Bottle rocket technical notes

I promised to write up some of the rocket details for a few people, so here goes…

I’ll start with some pictures of the actual launches, then talk about the technical detail.


Ready for pressurisation


Drogue parachute deployed perfectly, but no main

Probably a good job the main parachute didn't deploy, else we'd have ended up in the trees rather than the brambles. Also learnt that if you stand around pondering for long enough there's bound to be a sillier idiot than you willing to go climbing through the brambles.

Oh, and we did have some minor mishaps too...

Launch video (which I seem unable to embed)

Rocket body pressure chambers

The failed 'Anna Chapman' rocket

After the impressive sight, but inevitable failure of the Anna Chapman last year (launch video) I fancied building something similar but hopefully more successful. Last year we did find that 500ml bottles seemed to have a lot more bang for their buck than 2L bottles but this didn’t deter me too much. I’m still not really sure why it was, but I imagine it had something to do with cross sectional area versus volume. I surmised a longer rocket might well work better.

Inspiration from a rather impressive looking Air Command rocket using asymmetrically spliced bottles.

Air Command Water Rockets is an excellent resource, with some amazing launch videos which I wanted to try and at least go some way towards replicating. After some research I followed their Asymmetric Splice #5 Tutorial to build a number of spliced double-ended bottle sections. I shan’t repeat their tutorial, but I ended up using different glue as PL Premium is all but impossible to get hold of in the UK. I used Sikaflex EBT from Screwfix to seal the splice and a 2:3 Gorilla Glue:Polyfilla mix for the sleeving (more details on this Water Rocket Forum thread with links to Flickr pictures). Needless to say my garage is now covered in Polyfilla and glue; it gets everywhere.

IMPORTANT TIP: Wear gloves when handling Gorilla Glue; and replace said gloves if they break. Even the small amount I got on my thumb through a hole in my glove took over a week to wear off. It’s evil stuff.

The benefit of making a series of pressurisable ‘units’ rather than one long one is that they can be individually static pressure tested and the failures thrown away without affecting the rest. Having said that I am tempted to try and make a three or four bottle long ‘unit’ at some point.

Connected spliced bottles ready for launch

Bottle joins

The plan was to then join three or four of these together into a longer rocket using Tornado tubes from Hawkin’s Bazaar; but at £3.50 each they do work out rather expensive and the Northampton branch doesn’t seem to stock them.

Foreground left: Joined bottle connectors with various sized holes to allow air/water through. Foreground right: Gardena nozzles made to fit bottle top. Rear: Failed bottle top connectors.

From a thread on the Water Rocket Forum I came across a bright idea that Dan Brooker from British Water Rockets had to fuse two bottle tops into a coupling. He has made a demonstration video of how to do this. Despite his lack of testing at that stage I thought it worth a go. My success rate was less than one in three and I began to run out of bottle tops, but I found that some of my attempts withstood a static pressure test to 100psi. I’m not sure how much higher they’ll go, but Dan has reported some failures since. YMMV.

Talking of failures… when carrying out static pressure tests be wary of how much water leaks and air is let in – I had an inadvertent launch at 10 o’clock at night into next door but one’s garden.


My old Clarke cable tie based launcher

Martin's Clarke cable tie based launcher, with 21.5mm launch tube

Last year my Clark cable tie design based launcher sealed only with the assistance of PTFE tape on the outside of the bottles. Martin had used a carefully sized piece of pipe to get an internal seal, which seemed less hassle and allowed for an internal launch tube. So, I planned to use the internal seal and launch tube for the boosters, more on that in a minute.

New launcher with optional boosters (annotated)

Having read about possible options on Air Command Rockets, I decided to have a go at a Gardena (Hozelock) launch mechanism. Reasons were two-fold: a. it might seal better; b. the smaller nozzle may sustain flight longer (albeit at a lower force, but that comes in handy when using boosters). I could always revert to the old launcher if all else failed.


A safe distance from the launcher - hosepipe and string allows remote compression and launch

Oh, and being the sensible one, I elected to put a length garden hosepipe between the launcher and the pump. This reduced the risk of personal injury in the event of catastrophic failure and also kept the operator somewhat drier.

Others may not have thought ahead so much...

Martin getting a tad soggy. Note: Some of the photos I've used, including this one are from one of the 2010 Ballistics events


Air Command Water Rockets has a nice tutorial and videos about preparing a Gardena nozzle, so I shan’t dwell on that much. As they note, I had trouble finding a nozzle with a seal, so tried using some Sikaflex as a seal with mixed results. Ultimately, it was the nozzle that limited launch pressures to about 80psi, so this is where more work is needed to get a good seal between the Gardena nozzle and the bottle top.

Rather than the direct string pull I sought some mechanical advantage from a level arrangement for the launch release. In use we found that it still needed quite a tug to release and on one occasion the string snapped. Again, this needs more work.


As if I’d not created enough trouble for myself already, I thought it would be cool to have more than one stage. Direct staging mechanisms seemed a little complex for a first go at anything more than the plain bottles we launched last year, so boosters seemed a sensible bet. Again, Air Command Water Rockets provided much of the inspiration with their Drop Away Boosters and Cluster Launcher pages.

The principle behind the boosters was pretty simple, but also very critical with a number of single points of failure (no manned versions of this bottle rocket design). The boosters were set up such that they would produce more force than the main rocket until exhausted at which point they passively drop behind the main stage. That’s the theory at least, my design wasn’t quite as sophisticated or refined as Air Command’s. The pictures below illustrate the coat hanger hook and pen cylinder receptacles making up the mechanism. They didn’t operate particularly smoothly, though maybe a neater effort would have better results – I was pushed for time.

