Initiators - ~1 amp Test Data

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

RCBrust

Well-Known Member
Joined
Jan 18, 2009
Messages
193
Reaction score
101
Location
Coopersburg, PA
I decided to take a look at firing currents for ejection charge initiators and thought you guys might be interested in the results. A lot of flyers, including myself, often use 2, 3, 4 amps or more to fire these things. If you look at the specs for MJG Firewires, they recommend a nominal firing current of 1A. The Chinese initiators available on eBay list a firing current of 0.4A. With that in mind, I ran some tests to see how the initiators perform at around 1 amp. I started with the eBay initiators but I have Firewires on order and will report my findings with those when they arrive. My 3 goals for the tests were:

1. Use the altimeter, battery and initiators within their recommended specs (or close to it)
2. Have a configuration where a single altimeter pyro output could drive multiple initiators
3. Be able to have one (or more) of the initiators fail open or fail short and not have any affect on the others

To that end I decided to use a 2s LiPo battery and a 5 ohm resistor in series with each initiator. A fully charged 2S LiPo has a voltage of about 8.4V at room temp and the initiators have a resistance of about 0.8 to 1.7 ohms, depending on the brand. Max current per initiator should be about 8.4 / (5 + 0.8) or 1.45A. To guarantee a minimum of 1A, the battery voltage would need to be at least 1 * (5 + 1.7) or 6.7V. Even at a temp of 0C, a decent LiPo battery should be able to deliver at least 75% of its capacity before dropping to that voltage.

The battery that I used was a Turnigy 300mA 2S LiPo. It provides a constant discharge of 10.5A and a burst discharge of 21A so current is not an issue.

To simulate non-ideal conditions, I fully charged the battery and then discharged 150mAh from it so it had about half of its capacity left. I then began firing off the eBay initiators in series with a 5 ohm 1% power resistor and monitoring the current with a digital scope. The first 10 were done at room temp (~17C or 63F) and the current pulses ranged from 10.5ms to 11.8ms with amplitudes of about 1.25A before the initiator fused open.

I then cooled the whole setup (including the battery) to about 0C (32F) and fired off 10 more. The current pulses ranged from 13.3ms to 16.8ms with amplitudes around 1.1A.

Considering that many of the popular altimeters fire their pyro outputs for 1 second and can source around 5A, I'd have no problem using up to 3 of these initiator/resistor combinations on a single pyro output while staying well within the capability of the battery and the altimeter, and still firing within the first 2% of the 1 second window.

Here's a sample of what the current looked like during a firing:

current.jpg

I was actually a little surprised at how clean the current trace was. I was expecting it to be noisier. It should be noted that even with a 16ms wide pulse of current, a single firing pulls about (1250mA x 0.016s) / (3600s/hr) or 0.006mAh out of the battery, i.e. the 300mAh battery can do this quite a few times.

Randy
 
This is a very good test. It's a good reminder that a current source is a better way to fire an e-match than just shorting across a battery.

Your method works well because the battery can supply several times the current needed (it has a low internal resistance compared to the load resistance), and the 5-ohm resistor limits the current somewhat above the all-fire current of the e-match. This forms a simple (pseudo) current source. A short will see 8.4V/5ohms = 1.68A, and a worst-case 1.5-ohm e-match will see 8.4V/6.5ohms = 1.29A. Both are over the 1A all-fire current.

If you were to make a parallel circuit of two 5-ohm resistors each in series with an e-match, each branch would get a max of 1.68 amps without affecting each other in case of a short. This assumes you have a battery that can supply the total current without a significant drop in voltage (as is the case with a correctly-sized LiPo).

To see if this would work with a 9V alkaline battery, you need to include its ~1.5 ohms of internal resistance when the 9V is fresh (100x that of the LiPo). For each 1.5A drawn by an e-match, the battery will drop 1.5A*1.5ohms = 2.25V. With two e-matches, the battery is at 1/2 it's 9V at the instance of firing. Each branch of the circuit is now starved: 4.5V/5.6 ohms = 0.8A (I chose a 0.6-ohm e-match for a worse case). With two e-matches, the 9V alkaline is already supplying below the 1A all-fire current for regular e-matches to each branch. Most likely they would fire, but not guaranteed.

A better firing circuit is a true current source set to about 1.25A. You can easily make a current source from an LM317 adjustable voltage regulator plus a 1-ohm resistor. https://diyaudioprojects.com/Technical/Current-Regulator/
Put a current source in series with each e-match. This would also work for igniters if you test the required current needed to reliably fire.

Altimeters typically have a regulated firing circuit. You can also stagger the firing of ematches to reduce the instantaneous load on the battery.
 
