Mag switch and AMC7135 drivers

Hi All,

I hope to run an SST-90 at 8.4A using 6 parallel 4 x AMC7135 driver boards.

Instead of switching the 8.4A with the mag switch, I'm going to try switching only the connection to the Vdd pins.

Anyone tried this? It should save any problems switching the high currents.

That's the way to do it!

greencardigan said:

Hi All,

I hope to run an SST-90 at 8.4A using 6 parallel 4 x AMC7135 driver boards.

Instead of switching the 8.4A with the mag switch, I'm going to try switching only the connection to the Vdd pins.

Anyone tried this? It should save any problems switching the high currents.

Click to expand...

What batteries are you planning to use? I only ask because I built an SST-90 Mag using 7135's and I didn't really get great results. Click here

ti-force said:

What batteries are you planning to use? I only ask because I built an SST-90 Mag using 7135's and I didn't really get great results. Click here

Click to expand...

I'm planning to use 3 parallel IMR26500.

I also have some dx 25500 I might try too.

The 26500's (AW's?) should do better than the Batteryspace 26650's. Hopefully you will get a low Vf emitter. Anyway, could you go into more detail about how you're planning to cut Vdd?

I have read your SST-90 / AMC7135 thread and my build will be similar. Yes, I have AW IMR26500.

I have started stacking the AMC7135 boards and will probably have 2 stacks of 3 sitting side by side to reduce the height.

I have 3 parallel connections between each board. 1 for each pin on the AMC7135 chip (in from LED, out to battery -, and Vdd). My boards look slightly different to yours.

I'll run the battery + to the LED and also to the Vdd via the switch.

I have taken photos of my progress so far but don't have them with me at the moment. Watch out for a build log in the next week or so.

greencardigan said:

I'll run the battery + to the LED and also to the Vdd via the switch.

I have taken photos of my progress so far but don't have them with me at the moment. Watch out for a build log in the next week or so.

Click to expand...

I guess I'll have to see pictures or a diagram to get an idea of exactly how your talking about wiring it. I've seen someone on here mod a Mag switch with a Judco 10A (I think 10A) switch. You might want to consider that. Just something to think about. I'll keep an eye out for your build thread:thumbsup:. I'm planning an SST-90 with H6CC build when I get some spare time.

ti-force said:

The 26500's (AW's?) should do better than the Batteryspace 26650's. Hopefully you will get a low Vf emitter. Anyway, could you go into more detail about how you're planning to cut Vdd?

Click to expand...

What was the shortcoming with the Batteryspace 26650s when compared to AW's IMR26500s?

ti-force said:

I'm planning an SST-90 with H6CC build when I get some spare time.

Click to expand...

I'm waiting for the flex version (H6flex).

Justin Case said:

What was the shortcoming with the Batteryspace 26650s when compared to AW's IMR26500s?

Click to expand...

I've just had others tell me that the AW's delivered more current than the Batteryspace 26650's. Lower internal resistance maybe? What do you think? You certainly know more about this stuff than I do. If I could only learn half of what you know about this stuff.:thumbsup:

Hi guys,

I actualy just finished doing a build similar to this today. I used the Britelumens heatsink with a NW SST-90 5x7135 1400ma boards and a Shiningbeam 2.5 amp 3 mode driver.

I used 20 G wire for the whole project. I dont know a lot about the AW cells, but these drivers have a lot of internal resistance.

When I first assembled the light it was DD on 3 Accupower C cells. This drove the emitter to about 12 ish amps (its hard to say exactly since even with thick modified leads DDMs still have internal resistance that my tailcap does not). After I completed the driver it runs at only about 8 amps or so on high. In order to deliver the full 9.5 amps constantly I have to use 4 cells.

I think after some burn in time the VF may drop low enough on the led to allow me the full current on only 3 cells, but only time will tell. Good luck with your build.

My only suggestion is use thick wire, and get the resistance as low as possible.

vestureofblood said:

My only suggestion is use thick wire, and get the resistance as low as possible.

Click to expand...

Thanks for the info. I'm using 18AWG.

ti-force said:

I guess I'll have to see pictures or a diagram to get an idea of exactly how your talking about wiring it.

Click to expand...

ti-force, just for you. A quick sketch and a pic showing a test run using 2 dead AAs.

It works as expected and turns on when Vdd pins are connected to battery +.

vestureofblood said:

I actualy just finished doing a build similar to this today. I used the Britelumens heatsink with a NW SST-90 5x7135 1400ma boards and a Shiningbeam 2.5 amp 3 mode driver.

Click to expand...

How do you wire that combination?

saltytri said:

How do you wire that combination?

Click to expand...

Not trying to answer for Vestureofblood, but he probably wired it like the diagram in this post right here.

Click here for the link to the original thread. If you read that whole thread, you should know how to wire that setup.

ti-force said:

Not trying to answer for Vestureofblood, but he probably wired it like the diagram in this post right here.

Click here for the link to the original thread. If you read that whole thread, you should know how to wire that setup.

Click to expand...

Thats right.:thumbsup:

It was actually ti-force among others that handed me the key to unlocking this drivers potential.



I have some news you may be interested to hear as well ti-force.

