Spring bypass question

I used to braid both ends, but the results weren’t great. The braid would tend to fray and then the connection would break.

I get much better results using 22 AWG copper with with silicone wrap. I bend the top end of the wire so it can hook over the top of the spring. The bottom wire goes below the spring or is tucked into the bottom-most coil. I then solder the wire on.

I typically do both head and tail springs the same way: with the wire inside the spring.

To prevent the solder at the top of the spring from wearing out and breaking off, I now always solder a small copper disk to the top of the spring on top of the wire. This greatly extends the lifespan of the bypass and lasts much longer than solder braid.

I have some 18 AWG wire but have never used it. It’s so thick, I don’t think I could get it inside the small EDC lights I tend to mod.

For drivers with a particularly small or hard to access contact pad, like the H17F, I use the same technique as above, but might install the spring inverted so it’s wider at the top. I still have a copper disk for where it touches the battery though.

I do spring bypasses, but I hate them. They will break, it’s just a matter of when.

Which one is better and in what condition: spring bypass or beryllium copper spring?

Spring bypass is lower resistance so your lights should be slightly brighter. However, as a previous poster mentioned, they do tend to break after awhile and need to be redone.

In contrast, a beryllium copper spring should last for the life of the light.

Thanks. Guess I’ll try with beryllium copper spring. The reason I ask about condition is because I remember reading someone (forgot who it was) said he didn’t want to put too much stress to the LEDs so he didn’t do spring bypass.

I use lead free solder exclusively for >2years, mechanically it's much stronger and harder than lead based solder, so I recommend it for bypassing, reliability should be better than with lead PbSn solder which is very soft and more prone to break.

Gentlemen,

I didn’t abandon my own thread, that would be rude. I was out doing photography in an area that doesn’t have cell service… No texts, no internet, no phone. God, it was wonderful looking up and observing Perseid meteorites…

I understand and appreciate everything you’ve said, but I would imagine that contact resistance would be more of an issue that spring resistance. Only the smallest part of the spring actually touches the battery. I would rather see a pair of concentric springs than a single bypassed spring. Concentric springs would be less resistive, but the real benefit of concentric springs would be more spring in contact with the battery.

Bypassed spring may be more harmful than helpful. Consider that the bypass wire may cause axial misalignment between the spring and the battery. Less of the spring contacts the battery = increased contact resistance = more heat = loss of efficiency.

My profession requires that I use a good LCR meter. When I get a chance I’ll measure the resistance of several springs taking a 4-wire bridged measurement.

I appreciate the information and comments presented here, but I’m still a non-believer, until I prove that spring resistance results in a meaningful loss of efficiency.

Cheers

I’m just wondering why small brass pills aren’t soldered directly to the driver vs a spring (get ’em by the dozens at FT and the like). Have a big-ass bypassed spring in the tail, but the solid flat brass pill soldered to the driver should be seriously low-resistance.

1000% agree with you.

(Sheesh I haven’t used FT in years. Forgot all about them)

But there may be more (with me there usually is)…

Depending on the material used to make the spring, it may be difficult or impossible to make good bypass-wire solder joints. Assuming that the spring is solder-able I would prefer something like a #14 brass washer (or something like that) soldered to the spring (using a washer thin enough to allow the continued use of protect batteries, if necessary) much more contact area, which can’t be a bad thing.

Cheers

It’s not efficiency we are concerned about here, it’s output. If you look at djozz thread on springs you will see that a normal steel spring will have a voltage drop of near 300 millivolts. Add another 100 mV for driver and wire resistance if everything is done well. Taking into account the battery sag you now have a maximum of 3.7 volts available to the emitter.

An XPL-Hi at 3.7 volts will take roughly 4.5 amps producing about 1500 led lumens. Done properly, the same led given 4.0 volts will take about 6.5 amps giving 1800 plus led lumens.

For me candela is more important than lumens as I build primarily throwers and I can tell you from personal experience that the difference is substantial. For leds with a higher forward voltage like XP-G2 I like to use the 45% IACS springs because in my applications I can accept a roughly 11 mV drop.

By all means do the tests on various springs yourself and please post the results. We are always happy to have more resources.

Mountain Electronics sells brass buttons specifically for this purpose although I find it difficult to solder both the button and a wire bypass on the tail spring. If done properly a dome of solder can be applied to the top of the spring and sanded flat giving good contact.

The stock springs are solder-able as they need to be soldered to the switch board.

I respectfully disagree.

When efficiency is increased all kinds of good things can happen. Goodness = efficiency.

