Talk to me about switches and redundancy

[Wallace]

Twist & Tape for 17yrs....never, never failed. [when done correctly!!!!!]
I have had 3 switch failures......Key switch-screw switch-shurter switch.
Over the years figured I would try them.

T&T good for around mach2-2.3 [at least for me tape stayed put]
For everyday sport flying I like T&T for simplicity and reliability. [Tons of my friends use it]

For Xtreme flights I use the combat robot switch [fingertech robotics] which I believe I may have introduced them here, many years ago.

There is a bit of art to T-T...not done correctly you can bounce power or worse loose it during flight. YOU MUST cross wires first, then squeeze them while twisting BOTH around each other...like so: [2 altimeters]

View attachment 346378 Keeps the power positively contacted during twist...

...then tape to outside of payload/switchband so it can be disconnected if need be. My tape has held to around M-2.3 at stated. Depends on quality/stick-um of electrical tape used.
Note I start the twist with insulated portion also.

The incorrect way happens when one wire gets mostly twisted around the other. As improbable as it seems, done this way is NOT a good connection & can "bounce" or loose/regain contact several times during flight, causing many types of altimeters to "re-set" during flight...not a good thing..."crash" or stay disconnected, though looks like a good contact.

Scotch super 33 is the only tape..

 

[K'Tesh]

I'm planning on using a pull pin for arming mine, when I get that far. I've already scored the "Remove Before Flight" ribbon from eRockets.biz.

As for redundancy for the flight I'm planning on using two JLCR's secured to each other. for the main's release, and two altimeters (brand(s) yet to be determined) for the separation event.

Nice thing is that with the AV Bays from Binder, since they're identical, and not painted the airframe color, I'll be able to just transfer them from one rocket to the other. Though I'll likely build the 2nd one as a spare for the first (minus the electronics).

 

[kidrobo]

As for redundancy for the flight I'm planning on using two JLCR's secured to each other. for the main's release, and two altimeters (brand(s) yet to be determined) for the separation event.

This is wild, but I guess it makes sense. I've never heard of a flight with two chute releases plugged into each other. I dig it.

What size rocket and why not set it up for a traditional dual deploy? That feels like a lot of electronics.

 

[Cl(VII)]

I have both the single and double pull pin switches from Lab Rat. Very nicely made, even include quality mounting hardware all very reasonably priced. I also use an Eggtimer Quantum/Eggfinder mini combo for my nose cone av bay which technically requires no physical switch...arms over wi fi

Glad you like them. I'm printing another batch right now.

 

[Wallace]

they're fantastic. thank you, and keep up th good work...

 

[OverTheTop]

I have tried a lot of switches over the years. Magnetic, screw, pull pin, rotary, slide, WiFi. I now use pull-pins exclusively. They are so easy to use, not easy to forget if you have RBF tags on them, and importantly for me ZERO CURRENT DRAW WHEN OFF. That is very comforting when keeping the mass of the avionics down by using small batteries. Power consumption is particularly important when you prep well ahead of time.

Don't forget that you can put a lot of switches along a single pin.

I tend to use Allen-keys (screwdriver style) in 2 and 2.5mm as the pins. That's just because there were a heap getting thrown out at work. The RBF tag gets attached to the handle.



You don't even need to use a ladder on tall rockets. Just tie a long piece of string on each pin .

Listen to the opinions and go with what you think suits you .

 

[boatgeek]

I'm a fan of the Missileworks screw switches. Schurter switches were too fragile for me, especially when they are airframe mounted and not sled mounted. It was too easy to accidentally scrape the connectors off the switch body. I do use twist and tape for connecting ejection charges to wires coming from the AV bay, and I had one of those fail at my last launch. I chalk that up to user error, not leaving enough bare wire to twist together.

