Using an NTC ?

lctorana · Jul 28, 2008

LuxLuthor said:
Then I checked resistance in my 64°F room by running precisely calibrated 1.0 Amp into the NTC, and time required to reach stable resistance level

0.628 Ohms @ 1.0 Amp after 4 minutes

This is my experience, too.
I wanted to create the brightest-possible 3854L configuration, so went for a 8-cell (10 - 9.6V) battery with a 2.2ohm NTC.

It worked, but at switchon, was really dim, and warmed up so slowly that you would have turned the torch off and finished with it, long before maximum lumens were reached.

At 2A, the NTC takes MINUTES to reach low resistance.

I don't know what's "enough" current to make a NTC worthwhile, but I can say it's a lot more than the 2A of the 3854L.

LuxLuthor · Jul 28, 2008

Northern Lights !!! Great to see you doing well enough to dip your toe in the CPF pool!

lctorana, it is not an easy question to answer because the resistance is totally heat related. If you put it next to the bulb, it will pick up that heat and its resistance will drop fast, and related to the temp.

I posted some actual numbers in my first post, and a reminder for anyone considering the NTC, what determines its resistance is how physically hot/cold it becomes....to a certain end point.

I'm not as enthusiastic about using the NTC after seeing the resistance rise above 4 Ohms when I put it into the freezer...which approaches winter time use. Then with heat from heat shrink hot air gun, its resting resistance dropped to < 0.1 Ohm.

It only starts with a specified "resting" 1 Ohm resistance if the NTC is @ 25°C (77°F)--which is reasonable in summer. You end up needing to calculate the resistance of your entire light setup, and add the probable temp ranges the NTC will be exposed to. It is hard to get the right "all season" resistance balance using them.

The best solution will be AW's soft starting, regulated voltage driver.

2xTrinity · Jul 28, 2008

Years ago I saw a clever circuit using a NTC in parallel with a high current (low on resistance) FET. An RC circuit on the FET's gate gave it something like a 5 or 10 second delay before turning on the FET. When power was first applied, the NTC was there to limit current. After 10 seconds or so, the FET switched on, giving the load nearly 99% of the available voltage. Because current was now going through the FET, the NTC would cool off and be ready in short order when next needed.

In this circuit, was the FET being used as a switch somehow, or was a physical switch still being used to turn the circuit on and off? If the former, the light woudl never be able to turn off with an NTC in parallel.

I have tried to use RC circuit on the gate of an FET as a soft starter. When switched on, the FET Will then linger in a more highly resistiant region before the gate cap fully charges up. The problem there is that Unless you have a double-throw switc that toggles between positive, and ground, when the light is turned off it will take a long time for the capacitor voltage to completely discharge, leaving the light in partially on state for a fairly long time. In order to drive that, it's necessary to use a more complicated gate driver to allow the gate cap to drain when the light is switched off.

Northern Lights · Sep 13, 2008

Flipside said:
NL:

Did you ever hear back from that Engineer? My new FM85 instaflashes on fully charged batteries, and I'm hoping that an NTC could be the solution.

Certainly, I could charge the batteries to 4.0V instead of the normal 4.2V. But then, I'd be giving up run time and absolute brightness...

Had time to test all the samples. My original guestimates and NTC choices were best. The SL121R010 works best with the 5761 and similiar bulbs, if you need a little less resitance at steady state then the SL150R00 works good at 4-5 amps too.

I find that at the 5 amps of the 5761 it takes only 0.02 ohms to make 0.10 volt change. I was working with emolies at 7.6 V battery and needed 7.3 bulb voltage but after achieving that the lot of bulbs I have flash over 7.2 volts. I have been working on the edge. If the batteries set some the V battery settles at 7.5 volts.

I could not get the NTCs for read correctly, I had the light laid open on the bench and jumpers everywhere reading voltage.

I found the NTC to be very sensitive to temperature and changes in voltage and amps do to the resistance of the connections in the circuit. The NTC also have a huge variance in tolerances.

The best way to handle it is to set the light up and work a small wire onto the pins and take bulb voltage that way and see what the system is doing. I almost always put the NTC in the tail cap. I find the SL121r010 on 5761 4.5 to 5 amps reaches steady state at 0.060 to 0.080 ohms; right where it is supposed to be.