Do you have A Coles notes version for non electrical engineers. What was your take away?

My goal was to answer a few questions:

1. Do I feel confident firing initiators closer to their recommended firing currents? Yes.

2. Do I feel confident that I can reliably achieve this current level with a small 2S LiPo and 5 ohm resistor in series with each initiator, even at cold temperatures with the battery half depleted? Yes.

3. Do I feel confident using 2 or 3 of these on a single altimeter pyro output rated at 5 amps? Yes.

Again, these results are only valid for the Chinese initiators so far. Firewire results are yet to come.

Randy
 
Question: So what does this prove for the stupidheads out there such as me? When using a 2S lipo of any mah rating just stick on a 5 ohm resistor (of what wattage, 1/8? 1/4? etc) in a series with the initiator and one doesn't have to worry about blowing the FETs on whatever electronics they are using? If so, I'm game for it and thanks for testing. I got a fair share of those initiators. Kurt
 
Question: So what does this prove for the stupidheads out there such as me? When using a 2S lipo of any mah rating just stick on a 5 ohm resistor (of what wattage, 1/8? 1/4? etc) in a series with the initiator and one doesn't have to worry about blowing the FETs on whatever electronics they are using? If so, I'm game for it and thanks for testing. I got a fair share of those initiators. Kurt

Yes. Wattage of the 5-ohm resistor is not important because the action time if very short.

You can repeat for multiple initiator circuits up to the rating of the FET.
 
Question: So what does this prove for the stupidheads out there such as me? When using a 2S lipo of any mah rating just stick on a 5 ohm resistor (of what wattage, 1/8? 1/4? etc) in a series with the initiator and one doesn't have to worry about blowing the FETs on whatever electronics they are using? If so, I'm game for it and thanks for testing. I got a fair share of those initiators. Kurt

It depends on whether you want to add an inexpensive resistor to each initiator for single shot use, or have a resistor that's permanently mounted and used over and over again.

For single shot use, I personally would not use any old 1/8W metal film resistor unless it has some type of pulse or overload rating, or you test the heck out of it. You can find some carbon film resistors that have decent pulse ratings and are fairly cheap.

For permanently mounted repeated use, you can easily find pulse/overload rated resistors that will handle the full battery voltage for 1 second (worst case). They'll cost a few dollars but you only have to buy them once. This is most likely the way that I'm going to go. I'll make a PCB with appropriately rated resistors and terminals for multiple initiators.

Randy
 
Randy,
My apologies for jumping into your thread with more info. Your tests are the kind of thing people should be doing more often and not just "winging it"!

The safest thing to do is to use a 6W+ resistor, but not wire-wound (too much inductance limits the shape of the firing current). Probably $2 each.

The cheapest and "good enough" choice is a 1W non-precision metal film resistor, either axial lead or the "MELF" cylinder SMT package. The reason for using non-precision is that the process of laser trimming the resistor film leaves a weak spot that doesn't hold up to pulse currents. 25 cents. Use 5.1 ohm because it's a standard 5% value.

Better than a resistor, use an LM317D (SOT-223-4 package) rated at 1.5A continuous. Only 30 cents plus a 1-ohm 1W resistor (25 cents). This is how I would do it if I were making a little circuit board (or an altimeter).
 
Randy,
My apologies for jumping into your thread with more info. Your tests are the kind of thing people should be doing more often and not just "winging it"!

No problem John. The more opinions and ideas, the better.

Yeah, I actually started out with an active current source (a Linear Tech regulator similar to the LM317) but decided to go with a single resistor for the sake of simplicity. Six of one, half dozen of the other really. As with most things, there are umpteen ways to do this. If someone wants to play around with this method, here are some other suggestions.

For single shot use, i.e. an inexpensive resistor to wire in series with each initiator, the Stackpole CF12 should work well. It's available from Digi-Key in a 5.1 ohm version and 100pcs are about $3. They're axial lead 1/2W resistors pulse rated at about 11W for 20ms. If enough people are interested in this, I can get some in and run some accelerated testing on them to see how they do.

For repeated use (PCB mount), I'm going with the Bourns PWR163. It's also available from Digi-Key in a 5 ohm version and they're just under $3 each. They're 25W surface mount resistors and come in a fairly compact surface mount package. The worst case that I want to be able to handle is a fully charged 8.4V Battery directly across the 5 ohm resistor (initiator lead shorted) which is 14.1W. This resistor should easily handle that for 1 second, even with little heat sinking. I'll test them well beyond our use profile and report how they do.

