Sold/Expired Feeler - Programmable Hotwire Regulated Driver drop-in for D-M@g

LuxLuthor said:

Take your time, better to find bugs like this since you are so familiar with software.

Click to expand...

My plan is to ship it after it passes a full test with the WA1164. I should really drop it into a 2D M@g for the final check. I have a nice red, bored for 8xAA M@g that Will made to test in. This weekend I ran the regulator in a 3D M@g with 9xAA and the WA1164. There it ran in 4S voltage range. I should also test with J1 removed which will allow a 3S voltage range. I also have some D Li-Ion cells so I will test with 2 cells in the 2D. This is close enough to your configuration that it should be well tested.

The software is not quite in shape for distribution yet, after I ship the first article regulator I'll finish up the distribution software (and build myself a regulator for testing) and have the software ready for testing in a week or so. I have several weekend trips coming up so things may stretch out slightly.

So Lux can check out the bulbs that will take 7.3V first, then when he gets the new software he can try some other combinations. There will probably some surprises. Hopefully they will be software fixable.

Alan B said:

The software is not quite in shape for distribution yet, after I ship the first article regulator I'll finish up the distribution software (and build myself a regulator for testing) and have the software ready for testing in a week or so. I have several weekend trips coming up so things may stretch out slightly.

Click to expand...

I'd rather wait 3 months to have this right that buy a 'rush job' that has 'issues'. Take your time and make sure that you are 100% happy with it, then let us know and we will buy!

Any chance of testing it with 8S? Can anyone lend Alan an Elephant II body to test with? I would love to know how well this would work in an Elephant II with 8 x IMR 18650 cells using an Osram 64623 or 64458. I currently have 5 IMR 26500 cells running an Osram 64458 and would love to run the same bulb regulated to give me 6000+ Lumen until the auto-shutoff when the cells reach minimum voltage.


Also:
Will there be an option to run regulated until the voltage drops to a preset level, then just shut off? I don't need reducing output on my light, it is more a toy than a serious work light.

Thanks for the support, KiwiMark.

The way the low voltage battery protection works right now, when the voltage reaches the protection level it shuts the output off and drops the bulb voltage to half and then comes back on a little later, a distinct blink. If the minimum bulb voltage is set higher than that it won't come back on at all. So I believe you get what you want just by setting low and high bulb voltages such that low is more than half of high.

Further progress report - I installed the regulator in a 2D with WA1164 and a pair of Li-Ion D cells. All functions appear to work.

I'm tempted to ship this one to Lux and build up another to do the 3S testing. This one has been tested at 2S and 4S. It is programmed for 2S LiMN and 7.3 volts.

OK. S/N 001 is packed up and ready to go. I'll build up another one for testing 3S and so on. We're not done testing yet. So far all the errors seem to be software, the hardware has been good.

Lux should get this thursday or friday.

:thumbsup:

OK. SN 001 did get to the post office yesterday. So "it's in the mail".

Here's how that one is set up (roughly):

Osram Socket
40V MAX Parts

From Off:

Press to Low (Min Bulb Voltage)
Click to Memory
Double-Click to High (Max Bulb Voltage)
Triple-Click to Medium
Five-Click to Lockout

From On:

Press to Ramp (alternates Up/Down) (ramping sets memory value)
Click to Off

Max Bulb Voltage 7.3
Min Bulb Voltage 2.0

Maximum On Time 10 minutes
(ramps down and goes off after 10 minutes of no user input)

Overtemperature 60C

Battery Protection 6.0V
(Blinks and drops bulb voltage in half, goes off if this new voltage is lower than Min Voltage, does it again if battery stays below 6.0V). I plan to change this to drop to half power instead of half voltage in future versions.

Voltage Range 2S / 8.5V
Note that when overvoltage is detected the output is clamped below mid-range. So when I tested with 8xAA NiMH the output was limited to about 4.6 volts. If the supply voltage is raised higher the output will go up more since the micro cannot know as the ADC is saturated. I will probably adjust this algorithm to drop to half voltage instead of mid voltage in the future. In any case it protects the bulb for minor over-range conditions that might occur with overcharged batteries.

The package includes:

1) regulator
2) switch module
3) switch removal tool
4) programming adapter

The switch removal tool is a loop of wire. It is used to pull the switch out of the sled when changing to the programming adapter or removing from the M@g tube. The Sled is installed into the M@g without the switch. Slide it in from the battery end until it hits the locking ring (which should be left in place). Carefully center the sled in the M@g switch hole. A hex wrench is used through the M@g switch hole to tighten the sled in place. The switch is then plugged in through the M@g switch hole (against the stop ring). The rubber switch cover can then be re-installed.