Setting up the same circuit with low resistance jumpers, just a bulb and batteries and the NTC will give you readings way different than that of the set up light. It is very sensitive to the slightest changes in resistance in the circuit.

It is not too critical unless you are trying to work with in a tenth volt of insta-flash. You probably will have no problems if you work a half volt short of flash point.

I have one particuliar light that will flash 5761 at 7.3 volts and it will exceed that on freshly charged emolies; if they set 20 minutes or more the voltage reading is 7.2 or less bulb voltage.

I ran out of bulbs today. If I find that this is still too critical and is flashing then I will add a resistor to the circuit and shoot for 7.1 volts. My highest reading on the batteries was 7.6 volts under load, highest dreaw was 5.0 amps. Dropping a 0.2 volts more than the 7.3 uses 0.04 ohms and at .1 watt. I would use two half watt ohms resistors in parallel or 4 quarter watt .16 ohm resistors in parallel to achieve that 0.04 ohm resistance.

Northern Lights · Sep 21, 2008

After much experimentation. I found the NTC solution for Emolie and 5761, and that can be translated to a Li-ion build too.

First the NTCs, SL12 1R010, have a a 20% tolerance. That is huge.
Second the NTCs are very reactive and sensitive to temperature and slight changes in internal resistance and slight changes in circuit Amps.

The NTC will become more resitant when the Amps or temperature drop or other things decrease circuit resistance, like adding resistors or other NTCs in parallel.

Adding more than two in parallel or series you likely will defeat the purpose and they will not work.

Therefore I found that I had to test the lights used fully assembled. I mount NTCs in the tail cap with an inline fuse. Using two pieces of copper wire I wrap the bulb pins and stick them into the socket and from these extensions hook up a meter. Since I put charging jacks into the lights I could hook up to the socket and measure Vbattery quite easily.

I found that the slight differences from an assembled light and testing the circuit on just jumpers would cause significant differences in the NTC behaviour. In the light they acted more like the theoretical equations, outside they always were too high in resitance.

At 5 amps I was getting 0.06 resistance or close to it and that is spec.

With the emolie I found my current 5761 were flashing at just about 7.09 volts. Ouch, look at the pile at 5-6 dollars each.

I used many different Ametherm NTCs too.

Using a resistor in series was necessary to bring the emolie voltage down to Vbulb at 7.0 when the NTC reached steady state. 0.09 ohms and the NTC would flash and read 7.4 Vbulb. Vbattery was 7.6 volts. I needed to drop 0.4 volts more and add .08 ohms resistance. That figured to a 2 watt .18 (calls for a .17) total resistance in series with the NTC.

Put that in and it worked. The out put depends on the Vbattery but the drop is about .6 volts when fully charged.

It is by the eye brighter than the A123 driven 5761s and at the level of the D Lion driven 5761.

I do not use a NTC with that one but have a Batteyspace PCB on the pack after stripping the OEM PCBs. The PCB somehow behaves just enough as a soft start that I have no flash problems with that build.

So my suggestion is to suck it up and build you light. Start with the 12SL1R010 and a .10 resistor in series if you are using the 7.2 volts on Li ions or .18 ohms if on the 7.4 emolie and then take some readings off the bulb. Resistors are cheaper than bulbs, you can fine tune in this way. You may loose one or two bulbs but I have given you the starting point.

In all cases since I use the tail cap for the NTC, there is no spring resistance in these mods.

Good luck.

LuxLuthor · Sep 22, 2008

NL, I am one who can appreciate your sacrifices to "The Bulb Gods."

Interesting to see your 7.09V flash point in your setup vs. the flashing I had with my destructive testing. I only had 19 milliohms of resistance in my testing setup, but the key was the slow manual rotating dial increases that eliminated the inrush current spike.

The only viable solution I could justify with a bulb like the 5761 where such a small voltage tolerance window exists was using AWR's regulated Hotdriver. It's been a year since AW mentioned doing a regulated version of his multi-level, soft-starting driver, but it's encouraging to follow AlanB's project.

