with no load 5.427.    It appears the LEDs are under volted by a 10%, which has to decrease output considerably.   It is looking like powering this lamp from batteries may not be a good idea.

Next I looked for and found a 6V/1A a wall wart and put it in place of the batteries.    Actual voltage under load 6.1V, with a total current of 152mA.  That corresponds to average 19mA / 3.05V per LED.   

At this point, since the output has increased considerably, I am going to put the lamp together and give it a try for a while with the external supply.

If it does not give enough light for my eyes, I may look into increasing the voltage a bit.   I don't think there is a way to know the specs of the diodes, but I would think 19mA are on the moderate side.

Any guess as to what current these LEDs could take ? 

[Ian.M]

That's absolutely horrible.  It looks like each segment could be hacked to place all its LEDs in series and add a current limiting resistor^*, (because LEDs in parallel are evil  ), then you could use one of those cheap EBAY/Amazon boost converter modules to provide a 15V supply you can adjust (tweak) for the desired brightness.   I would certainly expect those LEDs to do 25 to 30mA, but whether or not they get significantly brighter at 30mA is another matter.   If the 6V 1/2A spec could be believed they would be supposed to handle 62.5mA each, but that spec is almost certainly just marketing blowing smoke up your ass.  Is there enough space in the base for one of these?
https://www.amazon.com/SUNROM-Step-Boost-XL6009-Piece/dp/B00V9Z2TJA

43mm * 21mm * 14mm (L * W * H)
Heck, if you can cram a pot in there as well, you could give it user adjustable brightness.

* Make the track cuts in red and add solder bridges and 100R 1206 resistor in green.  The result will pass about 30mA at 15V with  20% drop across the resistor for stability.  Drop to 14V it will draw about 20mA, push it to 16V it will draw about 40mA.

[jitter]

Instead of hacking the pcb, I would probably go for a similar module with adjustable current, essentially making it a constant current driver.
Then experiment a bit, and if necessary replace the LEDs with more efficient types.

[Ian.M]

@Jitter: LEDs in parallel!!    Call me old-fashioned, but if I know I'm going to have to maintain it, and am modding it anyway,  I much prefer rearranging them into series strings each with its own current limiter. 

[Back2Volts]

Instead of hacking the pcb, I would probably go for a similar module with adjustable current, essentially making it a constant current driver.
Then experiment a bit, and if necessary replace the LEDs with more efficient types.

As it is, there are eight parallel strings of two LEDs.   Do you mean driving constant current, lets say 200mA, through the parallel bundle and let the strings consume as they may ? 

[Ian.M]

As it is, there are eight parallel strings of two LEDs.   Do you mean driving constant current, lets say 200mA, through the parallel bundle and let the strings consume as they may ?

Add cheap and possibly flakey inter-board connectors to the mix and it isn't unlikely the nearest board to the stalk would intermittently get all 200mA, and at 100mA per string, probably burn out fairly quickly.

[Back2Volts]

@Jitter: LEDs in parallel!!    Call me old-fashioned, but if I know I'm going to have to maintain it, and am modding it anyway,  I much prefer rearranging them into series strings each with its own current limiter. 

What you are suggesting is a variation of the PCB hack above to lengthen the strings and then current drive each one.   That would be great but I have a few space constrains

• The flex goose neck is pretty thin, with OD of 9mm.   I am not sure what the ID, after discounting the corrugated wall, will allow to run multiple wires.   If I hack it as 4 strings, sharing ground, I would need 5 wires.   I think two strings, three wires, would work.
• The space in the column at the base is limited.   Think of it as a cavity that holds the standard 4xAA battery holder.   About 30mmx30mmx60mm.   Then the question is if there are small constant current boosters to fit two of them in that space.   May be an alternative would be to use one voltage booster to 28-30V and place two hand made limiting circuits in a piece of perf breadboard.   Any ideas about these small current drivers ? The best I found ready made are these from Germany http://www.ebay.com/itm/S1119-5-Pieces-Miniature-Constant-current-source-30mA-for-LEDs-on-4-24V-AC-DC-/331895438474?hash=item4d46827c8a:g:XlMAAOSwYmZXK25f
• Limited space to run wires in the ring.   Two strings should work

Could the simple current limiting use an LM317 in current source configuration for each string?   That would be very easy wire.

[Back2Volts]

As it is, there are eight parallel strings of two LEDs.   Do you mean driving constant current, lets say 200mA, through the parallel bundle and let the strings consume as they may ?

Add cheap and possibly flakey inter-board connectors to the mix and it isn't unlikely the nearest board to the stalk would intermittently get all 200mA, and at 100mA per string, probably burn out fairly quickly.

Very good point.  They aren't exactly precision connectors.   A bit on the loose side.

[jitter]

@Jitter: LEDs in parallel!!    Call me old-fashioned, but if I know I'm going to have to maintain it, and am modding it anyway,  I much prefer rearranging them into series strings each with its own current limiter. 

