Contest (feeler): 2-months, to design a NEW community boost driver!

[rhd]

Careful. Some people don't take well to truth telling.....especially around here...and especially as relates to the actual cost of producing flexdrives.

 

[LSRFAQ]

As some one who regularly advises on producing small lots of boards, and knows the hidden costs, you guys are dreaming. You want Chinese prices and US quality. What you do not want the producer to do is make a profit. If I'm only making 2$ a board, there really is NO incentive to produce boards, unless I'm in the 2000-3000 piece range every month. But wait, 2000 pieces times 2$ = 4000 * 12 = 48,000$ minus 1/3rd for taxes, = 36,000$
a year, Oh wait, overhead is more then 750$ a month, that comes out of the post tax environment so now I'm down to 27,000 a year, divide that by business days per year and I'm making ~ 17$ a hour. Or am I?

Sounds like good money, right? It is until you realize one thing..
Reality is there is not sales of 2000 pieces per month of hobby laser diode drivers. The LED drivers your buying for peanuts are piggybacks or overruns on commercial LED products sold in China for things like beer signs.

I mean come on, get a packaging catalog, I'm looking at 2.00$ a box + materials just to ship you a single driver with a decent chance of it surviving the trip. That is before postage.

If I have a big enough load, its 2-4$ a driver to make US made boards before I even get to stuffing the parts, if I want a quality board. I need 2000 boards just to get a SMD stuffing house to look at it, because of the small size of the board.

Set down and do the math.

We'll not even get into the thorny issues of having a Chinese company do it. You need to station a quality control guy of your own on site. Even if you contract that part out, its not cheap. You pay the initial costs of the machine setup, a production test board with pins might be several thousand dollars. (note, PTBs test all the drivers in a board lot at once) Don't get me started on the clone wars that might occur. Toss in the fake/ersatz/cloned parts issue and its a headache.

Plus eating the boards that customers kill, that are returned under warranty, just to keep your reputation up. That and tech support get expensive, fast.

Don't get me started on "Life time buy", end of life parts, component tolerances, constantly changing part numbers, etc.


Those who say this has to be done as a business are right. I'd love to see a factual engineering study of dead diodes vs diode lifetime vs driver costs.

You get what you pay for.


Steve

 

[jakeGT]

^^^ I do believe one person said that. I don't think everyone I'n this thread ia trying to make a million bucks pal. They are looking for an alternative I'n case the dr runs out of drivers again and their is a famine... That's just my take on it, but I think you should direct your comment directly at random person, the only person claiming we are paying too much for the drivers....

 

[rhd]

A while back I scoped out design services to have a constant current 1.2A schematic and PCB layout made custom. I had thought about investing in the services of someone who knew what they were doing with the design.

Anyway, long story short, the design they were willing to produce for several hundreds of dollars worth of design time, would have been commercially manufacturable, ie no soldering for me (I don't even think in China, though I forget the particulars) for $11 each at a QTY of 100 parts. In fact, as I got a little bit further down the line of discussion, this individual was willing to make his paycheck contingent upon me actually being able to produce his design at that particular price +/- 10%.

So for clarity here - I was talking with an electrical engineer who does contract design work. He wasn't the one who would have been creating the physical boards. Rather he offered to meet my specs for a board, in creating a design that would be producible at a QTY of 100, for $11 each. It would have been 2mm larger in each dimension than a flexdrive.

Anyway, long story short - I didn't go for it - didn't have the time or money to engage the design. But I absolutely don't doubt that it would have been doable for $11 per board all said and done. This guy knew the costs of each component, etc.

 

[steviedezie]

damn, thats pretty awesome... even at a modest profit of $5 a $16 1.2A board is great value. and with these things, you know if one guy can do it, theres gotta be few more able and willing to do it for the same price, you just gotta look hard enough.

 

[lasersbee]

A while back I scoped out design services to have a constant current 1.2A schematic and PCB layout made custom. I had thought about investing in the services of someone who knew what they were doing with the design.

Anyway, long story short, the design they were willing to produce for several hundreds of dollars worth of design time, would have been commercially manufacturable, ie no soldering for me (I don't even think in China, though I forget the particulars) for $11 each at a QTY of 100 parts. In fact, as I got a little bit further down the line of discussion, this individual was willing to make his paycheck contingent upon me actually being able to produce his design at that particular price +/- 10%.

So for clarity here - I was talking with an electrical engineer who does contract design work. He wasn't the one who would have been creating the physical boards. Rather he offered to meet my specs for a board, in creating a design that would be producible at a QTY of 100, for $11 each. It would have been 2mm larger in each dimension than a flexdrive.

Anyway, long story short - I didn't go for it - didn't have the time or money to engage the design. But I absolutely don't doubt that it would have been doable for $11 per board all said and done. This guy knew the costs of each component, etc.

First... several hundreds of dollars is cheap for a tested Boost Laser
Driver at the specs wanted...
That being said....
I would put more credence in an Electronics Engineer... and would only
put money on the table for a hands on tested actual circuit design...
I've known too many engineers in the past that for them all was possible
until you got down to the nity-grity and the design flopped for one
reason or the other...


Jerry

 

[rhd]

I would put more credence in an Electronics Engineer...

That's what he was.

 

[Wolfman29]

Hey, RHD, why not put that guy's name somewhere here? Or maybe even have him start up a business here and he could be making some profit Would benefit the community and him!

 

[rhd]

Hey, RHD, why not put that guy's name somewhere here? Or maybe even have him start up a business here and he could be making some profit Would benefit the community and him!

He didn't have any connection (or frankly interest from what I remember) in the laser community. I think I had sent him the datasheet for the Nicha 445 9mm from their site, and explained the project requirements and constraints.

