shining beam 2500mA driver wiring?

[TorchBoy]

The total current from an array of AMC7135s is much less of an issue than how many volts each AMC7135 is dropping and how they can get rid of their heat.

 

[ti-force]

I have run tons and tons of these with no "cooling" at 3.0 amps and they never have problems. I have done 1 sandwich that had a problem and I was never able to see that one again after I sent it off.

you would also think that with 3 SST-50s they would hit a thermal cut off before this light would....especially at over 4 amps.

What was your voltage in? When v in exceeds vf, the excess voltage is burned off when using these drivers. This creates heat which reduces efficiency. If his cells have a nominal voltage of 1.2v each cell, and he's using 4 cells, the total voltage is 4.8v. If vf is 3.35v, and v in is 4.8v, the driver has to burn off 1.45v. This is a rough estimate to get waste heat: 1.45v x 2.8A = 4.06 watts of waste heat.

*EDIT* Sorry TorchBoy, I guess you and I were both typing at the same time.

 

[TorchBoy]

Your example is much better than my explanation! If the cooling the board is getting is reasonable, perhaps it doesn't need a complete extra (4th) cell to get the volts for full output. That's a bit awkward. Perhaps four cells could be used, but with some extra resistance introduced somewhere. :shrug:

 

[Mettee]

I understand what you guys are saying.

I just dont think that is enough to make it do what it is doing...he has a problem with 3 cells. charged they should be over 4v and should easily supply the voltage to hit is mark of 2.8a. Something else is happening here that is being missed. I used 3 AA eneloops with the same board and a DxxxI P7 and it was over 2.xx amps for the whole test. With 4 it was still 2.xx amps the entire test. I can not remember what voltage it was for both. But I have also used 2 lion on these boards just to see what it would do and it never acted like described by the OP(admittedly not for long periods). And I would figure that is an over voltage above the VF if there was ever any.

 

[ti-force]

If the cooling the board is getting is reasonable, perhaps it doesn't need a complete extra (4th) cell to get the volts for full output. That's a bit awkward. Perhaps four cells could be used, but with some extra resistance introduced somewhere. :shrug:

If it were me, I'd switch to 1x Batteryspace 26650, but that's the setup I have, which is why I like it, plus I have a 2D Mag, which is shorter. Also, I understand that Li-xxxx's aren't for everyone.

 

[Justin Case]

I dare say that'll only get it to run in regulation for a short time when the 3 cells are freshly charged. Because four flat NiMHs will give fewer volts than 3 fully charged NiMHs, 4 cells will likely drop out of regulation at some point too. How much current does it provide when the 3 or 4 batteries are half or almost completely flat? Is either OK?

Justin is probably right about the thermal limiting, but I didn't get that in my headlamp. Can you cool them easily?

Silverfox's discharge curves for Accupower D NiMH show that those cells can hold ~1.25V for ~3A load for probably about 2 hours. So I would say that if you have a 3D Mag with low resistance and a suitable Vf P7, you should be able to run in full regulation for plenty of time.

The OP's Mag does not appear to be low resistance, however. He can get only 1.7A at the tail when using the SB AMC driver and 3xD NiMH. Or his P7 does not meet the I Vf bin spec. Or both. Whatever the reason, it seems clear that 3xD NiMH can't provide enough Vbatt in his 3D Mag, when they should be able to do so easily.

How is the SB driver heat sinked? To the LED heat sink? To the Mag tube? Hanging free?

 

[TorchBoy]

That graph doesn't look too bad. Did we have a figure for what the 3 cell battery voltage under load is in this situation?

 

[vestureofblood]

Hi jez,

I realize some of this may have already been covered, but here is my 2 cents.

Let me start by saying that it is possible you are getting a bit more current than you think when the light is assembled. Measuring current through the tiny pinpoints of your DMM leads, combined with the internal resistance of said device is going to cause a drop in current.

When you add a 4th cell what you are doing is pushing past all of this extra resistance with a higher voltage.

A quick example.

I hooked up an SST-50 DD on 1x imr 18650. Through the leads of my DMM I could measure about 3.6 amp after the emitter warmed up. I knew I was getting more current through the light with the tail cap on for reasons I wont take the time to explain here. So using some very thick copper wire ( like 14 G I think) I made a set of leads and just plugged the thick wire into the DMM and touched the other ends to the light and cell. Instantly the current shot up to over 6 amp where I had to disconnect.


