Latest Features
mCCR Face-Off: Gemini vs Sidewinder
There’s lots of buzz these days among sidemount rebreather aficionados: there’s a new sidemount mCCR in town, which has resulted in increased noise on the socials. Here geeky Floridian cave explorer Grant Tobin dives into the weeds and reviews Fathom Dive Systems’ new Gemini mCCR rebreather and contrasts it to his previous breather, the popular KISS Sidewinder.
By Grant Tobin. Images courtesy of SJ Alice Bennett unless noted.

I’d like to be able to give a fully independent review of the Fathom Dive Systems’ new Gemini sidemount rebreather, but I found myself unable to begin doing so without including a comparison with the KISS Sidewinder as well as my personal justifications for switching.
Disclosures: I have no affiliation with KISS, a division of Polish-based XDeep (although I am certified to service the units). I have no business affiliation with Fathom but am sometimes involved with Karst Underwater research (KUR). Charlie Roberson, the proprietor of Fathom is on the KUR board of directors. Also, I really don’t think any sidemount unit should be a diver’s first rebreather.

The Sidewinder
I can say, however, that the Sidewinder that was invented by Mike Young (and Edd Sorensen, and probably Matt Vinzant for the initial inspiration) has also been responsible for more cave line laid in the past five years than almost any other unit. I do firmly believe in dual oxygen monitoring. All opinions are my own and should be taken with your own grains of salt. [For detailed specs on the KISS Sidewinder see InDEPTH’s Holiday Rebreather Guide]
I did my initial training on the KISS Sidewinder in April 2019, first with Mike Young’s unit and then finishing on my own after it was delivered. I put approximately ~350 hours on the unit between April 2019 and Jan 2023. My max depth was 96 m/315 ft, and my longest dive was somewhere north of 6 hours. My lowest temperature dive was 1°C/34ºF.
Here are the modifications I made to the Sidewinder:
- Changed the KISS DSV to Golem Gear BOV to Golem Gear DSV (I dislike the KISS one. It was improved in the most recent update and is now easier to service, but I have a strong distaste for the circ-clips that hold it to the loop hoses, as well as the vertical DSV open close vs a horizontal twist).
- Added Loop covers (homemade from SuperFabric or whatever the armored milspec dotted stuff is)
- Added syntactic foam coating to the scrubber
- Changed the Omni-Swivel QDs for the ADV inlet to either QC6 or BC Inflator, depending on the team I’m diving with—Removed the ADV (it either needs an in-line shutoff, or to be removed, or to run on a lower diluent IP, or a stiffer ADV membrane, or two membranes)
- Removed the OPV (it’s in a great place to wet your unit if you happen to bump it into a restriction or unintentionally hit it with your elbow)
- Changed the Fischer-driven single monitor with Molex cells to dual monitor FathomHUD hardwired + Petrel 2/3 four-pin off an SMB splitter board (Fischer is silly—the argument for single monitor has improved with the advent of vibration in the Petrel 3, but I still am philosophically opposed).
- Changed the stock towers to Light Monkey towers (not only have there been various shades of manufacturing tolerances over the twenty years of KISS using these, but a few friends have popped theirs off grinding through restrictions)
- Changed the stock-blanked WellsMarine first stage and KISS single button oxygen add (GAV) to a Poseidon XStream MK3 on a KISS dual button orifice needle valve (remove depth limit, add needle flexibility, better access to first stage parts, integral OPV. The stock oxygen add is also prone to breaking when torqued by the wrong people)
- Switched from the plastic loop hose retaining clamps to metal hose clamps
What remained an issue with the Sidewinder unit itself after making these changes? [Ed. note: Similar to other CCR manufacturers, KISS/XDeep does not sanction the modification of their equipment, as it can create safety issues.]
Some of the things that make the split-backmount-not-really-a-sidemount unit great also create other issues
- Water removal. No water dump is available on the small lung. Brett Hemphill remains the only person I know who has used the DSMB style dump available on the larger lung, though I imagine Mike Young and a couple others have as well. Edd Sorensen speaks of intentionally flooding the unit and then proceeding to do a long cave dive on the unit “about 8 times” without ever having had a caustic, and he’s not the only one. The shape of the lung on the back makes it less likely for the caustic water to move from the exhale scrubber, through the counterlung, up the inhale scrubber, and into the cells (and eventually the inhale loop). I would not advise trying this yourself.
- Wet sorb. Lung butter and any moisture dumps from the exhale and onto the top of the exhale scrubber. In 22°C water, you can make a sandcastle out of the right-side scrubber after four-ish hours, and a partial sandcastle out of the inhale scrubber after five. Some of this is my drool, some is the byproduct of the scrubber reaction and lack of water trap on the unit.
- No gas inlet across cell faces. Some units (Fathom, for example, or a modified KISS Sidekick) have the diluent-in blowing across the face of the cells, allowing for instant cell verification. Others have cells oriented in a place that is less likely to get wet under normal diving conditions. In cold water, there is enough condensation on the inhale side of the hose that water drips down onto the back of the cells in the Sidewinder and into the Molex connectors. If the cell faces do end up wet, a few dil flushes may help stabilize, partially dry, and hopefully return the cells to function, but it’s often a roll of the dice if they all come back.
- Thread on scrubber canister heads. Some scrubbers do and some scrubbers do not have a line indicating max fill. Some scrubbers seem to have more and some fewer rotations of the head nut to lock down the heads on the scrubbers. There have been at least two instances of the user managing to grind through a restriction or otherwise and rotating the nut loose and knocking the head off. User error? Definitely possible. When the first reports of the tower to loop thread popping loose surfaced, the immediate reaction was to state user error and poor assembly. Having held a pair of the failed ones in my hand, I humbly disagree.
- Scrubber to counterlung connection. These use another pair of the circ clips.
- Internal mesh screen on the scrubber. For the first few years, the bottom of scrubbers had a glued-in circular disc with a mesh screen to support the sorb and keep it separate from the counterlung. I’ve managed to break two of them. Mike replaced them for me, though I still find the plastic crossbars thin.
- Water from the OPV can drip down into the back of the cells / into the cell head. Short people (and people that have been taught poorly) trigger the OPV each time they reach for their butt d ring or XDeep OPV with their right hand and it’s one of the few ways to get a quick caustic on a sidewinder.
- Various iterations of the SW MAVs and needles have had issues at extreme depths. Switching to higher durometer o-rings helped, but anybody that has taken their unit past 90 m/295 ft knows of these issues. Hint, they fail open and it’s a nice way to get yourself into a “boom drill.” It’s not a great trimix unit unless you need to squeeze through something at or on the way to these depths.
The above largely ignores any issues with parts that I’ve already replaced. An important thing to remember is that much of the Spirit series were simply built by parts on hand from the KISS Classic (and even the Sport). Have I been using this unit differently than intended? Certainly, or at very least differently than what Mike Young set out to make when he produced the units. With so many instructors screaming about how the SW is “the most versatile and simple rebreather made,” I did put it through its paces. By the time I retired it, my unit was, at most, half stock components.
The Gemini