Booster with hooks

Hook receptacles on main body bottle

Overall, the boosters didn’t work out too well. Even with PTFE tape around the booster launch tubes it was difficult to get a good, consistent seal on each of the three bottles; invariably, one let go and began to leak under pressure. And the minor mishap illustrated in the first series of pictures at the top of this post was due to one of the hooks detaching from its bottle.

As well launch efficiency, the reason for the launch tubes was to allow the compression line to remain dry/ This would allow the pressure to equalise between the boosters. However, looking at the failed launch with hindsight, we missed out a non-return valve. This meant that water from the main rocket filled the booster compression lines and prevented the pressure from equalising. I don’t think it was causal to the catastrophic failure, but it can’t have helped.


I went a little overboard on the parachute. In the past I’ve found charity clothing bags ideal for building parachutes, but while light, they aren’t particularly resiliant and the my choice of tether line (dental floss!) got tangled easily. So I splashed out on some Ripstop Nylon from Ebay and got sewing.

Leo Singleton‘s Bottle Rocket Handbook (PDF) had a nice section on parachute design. I chose to build an elliptical design, calculated using his Elliptical Parachute Calculator (36KB ZIP)

My first parachute took the defaults from the tool and ended up a rather too large 1M diameter, still it would be a nice gentle float down. As I didn’t think I’d be able to deploy this on its own, I built a drogue of 40cm.

Drogue (40cm) and main (1m) parachutes

Parachute mechanism

Again, drawing on ideas from Air Command Rockets I built a Parachute Side Deployment Mechanism, with their elastic band release modification. I didn’t have time to acquire a Servo Timer II nor much faith in guessing the timing accurately so decided to use a 35MHz radio control receiver to control the server.

Activated parachute mechanism

Parachute mechanism atop the rocket. Upper half of the top bottle contains the drogue chute and R/C release mechanism; the lower half the main chute which failed to deploy

As the main chute was too large to fit within the elastic deployment pocket, I bargained that the drogue would lift the whole mechanism from the rocket top and deploy the main. On both parachute-recovery launches the drogue deployed perfectly, but the main chute failed. With hindsight, I should have attached the drogue directly to the main and had a hole or flap for it to deploy through.

Next time, I think I’ll build a second deployment mechansim, attach a medium sized (70cm?) parachute and trigger it shortly after apogee with an altimeter. The existing R/C drogue parachute can act as a back up.

I’ve written enough for now, feel free to ask any further questions in the comments below, or in person.

Ballistics Day pics

Despite some initially poor weather, it dried out, brightened up, and the wind died down to allow us to launch some stuff!  Check the flickr feed here.  For those interested in the telemetry data from the payload we sent up, it looks like this.


Quite a short run as the rocket didn’t take long going up or down.  Acceleration data isn’t terribly readable – the peak acceleration is obviously off the +- 4g scale of the accelerometer.

One bad data point on the pressure causes the smoothed graph to have a small dip in it, but otherwise the data appears consistent (if unverified).

Temperature data isn’t useful, the sensor is cooling from a previous high of 35C from being wedged into the rocket nose cone.




Mat’s plane made a couple of trips up and down and around.  The temperature data again doesn’t appear to useful, as the sensor is much less responsive than the barometric one (they’re in the same package).  Currently it’s not used for altitude calculations.

Ballistics Day 2010

29th August 2010!

That’s when the Racecourse Park will echo to the bizarre sounds of our tennis-ball lofting machines. Meet at 2.30pm, and we’ll start cooking at 4pm. There will also be chemical and water rockets again, with material for those wanting to build water rockets.

There’s no specific plan for specific categories of launch. If anyone requires this, leave a comment here, and one will be written.

Events for August 2010

As well as the standard pub meet this month (Tuesday 17th August) NortHACKton is holding TWO other events. The dates for these events are up for selection by doodle poll until the 27th July where I shall pick the highest-voted days out of a hat of some kind, moderated by any mitigating factors that become apparent (site availability etc). Vote now!

Hack Evening Workshop

Our first workshop will be a series of short informal talks. These will not be held in the public area of a pub, but either a back-room or a community centre. Contact the mailing list or me directly if you’d like to give a talk. There will be time for questions and networking after each talk, and a projector for those with slides.

Ballistics Day

The previous rocketry day was a success. So we’re going to repeat it, but with the addition of trebuchets, or other catapults or methods of propulsion. Bring tennis balls, and then show how far you can propel them. Again, there will be a barbecue, and some raw materials for hacking on the day.

July 2010 meet write-up

A good meet today. We kicked off with some reminiscing about the rocketry day and some of the more entertaining lessons learnt (such as not launching chemical rockets into the wind). After this everyone agreed that we should do it again. Rather than merely repeat the rocketry day, it will instead be extended to also include siege weapons capable of throwing tennis balls, such as catapults, slingshots, trebuchets and the like. More information and a doodle poll soon.

Second event we would like to hold is an evening of talks, with an eye for holding this regularly. Location submissions are accepted for somewhere cheap and quiet. We’ve got a projector already, if a bit old. Again, I’ll release poll information tomorrow. Suggestions are also welcome on the list or as a comment here for places to advertise either or both of these events.

Finally, I am very pleased to announce the donation of a number of electronics boards. I’ll write this up in a separate post, as it’ll take a while to enumerate them and provide links to the right websites. But we now have toys that will go on the wiki and people can book them out and do what they like with them.