 

[Buckeye]

Ok, building my first AV bay.
Question #1. What type of switch? I am looking at either Featherlight magnetic switches or LabRats double pull pin switch kit. Thoughts?
Question #2. How many do redundant altimeters? I am wanting to go all out, and for the cost (~$50) it seems cheap enough to increase my chances of success. Do those that have redundant systems, do you use different brands? I have heard that if one fails, the chance of both systems will fail if you use the same brand?
Thanks for the feedback.

OK, while everbody is hung up on switches, I'll answer your second question.

I am adamantly anti-redundant altimeters. This is your first AVbay, don't over -complicate it. One is enough. I have never had a barometric altimeter (properly configured) fail to do its job. Nor can I recall many (any?) verified cases in this forum. Accelerometer altimeters can be wonky, though, and should not be used for deployments, in my experience.

Users are far more likely to screw up harnesses, BP charges, wiring, and batteries than to have an altimeter fail. Two posts upthread had flights which were "ballistic" and "lawn darts", ie, no deployment, despite redundant recovery activation.

Two altimeters of the same brand are not like a "bad batch"of Estes E9-6. I think the QC is better than that. The chances of two Brand X altimeters not working is miniscule. Plus, you checked the functionality of the altimeter using the built-in test features, right?

If I feel the need for backup, I use a 4-event altimeter (Raven) or use 2 ematches per charge.

 

[boatgeek]

I forgot that my post that got eaten in the Great Platform Change also had a short bit about redundant altimeters. I'm a huge fan of redundancy on the electronics side. As mentioned above, I had a recent flight where one of the charge's twist and tape came off. That could have been ugly if there hadn't been another altimeter (or another ematch per charge). I have seen other flights where one altimeter lost power or otherwise failed to work properly and the backup saved the flight. Granted, these were probably user error in setting up the AV bay. however, they happen often enough that I think it's totally worth having belt and suspenders for large projects. For small projects, I'll have motor eject as a backup to the altimeter. Simple 2-channel altimeters (Eggtimer Quark/Missileworks RRC2+ eg) are quite a bit cheaper than a Raven and dead simple to program.

If you do redundant altimeters, I slightly prefer having the same manufacturer, to reduce the chance that differences in programming will lead to both apogee charges going off at once. It's not a huge deal, but if you're starting an altimeter library now, it's something to think about. FWIW, YMMV

 

[OverTheTop]

I always run redundant altimeters, especially for main deployment. Way too dangerous to have something coming in ballistic for my liking.

 

[plugger]

I always run redundant altimeters, especially for main deployment. Way too dangerous to have something coming in ballistic for my liking.

Don't you mean apogee deployment, not main OTT?

For me anything 3 inch diameter or larger (including my L3) I run two fully redundant altimeter configurations. That's two altimeters, two switches, and two LiPos. If I'm using a cable cutter I use 2 on the single zip tie for redundancy there. For anything smaller (excluding MD birds) I use a single altimeter and a unmodified motor ejection as an backup apogee event. I, like you, don't like the idea of my rockets coming in ballistic. Have a "dumb" motor eject backup is normally good enough for smaller rockets, especially given my penchant for only using composites in builds. I suspect most of my rockets can tumble without a main deployment and require little if any repair work.

 

[OverTheTop]

Don't you mean apogee deployment, not main OTT?

No. I meant what I wrote.

I normally run single-ended dual-deploy. I blow the NC at apogee with avionics in the NC. If this fails then the dual altimeters (completely redundant, with separate LiPos for each eMatch) for the main have two shots available at pushing the main out and the NC off at the same time. Main altitude is set for 2000'. The extra height is to make sure I get a good GPS packet if something is sub-optimal. It also means the delay before the backup charge fires while it is still well in the air if the first main charge has failed to do the required job.

 

[plugger]

No. I meant what I wrote.

Well count me as confused then! It's the apogee event that ensures a non-ballistic return, not main deployment, right? So if you're running dual altimeters you'd have two apogee events and two main events, correct? While I think I understand how your main events provide triple/quad redundancy if you're coming in ballistic and separate the rocket at 2000' with a main deploy that zipper would be pretty rough!