You may be having more than just measurement problems though.

If you realy want to get this light to run at the full 2.8 amp on 3 cells it can be done.

I recommend using Accupower precharged cells because it seems to me like in the case of eneloops and Duraloops, that the precharged cells drive emitters harder and hold up under a load better (just MHO though).

I dont know anything about the brand of D cell you are using. If they are a lower quality off brand ( I am not saying they are) they may have very high internal resistance.

For now lets assume your cells are good, and take a look at some other problem areas.

From what I gather you are sure that your emitter is a CxxxI bin, so the VF of the led is likely not your problem.

The first thing I would do is get some thick copper wire. Go bigger than you think you need, like 20G or so.

I like to use deoxit gold solution on all metal to metal contacts where current flows as well.

If you haven't done it already modding the tailcap for lower resistance with some copper wire or braid will help too.

And last but not least. Even running at only 3.6 volts it wouldn't hurt to heat sink the driver (on both sides if using the SB driver) You could use thermal epoxy to stick it to the base of the led heat sink, and then put an aluminum disc on the other side. This will not effect the current at start up so much, but after the light has run for a few minutes it will help stabilize the current.


I hope this helps. Sorry if I am being long winded, but I am confident this problem can be overcome.

 

[jez]

Wow lots of replies to go through and great info in them.

I've not ruled out the cells being the problem. They are a German manufacturer and seem to have a good reputation.
http://www.ansmann.de/cms/businessdivision/consumroot/batteries-battery-packs/rechargeable-batteries-nimh/mono-d/10000-mah.html

The driver is just hanging free at the moment with no additional heatsinking.

It's frustrating for me having to ask all this as it seemed a fairly common and straight forward mod so problems should have been minimal.

For the driver heatsinking, do you glue the aluminium disc to the top of the drivers? No need to touch the actual circuit board? Now thats a newb question

I've got hold of a second DMM if that helps for measurements.

I think I can change the battery wires for 20 gauge but I don't think these will go through the heatsink holes. Would just the bare conductor be ok for testing knowing the heatsink is HAIII anodised?

Thanks

 

[Justin Case]

I have run tons and tons of these with no "cooling" at 3.0 amps and they never have problems. I have done 1 sandwich that had a problem and I was never able to see that one again after I sent it off.

you would also think that with 3 SST-50s they would hit a thermal cut off before this light would....especially at over 4 amps.

Thermal management is a tricky thing. If the setups are apples-to-apples, yes you would think that the OP's setup should work fine since higher-powered AMC setups have shown no problems.

Based on the scant evidence, this is probably not an apples-to-apples situation. Hence the OP's problems.

Possibly the AMC chips' mounting tabs are poorly sinked.

I'd run the light with its guts out in the open. Periodically feel the temp of the 8 AMC chips. If the light essentially cuts out within 15 sec, the test should be quick to perform.

With all of the various changes, re-measurements, and so forth, it's hard for me to keep track of the current status. Jez, am I correct that you are now getting 2.1A at the tail with 3xD NiMH? And you believe that you have a Vf of 3.35V at 2.8A, with the Vf and tail current measurements done separately? If that's correct, then you have a relatively low Vf P7.

 

[Justin Case]

I understand what you guys are saying.

I just dont think that is enough to make it do what it is doing...he has a problem with 3 cells. charged they should be over 4v and should easily supply the voltage to hit is mark of 2.8a. Something else is happening here that is being missed. I used 3 AA eneloops with the same board and a DxxxI P7 and it was over 2.xx amps for the whole test. With 4 it was still 2.xx amps the entire test. I can not remember what voltage it was for both. But I have also used 2 lion on these boards just to see what it would do and it never acted like described by the OP(admittedly not for long periods). And I would figure that is an over voltage above the VF if there was ever any.

If the SB driver has to dissipate 4W of waste heat, that's a lot and the board should be heat sinked.

There appears to be at least two different problems, which manifest themselves when the OP runs with either 3xD or 4xD.