I ordered more of the parts yesterday.

I was researching FETs this morning. I found some more good ones, and also some good looking High Voltage ones. Will have to put in another order soon for FETs.

Looks like the project is going smooth. :thumbsup:

I would be glad to fund some high voltage parts for experimentation. Let me know what I can do to help. :naughty:

I wonder if anyone would send me a 64623 bulb so I can do some high current testing here. I can use a couple of 12V 50AH AGM batteries to power the regulator. Then I could make some measurements of the losses, etc at 10 amps and see what heats up. Perhaps there is another bulb worth trying that operates in the under 20V region.

Gimme your snailmail addy, and I'll send you a 623.

sylathnie said:

I would be glad to fund some high voltage parts for experimentation. Let me know what I can do to help. :naughty:

Click to expand...

I'll probably include some HV FETs with the next order. Still will need HV capacitors and voltage regulator.

donn_ said:

Gimme your snailmail addy, and I'll send you a 623.

Click to expand...

Email sent. Thanks.

Got it. Lamp ships tomorrow.

donn_ said:

Got it. Lamp ships tomorrow.

Click to expand...

Perfect. Thanks.

Now, assuming I don't blow it up in testing, should I get a 6D M@g in bright red to fill with A123 M1's and I suppose it will require a Borofloat lens rather than UCL... Six A123 M1's fit okay into 6D??

You'll need an extension to fit 6x 26650 A123 cells in a 6D tube. Each M1 cell is ~5mm longer than a D cell.

Donn, thanks for sending him the 64623.

I got the sled from Alan today!!! :twothumbs

It is all extremely nice looking, and the sled is very high quality, and very easy drop in that tightens with same 5/64th" allen wrench that Mag switch uses. I talked to Alan, and am going to start a new thread in the Custom Homemade & Modified Lights section to report on it.

Briefly, it fits well in 8 different Mag D's, pushing up from bottom and against the retaining clip with perfect buttom position in standard mag hole, with rubber boot covering over button nicely.

It also fits just right in FM's old and new Elephants, Modamag's Collosus, and H22A's Mammoth. I took a bunch of pix comparing it to other switches we have all been using. My testing will be with all the bulbs that can take up to 7.3V, particularly the 5761. Using 2 x Emoli IMR-26700.

I only have partially charged them, but it is working absolutely perfectly on all initial tests of the clicks that Alan programmed. Damn....a regulated 7.3V 5761 !!!!!!! Just the thought of that knowing it is such a hard bulb to get the optimal performance without flashing it. Wait until Northern Exposure sees this. I think between him and me, we blew about 25 bulbs screwing around with NTC's and parallel resistors. LOL!

I forget how many I said to reserve, but minimum of 6-8. I'll post review link with tons of pictures tomorrow, as I have a lot of things to test. This is more than I hoped it would be. It is a Drop In Dream!

LuxLuthor's review is here.

Alan B said:

How do folks feel about SAE vs Metric on the bulb mount threads?

Click to expand...

If the KIU/AW stuff is metric, I vote for metric. I have a baggie full of various extension bits.

The KIU stuff is metric, I don't have AW but have heard it is metric also.

Through-hole vs Surface Mount:

I was updating post #1 of this thread and I mentioned there that the PC board was all through-hole. At this time I am planning to move to a hybrid design that will have mostly surface mount. This is helpful to improve clearances, reduce assembly time and effort and save a little cost on parts.

It makes it a little more difficult to service in the sense that more equipment is required, or at least different equipment. The most likely devices to be damaged are the fuse, regulator and FET and those are still through-hole. The microprocessor was in a socket, in the surface mount case it will be soldered in. The micro was the first part I changed to surface mount because I was running out of clearance there. The first article proves that my design with a socket works, but it is very close to the body of the m@g and the socket pins must be clipped below the board to keep it from shorting. That adds time and the socket adds cost. The price kept coming out over $100, this should be enough savings to keep it under, and it will have more clearance against shorts and be less time and effort to assemble.

So the question here is - does anyone mind this change? I was reluctant at first because it adds time to the project, but I did the design changes while waiting for other things. It also adds equipment cost to me as I'll have to set up to handle SMT. But buying tools is OK with me as long as it fits the budget. I think I'll get a stereo microscope and a hot air SMT rework station to start with. Not first class models which would cost several thousand dollars, but something serviceable. I'm also lining up some production capability so I won't have to build all of them myself. All in all it won't add much time to the project in the short term and in the longer term it should help us get to #99 a bit quicker.