EEVblog #339 – Cypres Parachute AAD TeardownPosted on August 22nd, 2012 19 comments
Forum Topic HERE
A teardown of the Airpres Cypres reserve parachute Automatic Activation Device
I think the pressure sensor must be an absolute type, not gage. Gage type measures the applied pressure relative to ambient atmosphere. When you WANT to measure the ambient atmospheric pressure (like an altimeter would), you need an absolute type sensor, which are available with either a vacuum reference or a standard 1 atm. reference (commonly used for barometers).
How about firing that pyrotechnic line cutter for us?
I have one mounted on my rig. The adjustment is to fire if I pass 1500 meters above ground level in terminal velocity (free-fall) or close to it.
I´m not sure on how the pressure sensor can actually work inside such an hermetic casing.
I checked mine and I don´t see any venting ports. It´s all sealed. Maybe the entire case acts as a sensor since it can deform under pressure change. With a well calibrated ADC you can achieve great sensitivity levels.
But it works.. Mine never deployed.. But I saw a few deployments… They are the last resort in case the skydiver becomes incapacitated and can´t open the main nor the reserve chute. It was invented after the death of a very famous skydiver that became unconscious in the middle of a jump due to fatigue. He was carrying very heavy filming equipment and did many jumps that day.
Many people were save by this device since it was made available.
If your rig is not equipped with this device.. DON`T JUMP.. Simple as that.
Concerning the firing mosfets, they are probably arranged in two paralell fetss on the – side to GND and two paralell FETS on the + side to +batt.
That way, any single fet of the four can fail, open or short, the device is still reliable.
It’s the same kind of circuit in airbags, it should have also diagnostic capabilities (at least testing the blast cap resistance, perhaps even the mosfets)
The cap on the btt is probably to ensure enough current spike capability for the blast cap.
These things cost like $1,500 USD each!
Most reading this Blog have a pretty good idea what it costs to build and sell one of these; nowhere even close to $1,500 a unit, especially in thousands volume. Except for one thing: Yup – licenses, permits, insurance, Mega-CYA.
If you sold these for parachuting pigs, they would probably cost around $150, not $1,500.
Now you know the how the cost of things sky-rocket due to out of control government regulations and greedy Trial Lawyers. Sheeesh.
I would presume that if you added the cost of everything else required to skydive, the $1,500 for this (amortized over 10 years, remember) would be minimal.
And I would guess if you were an avid skydiver, you would make the investment because it’s your life on the line, and it makes sense.
Government regulations have nothing to do with it – they aren’t devices that MUST be on your parachute. You’re free to save that $1,500 and not use it (I presume most good schools offer them to rent if not part of the standard rental kit already).
The parts are cheap, the manufacturing is fairly cheap, but the engineering isn’t. And building for reliability is extremely difficult. It’s part of a safety-critical system.
Hell, the parts may NOT be cheap if you go for aviation grade parts with full tracability so if a part goes bad, you can find out where the part came from and who made it, what other parts were part of the batch, follow the chain down and recall those units as well. Of course, that paperwork turns a 5 cent bolt into a $5+ bolt, but if it’s a failure in manufacturing, it’s possible to trace what other units are affected and recall them for fixing.
The biggest expense for constructing a unit like this is R&D and testing. And I am talking both about electronics, software, and cutter design. Search around for Airtec’s testing stuff and it is absolutely ridiculous. One time a user complained that he had a unit fire ‘high’ (his own error) and the manufacturer requested his unit back and did 40 test flights with it, finding no abnormality. Additionally, I believe each of their cutters is checked for correctness of assembly via X-ray. You’d think that’s ridiculous, but there was a recent incident where a person (using a different brand cutter) died when their AAD failed to release the reserve. The cause? Someone had forgotten to install the blade into the cutter during manufacture. Later information was published on how users could use a magnet to verify that their cutters had blades installed (if the cutter came from that specific lot).
Well. One reason for use only a pressure sensor as input. Will be that if it fails, you won’t survive for telling anyone!
I was hoping to see a demo with a small cargo chute thrown from the EEVblog helicopter before the teardown.
Add my vote in favor of firing the thing sometime.
Manufacturer’s test firing demo:
There also exists a video somewhere of a fellow who ripped the metal end off the pyro unit to see how far the blade would launch. I believe it flew across the room through and embedded itself in his desk. Can’t find the link, though.
The cost X The Price X The “Willing to pay”…
The good, the bad and the ugly rsrsrs
It saves lifes… 1500 is cheap… The full rig with main+reserve+cypres costs US$5000 new.
Companies are built to profit. The trick is to charge as much as you can and remain in the price zone the customer is willing to pay for your product. Right now Cypres rules. No competition.
Making a product that needs to meet such high standards in quality, performance and long life durability plus regulations compliance is something that is expensive.
Normally in these type of products the cost (cost and not price) will be influenced by the knowhow, testing and high quality of implementation other than the BOM cost. This cannot fail. By far this is the most important characteristic considered in the design.
Another example would be the traffic lights. Never, never never the condition of green lights for both streets in the intersection should occur. Then you design and develop arround this.
It was no surprise for me to find highly proven components, that although dated, it is known that they will perform. No unknown bugs or surprises.
You can see top quality component such as the Vima caps.
There are several competitors, though Airtec is the market leader.
98xx and 93xx date codes, mfd “99″, no wonder it looks old school
Quick suggestion Dave – knock off the William Shatner impression, please :).
That is a Cypres 1 unit. It is basically of late 80s design. These things have a shelf life of appoximately 12.5 years, at which point they lose airworthiness and can be torn apart.
The cutter connection housing has four pins because certain parachute container designs require two cutters.
The pressure sensor samples ambient pressure at regular intervals to calibrate itself against ground level, with some heuristics to detect takeoff. The altitude estimation algorithm is actually quite sophisticated, attempting to estimate the user’s body attitude (falling stable vs tumbling) to determine whether to apply a vertical offset of ~300ft as a result of the pressure difference from the burble created by falling face-down with the unit mounted on your back.
There is a block of jumpers on the main board, presumably to switch the activation parameters based on the unit type (student jumper / expert / tandem / etc). The unit will not “arm” the cutter until it has detected a climb to at least 1,500 feet.
Older Cypres1 designs used a much larger pressure sensor from National Instruments, with the pressure port tube trimmed off to fit in the housing (and a corner slightly melted to make room for an electrolytic cap sitting nearby).
The LCD / control unit is actually a simple up/down pulse counter, with the main unit sending 10,000 count-down pulses to make the display advance from 9999 all the way down to 0 during self-test. A separate wire in the hardness is used to set the up/down counting direction and the arrow indicator. The display is normally used to set the dropzone-runway offset where applicable. I’ve got the pinout for the thing somewhere.
Leave a reply