Please help n00b diagnose puzzling Mag623 problem (Problem fixed! Beamshots!)

[ejot]

Hi guys. :wave:
My light has a problem and it would seem that .

Finally got around to putting together my Mag623 tonight. Pretty standard recipe: 3D host, Osram 64623, AW soft start 3-level driver, 4x AW IMR26500 w/ PVC sleeve, stock tailcap spring trimmed, bent and sunk into bottom of tailcap (I machined a small flat groove at the mating surface to ensure a very solid contact).

Popped in the IMR's fresh off the Pila (4.19-4.20V), took her outside and hit the button. It worked magnificently, absolutely stunning. Left it on low for about 30 sec, then on mid for another 20s or so, then to high. After maybe 20 seconds on high*, the light cut off completely and suddenly. I immediately pulled the batteries and let it cool off.

Put in another bulb (latex gloves worn at all times during bulb handling), and tried again. No go. Tested both bulbs and a third NIB bulb, and all had ~0.4Ω between the pins. So it seems they didn't

Loosened and re-tightened the driver/switch set screw several times and really torqued it down to make sure of good contact. Measured resistance between host body and switch body and got ~0.1Ω ... put everything back and tried again, nothing.

Pulled the batteries and found two at 4.09V and two at 4.10V.

Tested continuity between tailcap/spring assembly ... same as before first assembly at ~0.3Ω.

Put everything back together except the bulb and head/reflector. Turned switch on and measured voltage between pin receptacles -- nothing, 0.0V.
So I did a few tests involving the switch. I don't really understand the workings of the driver, so these were pretty much random tests. With the switch alone and out of the body, I measured resistance between the "+" batt contact point and aluminum switch "body".... infinite. Then I measure voltage between those points and found about 50mV. Just measured this again (about an hour later) and found 42mV. Maybe some sort of slow discharging capacitor in there? Or is my cr@ppy Craftsman DMM being screwy? --right now all my good meters are in my lab at work.

Finally, I put the switch and batteries back in and measure voltage between the negative of the bottom battery and the switch body. And to my surprise, I find the full battery voltage of ~15.6V. Likewise 15.6V between batt "-" and host body. That seems like a short through the switch => current would not flow through the bulb. Yet there is no continuity through the switch when tested alone, as mentioned above.

Can anyone figure out where I've messed up here? Sorry for the lengthy post, just trying to give you all the info I can think of.

Cheers,
-ejot

 

[Chodes]

Re: Please help n00b diagnose puzzling Mag623 problem

I can't see where you have screwed up.

I'd say test the AW sofstarter out of the light. You can use a low wattage bulb to make it easier. If it won't work with a 20W bulb , no need to check with 100W.

 

[ANDREAS FERRARI]

Re: Please help n00b diagnose puzzling Mag623 problem

I can't see where you have screwed up.

I'd say test the AW sofstarter out of the light.

If the bulb and the batteries are OK then it has to be the switch(or a bad ground).Maybe someone smarter than me and Chodes can figure it out!If I were a betting man I would say the connection between the bulb and the switch!

 

[Patriot]

Re: Please help n00b diagnose puzzling Mag623 problem

I can't imagine what else it could be besides the switch, especially since initially it worked fine. Your batteries are good and your bulbs are good. The failure occurred while it was running so I don't suspect and contact issues.

I'm leaning toward the switch too.

 

[ejot]

Re: Please help n00b diagnose puzzling Mag623 problem

Thanks for the feedback. I was leaning toward the switch as well but was hoping there was something simple and obvious I could have been overlooking.

Should be able to dig up an appropriate power supply at work to test the switch alone with a 20W bulb.


Oh, and ....... :wow: this thing was really bright. :rock:
Briefly. :mecry:

 

[ejot]

Re: Please help n00b diagnose puzzling Mag623 problem

I'd say test the AW sofstarter out of the light. You can use a low wattage bulb to make it easier.