 

[OverTheTop]

Apogee is set up as two charges but only one altimeter. Mains are fully independent.

Only had one flight with non-nominal recovery. About a year back I flew the Apache on an M840 to 17k'. As she rolled over the top the NC blew as per normal, but the harness then slipstreamed behind the airframe, so the usual 100fps turned into nearly 300. Altimeters hatched the main as usual and pulled the whole thing up quite quickly. Busted a transition and cracked my NC. NC was repaired and the transition reprinted. Easy fix
Details here: https://forum.ausrocketry.com/viewtopic.php?f=6&t=5940

 

[Tonimus]

I am a big fan of the Eggtimer wifi switches. Things work great, have continuity hookup if you want it, and draw remarkably little current for what they do. They draw about 85mA.

 

[Wayco]

I replaced the rotary switches in my Ultimate Darkstar with Eggtimer WiFi switches, now I don't need to bring a ladder to the launch pad. They work great for avbays that have the space. When you want a redundant system in a 54mm MD rocket though, magnetic switches can't be beat.

Low quiescent current draw of under 3 micro-Amps. This switch can be left in standby for over a week without significantly draining even a tiny 135 mAhr lithium polymer battery.

 

[Bobfly]

MAC Performance sells a nice quality push button switch. They are easy to install. The switch has a large push button head so alignment with a vent hole isn’t super critical. The head of the switch can be depressed anywhere on its surface area.

I use key switches with remove before Flight red flags on my larger projects. You can order the switches where the keys can only be removed when the circuit is armed.

 

[roytyson]

I wanted to thank everyone for the feedback. What I discovered was asking this question is like asking what kind of glue or paint to use
I ordered my Eggtimer Quantum, and I decided on the wifi switches. Connor is assembling them for me, as I don't have the time to teach myself at this point in my life.

 

[Nathan]

. . .As for redundancy for the flight I'm planning on using two JLCR's secured to each other. for the main's release, and two altimeters (brand(s) yet to be determined) for the separation event. . .

I have ground tested using two JLCR's for redundancy. I found that there seems to be a significant risk of the chute becoming entangled with two tethered JLCR's if one of the JLCR's fails to open. You end up with the two JLCR's attached to each other with a rubber band and both attached to tethers which creates a loop that can the chute can get hung up on. So I never flew that configuration.

 

[Dad Man Walking]

OK, while everbody is hung up on switches, I'll answer your second question.

I am adamantly anti-redundant altimeters. This is your first AVbay, don't over -complicate it. One is enough. I have never had a barometric altimeter (properly configured) fail to do its job. Nor can I recall many (any?) verified cases in this forum. Accelerometer altimeters can be wonky, though, and should not be used for deployments, in my experience.

Users are far more likely to screw up harnesses, BP charges, wiring, and batteries than to have an altimeter fail. Two posts upthread had flights which were "ballistic" and "lawn darts", ie, no deployment, despite redundant recovery activation.

Two altimeters of the same brand are not like a "bad batch"of Estes E9-6. I think the QC is better than that. The chances of two Brand X altimeters not working is miniscule. Plus, you checked the functionality of the altimeter using the built-in test features, right?

If I feel the need for backup, I use a 4-event altimeter (Raven) or use 2 ematches per charge.

I'll take the bait...

I've seen this idea posted before, not sure if was your post or another flier with similar views. I hold the opposite opinion (respectfully). I agree that commercial altimeters are robust and not likely to be the point of failure. But they are part of a system that includes the battery, wiring, a switch, the e-matches, and and (typically) a BP charge. In this system, the altimeter is the least likely thing to fail. But a whole lot of us have seen failures in just about every other part of the system. I've seen:

