Can I use 8.4V with this Lambda driver?? <----YES I CAN!!!!

vob, the “millipede” is not an MCU, it’s what’s called a timer. You don’t program those. I guess they are considered a discrete logic component, although I’m not really sure if that’s the proper term. Anyway their function is not programmable, they are a building block for simple circuits. (555 timer IC - Wikipedia) If I were to implement LVP it would be with an MCU such as an Atmel ATtiny13A (which does need to be programmed).

As to what’s in the picture you referred to, it’s definitely not a 2-cell NiMH driver. It’s just a miniature version of the Varapower DD driver in your OP here.

Here is an update. I stuck this light in my make shift sphere to get a general idea of the results.

Turn on.......... 3222
30 seconds .....2878
1 min..............2718
2 min..............2552
3 min..............2241

I have to say I am quite impressed with this. Less than a 400 lumen drop from turn on to the 30 second mark, and a steady output even minutes into the test. At 8.3 amp I was expecting the drop from turn on to 30 seconds to be more than that. I guess there really is something to say for a large mass of copper.

I double checked to see if there was really anything more to gain going above 6 amps because previous tests on a smaller heat sink had indicated there was not much. At the 1 minute mark at 6 amps the light tested at 2224 so the 2.3 amps of extra current are yielding almost 500 extra lumens.

Hey, guess who just got an Email from Lambda. Yep, thats right. His sister No wait it was me!

He said that the max for this driver is actually 16V at 30A. WOW huh...

I sent him everyone's best wishes from BLF and CPF and told him how much we all miss him, and that the praises of his work still ring though our hallways..

Sorry for taking so long to get some pictures. But here you go. Glad you were able to get ahold of Kevin. I had emailed and didn’t hear anything back, glad he’s doin alright.


Lance

Thanks lawallac! Would it be possible for you to post the top markings that are on the components?

The two small black resistors are easily legible, as are the two large yellow capacitors. The large inductor is also clearly marked. What’s left is:

  • Large(est) black FET.
  • Medium sized black diode next to FET.
  • Medium sized 3-leg+tab (SOT89?) component between that stuff and the red/black wires.
  • 8-legged component between black and green wires.
  • 8-legged component between green wires.

I’ve had some difficulty reading the components but here is what I make of it.

Good work lawallac, thanks!

  • Unfortunately I think I just can’t get enough traction on that boost driver with the info we are going to have here. 2/3 FETs are readily identifiable from the markings you posted. The SOT-89 part you marked with a green line is a mystery to me, as is the FET marked with the grey line. I really don’t understand why the boost converter needs so many FETs.
  • I assume these lights also feature a pot? If so I also don’t necessarily understand how that part of the driver is hooked up (it might explain some of the extra FETs though…)
  • I have not attempted to diagram out the driver.

I think I’ll have to bow out of attempting to clone this boost driver. Maybe I could ask Lambda for help understanding it if he’s willing. Not right now though.

No pot on this light. I don’t know if you had a chance to read through the thread Here , but he did mention difficulty in finding a FET that operated in the 2.4 voltage range. I will have to also peak inside the 2c versions.

Hmm, it’s starting to come together now. A little.

I’ve only skimmed the first page of that thread once or twice. At your suggestion I read the entire thing. Lambda doesn’t say that it’s voltage-controlled, but it is. 331C1 is a voltage controlled (CV) boost controller for 3.3v. I have been unable to find a datasheet, but 501C1 also exists (5v instead of 3.3v). Here is a clear photo of 501C1 from this thread on elektroda.pl. I see some more discussion on 501C1 over in this thread on the same forum.

Voltage controlled seems to be able to work acceptably, at least in some cases. DerWitchel also used CV in a driver (step-down though) for SST50/SST90.

In this case we’d see a little lower current on XM-L2 or a very low (<2A) current on XP-G2, based on djozz’s charts (post #51). Of course MT-G2 or strings of LEDs are both completely out of the picture, but that was really true anyway on 2s1p NiMH - regardless of driver selection!