A note on the MJG Firewires. I just got an email that they are delayed until near the end of the month. As soon as I get them, I'll be testing them.

Randy
 
Ignorant question from a total noob: are there any of the common altimeters where such a thing could be stuffed in place of a lesser component? A minor mod to an Egg*, for instance.

Sent from my Pixel 2 using Rocketry Forum mobile app
 
Guys, a little extra information for you. I've seen discussions regarding the plasma created by the charge going off creating a short across the igniter leads. So, for the heck of it, I loaded up two 2 gram charges and fired them using the same setup and I didn't see any noticeable difference in the current waveforms. The charges were packed into PVC canisters with dog barf packed on top and then sealed with multiple layers of masking tape.

Randy
 
plasma created by the charge going

The ball of still-conductive gases referred to as the "plasma ball" is created by the pyrogen on the match and is NOT reliant on the BP charge.

To test, strip off all the pyrogen from and see if you see a difference between the coated head and a naked head.
 
The ball of still-conductive gases referred to as the "plasma ball" is created by the pyrogen on the match and is NOT reliant on the BP charge.

To test, strip off all the pyrogen from and see if you see a difference between the coated head and a naked head.

I don't know that it's worth the effort considering that the amplitude of the measured current pulses fall right in line with Vbatt / (5 Ohm + Rinitiator).

Randy
 
I don't know that it's worth the effort considering that the amplitude of the measured current pulses fall right in line with Vbatt / (5 Ohm + Rinitiator).

The effect you'd see with your setup would be a ~Vbatt/5 Ohm current for an instance. If your sample rate were high enough, you'd see a current spike when the plasma ball forms. The inductance of the wiring might be enough to suppress the current spike. So, not really a practical thing to worry about.

Without the current limiting resistor, various e-matches (and igniters) will each have a unique current waveshape. A dead short isn't necessarily the best way to put energy into a bridge wire, and it will show in the current flow pattern. The internal impedance of the battery and the wiring will be part of the dynamic signature. Much better to have 1.25 to 1.5 times the all-fire current as a limit (or from a current source), and "gently" burn the bridge wire in 10-30 msecs.
 
Guys, a little extra information for you. I've seen discussions regarding the plasma created by the charge going off creating a short across the igniter leads. So, for the heck of it, I loaded up two 2 gram charges and fired them using the same setup and I didn't see any noticeable difference in the current waveforms. The charges were packed into PVC canisters with dog barf packed on top and then sealed with multiple layers of masking tape.

I've seen that effect in the past, but this was at higher voltages, when playing with a small capacitor bank (2000µF @ 400V). The current became significantly higher than what would have been observed in an ohmic conductor (1-2kA vs. ~400A otherwise). If I remember correctly, the voltage dropped to about 100V before the arc discharge stopped.

I'm wondering when/how the arc discharge can be kept alive at lower voltages by the combustion. In particular, I'm asking myself if precautions should be taken for igniters used in head end ignition of motors.

Reinhard
 
Exactly.

Do you think the conductive effect of the plasma ball could be that short lived, i.e. in the microsecond range?

Randy

Yes, without current limiting, a bridgewire current looks nasty. Spikes are in the 10's of microsecond range.

Here's an old paper from 2000 by the GWhiz altimeter designer: https://publicmissiles.com/IgnitersWhitePaperbyG-Wiz.pdf
Important to notice that this setup was on a lab bench with less resistance & inductance than a launch controller. He used a 0.15 ohm resistor for his current sensing which gives very little current limiting.

I've measured firing currents for higher-voltage NASA/Mil standard initiators many years ago. The response is 90% dependent on the firing circuit and temperature (I had to test down to -40C).
 
I've measured firing currents for higher-voltage NASA/Mil standard initiators many years ago. The response is 90% dependent on the firing circuit and temperature (I had to test down to -40C).

Very interesting, what 'rules of thumb' did you uncover?
 
Ignorant question from a total noob: are there any of the common altimeters where such a thing could be stuffed in place of a lesser component? A minor mod to an Egg*, for instance.

Sent from my Pixel 2 using Rocketry Forum mobile app

[I apologize ahead of time if this isn't non-EE speak...]

You could limit the output current of any Eggtimer altimeter by adjusting the size of the current limiting resistor between the processor and the optoisolator driving the BJT transistor. The optoisolators have a CTR of just about 100%, so with a hfe of about 3000 @ 2A for the transistors, you'd need about a 5K resistor between the processor and the optoisolator to limit your output to 2A (assuming a 2S LiPo). Typically, we use 1K... this gives you a 10A driving capability. If you're interested in doing this, email me and let me know which altimeter you have, and I can tell you which resistors to play with. Note that if you have an "A" version TRS, these resistors are underneath the processor, so you're pretty much resigned to the as-designed configuration unless you feel like desoldering the processor.
 