1. batteries that were not secured properly and broke free during the flight (probably at the apogee deployment event in the case I'm referring to)
2. ematches that tested fine on the bench immediately prior to flight, gave happy beeps on the pad, but failed to light. Post-mortem on the igniter was that it looked brand new but did not have continuity, so it broke during the flight, maybe when the altimeter hit it with juice
3. switches that weren't suitable for the loads we subject them too...worked fine for a while but failed at a lousy time
4. a primary charge that failed to separate the rocket at apogee, and there was no larger backup charge. A main coming out during a ballistic descent makes an unforgettable sound...it's as if fabric could scream.
5. e-match wires that were torn away from the terminals (probably at boost) because the recovery gear shifted and pulled a wire with it

That's five failure modes I can think of off the top of my head, I know I've seen more. (And, for the record, number 2 and number 5 happened to me, I was a witness to the others.)

None of these have anything do with the reliability of the altimeter. Numbers 1, 3, and 5 are 100% caused by the builder/flier's errors or lack of experience, and are preventable but still happen to people. Number 4 could have been because the charge was not properly prepared, or maybe because the coupler got jammed in the body tube for some reason. Number 2 was some sort of product defect that was not detectable even though we prepped by-the-book...measured resistance at the bench, confirmed continuity of the complete circuit at the bench and again on the pad.

So my experience is that bad stuff happens, sometimes to people who are doing all of the right things. I now use two ematches in every charge, and avoid flight profiles that don't have a backup event for apogee, at a minimum. I only have a couple of rockets that are DD but not redundant, and I keep the motor selection to ones that have a delay long enough to serve as the backup to the altimeter and/or low enough to track visually the entire flight, so that if something does go wrong we can see it coming long before it crashes into (hopefully) the ground.

Safety first (and second, and third). My two cents, anyway.

 

[mpitfield]

Redundant systems have saved me twice now. For the cost of a Quark, it's hard not to justify it.

 

[OverTheTop]

OK, while everbody is hung up on switches, I'll answer your second question.

I am adamantly anti-redundant altimeters. This is your first AVbay, don't over -complicate it. One is enough. I have never had a barometric altimeter (properly configured) fail to do its job. Nor can I recall many (any?) verified cases in this forum. Accelerometer altimeters can be wonky, though, and should not be used for deployments, in my experience.

Users are far more likely to screw up harnesses, BP charges, wiring, and batteries than to have an altimeter fail. Two posts upthread had flights which were "ballistic" and "lawn darts", ie, no deployment, despite redundant recovery activation.

Two altimeters of the same brand are not like a "bad batch"of Estes E9-6. I think the QC is better than that. The chances of two Brand X altimeters not working is miniscule. Plus, you checked the functionality of the altimeter using the built-in test features, right?

If I feel the need for backup, I use a 4-event altimeter (Raven) or use 2 ematches per charge.

I will stick with redundant altimeters. I know it has saved me once.

As for
"Two altimeters of the same brand are not like a "bad batch"of Estes E9-6. I think the QC is better than that. The chances of two Brand X altimeters not working is miniscule.",
it is the same firmware in each altimeter (assuming two). If there is a bug in the firmware then both altimeters are likely to act the same. Having had some fairly serious software testing work in the past I can say that for a program to be "bug free" is very hard to test for. The units will likely behave very similar, or misbehave very similar. The Boeing 787 has triple-redundant Flight Control System computers. All three behave similarly and if one misbehaves it it voted out by the other two. What if all three misbehave together? I have read that there is a known issue with them that requires them to be rebooted every small number of days (three I think). If not then there is a chance that all three will fail in the same way at the same time. Such is the joys of firmware...

 

[cwbullet]

Moved thread to electronics

 

[Buckeye]

I'll take the bait...

I've seen this idea posted before, not sure if was your post or another flier with similar views. I hold the opposite opinion (respectfully). I agree that commercial altimeters are robust and not likely to be the point of failure. But they are part of a system that includes the battery, wiring, a switch, the e-matches, and and (typically) a BP charge. In this system, the altimeter is the least likely thing to fail. But a whole lot of us have seen failures in just about every other part of the system. I've seen:

Yes, I suffered two or three of those problems as well as I experimented with different DD techniques. They occurred in cardboard-constructed MPR and low-end HPR, so the damage/cost was not very significant. That's where I honed my skills. I do strive for motor backup, as long as the delay is at least 3 seconds greater than the optimal apogee time.

