Spring Resistance Test -Real World

I am planning a driver spring test, not for resistance as the title states, but for light output with different springs installed. I will be using my P60 drop in with interchangeable contacts (pictured below) for the test because I will be able to swap springs easily…this means the spring being tested will be the only variable. I have seen other spring tests done where the resistance of the spring was being measured…I know lower resistance is better, but how much difference does it really make.
I’m letting everyone know I will be performing this test in case, someone has a certain spring they want tested. If you want to add a spring to the test, PM me for my address and you can mail it to me for the cost of a stamp. I will postpone the test if someone wants to send me a spring.
I will be testing a plated steel spring, so I don’t think testing a different plated steel spring will be much different.

I plan on testing:
Stock Nanjg short spring
Solid contact (pictured)
Plated beryllium copper spring (IO/Mountain Electronics)
FT’s carbobronze spring (pictured)
Plated Steel spring (Convoy driver spring)
Plated Steel spring with copper braid (Convoy driver spring)

The reason I’m testing the Solarforce switch spring…a spring that would rarely be used on a driver…It will be an easy comparison of the performance of a switch spring braided and unbraided.

Test procedure:
My PFlexPRO P60 XML2 U3-1A 3.8A drop in with interchangeable contacts will be used. The drop in will be installed using a copper wrap. The same wrap will be used when the springs are swapped.
Host will be a Surefire 6P (bored) with the switch spring copper braided and capped.
I could run this test using a lab power supply, but I want to keep it as ‘real world’ as possible. I will use an Efest 35A 18650 battery. After each run, the battery will rest for 10 minutes, then will be recharged for 2 hours (recharging will only take about 10 minutes, the additional time, the battery will simply set in the charger resting).
I will run each test using my integrating sphere to read the output. Each test run will last 5 minutes. After all test are finished, I will plot the results as OTF lumens.

I plan on running the spring test soon, so let me know if you want to add a spring to the mix.


The test is complete…since this is really a driver spring evaluation, I decided to test the Convoy driver spring instead of the Solarforce switch spring.
Because of an anomaly in the initial run yesterday, I decided to run the test again today and it yielded the same results. Since I had 2 comparable runs, I decided to average the output (that’s why the graph lines will be jagged).
Rufusbduck had suggested running the test direct drive for a more pronounced difference, but my reason for the running the test was to determine the best spring(s) to offer with my interchangeable spring driver. Since it’s P60, I like to keep the power moderate (maximum 3.8A).
First, I will show the output graph and then give my thoughts on the results of each spring -I have zoomed in on the plot so the variations are more obvious…:
The solid brass contact was the clear winner…it’s heavy enough so heating is not a problem and provides a very short electrical path. I make the solid contact to provide more room for protected batteries, but also gain better performance. In this test, I had to stack 3 solid contacts to gain enough length for the shorter battery used in the test.

The Convoy driver spring - braided -If you don’t mind braiding a small spring, this a very good choice. I braided this spring with 22ga wire…a heavier gauge should give better performance. As you can see in the graph, the unbraided version had issues.

The Beryllium Copper spring (Mountain Electronics/IO). This is a great spring, but it isn’t driver friendly. It’s very short with a very wide base. It doesn’t fit a typical spring solder pad. If you make a few bends, it may work well with a single sided driver. The wire gauge is heavy but because the wire slope is shallow, it’s not too stiff.

The Nanjg driver spring This is short, stiff plated steel spring. Because of the short length, heavy gauge wire…it didn’t perform badly…but not good either. At least, heating wasn’t a problem at 3.8A. My main problem with this spring…I have hundreds of the on my workshop floor and when I step on them, they get wedged in the soles of my shoe and they’re a pain to pick out.

The Convoy driver spring This one made me run the test twice. You can see the drop off around the 3.5 minute mark…it doesn’t look like the other springs. The drop off is due to the spring heating up…when metal heats up, the resistance also goes up. If you look carefully, you can see the effect of higher resistance at 1 minute…the output curve should follow the Nanjg spring, but instead, it crosses over the Nanjg output. Without braiding…I don’t recommend using this spring (unless your power is well below 3A)

The Fastech ’Carbobronze’ Spring…what is carbobronze anyway? Since this spring is a copper alloy, I hoped for better. The output with this spring started very high, but because of the small diameter wire, it started heating and the output dropped quickly. This spring, if made of a heavier gauge wire would be a good spring, but at 3.8A…it finished next to last. I place the Convoy spring last in performance because of the steep the drop off.