In series/parallel arrangements, I don't really think of it as much of a problem seeing that integrated LED luminaires running on SELV drivers are widely used. And this necessitates parallelling to keep the drive voltage low enough (SELV must not exceed 60 VDC).
I know the theoretical drawbacks of parallelling LEDs, especially if they fail open circuit, but in practice and on relatively underdriven LEDs, it doesn't appear to be a problem.
The attachments show two such LED-modules.

The first one is a Philips myLiving Denim, 8 W, made up of 7 groups in series, each group consisting of 3 parallelled LEDs, 21 LEDs total. Don't mind the driver, normally it's enclosed in its own housing.
The second is the LED-module of a Noxion bulkhead, 14 W, consisting of 8 parallelled strings of each 12 LEDs in series, so 96 total. As an aside, in the photo you can see I was mapping out the layout as I needed to short out one LED in each string to bring down the light output. "x" denotes the 'start' or positive of a string, "/x" the 'end' or negative. The use of a CC driver meant that the resulting LEDs continued to be run at the same current.

[jitter]

Instead of hacking the pcb, I would probably go for a similar module with adjustable current, essentially making it a constant current driver.
Then experiment a bit, and if necessary replace the LEDs with more efficient types.

As it is, there are eight parallel strings of two LEDs.   Do you mean driving constant current, lets say 200mA, through the parallel bundle and let the strings consume as they may ?

Yes, that's what I was saying. 200 mA divided across 8 pairs would more or less result in 25 mA per pair give or take a few to account for differences.
Of course you will need  to start lower and work your way up slowly while keeping an eye (or rather thermocouple  ) on the temperature of the bigger copper pads under the LEDs.
The Philips luminaire in above post runs around 50-60C/122-140F on the pads after several hours operation. So I guess that's a reasonable temperature to stay under, but retrofit bulbs will probably run the LEDs quite a bit hotter and get away with it, so you might too.

However, the difference between the above mentioned luminaires and this solder lamp is that the latter has a relatively low number of LEDs in series/parallel. So differences between the different strings may work out to be relatively large. In that respect I agree with Ian.M that putting all LEDs in series may actually be better.

16 x 3 V = 48 V, still within the realm of a SELV driver. However, I would still go for an adjustable CC driver with large enough a compliance voltage to drive the LEDs.
This type of driver negates the need for current limiting resistors (and thus no energy wasted in them) and it adjusts for changes in the LEDs characteristics as they heat up. Even an LED failing short circuit won't lead to a cascade effect with such a driver (within the compliance voltage range). Another bonus: you only need to feed two wires through that small ID.

Instead of a dedicated LED-driver you may use a (boost) converter with an adjustable current limit.
"Compliance voltage" is the voltage range of a constant current driver in which it can maintain that current.

[Back2Volts]

Thank you guys for the ideas.   I am going to be busy for the next week or so.   I think, the first thing I will try when I have time is to play a bit with the voltage to see the effect in current and luminosity.   From there I may hack the boards as Ian suggests or may be rearrange it in four strings.   I have been thinking that I could possibly run four 30AWG KYNAR wires, for as many strings.  I assume these wires should be able to handle 50mA.   The challenge would be to get small drivers that fit into the base column.
 

[Ian.M]

The physical difficulties with rethreading the gooseneck and fitting multiple drivers are why I proposed the CV boost module + hacking each quadrant to be a series string with its own resistor.

So it wastes some power - So what?  When its on mains power, even if you push it up to 50mA If, it will only waste 1W.  When its on batteries, the DC-DC converter's minimum input voltage of 3V will let it drain them fully while maintaining brightness till the bitter end, instead of barely touching them before it dims to unusability,  more than making up for the 25% to 30% extra power the resistors and DC-DC converter will waste.

If you find you don't like the cheap and easy CV option, it would be easy to convert to one of the CC options, or even to revert it to its original circuit . . . .

[Back2Volts]

I am not too worried about efficiency.   It does not have to be the best technical solution as long as it performs.    I have a brand new C-DC converter, a variation of the one posted, already on my bench, ready to go.   Is was in a drawer, waiting for the proper occasion.   

I will start by over volting progressively up to 6.5V or so.  I need much higher luminosity and after seeing the lamp perform at 6.1V/150mA, I am starting to think that the set of existing LEDs will not cut it, even pushing them to 30mA.   If that is true, then I may need to replace the LDEs for higher power ones, but that brings other issues to solve.

• As Buriedcode said, the lamp is not designed for higher power (PCB..).
• The plastic lens which encloses the LED assembly prevents air circulation (cooling).   It cannot be removed because it holds the heavy magnifying glass.   There is the possibility of cutting some of the plastic in between the four holding screws, while trying to preserve the structural strength to hold the glass.
• Removing/replacing the LEDs will most likely require a hot air station, which at the time I do not have.