Very interesting, what 'rules of thumb' did you uncover?

I didn't uncover anything new that wasn't already in the NASA specs for driving a bridgewire. Used a current source to deliver the recommended energy in the correct pulse shape. But, I did verify my firing circuit design for the thermal and vacuum requirements. (It's probably still being used?).

Note: Instead of series resistance as a current sensor, a better way is to use a Hall Effect current sensor. It is isolated, has almost no load on the circuit, and is very fast <100 nsecs. Example here.
 
Referencing Figure 4 in that paper, perhaps I'm not looking far enough out in time. I'll capture some more traces and look for 100-200ms.

Randy
 
[I apologize ahead of time if this isn't non-EE speak...]

You could limit the output current of any Eggtimer altimeter by adjusting the size of the current limiting resistor between the processor and the optoisolator driving the BJT transistor. The optoisolators have a CTR of just about 100%, so with a hfe of about 3000 @ 2A for the transistors, you'd need about a 5K resistor between the processor and the optoisolator to limit your output to 2A (assuming a 2S LiPo). Typically, we use 1K... this gives you a 10A driving capability. If you're interested in doing this, email me and let me know which altimeter you have, and I can tell you which resistors to play with. Note that if you have an "A" version TRS, these resistors are underneath the processor, so you're pretty much resigned to the as-designed configuration unless you feel like desoldering the processor.

The transfer function across an optoisolator is usually very non-linear and dependent on temperature. It's best to drive it into saturation for a firing circuit and do the current limiting on the output side.
 
The transfer function across an optoisolator is usually very non-linear and dependent on temperature. It's best to drive it into saturation for a firing circuit and do the current limiting on the output side.

Agreed, but I was answering the question whether or not you COULD do it by subbing a component on the board. I've recommended adding external current limiting resistors for certain applications (HPR airstart igniters, for example) for some time. For most applications, it's not necessary, as long as you check the resistance of the ematch before you fly them to make sure they're not dead-shorted out of the box.
 
The safest thing to do is to use a 6W+ resistor, but not wire-wound (too much inductance limits the shape of the firing current). Probably $2 each.
Unless you are getting way up in frequency then the inductance is of no consequence. It is of no consequence in this application.

he cheapest and "good enough" choice is a 1W non-precision metal film resistor, either axial lead or the "MELF" cylinder SMT package. The reason for using non-precision is that the process of laser trimming the resistor film leaves a weak spot that doesn't hold up to pulse currents. 25 cents. Use 5.1 ohm because it's a standard 5% value.
Metal film resistors can actually act as fuses, especially the surface mount ones, depending on how they are manufactured. The MELF resistors you mention (Metal Ended Lead-less Form) from Phillips are among the best around for pulse/surge rating due to their construction. The leaded version are generally better because their power rating starts off a bit higher than most surface mount ones anyway.
 
Referencing Figure 4 in that paper, perhaps I'm not looking far enough out in time. I'll capture some more traces and look for 100-200ms.

Well, that did the trick. I fired off a few more initiators last night, and sampled the current for 1 second at a 10MHz sampling rate. I was getting the standard ~12ms current pulse as shown in my first post (let's call the beginning of that pulse t=0), and then at about 70ms, the current begins to increase linearly again and peaks at around 500ms, then exponentially decays away and is almost back to zero at the end of the 1 second interval.

What's interesting is that the peak current due to the plasma ball is only around 50mA which means the impedance of the plasma only dropped to about 150 ohms. The plasma created by the pyrogen firing off seems to behave just like the plasma produced by an electrically induced arc, i.e. it exhibits a non-linear (negative) impedance. The 5 ohms of source impedance stops the runaway effect you get with a negative impedance where the more current it draws, the lower the arc impedance gets and the more current it draws... etc. The voltage across the plasma never collapses down to create a short. So, a simple 5 ohm resistor works well in limiting current draw in both the initiator's resistive element and the plasma ball which follows.

BTW, the MJG Firewires should be shipping today so I should be able to test them soon.

Randy
 
Unless you are getting way up in frequency then the inductance is of no consequence. It is of no consequence in this application.

