Funny you should ask.. it is absolutely automatic.. you can take the head off, change lamps and put the head back on.. turn on the light.. blamo.. done. It's one of the best parts of this design, and it's a really cool twist of the fate of physics and battery chemistry how it works.
Explanation: Since the batteries are notorious for dropping voltage as the current goes up.. if you want to get more POWER out of the batteries, it necessitates that the voltage will be lower and the current will be higher.
So.. the brilliance of the setup with the hotdriver is this:
It has two settings.. independent of each other.. CURRENT and VOLTAGE.. whichever hits first wins.
So.. the HDM6 will be set up for about 4.95A max and about 7.3V max..
When you put in an MN20 lamp.. the voltage will go right up to 7.3V.. (and slowly decline as batteries die.. but it will stay at about 7.2V most of the battery life.. you will not see much dimming for the first 60-70 minutes.. estimated runtime is as much as 86 minutes with LOLA..
When you put in an MN21 lamp.. the voltage will attempt to get to 7.3V.. but when it gets to about 6.7-6.8V.. it will find that it's at 4.95A and stay right there.
the startup is much faster with the LOLA but i'm still fine-tuning the startup solution with the HDM6.
I had to invent a different solution for shutdown and have the circuit board design done, just need to triple check all the traces and it's a done deal!
here are a few images from my last night allnighter fixing startup issues with hotdrivers:
Here is the LOLA startup.. vFET vs vLamp
Here is the LOLA startup.. vFET vs current..
Here is the HOLA startup curve.
It's a bit tricky to read the current on the LOLA.. current starts very low and goes up and ends up being the top trace...it might seem like that's an excessive spike, but it's about 50% lower than w/o soft start, and probably about 1/4 to 1/6th the POWER spike.. because power is current x voltage.. and where the current spike is high, the voltage is very low... it's nearly ideal to start this way because it gives you a very rapid start while still bulb-saving.
Since the driver didn't have enough voltage over head to run the HOLA properly (too many long wires on the bench.. i need to make some 14ga jumpers!).. I'm going to show you the osram 62138 startup curve:
You can see that the current limit is 'fuzzy'.. you can actually hear it.. there is an oscillation that happens at startup, probably having to do with the way that voltage coming in changes from the current demand difference.. you can hear a bit of a 'squeak' during that phase if you listen very close.
It's absolutely nothing to worry about.. every hotdriver ever made has done this.. there were a couple version (mark6 comes to mind) that had an instability issue like newbie mentioned... mark 7 introduced a capacitor on the gate of the transistor that eliminated the instability, and it's worked awesome ever since.
Even with the model that had the instability.. it could only be found one way.. by turning off the light, and then before it was done cooling off.. try to press the light back on.. about 1 out of 10 times you could get it to try to regulate at about 1/2 voltage.. of course.. with a mag85, it could probably do that for a minute or two before heat became an issue.. and it was a very unstable equilibrium and would last 5 seconds usually and turn on fully.
I tried to let the mag138 just work its way out of the stuck position and it never did.. the FET got so hot it desoldered itself! that was the motivating issue i needed to fix it and born was mark7.. and what became the vOne hotdriver.
I've never seen a LOLA running before so i can't really compare. My understanding is that you'll get nearly 7.4V with fresh batteries and LOLA, so the hotdriver might be slightly dimmer, but the light was as white and brilliant as i could ever have expected of any light when running on the bench. (so far the LOLA tests were done from a driver that is outside the body).. but the lamp was inside the head, so i could aim the beam around and it was gorgeous!
I hope i can sell a bunch of these, i think i just might buy an M6 now.. never in a million years would i have before a solution like Hotdriver M6. So much for profits on this project :-D
Oh.. i have a decision to make.. I can either keep the temp-sensor and use a big of a 'hack' to make low-voltage cutout work.. i still have to run down a set of batteries to see how it really behaves.. turning a dial on the power supply doesn't really behave how a battery does.. or i can kill the temp sensor and replace it with the really kickbutt low-volt sensor from the BAM!
I will be doing testing of both possibilities and i've already designed the vTwo hotdriver board to accommodate either possibility. The power drop on the FET is insignificant at HALF of that with the 1274 solutions, so i'm confident it wont be a problem, totally a non-issue... other than ONE possible example.. use in extreme heat, like desert conditions.. so.. I'm leaning heavily toward making them all with the precision low-voltage and no high temp cutout.. since they will all be used in the same device unlike the standard hotdrivers which have myriad possibilities including one with runtimes well over an hour.
If somebody needs high-temp because of climate, i can easily make a variation.. there might be some 'training' involving paying attention to the light running out.. with the hack.. the lamp will dim DRAMATICALLY but it will try to run at like 1/2 power.. putting a LOT of heat on the FET.. i still need to test the hack out in a real world test, so it might be a moot point.. the hack might work just about as good as the precision LV chip and i can keep the high-temp on all of them.
You won't have to decide when you order.. i will get a 'details' form after i make a batch of a few tester models that can be used for a good month while waiting for the machined parts and batteries and such to come in.. that way if there are any tweaks needed to be done, that can happen before the mass production, which will start about 3-4 weeks after the opening of the sale. (it should be about 5-6 weeks for the machined parts to get back).
Thanks to wquiles for machining the tailcup prototype.. and lending me his cherished M6 for fitting the HDM6. I think that machining job might have been the first one he did 'for real' almost certainly. it came out amazing.. i needed to tweak a bit of it.. was too tight.. but that was perfect.. i got to be able to update my model for the full run!
So.. 'game on'. .. get those paypal triggers ready.. i'm planning to get the pack back into the light by this weekend, and once the operation is confirmed, open up a sales thread on BST.
-awr