Also, while I am looking for a simple solution, it seems that a PCB hack similar to yours above, with proper current driving would get the most of each LED.   That may get a bit more light, but not sure is worth the effort.

At the end I wonder If I should attach three small ears to the ring with some real power LEDS.

[jitter]

There's only so much you can do with what you've got now.
If you're going to increase the voltage, by all means use a current limiting resistor. A small increase in voltage could result in a big increase in current drawn without it.

While modifying some CC LED-drivers to a lower current I found that halving the current did not result in half the voltage. That was expected, but the amount of voltage the CC driver needed to drop to adjust to the lower current was way less than I would have expected. So going the other way around, some sort of runaway might occur without any limiting (and the heating of the LEDs worsens this).

The Noxion module in the post above (8// strings of 12 series LEDs) was originally driven @ 300 mA. To get that LED-module to draw 300 mA, the CC driver needed to put 33.5 V across the terminals (experiment here). Dropping it to 170 mA still resulted in 32.8 V across the terminals.
So a 2.1% decrease in voltage resulted in a 43% decrease in current, not particularly directly proportional (and the reason why CC drivers are by far the most applied technique).
You can imagine what might have happened if I had hooked it up to a lab supply with the current limit set to max. and upped the voltage just a little too much...

[ebclr]

https://pt.aliexpress.com/item/6W-12W-15W-18W-24W-33W-LED-PANEL-Circle-Ring-Light-SMD-5730-LED-Round-Ceiling/32699233289.html?btsid=eb468b5c-e7c4-4cc8-811b-249c4d045812&s=p&ws_ab_test=searchweb201556_0%2Csearchweb201602_3_10057_10065_10056_10068_10037_10055_10054_10069_301_10059_10033_10058_10032_418_10073_10017_10070_10060_10061_10052_10062_10053_10050_10051%2Csearchweb201603_6&spm=2114.02020208.3.19.6G7l1v

[jitter]

Brrr... I am very suspicious of those things on AliExpress that only name CE and RoHS as certifications (and CCC is only relevant to the Chinese market).
Use the LEDs and replace the driver with something really (and believably) certified.

And what are you going to get 2700 or 3500 K or anything in between? Guess what was on special offer on the Shenzen Market this week... Not good enough.

Edit: note how the 15 W only gets a Wun Hung Lo driver, but the 18 W has the "High-end" LED-driver... 

[ebclr]

Buy one certified in Netherlands for 3K

[jitter]

Buy one certified in Netherlands for 3K

Not sure I understand what you're saying.

E.g. MeanWell drivers, a reputable brand, starting at less than € 10 over here, and that's including sales taxes and bought in a local webshop.

[Back2Volts]

https://pt.aliexpress.com/item/6W-12W-15W-18W-24W-33W-LED-PANEL-Circle-Ring-Light-SMD-5730-LED-Round-Ceiling/32699233289.html?btsid=eb468b5c-e7c4-4cc8-811b-249c4d045812&s=p&ws_ab_test=searchweb201556_0%2Csearchweb201602_3_10057_10065_10056_10068_10037_10055_10054_10069_301_10059_10033_10058_10032_418_10073_10017_10070_10060_10061_10052_10062_10053_10050_10051%2Csearchweb201603_6&spm=2114.02020208.3.19.6G7l1v

I had looked at the rings on eBay, which are mostly 12V designed for cars.   There is a 110mm that could possibly fit.   Then I went to Amazon to see what listed there.   The feedback was horrible.

These ones look a little better, but I think the band size in the 110mm is too wide.   Unfortunately for some reason they do not have the specs for the 110mm.

I will keep it in mind.   Thanks

Note:  The manufacturing tolerance in the 18W is so bad that the ID of the ring is bigger than the OD !   

[Back2Volts]

Buy one certified in Netherlands for 3K

Do you know what internal diameter they have ?     

[Back2Volts]

There's only so much you can do with what you've got now.
If you're going to increase the voltage, by all means use a current limiting resistor. A small increase in voltage could result in a big increase in current drawn without it.

While modifying some CC LED-drivers to a lower current I found that halving the current did not result in half the voltage. That was expected, but the amount of voltage the CC driver needed to drop to adjust to the lower current was way less than I would have expected. So going the other way around, some sort of runaway might occur without any limiting (and the heating of the LEDs worsens this).

The Noxion module in the post above (8// strings of 12 series LEDs) was originally driven @ 300 mA. To get that LED-module to draw 300 mA, the CC driver needed to put 33.5 V across the terminals (experiment here). Dropping it to 170 mA still resulted in 32.8 V across the terminals.
So a 2.1% decrease in voltage resulted in a 43% decrease in current, not particularly directly proportional (and the reason why CC drivers are by far the most applied technique).
You can imagine what might have happened if I had hooked it up to a lab supply with the current limit set to max. and upped the voltage just a little too much...

I had the feeling something like that would be happening.    What resistor value would you suggest ?   50-100 ohms ?