You’re assuming that supply voltage and emitter output are directly proportional. The electronics in between the battery and emitter negate the possibility of the proportional relationship you’re assuming. Of greater consideration should be emitter efficiency (lumens/watt) and/or efficiency of voltage/current regulators and/or efficiency of the processor itself along with the quality of PWM and PWM pulse integration and/or the on resistance of the low-side mosfet switch, along with transfer of heat from the emitter to the outside world.

Cheers

Ok, correct me if I’m wrong but the battery is connected directly to the emitter on the positive side (at least in the lights I build) and the electronics follow after on the negative return from the emitter. Or am I missing something here.

Edit: for clarity

As I mentioned earlier we are not concerned with efficiency when we do a bypass, just raw output.

V+ goes to voltage/current regulators . From there, regulated Vcc (usually 3.3 volts) goes to the processor. Buck regulation is easy. The processor looks at Vcc through a simple voltage divider. Next, the processor delivers PWM based on battery voltage and user input. That is the best case scenario, based on buck regulation for the electronics and an emitter that can be directly driven by the battery. All of that goes out the window when the emitter requires a supply voltage higher or lower than what the battery provides.

Next comes thermal sag. This is a very big deal as flashlights become brighter and smaller.

The point of diminished returns can happen when emitters are over driven. It can become impossible to remove heat fast enough, and the emitter dims as a result. I’ve proven this by driving a loose XHP70 on my bench. Driving the XHP70 at 12.2 volts resulted in more heat than I could remove using a fan-cooled heatsink. This is why turbo is usually limited to just a few minutes, but it doesn’t end there. Thermal sag must always be factored in.

Getting this stuff right requires a thoughtful, sensible design, especially now when flashlights are sold based on output lumens. When all of this is considered, soldering a bypass wire across the spring becomes a non-issue. The watt or two that’s lost across the spring is not a big deal when overall input wattage is considered.

Battery voltage stays flat up until the point when voltage drops very quickly. Bypassing the spring may result in an additional handful of seconds of useful light. IMO not enough justification for me to trash my springs.

So there you have it. Now harry steps off soapbox and slithers away. Cheers.

For a FET/direct drive situation bypassing the springs can result in significant performance increase, which is what Swib and others are referring to when they talk about performance increase. The spring resistance is a significant portion of the total circuit resistance. A good 18650 cell has 15 to 30 mOhms, FET has ~5 mOhms. So when you take out the 70+ mOhms of the springs it makes a big difference. I don’t think contact resistance is usually more than a few mOhms.

Edit. To summarize, bypassing the spring can result in more current to the emitter, which usually means more output.

Contact resistance is usually in the 1-4mOhm mark depending on materials used.

In that case, gold is usually 1mOhm due to its high conducivity and softness, nickel is usually 2MmOhm, and steel can be even worse at up to 3,3-3,9mOhm, or even higher in some cases.

Or even worse. A copper alloy spring which is oxidized can be worse than that with no coating

This is why I chose a nickel plating: most cost effective while still having low contact resistance.

In most cases, like the BeCu spring I made, I chose mechanical properties over pure conductivity. I could have went with BeCu C17500 instead of BeCu C17530.

I could have gotten an even lower voltage drop at 0.045V at 6A had I changed the material, but at that point, it would be way worse mechanically, and a 0,007V difference is quite small.

That’s the whole point of why I made this specific spring: to get a spring good enough for 95% of cases where you don’t need more than 13-14A of current, and easy to bypass copper alloy spring.

Using higher conductivity parts is also crucial in FETs, but also regulated linear drivers, but most importantly, boost drivers.

TLDR: A nice phosphor bronze spring, or even BeCu spring if you want even higher power, is usually great in most cases.

If you want uttemost performance, while sacrificing mechanical properties, go with a spring bypass!

Take a look here and you will see that in fact V+ does go directly to the led+ in the builds I do.

We all understand the issues with heat and this has been addressed by one of our own members (Toykeeper) who has done a phenomenal job programming driver software to include thermal management.

“In the builds you do”. That’s what you said. I thought you were referring to products designed by engineers. The products that we pay good money for, I was not referring to designs created by electronics enthusiasts. How, exactly, would you drive a 6 volt XHP70 from a single cell? How would you make that regulator small enough to live inside a flashlight tube? I commend you for getting your design to work. Did you us a a spring bypass wire?

You know a lot more about this stuff than I do, stuff for me to learn from, so thank you kindly for the information

Question- my guess is that a small LDO regulator is used only to provide the 3.3V or 5V the electronics require. I can’t imagine a linear regulator anywhere near the emitter.

Cheers