It would still be worthwhile to me to find out exactly how this driver operates, but I don’t think that actually cloning it will produce great results.

Nice detective work wight. What your doing is black magic to me.

We’d be in pretty good shape if it wasn’t black magic to me too, eh?! :smiley:

Looks like the Russians have some different thinking on the subject of those boost controllers. Somebody over on radiokot.ru is saying that it looks like MC33463H in an SOT89 package. Another says ELM92501C-N (well, ELM92xxxx) in the SOT89 package.

The Polish folks thought that it was it was SOT223, but I’d been assuming SOT89. SOT89 has a tab which is flat and connects back to pin2 (same package as our 7135’s). I don’t know much about it, but it looks like SOT223 has a tab that comes out and then bends down - I’d say the pictures make it clear that that is not what we’ve got here. So we’ve got SOT89!

Moving on, the difference between the two current proposals is really just the location of Vout vs GND. They are transposed between the two components. Oops. Looks like the ELM product is only available with an internal FET, we need one with an external FET. MC33463H–XXLT1 still fits the bill I think!

Here is what is inside of a 2C. Tight quarters in there!

I have 2 of these 2Cs. In this one Kevin made me a nice adapter to run an 18650.

Update.

Just to be on the safe side I did ask Lambda about how he would feel if we cloned this board ( I wanted to make sure he was not offended or felt ripped off in any way) and he said he had no problem with it. In fact he offered this piece of advice.


"Hi Matt,
I don't have a problem with you guys cloning the board.
I'll even pitch in and help if you want. I probably still have some inductors and FETs.
The capacitors is what gets expensive. You'll need to get X5R rated 100uF (+-10%) caps in 1210 SMD package; three per board. Don't skimp on the caps, it's important they stay near 100uF over the full temperature range. Cheaper caps can drop 80% in value at elevated temps, so these are critical for good performance.
At 100pcs, they work out to around $3 each, so almost $10 in caps per board.
Let me know if you need help with the clone.
However, I won't be around much until January.
Good luck, and enjoy the Holidays!
Kevin"

As I said before it is sad for us that he is no longer in the businesses ( Hopefully he is happy out on the farm though), but it great to see his legacy will live on, and who knows perhaps he will come out of retirement and join us again :) Old habits you know.

That’s great! As I said earlier, I think that a direct clone of this circuit is much less useful today than in the day of the XM-L. It’s a voltage-controlled boost circuit targeted at 3.3v, too low for “big power” applications with modern high Vf LEDs such as XM-L2 (and especially the latest batches it seems).

That said, I think we can learn a lot from Lambda’s design and hopefully come up with something good.

A components list would be great. A datasheet for the 331C1 would be really good too.




Do you think that I could use 3 li-ions to power 3x nine volt MT-G2s in par using this method? If so what resistor would I need to drop off all that extra voltage to keep the current down to about 12-14 amps at the tail with this setup?

Eh, gotta rewind back to what we were talking about here…

  • First of all, I certainly can’t tell you how much current you’ll get out of a DD driver in this setup. I hope you understand that with this DD driver we cannot limit based on current, only change the duty cycle. The duty cycle will reduce the amount of time the emitter is getting power, giving us a lower average current. This will be based on the duty cycle, which is just a percentage of on-time to off time. Max current depends on your emitters, cells, and wiring - same as it always has. [This is exactly the same as the Varapower always was.]
  • Measure the resistance of the potentiometer (from one outside pin to the other outside pin), then we can start to work on what value resistor to use.

Turns out I didnt end up needing a resistor. I wired 3 9V MTG2 in par with some modest wire and the current total for the 3 of 2-lion was only 10 amp at the tail. I used 3 strands of 26G wire for the emitters and 18G below the driver. I am actually thinking about going to some 24G so I can get up around the 4 amp per die mark.

Oddly enough even when I went to 3 strands of 20G teflon wire to the emitters I got only 10.5A at the tail so I guess direct driving 9V emitters on 3 li-ions is ok after all.

  • You may be limited by the FET at this point. Is it getting hot?
  • Did you mention which cells you are using?