We're not talking about steady-state impedance in the frequency domain. The inductance of a wirewound resistor will limit the di/dt of the igniter current. It will return this stored energy back to the circuit with possible higher-voltage overshoots&undershoots. The way that energy is transferred into the e-match will change. Simple view for the EE-challenged:

resistor-wirewound-risetime-01.png


Metal film resistors can actually act as fuses, especially the surface mount ones, depending on how they are manufactured. The MELF resistors you mention (Metal Ended Lead-less Form) from Phillips are among the best around for pulse/surge rating due to their construction. The leaded version are generally better because their power rating starts off a bit higher than most surface mount ones anyway.

That is why I mentioned MELF smt's early on in this thread. And non-precision (5.1 ohm 5% instead of 5.0 ohm 1%) because the process of laser trimming the film reduces the surge current capability of the resistor.

Surface mount components are generally better for bulk heat dissipation due lower thermal resistance and conduction into the circuit board. Leaded components rely and larger surface area and convective cooling. For this applications, the difference would be important if an onboard limit resistor needs to handle a shorted output for any length of time.
 
We're not talking about steady-state impedance in the frequency domain. The inductance of a wirewound resistor will limit the di/dt of the igniter current. It will return this stored energy back to the circuit with possible higher-voltage overshoots&undershoots. The way that energy is transferred into the e-match will change. Simple view for the EE-challenged:

Thanks for the reply John. I still stand by my comments regarding the inductance. The inductance is small (unless you use particularly high-value WW resistors) and in this application is of no consequence. You are right theoretically, but the real world effect is insignificant and of no consequence. Fit for purpose whichever you choose.

It is not linear rise in current anyway. Other second-order effects kick in and distort the waveform. Resistance of the filament increases as the current goes up of course, and with the source impedance of a battery being quite complicated the rise is highly unlikely to be the theoretical linear ramp anyway.

If you are worried about inductance don't forget to add the inductance of the wiring to the igniter which can be significant.

Some things are worth dealing with, some not. Fit for purpose is the aim, not perfection. If you want to get really keen you could design an experiment with multiple igniter types, multiple resistor types, and a bespoke drive circuit for a repeatable trigger supply and run the statistics on a stat-significant quantity. The efforts would probably show more variability in the igniters than would be due to the way they were driven. Remember we are not putting significant millihenrys of inductance in series here.

Happy to be proven wrong if you have some data :)

That is why I mentioned MELF smt's early on in this thread. And non-precision (5.1 ohm 5% instead of 5.0 ohm 1%) because the process of laser trimming the film reduces the surge current capability of the resistor.

Depends entirely on the resistor. I won't comment on other brands of MELF as I have not used them. The Philips ones are spectacular with their pulse handling. The laser trimming method does not cut across the resistive element, but helically in line with it. This results in a superior product. There is no difference between precision and non-precision in their range as far as surge capability. Interestingly their noise figures are about twice as good as other types of SMT resistors. That's why I use them in the detectors of our spectrometers in critical locations.

Surface mount components are generally better for bulk heat dissipation due lower thermal resistance and conduction into the circuit board. Leaded components rely and larger surface area and convective cooling. For this applications, the difference would be important if an onboard limit resistor needs to handle a shorted output for any length of time.

Based on my experience if a SMT resistor, particularly 1206 and above, is used at its rated power the PCB will end up discolored over time. This can be mitigated by larger copper pads and coupling the heat to the other side of the board or to ground planes. Personally, if I were looking at that I would use a resistor rated higher than required to avoid PCB damage if long-term exposure to that power is required, or use a through-hole part. For a relatively short pulse, up to a second or so, the SMT would be fine.



BTW I am not EE challenged. I do this for a day job :wink:
 
BTW I am not EE challenged. I do this for a day job :wink:

I didn't think you were EE challenged, but the audience here is. The idea is not to over-analyze Randy's test results, but to help people understand the limitations of testing e-matches and bridgewires. And ultimately, to improve how we fire them (for reliability across types, especially the cheap ones, and to be nice to the firing circuit in the altimeter). You may have missed his original intent. Lecturing me is taking away from the focus of the thread. :)

The inductance of the wiring in a launch controller on the ground is a different parasitic circuit than the one in the altimeter + altimeter wiring. The inductance added from a 5-ohm wirewound resistor is the majority of the inductance, considering that the altimeter has a large capacitor, typically, ahead of the firing circuit and not much on the board up to the terminal block. The type of resistor does make a measurable difference. I have measured it and have designed around it.

Regarding your experience with PCB damage from heat from a SMT resistor, that is not normal. It's a design flaw. ;) Now, consider how would you remove the heat from a danging, leaded component with no conduction and only convection in an enclosed space... or in the vacuum of space. I have significant experience doing both of those things for industrial, mil and aerospace applications.

Now back to your regularly scheduled thread, awaiting Randy's new results.
 
Back
Top