Patriot · Sep 22, 2008

Very interesting read.

As it stands now is the NTC still the most practical option to block start-up voltage spikes.....at least for the casual moder?

lctorana · Sep 22, 2008

I definitely say so. I'm a total NTC convert.

Northern Lights · Sep 23, 2008

LL-It seems to me the 5761 varies a bunch in the different lots as far as some of he parameters are. When AWR was still communicative he had found a flash point of 7.1, and yours is higher.
I found different dimensions too, just a little difference in the envelope. Once I had a light with a reflector just fitted to teh 5761 I had on hand, next lot I screwed the head on and crushed the bulb as it had some "ears" where the pins were crimped. I have also seen some slight difference in the distance between the element and the bottom of the bulb. This will make a difference in the focal point. I have one light that the point is so low that when a bulb with a little longer space between bottom and element is used I find I have to set the bulb a bit high in the socket.
I think these differences were much the problem in analyzing the NTC problems.
It is also the reason you need to work with a meter and the actual light to achieve the results in the mod you want.
Thanks for the support, LL, soon I will be moving onto the IRC 65 and I will need you to build me that battery pack...

jimjones3630 · Sep 24, 2008

Northern Lights said:
After much experimentation. I found the NTC solution for Emolie and 5761, and that can be translated to a Li-ion build too.

First the NTCs, SL12 1R010, have a a 20% tolerance. That is huge.
Second the NTCs are very reactive and sensitive to temperature and slight changes in internal resistance and slight changes in circuit Amps.

The NTC will become more resitant when the Amps or temperature drop or other things decrease circuit resistance, like adding resistors or other NTCs in parallel.

Adding more than two in parallel or series you likely will defeat the purpose and they will not work.

Therefore I found that I had to test the lights used fully assembled. I mount NTCs in the tail cap with an inline fuse. Using two pieces of copper wire I wrap the bulb pins and stick them into the socket and from these extensions hook up a meter. Since I put charging jacks into the lights I could hook up to the socket and measure Vbattery quite easily.

I found that the slight differences from an assembled light and testing the circuit on just jumpers would cause significant differences in the NTC behaviour. In the light they acted more like the theoretical equations, outside they always were too high in resitance.

At 5 amps I was getting 0.06 resistance or close to it and that is spec.

With the emolie I found my current 5761 were flashing at just about 7.09 volts. Ouch, look at the pile at 5-6 dollars each.

I used many different Ametherm NTCs too.

Using a resistor in series was necessary to bring the emolie voltage down to Vbulb at 7.0 when the NTC reached steady state. 0.09 ohms and the NTC would flash and read 7.4 Vbulb. Vbattery was 7.6 volts. I needed to drop 0.4 volts more and add .08 ohms resistance. That figured to a 2 watt .18 (calls for a .17) total resistance in series with the NTC.

Put that in and it worked. The out put depends on the Vbattery but the drop is about .6 volts when fully charged.

It is by the eye brighter than the A123 driven 5761s and at the level of the D Lion driven 5761.

I do not use a NTC with that one but have a Batteyspace PCB on the pack after stripping the OEM PCBs. The PCB somehow behaves just enough as a soft start that I have no flash problems with that build.

So my suggestion is to suck it up and build you light. Start with the 12SL1R010 and a .10 resistor in series if you are using the 7.2 volts on Li ions or .18 ohms if on the 7.4 emolie and then take some readings off the bulb. Resistors are cheaper than bulbs, you can fine tune in this way. You may loose one or two bulbs but I have given you the starting point.

In all cases since I use the tail cap for the NTC, there is no spring resistance in these mods.

Good luck.

NL Thanks for the work and expense. Will benefit many devoted 5761 fans.

Using an NTC ?

lctorana

Flashlight Enthusiast

LuxLuthor

Flashaholic

2xTrinity

Flashlight Enthusiast

Northern Lights

Flashlight Enthusiast

Northern Lights

Flashlight Enthusiast

LuxLuthor

Flashaholic

Patriot

Flashaholic

lctorana

Flashlight Enthusiast

Northern Lights

Flashlight Enthusiast

jimjones3630

Flashlight Enthusiast

Similar threads

Latest posts