The first problem seems to be that Vbatt isn't enough to reach full regulation. That problem shows up when running with 3xD, since Vbatt has less margin. However, efficiency is basically 100%, so the driver has to dissipate very little waste heat. If the P7's Vf at 2.8A really is 3.35V, then the source of the problem has to be the Mag host. The OP has done a tail spring low resistance mod. I suppose the switch is next. Another easy thing to check are all of the solder connections. I doubt it is the 24 ga hookup wiring. Resistance per 1000 ft for 24 ga is 25.67 ohms. The OP previously estimated a total length of about 3.5" (0.29 ft) of wire. Thus, total resistance is about 0.0075 ohms. If the current is 2.8A, then the voltage drop is 0.02V, which is negligible.

The second problem is the light essentially cuts out (tail current drops from 2.8A to 0.7A) after 15 sec when using 4xD. Lacking any additional information, the tail current numbers suggest that 6 of the 8 AMC chips have shut down, perhaps due to thermal considerations. ti-force's calculation explains the issue -- a potential of 4W of waste heat that needs to be dissipated. Fifteen sec seems pretty quick to me to shut down, which possibly hints at a defective board. That's why I suggest checking the sinking of the mounting tab.

If you are going to get into mod'ing, I would recommend getting a bench power supply. You can find versions on the web for about $100 that can deliver up to 30V/5A. It makes it a lot easier to test your setups with known voltage/current inputs before you complete the assembly.

 

[Justin Case]

For the driver heatsinking, do you glue the aluminium disc to the top of the drivers? No need to touch the actual circuit board? Now thats a newb question

I would heat sink to the wide tab of the 7135 case. Sinking to the plastic case itself isn't going to be very effective since the case is a poor thermal conductor.

I think I can change the battery wires for 20 gauge but I don't think these will go through the heatsink holes. Would just the bare conductor be ok for testing knowing the heatsink is HAIII anodised?

I doubt that switching to 20 ga wire is going to matter, but it can't hurt.

I personally would not take the chance with using a bare conductor unless I knew for a fact that the holes were also HA.

If you are just testing with 20 ga and the wire is too thick to fit through the heat sink feed holes anyway, then why not just go around the holes. Run the wire from the driver directly to the LED. If you plan on a permanent installation with the 20 ga, then I'd probably drill out the feed holes.

 

[vestureofblood]

http://www.ansmann.de/cms/businessd...argeable-batteries-nimh/mono-d/10000-mah.html

The driver is just hanging free at the moment with no additional heatsinking.

It's frustrating for me having to ask all this as it seemed a fairly common and straight forward mod so problems should have been minimal.

Trust me we have all been there. While some mods go just as planned, other that seem simple can often turn up unexpected issues. But that is how we lern. Hmmm what is the saying.. Great salors are not made on a clam sea?

For the driver heatsinking, do you glue the aluminium disc to the top of the drivers? No need to touch the actual circuit board? Now thats a newb question

Correct just use a thin layer of arctic alumina or similar to glue it to the top of the chips, the heat sink does not have to contact the PCB.

For the amount of heat sinking you will need you could probly even just glue a couple of thick chunks of aluminum to the driver and not worry about attaching it to the body of the light.



This is a pair of the SB drivers wired in parallel for an SST-50. Just attach the wires so they lay flat and sticks out the sides.

You will have to remove the spring from the bottom of the driver, or drill a hole in the center of your heat sink.

One other thing I do to aid in better current flow, and to drastically reduce the risk of tearing off the electrical pads of these drivers is to attach the led pos wire to the end of the diode as well as to the + pad.

Here is a pic, the + end of the diode is right next to the + led pad.

I've got hold of a second DMM if that helps for measurements.

You could possible use both DMM in parallel to get a better idea of the current as mentioned by another member in this thread.

Here is vid that may help you understand what I mean about DMM resistance.

I think I can change the battery wires for 20 gauge but I don't think these will go through the heatsink holes. Would just the bare conductor be ok for testing knowing the heatsink is HAIII anodised?

I agree with Justin on this, bare wire is not the best idea. If 22g will fit you could probly get by with that.

What gauge wire are you using now?

 

[jez]

I've just tried 4 c's with it all hanging free and as the current started to drop I touched the regs on both sides of the driver and almost burnt my finger so they are getting damn hot. Guess that's that for 4 cells. Can't imagine any heatsinking stopping them cooking.

2.35A (tailcap) and 3.3 Vf (led) with 3 d's measured one after the other. Seems my d's are either getting better or something else is settling in?
regs are fine at this current to touch.

 

[vestureofblood]

I've just tried 4 c's with it all hanging free and as the current started to drop I touched the regs on both sides of the driver and almost burnt my finger so they are getting damn hot. Guess that's that for 4 cells. Can't imagine any heatsinking stopping them cooking.

