MG X-Thrower driver...

My X-Thrower drops out of high after about 5 minutes. It has the stock driver and Vinh swapped out the SST-50 for an XML. I was going to try and swap the driver out. Would another 2.8 driver from Shiningbeam be the best swap? How do I remove the old one? I have the "pill" removed but not sure how to remove the driver from the bottom.

Sorry for the newbie questions but it will be my first "mod" and need guidance,

Thanks,
Barry

Anybody?

If the driver can't be pried out from the edge, you can try soldering a piece of thick copper wire to the ground ring of the existing board. When it has cooled enough to touch, grip the pill and pull the board out by the wire. This method has worked for me countless times.

Is it possible that your existing driver is stepping down due to thermal protection kicking in? The 2.8A board from Shining Beam is the same as in the MG X-thrower. It also has thermal protection that will step down, or shut off when overheated. The AMC7135 regulator ICs need to have a direct path to the heatsink for best performance.

Thanks DatiLED!!!

As datiLED states, soldering a wire to help pull out the driver is a good method. I've used it as well, successfully. Drivers are typically press-fit into the pill. The fit can be tight enough to make prying the driver's edge a hard task. Instead of soldering to the ground ring, however, I've made a wire loop by soldering to the center Batt+ pad. But it's all the same concept.

However, it's not clear to me how much help a driver swap will be if you simply go for a different 7135 linear regulator driver. The problem is that your battery source (I assume 1x18650 or similar) just can't deliver enough voltage to stay in full regulation when running in High. At 2.8A forward current, an SST-50 has a forward voltage of about 3.5V. IMO, that's way too high and it's no wonder your light fell out of High mode very quickly. At a minimum, your 1x18650 has to provide about 3.62V to reach and stay in full regulation. But that ignores additional voltage needed to overcome parasitic resistance from spring and battery contacts, switch contacts, etc. Even if the total parasitic resistance is say 0.1 ohms, that means another 0.28V, for a total of about 3.9V.

If you search CPF for HKJ's 18650 discharge graphs, you'll see that 1x18650 at 2A draw (not even 2.8A!) falls below 3.9V basically at the start, which is consistent with your observation that your light falls out of High in ~5 min.

Of course, the 3.9V estimate is just a guess, but the point is that it shows you what the challenge is -- a high voltage to stay in regulation due to a high LED Vf, an additional 0.12V overhead from the 7135, and parasitic resistance from the various contacts.

Switching to an XM-L will definitely help. An XM-L has a forward voltage of about 3.2V according to the Cree datasheet, so you save about 0.3V, which is a lot. Now maybe you need about 3.6V to reach and stay in regulation. I would guess that you might get 20 min or so of full power run time.

However, as I stated above, switching to another 7135 driver isn't going to help, since you haven't changed anything relative to the required input voltage (unless the 7135 driver is a lower-powered driver, such as one with 2.1A or even less of drive current for High mode).

if you swaping leds, might as well go for xml2 they have even lower vf

Barry what cells are you using? For example some cells such as Panasonics NCR18650A/B series do have high capacity, but sag more under load compared to a Sanyo UR18650FM. For single cell 18650 setups with AMC7135 linear regulators a higher voltage under load is more desireable than greater capacity.

Checkout HKJ's Battery comparator

Vinh swapped out the SST-50 for an XML a while back but I wanted not to have to worry about any heat issues so he put the max at about 500 lumens. Heat shouldn't be an issue with an under driven XML shoud it?

I use either AW 2900ma or I have some Ultrafire 300ma that I try to rotate between some of my other lights. I have a Pocket Rocket with an XML and have been able to run on high till I couldn't hold the light anymore with both brands of batteries. Also use both types in an EDC-MCE.

Barry

alpg88 said:

if you swaping leds, might as well go for xml2 they have even lower vf

Click to expand...

The datasheets don't appear to support that claim. XM-L at 25C is rated at 3.35V at 3A. XM-L2 at 85C is rated at 3.3V at 3A. Since LEDs are NTC devices, this slight Vf difference seems irrelevant. XM-L temp coefficient is -2.1 mv/C. So a temp delta of 85-25 = 60C gives a Vf drop of 126 mV, or 0.18V. So in fact, the older XM-L at 85C could have a Vf of 3.22V.