Next, I tried to determine the change in resistance due to heat. I consider the solid contact my benchmark. For each spring, I took an average output from seconds 5-9 and took an average output from 296-300 seconds (beginning and end output) and compared them as a percentage. The higher the percentage drop, the more the spring resistance changed. I think this is a reasonable assumption.
Solid contact: 5.63%
Beryllium Cu: 5.79%
Nanjg: 6.50%
Convoy Braided: 6.56%
Carbobronze: 7.36%
Convoy: 7.64%

The percentages do not look very substantial, but with longer runtimes, the difference will become greater and greater until a resistance/current equilibrium is reached …or… at higher powers, the spring will fry.

Morale of the story…we need a good beryllium copper spring

Here are the springs tested -I had to make a longer solid contact. The Nanjg spring is stretched.

sounds like a great idea.

One question, is the “Plated beryllium copper spring”

the same as:



It is the beryllium copper spring from Mountain Electronics, which is the same as the one from IO.

I’ve edited the original post for the following reasons:

The ZeroRez switch is not compatible with a solid contact or the short Nanjg spring, so I will be using a regular switch with the spring copper braided and copper capped.
I will be running each test for 5 minutes instead of the original 40 seconds…I want to see if there is a difference when the spring heats up and the resistance changes.

Hmmm interesting! :wink:

This will be interesting, subscribed

Is it just me? or is one of the major functional difference runtime. using juice to heat a spring has got to affect runtime.

If this is true does it make sense to make runtime tests t see how much a difference (in minutes) one spring make over another. From practical stand point that could be much more noticeable as a real world test then a few lumens.


This might be a stupid question but why is the ZeroRez not compatible with a solid contact or the short Nanjg spring?

You’re right…but there is a problem using runtime to measure spring resistance. If a spring heats up, it’s resistance goes up. All of the current isn’t used or lost heating the spring. The spring, while hot will have higher resistance…this higher resistance will cause the output will be lower…this lower output (and higher resistance) will increase runtime.
I’ve seen this phenomenon while testing tail cap braids. With the switch spring braided, the output is higher, but the runtime is considerably shorter.

No question is ever stupid. The ZeroRez switch is a very simple brass cone. The driver spring presses the battery into the brass cone and in the process, presses the piece of brass away from the end of the battery tube. To turn the switch on, the brass cone is pressed down so it makes contact to the end of the battery tube (it’s always making contact with the battery). While pressing the switch down, you are actually pushing the battery forward compressing the driver spring. The system is a little picky on how much spring action the driver spring provides. It’s a little hard to explain, but if you ever see it and figure how it works, it makes sense.

If it’s a regulated driver then run time will be the only difference since differences in spring resistance will only result in different required regulation. Output differences should only show up with FET DD drivers where tiny differences in resistance result in much larger differences in led current. At least that’s my impression of what should happen.

Is the solid contact brass?

The solid contact is a brass flat head screw with a flat brass disk soldered to it.

The driver is a 7135 based driver. There will be a difference…at anything over 3A, there is a noticeable (easily measurable) difference when you braid the switch spring. The driver spring is just another bottle neck in the system.

Test results have been added to the original post.

I just saw this, screw in spring or screw in flat contact.
What a porn :smiley:

I don’t have a whole lot to add other than saying thank you for putting this out there :beer:

that and I will keep using the beryllium springs for my drivers!

Since my drop ins are potted, I can’t solder on something different if the spring gets crushed or if I want a solid contact. This was the only way I could think of to do it.
The inspiration came from member KawiBoy1428. I built him a drop in, he needed a stronger spring for a zerorez, so I made him one with the interchangeable contacts and I’ve been building them that way ever since.

…or did I miss a joke?

No joke, I really like the idea and how you implemented it, looks very nice and useful thing, didn’t see that before.
It’s one of those moments when you say “why didn’t I think of that”. :slight_smile: