Teardown of the K1

Teardown of the K1
By Karl Denninger


Ok, here we go! I'm going to do this in reverse order - that is, as a build-up rather than a teardown, since that's what you're going to do before you dive anyway, right?

We start with the bare cannister. It has attached to it a standard steel STA and two bottle mounts ("Pony Tamers".) The downtube is Schedule 40 PVC. The tube itself is PE3408 water pipe, SDR-11, 6" nominal inside diameter.


The first step in assembly is to insert the bottom piece for the scrubber cannister. This piece is an assembly made up of three pieces of Delrin. The outside and inside cover pieces are on the bottom, and the lower scrubber media scrim is next; the scrim itself is drilled delrin with a polypropylene mesh screwed to it. This allows very inexpensive and quick changes of the mesh if it gets damaged in some fashion. All three are connected together with through-bolts. The bottom cover piece has both countersinks (to keep the bolts from digging into a boat deck or kitchen table surface!) and chamfers at the end of the boltholes. An O-ring is used on the bolts which fits into the countersink much like a tank neck O-ring to provide a water and gas-tight seal. Note the face-seal O-ring (much like a SS scooter) for the can base.


One of the oddities of SDR PE3408 pipe is that its not actually round! It only looks that way. And - the errors are as much as a tenth of an inch - too much to clean up nicely on a lathe. So - all the pieces that fit into the can are "one-offs" - they are laid up as patterns on stock, cut, and then individually sanded to fit the contour of the actual pipe. What this means in practice is that the fit is specific to the pipe section used AND that it must be assembled with the pieces aligned properly, or it will not go in. This shows the alignment marks on the bottom scrubber scrim to make insertion possible - the reference is the downtube. The face-seal O-ring is clearly visible here and bears directly on the tube. 4 locking Nielson-Session latches close the bottom of the can.


And, the assembled bottom scrubber - note the countersunk machine screws so you can set the can upright on the a counter (or a boat!) without damaging the surface. (One of the design goals was that the system had to sit upright, on its own, once assembled)


This is a direct-fill cannister. What this means is that there is no scrubber insert - the can IS the scrubber. This works well because PE3408 SDR-11 pipe is more than a half-inch thick and plastic is an excellent heat insulator, keeping the reaction temperature in the scrubber bed high. It also makes for a much simpler system with fewer ways for gas to channel around the media. So we'll take it outside and fill it to capacity - approximately 7 lbs of either 4-8 or 6-12 Sodasorb or Sofnolime. Breathing resistance is fine with 6-12, but its more expensive and not necessary for shallower work.


Once the can is full and "shaken, not stirred", we place the top scrim pieces in. There are two - a "dust pad" made out of polypropylene fiber as a dust filter (breathing dust = bad!) and then a top structural scrim made out of drilled PE plate. Again, the same polypropylene mesh is used as a structural factor, although with the pad it is not really necessary unless you intend to bash the unit around once charged. The top plate, not being tied to the structure of the unit directly, does not need to be as strong as the bottom and thus a thinner and lighter PE plate is used rather than a second Delrin piece. A "half cannister adapter" is simply a piece of 3/4" CPVC that fits over the top of the stalk and allows the spring to load properly with only half the "rocks" loaded - great for pool use or a couple of SHORT recreational dives. With the half-pack cannister set up you would not want to exceed roughly 90 minutes of time or decend to more than 100', lest the scrubber not have enough "dwell" to get all the CO2 out!


Now we have to fit the anti-channeling spring and pin. The spring applies pressure on the scrubber media through the top scrim, so if you do less than a perfect job of packing and the scrubber media shifts it will "autopack" itself rather than develop channels in the media. This spring fits over the top of the central stalk, and then a cotter pin is inserted through the center stalk to pin it in place. Finally, to keep dust and crud out of the center stalk, a pipe cap is (loosely - you want to be able to get it off later!) put on top of the entire assembly. We now have a loaded scrubber and are ready to proceed with the electronics.


Next up is the electronics pod and head. The construction is almost identical to the base. Less-beefy screws are used as the structural load is far less, and there is no need for countersinking the screwheads. The top and bottom lid pieces are cut from Delrin as is the bottom, and shaped to fit the eccentricity in the tube. The sensor plate is cut from white PE sheet, then drilled and threaded as required for the sensors.

The solenoid is mounted directly on the head, using a 1.5" length of 1/8" brass pipe. A 90 degree elbow is attached on the top and again, the penetration has a chamfer to allow an O-ring seal to be made. On the end is a standard BC connector. With judicious use of teflon tape this, when assembled, is gas-tight. The outlet of the solenoid has a second section of 1/8" diameter brass pipe which is aligned with a hole in the PE sensor holder, directing oxygen injection below the sensor plate but above the scrubber media.

This design decision is intentional - the electronics perform proportional injection based on the error (difference between desired setpoint and actual measured PO2), including the depth of the diver at the time. As a consequence it is desired that the electronics see the O2 rise as soon as possible - thus, we want the O2 to go in right near the sensors in this design! This is contrary to most other rebreather systems but in practice it leads to VERY SMALL injections of O2 (on times as low as 50 milliseconds are possible) and extremely tight control of the PO2 at a constant depth - typically, the PO2 will run between .01 and .03 below setpoint.

The amplifier and digital conversion board for the sensors is visible here, encapsulated in potting compound to be impervious to both pressure and water. The depth sensor is visible in the first photo - it is inside the head and, since the loop is at ambient pressure, it reads the depth of the system. The third picture shows the space avaialble for a 4th sensor (e.g. VR3). All sensors can be inserted and removed without having to disassemble the sensor carriage if desired. Note that the protocol between the head and CPU is 100% digital - analog signals from the depth and oxygen sensors only have to run down a few inches of wire, leading to EXCELLENT noise immunity and fewer required wires in the cable. The head to CPU cable requires seven wires (+12V, +5V, Ground, Clock, Data, Select and Solenoid.) The protocol between the head and CPU board allows for data conversion integrity checking as the ADC used has zero, half-scale and full-scale internal registers - these are read on each conversion to insure that data integrity exists between the two devices.

The solenoid and power connections are removable - they are sealed with liquid electrical tape once put together, which provides water impervious connections, yet it can be taken apart if desired. Beware that after using this up to a week is required for all the volatiles to offgas - you don't want to be breathing that stuff unless you're into sniffing glue!

All wire in the head other than the E/O power cord is 100% tinned boat cable. NO bare copper wires, no "black crud of death" wicking problems.


Put together, here's the entire scrubber and head assembly, ready to mount the wing, plate, bottles and counterlungs.


We attach our standard wing and SS plate, along with the battery pack (a standard 4.5AH NiMH light can battery - 12V supply.) The supply power has an E/O cord attached which provides both disconnectability and total waterblock isolation between the head and battery cannister. The control/display head is also shown here, plugged into the Fischer connector from the head.


Now we add the bottles - oxygen on the exhale (right) side, diluent on the left. The oxygen bottle regulator is a DS4 with the IP turned way down - as it turns out, it is stable with an IP of just 60 psi! This is a huge bonus because it allows the use of a larger orifice solenoid (less likely to clog!) and extremely fine metering of O2 content, and also better manual-add control. Also, lower pressures reduce (but do not eliminate!) the risk of O2 fires. An OPV is fitted "just in case" on the oxygen regulator.

Note the three-way connector on the diluent regulator. We run with one port plugged for either wetsuit diving or a drysuit dive where you need external suit inflation gas due to Trimix diluent. Since the DS4 only has two ports on each side, this is required to get reasonable hose routing. Deleting the BOV would get rid of the need for this, but the BOV, right now, is something I want - so there you have it.


Now we're ready to add the counterlungs and T-Pieces. The T-pieces are made out of Schedule 80 PVC parts; the downtube into the counterlung is drilled to provide ventilation and uses a triple O-ring seal - two barrel O-rings and one face ring (not visible easily) above it, which acts as a crush seal between the flange and insert. The exhale lung has a Viking "hazmat" dump fitted (will not flow water backwards even if depressed) to get rid of water from mouthpiece "oops"es or other minor leaks in that side. In an emergency you can dewater the cannister by rolling to run the water into the exhale counterlung, then expel it through the dump. Of course whether the scrubber would still be breathable after that is open to debate - it would likely be safer to bail! The separators in the T-pieces are made out of PE and epoxied on both sides to provide an "epoxy clamp" to keep them in place.

To flood the can you'd really have to try - the T-pieces provide excellent water trap capabilities. Taking the mouthpiece out without shutting it first results in lots of water in the exhale lung, but it does not get through to the scrubber due to the geometry of the system.

This is one area that I may rework, depending on my experience with it - not for functional reasons but due to size - the T-pieces are BIG! But - they work!


Here is the kit with the counterlungs and T-pieces attached...


And finally, with the DSV/BOV attached and ready to dive...


(Note - hose clamps for the hoses, which are covered in rubber "shrouds" so they do not cut into the hoses, have been intentionally omitted from the pictures!)

There 'ya are - ready to jump in and go.

The K1 will be at the DEMA rebreather party on Friday night - come and see it! It'll be coming to the party charged and ready to dive, although I don't know if we'll get to actually see it "in anger" or not.....

Before I take this beast into decompression, or anywhere that I cannot bail if there's an electronics problem, the 4th Cell will be fitted for my VR3, so I have a means of controlling the PO2 in the event the electronics take a dump.

I am also considering fitting a homemade HUD which will be totally self-contained, using only a junction block for it and the PO2 sensors for the main electronics. Thus, the only thing the HUD will share with the primary will be the actual O2 sensors. This has its share of potential issues, the most important of which will be the location of the electronics to do the job - I suspect I will place it in the head if I can make it all fit.


Discuss this article HERE