I have found that the xml2 has a higher vf than the XML. Only with IMR cells I can push them to 4.5A with 7135drivers, and only for a short time.

Justin Case said:

The datasheets don't appear to support that claim. .

Click to expand...

they sure do.

http://www.cree.com/~/media/Files/Cree/LED%20Components%20and%20Modules/XLamp/Data%20and%20Binning/XLampXML2.pdf
http://www.cree.com/~/media/Files/C...d Modules/XLamp/Data and Binning/XLampXML.pdf

alpg88 said:

they sure do.

http://www.cree.com/~/media/Files/C... Modules/XLamp/Data and Binning/XLampXML2.pdf
http://www.cree.com/~/media/Files/C...d Modules/XLamp/Data and Binning/XLampXML.pdf

Click to expand...

Please point out where in the linked datasheets that the XM-L2 Vf is lower than that for the XM-L. As I cited above, at 85C, the XM-L2 Vf at 3A (which is close enough to the 2.8A application here) is 3.3V. That is straight from the XM-L2 datasheet, p. 3. For the XM-L, Vf is 3.35V at 25C. This is also right out of the XM-L datasheet, pp. 3-4. Since the XM-L temp coefficient is -2.1 mV/C (another datasheet value, p. 3), the equivalent Vf at 85C for the XM-L is thus 3.35V - 2.1*(85-25) = 3.22V. Unquestionably, the datasheets show that the XM-L Vf is LOWER than that for the XM-L2.

You can also work it in reverse. The XM-L2 datasheet, p, 3, shows a temp coefficient of -1.6 mV/C. Thus, at 25C, the XM-L2 Vf calculates to 3.3V - 1.6(-60) = 3.40V. Again, HIGHER than that for the XM-L.

Epsilon said:

I have found that the xml2 has a higher vf than the XML. Only with IMR cells I can push them to 4.5A with 7135drivers, and only for a short time.

Click to expand...

than you got resistance issue somwhere in your xml2 set up, or some other problem, cuz cree says xml2 vf is lower,
http://www.cree.com/led-components-and-modules/products/xlamp/discrete-directional/xlamp-xml
http://www.cree.com/led-components-and-modules/products/xlamp/discrete-directional/xlamp-xml2

Justin Case said:

Please point out where in the linked datasheets that the XM-L2 Vf is lower than that for the XM-L. .

Click to expand...

xml2

Forward voltage (@ 700 mA, 85 °C) V 2.85
Forward voltage (@ 1500 mA, 85 °C) V 3.05​

Forward voltage (@ 3000 mA, 85 °C) V 3.3

Typical forward voltage (V)

2.85

xml

Forward voltage (@ 700 mA) V 2.9
Forward voltage (@ 1500 mA) V 3.1​

Forward voltage (@ 3000 mA) V 3.35

Typical forward voltage (V)

3.1

alpg88 said:

than you got resistance issue somwhere in your xml2 set up, or some other problem, cuz cree says xml2 vf is lower,
http://www.cree.com/led-components-and-modules/products/xlamp/discrete-directional/xlamp-xml
http://www.cree.com/led-components-and-modules/products/xlamp/discrete-directional/xlamp-xml2

Click to expand...

Cree can say a lot, the specimens that I have are clearly not. The setup is the same and resistance neglictable in the wires I use (0.25mm2 for a grand total of 5cm long, please calculate the resistance for me).

You post the Vf @ 25c vs 85C, which are different as Justin points out. The 60deg difference is good for 0.18v, which can be enough.

edit: Nice collection you have there by the way

While the Vf of the LED will certainly dictate how long the light will run in regulation with a AMC7135 linear regulator, I feel that the cells Barry is using are not ideal for the application. The AW2900 (NCR18650) he is using according to HKJ's discharge graph doesn't hold its voltage well over time. I would forget about using the Ultrafire crap brands. Using a cell such a Sanyo UR18650FM would be more ideal for his setup.

Try two CR123 primaries and see if that helps. The X-Thrower is supposed to handle those on high for 15 minutes before heat becomes an issue (and with your under-driven XM-L swap it may not get hot as quickly). If it does boost the runtime before dropdown, I'd say you want some different 18650s. I've become partial to IMRs, personally (runtime not a big issue for me).

Epsilon said:

Cree can say a lot, the specimens that I have are clearly not. The setup is the same and resistance neglictable in the wires I use (0.25mm2 for a grand total of 5cm long, please calculate the resistance for me).

Click to expand...

sure, lol, but i'll take cree word over anyone elses when it comes to cree leds.

i don't calculate resistance in such cases, i mesure it, on the set up at hand, if i come across issues. there could be more factors at play, besides wire guage.

alpg88 said:

xml2

Forward voltage (@ 700 mA, 85 °C) V 2.85
Forward voltage (@ 1500 mA, 85 °C) V 3.05​

Forward voltage (@ 3000 mA, 85 °C) V 3.3

Typical forward voltage (V)

2.85

xml

Forward voltage (@ 700 mA) V 2.9
Forward voltage (@ 1500 mA) V 3.1​

Forward voltage (@ 3000 mA) V 3.35

Typical forward voltage (V)

3.1

Click to expand...

I don't know where the spec "Typical forward voltage" comes from in the datasheets (when I do a "find", that text string does not appear anywhere in the docs), but all you've reproduced are the Vf numbers from P. 3 for each datasheet. Also, your XM-L2 typical Vf is at 700mA forward current, but the XM-L typical Vf is for 1500mA. Apples-oranges. The comparison for 700mA forward current for both LEDs would give 2.85V vs 2.9V.

BUT...

The XM-L2 Vfs are given at 85C, while XM-L Vfs are at 25C. If you factor in temperature coefficient of voltage, XM-L is LOWER at every forward current value, whether for 85C or 25C.

At 85C:
XM-L Vf
@700mA = 2.9V - 0.0021(60) = 2.9V - 0.126V = 2.77V -> lower than the XM-L2's value of 2.85V
@1500mA = 3.1V - 0.126V = 2.97V -> lower than XM-L2 3.05V
@3000mA = 3.35V - 0.126V = 3.22V-> lower than XM-L2 3.3V

At 25C:
XM-L2 Vf
@700mA = 2.85V - 0.0016(-60) = 2.85V + 0.096V = 2.95V -> higher than XM-L 2.9V
@1500mA = 3.05V + 0.096V = 3.15V -> higher than XM-L 3.1V
@3000mA = 3.3V + 0.096V = 3.40V -> higher than XM-L 3.35V

LilKevin715 said:

While the Vf of the LED will certainly dictate how long the light will run in regulation with a AMC7135 linear regulator, I feel that the cells Barry is using are not ideal for the application. The AW2900 (NCR18650) he is using according to HKJ's discharge graph doesn't hold its voltage well over time. I would forget about using the Ultrafire crap brands. Using a cell such a Sanyo UR18650FM would be more ideal for his setup.

Click to expand...

With the 7135, you need to hold a high enough voltage to reach and stay in regulation while under load, but not be so high as to have to burn off a lot of excess voltage as heat.

The approach would be to use HKJ's battery comparator at the current draw of interest to see which cells hold their voltages under load above Vreg for the longest time. The hard part is knowing what Vreg is. Certainly, at a minimum it is the LED Vf + 0.12V (from the 7135 overhead). But there is also voltage due to parasitic resistance, as I mentioned earlier, due to things like spring contact resistance, battery terminal contact resistance, and switch contact resistance.

When you start dealing with relatively high current draws, parasitic resistance becomes more and more important. If your light draws only 1/2 amp, then even if you had 100milliohms parasitic resistance in your light, you are looking at V = IR = 0.5A * 0.1 ohms = 0.05V drop. That gives you a total Vreg = Vf + 0.12 + 0.05. If your Vf is say 3.3V, then Vreg = 3.47V. But at 3A draw, you are looking at 6X the voltage drop, or 0.3V, giving Vreg = 3.72V (let's call it 3.7V).

The above numbers are all hypotheticals, but they are probably reasonably close to the actual case at hand. Referring back to HKJ's comparator for the AW2900 cells, they fall below 3.7V at about 0.5Ah. At 3A draw, that gives a calculated run time in regulation of about 0.5Ah/3A * 60min = 10 min. I think the OP cited a 5 min run time in regulation in High mode. HKJ's comparator suggests that Vreg could be closer to 3.75V.

If you go with a lower Vf LED and Vreg is closer to 3.5V, then run time could be 1.1Ah/3A * 60min = 22 min.