This was some time ago, before the rarity of Flexdrives had spurred on a lot of the driver creativity we've seen since. At the time, my objective was to release the schematic and PCB files, and let anyone use them. I would have asked for a few bucks (honour system) if anyone sold them, at least until I recouped the design cost I would have paid him.

I didn't go for it, and I wouldn't now, largely because we've come a ways since then, and I'm not sure it's necessary at this stage.

We've got a number of driver alternatives now, both in the form of some pretty awesome homebrew linears, and some repurposed LED drivers. I think the debate on LED drivers has largely become a philosophical / religious one. There's chatter about whether LED drivers are safe long-term, but I think there is so much practical evidence that they work reliably, that the conversation about their viability as laser drivers has become largely hypothetical.

The reality is, LED drivers (if chosen correctly, tested per driver, and built with protective caps on the diode) just plain WORK. Whether they may be doing some small-scale damage to the laser diodes that will shorten their life, I don't know. They may, or they may not, but practically speaking, so many users are using LED drivers with so much success, that I'm totally comfortable considering them part of a "proper" build. We aren't seeing the diode deaths that many predicted would result from cheap Chinese LED drivers.

I have some 445s that I know will never leave my personal collection - and I have no concern about using LED boost drivers in them. If they die in 4 years instead of 5, I'll survive - and as far as I'm concerned, that's a big IF. When I look at this situation with pure logic, I can't see any compelling reason to go for a $25 driver over a $5 option, when the cost of a diode is $40.

So to be honest, I just don't feel boost drivers to be scarce at this point. There really aren't a lot of build scenarios where I can't think of a reliable DIY driver solution that costs around $5. We're covered on the boost front by good LED drivers. We're covered on the linear side by awesome drivers like what Jib has done, and the ambitious creations of relatively new members like Jufran. I just think that all-around, we're doing fairly well for drivers now - a year after the 445s arrived.

Ironically, the one build-scenario that I still feel we're lacking a decent DIY driver solution for is not a new or cutting edge diode - specifically, it's the single-cell RED build. Tough to do well with a linear, incompatible with an LED boost (generally). You CAN run a red with a single cell and an LDO linear, but as of yet, nobody has been able to make this hit the highest current capabilities of an 815 or 826.

That's my driver rant for the night.

 

[steviedezie]

yessss rhd, you are the MAN... you hit it right on the head...

there is no reliable steady 1 cell red driver available... I have had many ideas for a high powered red build using only 1 cell...but its hard. The only success ive had using a 1 cell red build is by using a boost driver, and connecting smd 1n4001 diodes in series with the output to make the diode Vf higher than the input so the driver can boost correctly...that is as far as I know the only way to run 1 cell red build. THe only other (but costly) option is to use a flexdrive, which will buck. but 23-26 dollars for a driver in a red build is very costly... usually, the diode and host/heatsink wont even hit that amount so having the driver at 50% of the lasers cost isnt very appealing.

sorry if i dont make sense right now, i just came of a wedding and am definitely under the influence...i will check back in the morning to make sure i didnt say stupid crap....thank you and goodnight yall!!

 

[Wolfman29]

You use a cap on the output, but in my experience, they are stable enough that I don't even need a cap on the output. I don't have one on either of my favorite builds (I use this, it's my favorite: 3.0~9.0V 9W 1-Mode Regulated LED Driver Circuit Board 20mm - DinoDirect.com) which are a 12x (508mA) and a 900mA 445nm, and both are running smoothly (maybe? my LPM tests may not be very reliable... but my DIY LPM seems to be indicating that my 405 is stronger than my 445?).

 

[rhd]

but 23-26 dollars for a driver in a red build is very costly... usually, the diode and host/heatsink wont even hit that amount so having the driver at 50% of the lasers cost isnt very appealing.

If you're talking about a LOC diode, then a $25 driver is like 300% of the diode's cost - even worse

 

[rhd]

You use a cap on the output, but in my experience, they are stable enough that I don't even need a cap on the output. I don't have one on either of my favorite builds (I use this, it's my favorite: 3.0~9.0V 9W 1-Mode Regulated LED Driver Circuit Board 20mm - DinoDirect.com) which are a 12x (508mA) and a 900mA 445nm, and both are running smoothly (maybe? my LPM tests may not be very reliable... but my DIY LPM seems to be indicating that my 405 is stronger than my 445?).

Interesting - I've somehow missed that driver. Going to order a handful tonight and give them a test.

 

[steviedezie]

so far my single cell lpc-826 build is using this driver here:

$2.19 - 1.5V~4.2V 3W Cree Circuit Board for Flashlights (16.7mm*5.6mm) - Flashlight Parts and Tools

Its a boost driver, so i soldered 2 1n4001 diodes in series with the output and it boosts up the voltage fine, im using a 14500 cell and the current to the diode is 515...which i think is perfect for long life with a lpc-826 diode.

 

[Wolfman29]

Yeah, rhd, it's a powerful little bugger, pretty efficient, VERY stable (don't have an O-scope, but detected no ripple on my DMM).

It's my favorite driver so far, and it's damn cheap (and hard to screw up).

@stevie: I used to have a couple of those drivers... but I never though to stick two 1N4001 diodes in series with the output to get it to power a red! Good thinking!

 

[steviedezie]

yea...the host i wanted to put the lpc-826 in is very small and can only take 1 cell...i love this host for red though...that driver combined with an 826 yeilds 310mW after a AR coated acrylic for red... my best red build

theres not much room in the host...but somehow i managed to solder in an 2 additional 1n4001 smd diodes... and you the crazy thing is....the heatsink i made for this thing was completely done on a drill press...no lathe at all... it was a PITA but worth it.