AMC7135 more and more, no effect

I currently am adding AMC7135 to my driver. With 4 I get about 2.95v at the led, which is about 1.2A according to the current/voltage graph of a XM-L.
With 6 I get barely 3.03v at the led, and with 8 I don’t get more, which is about 1.5A.
I am using thick and short wires, the XM-L is on the large heatsink for the test.

What is the issue ? Why can’t I get more than 1.5A ? :~

I use redilast 18650, so the issue is not on the battery technically.

Did I miss something ?

Measure current, not voltage. Voltage is not very exact to determine the current.

The graph is pretty clear though

And measuring the current doesn’t give me more.

According to the graph I am far from having 2A with my 3.03v, and the led is not as bright as it should be.

You should perform a tailcap reading. Those graphs arent exact because Vf varies from emitter to emitter.

Have you examined your soldering under magnification. Is the battery capable of 3A current draw? Also, that graph was generated at an emitter temp of 25C. Vf goes down as led junction temperature rises. To find out the current you have to measure the current and the EE’s seem to concur that really accurate current measurements are beyond the capabilities of inexpensive DMM’s. I use mine more to spot problems than for exact data.

No I added 2 more AMC7135, so I get 8

At tailcap I get about 1.4A.
I know with the multimeter connected I can’t get 3A, but even with no multimeter the flashlight is less bright than my fenix rated 468 lumens, while it should have about 900 lumens.
This is my second driver nanjc with 8 AMC7135 I mod and I still can’t get 3A.

Do you think the spring can reduce the current by being too thin ?

It’s certainly possible, some people add a braid from top to bottom of the springs, essentially, the braid carries the current, the spring then just applies tension to the battery connections.

Any bad connections, poor 0v path, poor positive path, high resistance switch connections, all will limit current supplied to the driver and therefore emmitter.

I would go back and check all your soldering joints to the 7135’s too, many people have successfully added more chips and seen the current increase at the tail cap, so something is playing up in your set up.

Hope you get it sorted, its bound to be something silly, have you repeated the tests with different power supplies - batteries, bench supply, what ever? It’s more a case of eliminating each point and see if you find a difference than anything.

Well, what kind of cells do you use?

The type of dmm leads and how they connect the circuit makes a HUGE difference. When I changed one long thin lead with a probe on the end to a short fat awg stranded wire with an alligator clip current jumped from 1.4A to 1.9A. When I changed the other it jumped again up to 2.4A. The alligator clips help but are not as good as a solder connection. Are you still measuring voltage? I have measured 2.8A through an led with a Vf at that current of 3.1V. The Cree curves are based on The average values of all the different bins so can’t be used to infer from I to V or from V to I. A short between the legs of any chip would short the output of all to zero so either the legs on some chips aren’t connected, or some chips were overheated, or the measurement is not accurate. Can you post some close ups? Image shack is free. A poor connection somewhere, either on the board or to the meter is a likely culprit. One other thought, measure the led voltage drop and then measure the battery voltage under load(with the led on high). If they are ~ the same then voltage sag in the cell is the problem. With a fully charged cell the voltage under load should be greater than the led Vf with the excess being converted to heat in the driver.

I use redilast 2900mA 18650

Here are some images.
I know soldering looks dirty here but without the macro zoom and the flash of the camera everything looks clean (or at least almost)
There are some solder on the back due to testing.

In the last image the bottom chip above R2 looks heat damaged. It looks like you are doing it correctly but need a bit more speed to avoid overheating. Also, the chips may handle heat better when in use if the larger outside ground tab is soldered rather than the middle pin. You can actually remove the middle pin(dremel) to make more room for the iron. Electrically they are the same. I hate to say it but welcome to the learning curve. You may have to start over.

Btw, thanks for posting the pics. It helps and even under mag it looks better than my first attempts. Some people use a tiny drop of AA adhesive or superglue(keep the pins clean)to hold the chips in place prior to soldering. You might be able consider reusing still functional chips but I would not as they would first have to be identified and then have to survive desoldering and resoldering.

I can measure 4.2A very well with the stock spring on the 8*7135 driver modded to 12*7135, so it depends on the spring.