Some odd test results of HX1175B and MT-G2
I tested three different HX1175B with MT-G2. Here are my results.
With 2 x 18650
Driver A - 3.5A
Driver B - 4.2A
Driver C - 4.8A
With 3 x 18650
Driver A - 7A
Driver B - 13A
Driver C - 6.8A
These are my results. I can't explain them, especially the 13A. All the modes were functioning, and I could feel it heating up fast. It kind of scared me.
Please check the markings on the sense resistors of each driver (A/B/C) and post your results.
B and C have R068 resistors. A is already wired into my Boxer, but probably the same.
Thanks. If possible I’d like to know the voltage drop across the sense resistor bank of each one while using 3s batteries. You can simply put your meter in the 2V range and put a probe on either side of the sense resistors. I’ll admit that this may not tell us much. The behavior does seem odd and that is the first place I can think of to check. The idea is to help narrow down the problem. If the voltage drop is ~0.25v at 13A then we can assume that the sense resistors are low quality and that the values do not match the markings. If the voltage is significantly higher than 0.25v then we will say that something else (the buck controller) is acting up.
I can’t do it right now. Everything has been desoldered and put away. All my work is done on my kitchen table…and I have to eat. If I have time, maybe tomorrow.
Do I test each resistor separately, or bridge both at the same time with my probes.
You’re luring me into realms that are beyond my understanding.
Edit: I assume this should be done while the emitter is blasting away? The driver might not last long enough to do the test going at 13A. It got real hot almost instantly.
The resistors are all in parallel so any points on either side should be equivalent. You would have to remove them to get individual values. I think what you’re looking at is how each set compares with each other set to see if that’s where the difference in output lies.
I’m definitely looking forward to your results if you decide to take the time to do the test.
EDIT: RBD is correct - I just want to know how the sets compare to each other as a whole right now, no need to remove anything.
I’m also worrying about frying my MTG at 13A
Edit: And if it’s a 7A buck driver, why not 7A with two cells? LCK 5A buck gives 5A with two cells.
If it starts on a room temperature heatsink of decent size it should survive a few seconds to get the voltage drop for sure. You’ve got to be strict with yourself, if you fumble the probes - stop and turn the thing off. I wouldn’t waste time fiddling with the thing while it’s on, despite djozz showing that his emitter survived 16A. Do not do any testing without a heatsink. You should get a stable enough figure immediately, so it’s only a matter of the time it takes you to look at the DMM and note the reading.
I have it clamped to a ten pound slab of copper.
In that case I think you’re golden :-).
Sounds like it’s dropping out of regulation. That’s a huge dropout voltage though (>1v?). Above a certain point I assume it regulates pretty cleanly. Maybe this is related to the Zener used to power the QX9920, note that SF uses the QX9920 without an early dropout (XM-L2 driven using 1s) but they do not use the Zener to power teh QX9920, they do it directly off of battery voltage (not feasible here).
Then again, the LCK driver will run 3 x XML in series. I guess they’re different animals.
Note: Power Failure…there goes supper.
I don’t understand what you are getting at there. Sorry about dinner :-/
I’m sure I don’t understand either.
What I mean is, the HX1175B isn’t made to drive emitters in series (at least I don’t think so), whereas the LCK will drive 3@5A per.
I was thinking that they don’t function in the same general manner.
EDIT 7/2/15: I wrote this before I understood a lot of the math and operation of these converters. I would now recommend a different setup than this---especially the inductor. Even though this setup works, it is very inefficient.
I have a few updates that I will post here soon, I may need to revise some of my prior thoughts. The biggest help in driving dedomed XM-L2s has been to change the inductor and to add some low ESR output capacitance. You really need a good 47uH inductor. I went with a 10.5A peak shielded inductor (the stock inductor is heavily influenced by its position relative to the output wires.)
The quality of the capacitor matters a lot in this type of application. You need low ESR. If you don't have good quality caps, then multiple low farad caps in parallel will be more effective than one large capacitor. Ceramic capacitors have the lowest ESR, but the capacity is low and if used by themselves can cause some issues with ringing (especially on the input side.)
This is from a driver with the 47uH shielded inductor, stock sense resistors, 70N02 FET, 1x150uF and 1x470uF low ESR caps, about 7A output. This is MUCH less ripple than with the stock set up. This will run on a dedomed XM-L2 without any issues--try it with the stock setup and you'll get a very brief flash.
Good update RMM. Thanks for keeping us posted. Looks like I need to eat my words a little about the output capacitance not helping ;). Low ESR seems to be the key to a lot of things. I assume that a shielded, 10.5A, 47uH inductor must be very large physically? Between the two large caps and the giant inductor, you seem to have assembled a real brick :p.
Can you try increasing the sample rate the next time you capture a trace? The outlier samples seem to be a long way out.
I see what you were thinking now. The HX-1175b is constructed using a chip meant in part for controlling high voltage LED strings. The HX-1175b itself is marketed as a driver for SST-90 of course, but I’ve seen little to indicate that it shouldn’t work for higher voltage loads. Thinking more about what I already said, I can’t really see how the PSU section (zener and resistor) for the buck controller would be the problem. I’ll probably pick one up sooner or later to try to learn a little more about it.
Took measurements with 2S and 3S driving an MT-G2. Cells were King Kong ICR's charged to 4.2ish volts. The cells are older and have internal resistance of about 40. Current to the MT-G2 with no driver was 6.5 amps.
| 2S | 3S | ||||
| High | Low | High | Low | ||
| Input Current (Amps) | 5 | 2.5 | 5 | 2.5 | |
| Emitter Current (Amps) | 5.1 | 2.5 | 6.7 |
3.4 | |
| Sensor Bank (mV) | 162 | 77 | 166 | 82 |
The driver used is completely stock. The HX-1175b1 only has High, Low, and Strobe modes.
EDIT: Fixed typos in table.
Ouchy,
If you get a chance, please post a pic of the 13amp driver. Also, it would be helpful to know what current 2S of your 18500's deliver when connected directly to an MT-G2.
How do I increase the sample rate on the DDS? I haven't seen any place to do that.
The inductor is actually the same size or smaller than the stock one, although I think it has higher DC resistance (more heat losses.) so it may not be the most efficient but it works well, is small enough to fit in most lights, and runs smooth. The saturation point is what I was really after since I think what happens with the stock inductor that wreaks havoc. With the stock inductor you can run pretty smooth up to a certain point then it just spikes like crazy! Sometimes you can have an LED working fine then you move the LED wires by the inductor and then it blows.
The driver fits well in the TK61 and my little mini intimidator. If you are happy with 6.5A-7A you can get away without the output capacitors but I've added them for extra insurance.
I just corrected a couple errors in the table in Post 155.
EDIT: Fixed again. The table is so small when I edit it that I can't really see it until it is posted. Numbers are all good now.