I guess my main point is I fixed those problems instead of sloppily doubling them and hoping one system works. I did a lot of ground testing, altimeter testing, and "reverse" dual deploy. I am to the point where I am confident in my techniques that I am 99% sure they will work. Cost of another altimeter is not an issue. The time constructing and prepping dual systems is not worth it to me.

 

[Voyager1]

Yes, they are absolutely necessary. And no, plugging the battery in at the pad would not meet the requirement, even if this was possible.

A typical AV bay lives in a coupler tube joining two sections of airframe tubing and is bolted shut. And the batteries are usually not just plugged in on high power rockets, they are secured with additional reinforcements like cable ties, electrical tape, brackets or compartment lids to keep them from bouncing out of the battery box under extreme loads. All of this prep is done at the table...all you do at the pad is arm the electronics with some sort of switch.

And you need to think about the disarming sequence if you needed to scrub the flight. If you have to do anything more that open a switch, you are working on an armed rocket.

Yep!

 

[OverTheTop]

I am to the point where I am confident in my techniques that I am 99% sure they will work. Cost of another altimeter is not an issue. The time constructing and prepping dual systems is not worth it to me.

If you have thought of all the possible off-nominal situations and have them covered this philosophy is probably ok. Design for off-nominal and design so it will not hit the ground at speed. Trashing a rocket at altitude and having it hit the ground in a controlled manner is an acceptable outcome IMHO.

 

[Andrew_ASC]

My favorate switch is the lowly Aerocon PCB switch. These are glued onto the sled is some manner and aligned with a hole in the air frame. The wires are soldered on and then potted in epoxy. There is literally nothing that can fail.

https://aeroconsystems.com/cart/switches/pcb-screw-switch/

I like the magnetic switches, but my experience with them is that after a time, they fail.

Jim

Thank you!

 

[Andrew_ASC]

How are you guys configuring the backup altimeter if it has ematches for main/Apogee also?

Let’s say I have a Telemini configured with main at 900ft and Apogee is auto detect with a charge each as a primary altimeter.

Now I want a backup altimeter such as a stratologger? Do I set the stratologger to an predicted Apogee altitude plus a Apogee time delay lockout for a second to prevent the primary altimeter from overpressuring it if primary works as nominal.

Then can I set the backup main at 800ft?

Project is 54mm MD, first time using redundant altimeters. Trying to grasp the programming logic here in mind first so during primary altimeter functioning and firing charges that the charge event won’t cause a altimeter conflict on the secondary backup altimeter also causing it to fire (if that’s a possibility).

Figured this was something I should ask.

 

[roytyson]

How I do it is: I set up my primary altimeter as I want. Then my secondary altimeter I set at apogee + .5 seconds. Then set the main at 50’ less than the primary altimeter.

 

[OZRoc]

How are you guys configuring the backup altimeter if it has ematches for main/Apogee also?

Let’s say I have a Telemini configured with main at 900ft and Apogee is auto detect with a charge each as a primary altimeter.

Now I want a backup altimeter such as a stratologger? Do I set the stratologger to an predicted Apogee altitude plus a Apogee time delay lockout for a second to prevent the primary altimeter from overpressuring it if primary works as nominal.

Then can I set the backup main at 800ft?

Project is 54mm MD, first time using redundant altimeters. Trying to grasp the programming logic here in mind first so during primary altimeter functioning and firing charges that the charge event won’t cause a altimeter conflict on the secondary backup altimeter also causing it to fire (if that’s a possibility).

Figured this was something I should ask.

I'm just checking here but are you using separate charges for each altimeter? You'll have a total of 4 charges (1 x drogue and 1 x main) for each altimeter?
Otherwise the idea of redundancy doesn't really apply.
Cheers