I thought about what you said about the sag after only a few minutes runtime on you build, so I went out to check to see how long the 4 nimh cells would hold up to the 10 amp load.

Just like in your light it was super hot out of the gate, and after only a few minutes the current would sag to a disappointing 7 ish amps.

Suspecting I already knew the cause of this I did a similar test with the 5 amp SST-50 I have. On 3 cells the current would sag after a few minutes to around 3.5 amp.

Now here the iron clad proof that my suspicions were true. On 4 cells the current out of the gate was 5.56A, and very quickly (even faster than on 3 cell) the current plummeted downward eventually reaching 1.8 A.

The solution for this whole problem (at least for me) was heat sinking for the drivers. I disassemble the SST-50 driver made of 2 SB drivers sharing one controller board, and added 3 thick slices of aluminum flat bar and some arctic silver.




After doing this, the current from start on 3 cells was about 4.5, and from there ........... it climbed . After a short period of time the current stabilized at around 4.67. It also held up under the 4 cell load far better than before.

I will post more on this later, I really just wanted to get this info on the table so the OP could take this into consideration before his light was fully assembled. I also thought it may solve our current issues if we can come up with a way to sink all the boards it takes to do the 90.

to answer the original post, i thought that was the benefit of running one of these drivers? you wire it such that the switch only sees a small switching current to flip the relay, not pass the full current that goes to the LED.

vestureofblood said:

Thats right.:thumbsup:

It was actually ti-force among others that handed me the key to unlocking this drivers potential.



I have some news you may be interested to hear as well ti-force.

I thought about what you said about the sag after only a few minutes runtime on you build, so I went out to check to see how long the 4 nimh cells would hold up to the 10 amp load.

Just like in your light it was super hot out of the gate, and after only a few minutes the current would sag to a disappointing 7 ish amps.

Suspecting I already knew the cause of this I did a similar test with the 5 amp SST-50 I have. On 3 cells the current would sag after a few minutes to around 3.5 amp.

Now here the iron clad proof that my suspicions were true. On 4 cells the current out of the gate was 5.56A, and very quickly (even faster than on 3 cell) the current plummeted downward eventually reaching 1.8 A.

The solution for this whole problem (at least for me) was heat sinking for the drivers. I disassemble the SST-50 driver made of 2 SB drivers sharing one controller board, and added 3 thick slices of aluminum flat bar and some arctic silver.


After doing this, the current from start on 3 cells was about 4.5, and from there ........... it climbed . After a short period of time the current stabilized at around 4.67. It also held up under the 4 cell load far better than before.

I will post more on this later, I really just wanted to get this info on the table so the OP could take this into consideration before his light was fully assembled. I also thought it may solve our current issues if we can come up with a way to sink all the boards it takes to do the 90.

Click to expand...

So it looks like the higher voltage (4 cells vs 3 cells) is causing the driver to heat up a lot from a lack of dissipating the extra heat, due to the higher voltage. I certainly wouldn't have thought it would make that kind of difference though. Thanks for pointing that out. Like you said though, it will be difficult to get the heat out because of the number of boards required.

I wonder if you could use an idea that Justin Case came up with to get the heat away. See his post/thread right here.
He used copper strips to wick the heat away from the components and into an aluminum sink. You might be able to make one strip for every 2- 7135 chips and run it out to a sink. So that would be 2- strips of copper per board. Just an idea, it might not work out.

irv_usc said:

to answer the original post, i thought that was the benefit of running one of these drivers? you wire it such that the switch only sees a small switching current to flip the relay, not pass the full current that goes to the LED.

Click to expand...

These drivers are linear regulators. Each 7135 chip(not board, but chip) allows a (roughly) 350-400mA increase in current from the battery to the emitter, so the amount of current depends on the number of 7135 chips. I'm certainly no expert at this stuff, and there's still a LOT that I don't know, but I thought current in was current out. In other words, the whole circuit, using these type drivers at least, would see the same amount of current. I may be wrong though.

greencardigan said:

ti-force, just for you. A quick sketch and a pic showing a test run using 2 dead AAs.

It works as expected and turns on when Vdd pins are connected to battery +.

Click to expand...

greencardigan, I am not sure, but in that picture it looks like your LED- may be making contact with the ground, on the drivers I have those tiny holes are all connected to ground (batt in). If that is the case the light may still come on, but it would be DD.

ti-force said:

So it looks like the higher voltage (4 cells vs 3 cells) is causing the driver to heat up a lot from a lack of dissipating the extra heat, due to the higher voltage. I certainly wouldn't have thought it would make that kind of difference though. Thanks for pointing that out. Like you said though, it will be difficult to get the heat out because of the number of boards required.

I wonder if you could use an idea that Justin Case came up with to get the heat away. See his post/thread right here.
He used copper strips to wick the heat away from the components and into an aluminum sink. You might be able to make one strip for every 2- 7135 chips and run it out to a sink. So that would be 2- strips of copper per board. Just an idea, it might not work out.

Click to expand...

That's actually not a bad idea. At the very least I think it would give some improvement.

I am fairly confident that I have come up with a reasonable solution of my own. Sadly this means another 28$ worth of parts will be needed, but I may give it a shot.