I played with it some last night and the switch does seem not seem to be working right, though I did get a completely unexpected result as well.

This is a 20W MR16 bulb. I sawed a D battery holder in half and used it to hold the switch/driver. 12V computer power supply is all I could dig up. It couldn't deliver the full 12V under load, but did give me about 9.6 which should be sufficient for testing. (Driver should give Vin = Vout for 6V - 30V)

When I first put power to the driver, the bulb lit up ....... without touching the switch at all. :thinking: And wouldn't change levels or turn off.
Fluke is showing current and Agilent is showing Vin.

After playing around with some other bulbs, trying the switch in the light body again, etc., I tried to duplicate the above and could not. Same exact setup, bulb wouldn't light. :thinking: No response from switch.

Bulb would of course still light when driven direct:

:shrug: :shrug:

 

[Chodes]

Re: Please help n00b diagnose puzzling Mag623 problem

Well , the good news is that AW's service is fantastic , I'm sure he'll look after you.

 

[StarHalo]

Re: Please help n00b diagnose puzzling Mag623 problem

How does somebody with that many flashlights/much hardware/knowledge have only 25 posts?!

 

[ejot]

Re: Please help n00b diagnose puzzling Mag623 problem

Well, I only somewhat recently discovered "good" flashlights. I may have gone a bit... ah... overboard. But I do really enjoy all of them.

As for the other stuff -- just the result of other hobbies and tinkering, and twenty years of taking apart whatever I can get my hands on. The Agilent multimeter has a funny story. Found it in the garbage of an electronics lab when I was in school, and asked the professor if I could have it. Said sure, didn't work.

First thing I try is replacing the fuse. Voila, like new. :nana:

Knowledge? I have been reading these boards a lot, but I can only say that I have much, much more to learn. There is a tremendous wealth of information around here. :twothumbs

 

[KiwiMark]

Re: Please help n00b diagnose puzzling Mag623 problem

Well , the good news is that AW's service is fantastic , I'm sure he'll look after you.

I had a problem with one of my AW softstarter switches - I returned it to AW, he fixed it and sent it back. No problems since. AW is a top guy and stands behind his products. I have 2 switches and a bunch of Li-ion cells (14500, 16340, 18650, IMR 18650, IMR 26500). I would happily recommend AW to anyone :thumbsup:. One of my switches handles somewhere around 180W (64458 + 5 x IMR 26500) and works flawlessly (not too sure if it is the repaired one or not). The other one is currently running a 64623 + 4 x IMR 18650.

ejot:
Contact AW and let him know of your problems. Probably you will end up sending back the switch and he will sort it out for you.

 

[Polar Light]

Re: Please help n00b diagnose puzzling Mag623 problem

Hello ejot!

How did this end up? Did you solve the problem?

I am asking this because I have very much same kind of symptoms. The only difference is that I am using 4c M@g.

Checking the bulb, batteries and contacts didn't help.

PL

 

[ejot]

Re: Please help n00b diagnose puzzling Mag623 problem

AW's service has been great and a new driver/switch is on the way. :twothumbs I will update with results when I get it all back together.

 

[ejot]

Re: Please help n00b diagnose puzzling Mag623 problem

It's alive!!

As you guys said, AW's service was top-notch. Sent back the old switch no questions asked and received a new one today.

Put everything together, went outside and fired her up. I forgot how long it takes the filament to get up to temperature, so it went something like this:

:sigh: ... :shrug: ... ... :mecry:... oo: ... :huh: ... ... ... :naughty: ... :laughing: ... ... lovecpf









A couple shots:

And of course I had to set some paper on fire :devil:

Thanks for everyone's thoughts and advice.
lovecpf

 

[LuxLuthor]

Re: Please help n00b diagnose puzzling Mag623 problem

Yeah, AW is one the all time best sellers here at CPF.