2.35A (tailcap) and 3.3 Vf (led) with 3 d's measured one after the other. Seems my d's are either getting better or something else is settling in?
regs are fine at this current to touch.

Actualy you would be surprised how far an 1/8 of an inch if aluminum will go as far as heat sinking.

I use 3 of these driver in parallel sinked to good size aluminum disks, and it pegs out with a stable current of 8.4 amp on high with the use of 4 C cells.

 

[jez]

I've just mangled together 2 aluminum heatsink discs a little bigger than the diameter of a c cell and taped them tightly either side of the driver.

With 4 c cells the current held for a little longer then dropped again but slower.
The sinks got very hot too.

Vestureofblood, how you got two of these with a couple of sinks running amazes me.

Also I checked the soldering of the mounting tags.

 

[Mettee]

So you got 2.35amps so far with 3 cells. If you were measuring that with a DMM with small leads that might actually be 2.8amps but it is unable to read that due to the smaller leads on the DMM.

So it seems everything could be fine, just a misreading throwing you off. What is the brightness like. Seems no one has asked that.

And yes you might get "settling in" as you burn in the led the VF lowers slightly.

 

[ti-force]

According to the specs at Flukes website, your 179 meter should be pretty accurate when measuring DC current. Has anyone had any negative experience with a Fluke or other nice meter that would suggest otherwise? Just curious. The specs for DC current accuracy with the 179 meter are: +/- 1.0%+3, so just say you're getting 2800ma. Using the accuracy specs, wouldn't 1.0%+3 = 2800ma x .01 = 28ma, 28ma + 3 = 31ma, 2800ma - 31ma = 2769ma. If I'm figuring this correctly, that's not much drop from this meter.

You stated earlier in this thread that you're using alligator clips during current measurements right? Are they alligator clip leads, or alligator clips that push on to the test leads? If they're the push on type, remove them and take another reading to see if the reading goes up any with 3xNiMH's. I don't know how much voltage drop those clips create, but these things add up. Also, I can't remember if you said or not, but did you clean all of the contact points with alcohol? The Mag body (where the tail cap makes contact with the body), the tail cap (where it makes contact); both ends of the spring and the contact point for spring to tail cap.

Don't feel bad about asking these questions; you can't learn if you don't ask right. I had issues with my first SSC P7 build too, in fact, that's where I cut my teeth, and I've been learning ever since, although, the basics make a lot better sense to me now. I have to thank Linger, JustinCase and TorchBoy for helping me a LOT with my problems. If not for them, I'd still be lost:thumbsup:.

 

[Justin Case]

I've just mangled together 2 aluminum heatsink discs a little bigger than the diameter of a c cell and taped them tightly either side of the driver.


With 4 c cells the current held for a little longer then dropped again but slower.
The sinks got very hot too.

If you don't mind the mess, try adding a little thermal compound to the mounting tabs to put them in thermal contact with the aluminum disks. It's not the ideal way to heat sink those contacts, but it should give you rapid feedback in terms of effectiveness. If it works outright, you are all set. If you get longer run time but still suffer from early shut down, then you'll have to improve the thermal link to the aluminum disks, perhaps by fashioning some sort of metal standoffs that thermally connect the mounting tabs and the disks.

The mounting tabs are solderable, so you could solder a piece of large gauge wire to the tabs. You will need something that is a little more than 1mm thick/tall to rise above the top face of the AMC's case. 16 ga wire will do that. 18 or 20 ga plus some solder thickness could also work. Maybe even just a solo solder blob. Carefully file it flat for good contact to the aluminum disk and add some thermal compound.

 

[vestureofblood]

I've just mangled together 2 aluminum heatsink discs a little bigger than the diameter of a c cell and taped them tightly either side of the driver.

With 4 c cells the current held for a little longer then dropped again but slower.
The sinks got very hot too.

Vestureofblood, how you got two of these with a couple of sinks running amazes me.

Also I checked the soldering of the mounting tags.

What you have there is not bad. If you have any arctic silver 5 or similar compound you could add just a tiny bit to the top of the driver components on each side. This would help transfer the heat to the sinks more rapidly.

I am glad you have not gotten discouraged, you have all the making of a true flashaholic/moder IMHO. At this point I would say at the very least you have gained both knowledge and experience.