I saw this thread too late, but a couple other ideas I had were to lightly sand the bottom of the switch contact for V+ of battery. Then put either legs of paper clip, or lengths of the center wire from spare cable coaxial cable cleaned with sandpaper into the two bipin holes and clip your leads on them and measure voltage as you press the low/med/high/off settings. If neither of those two made any output difference (in addition to what you already tried), then a component/connection failed in the switch.

Which Fluke model is that, as the Pulse Width Modulation output from AW (& AlanB/wquiles/JimmyM's) switch is most accurately measured with the AC + DC mode, and why I upgraded from 179 to 189.

Good show on the final image...as the absolute test of the '623 !!!

 

[ejot]

Those are both very good tips. I did actually do the paperclip test with equally confusing results, but didn't try sanding or DeOxit'ing the contact. Should have thought of that, as I've played with vintage 6V car electronics long enough to know how touchy a seemingly clean connection can be.

The Fluke is an 81438 ... the Sears-only model. Not the most capable. I do have a 179... but a 189 would be better, eh? I like excuses ... er, reasons to buy new toys.

 

[LuxLuthor]

Those are both very good tips. I did actually do the paperclip test with equally confusing results, but didn't try sanding or DeOxit'ing the contact. Should have thought of that, as I've played with vintage 6V car electronics long enough to know how touchy a seemingly clean connection can be.

The Fluke is an 81438 ... the Sears-only model. Not the most capable. I do have a 179... but a 189 would be better, eh? I like excuses ... er, reasons to buy new toys.

I bought my 179 from a member here, and it has been perfectly fine....until I needed to start doing testing and reprogramming RMS measurements for these PWM regulated driver/switches. It was misleading that the 179 says "True RMS DMM," but the AC + DC mode was missing. I got the 189 from EBay for a great price--I think just over $200. Check EBay photos, ratings, and all the other steps to protect yourself if you go that way.

I don't think it would have made the difference on answering Yes/No on that switch working, so don't waste your $$$ if you won't use extra functions/accuracy of 189.

 

[csshih]

Re: Please help n00b diagnose puzzling Mag623 problem

As for the other stuff -- just the result of other hobbies and tinkering, and twenty years of taking apart whatever I can get my hands on. The Agilent multimeter has a funny story. Found it in the garbage of an electronics lab when I was in school, and asked the professor if I could have it. Said sure, didn't work.
First thing I try is replacing the fuse. Voila, like new. :nana:

................

YOU LUCKY *obscenity* !!!

hehehe.

 

[RichS]

Re: Please help n00b diagnose puzzling Mag623 problem

Put everything together, went outside and fired her up. I forgot how long it takes the filament to get up to temperature, so it went something like this:

:sigh: ... :shrug: ... ... :mecry:... oo: ... :huh: ... ... ... :naughty: ... :laughing: ... ... lovecpf

I was not aware that there was a "warm-up" period to get to full brightness on the 623 (or any other incans) other than the couple of seconds for the AW soft-start. I know it takes anywhere from 30-60 seconds on some HIDs, but I've never experienced this with an incan. Of course, up to this point my brightest incan was a Mag85, but I'm helping a friend finish a Mag623. We are using the same 3D w/ 4xIMR 26500s + AW soft start config, and I was just curious about this delay you mentioned.

Also - thanks for the great beamshots!!

 

[LuxLuthor]

Re: Please help n00b diagnose puzzling Mag623 problem

I was not aware that there was a "warm-up" period to get to full brightness on the 623 (or any other incans) other than the couple of seconds for the AW soft-start.

There isn't any warm up period of the bulb as far as lumen output...as you say, only a soft starting delay built into the switch to avoid flashing.

 

[ejot]

That makes a lot more sense! I think I was too amped up about the light to think that through correctly the other night. With all the anticipation, that second or so seemed to last foooooooooreeeeeeeeeeeevvvver. :sweat:

RichS, glad you liked the shots.