I began hearing rumors that Fathom was working on a prototype sidemount unit in early 2022. Charlie and the team he dives with are pretty experienced at home builds and have used various rebreathers over the years, from modified Fathoms to Billy Gambrill-esque Lowriders, to Sidekicks and SF2s and Choptimas and others, so I was eager to see what would pop out. Rumors beget rumors and poof, we got to the Diving Equipment and marketing Association (DEMA) trade show in Orlando, Nov 2022, and the Gemini was on display to the public.
According to the Fathom Gemini CCR Spec Sheet:
- $7695
- Dual scrubber (hence the name)
- Axial
- potted head using SMB connectors (JJ/Fathom cells)
- No ADV
- OPV/water dump on the lung itself
- Easy integration with GG BOV, GG DSV as stock
- Fathom dual button needle valve with blocked, stiff spring Apeks DS4
- Swappable scrubber baskets
- Integrated HUD
The Gemini and the Sidewinder are obviously units of a similar skeleton (not unlike an xCCR vs a JJ-CCR vs a BMCL Liberty etc), and a few other comparisons can be drawn. That said, the stock and early manufactured Sidewinders (2018, 2019) are probably a Honda Fiat to the 2023 Gemini’s Toyota Corolla (or even the old Lexus LS 400)? The 2022-ish Sidewinders, properly configured with a bunch of aftermarket work, (imo) bring them closer to a Gemini. Some gaps are inevitable, though.
Here’s some information that’s not in the spec sheet:
- Prelim EN14143 testing with 812 sorb has given a 2.5 hour duration at a breathing rate of 40 LPM with 1.6LPM CO2 injection on air at 4°C. My personal breathing rate is a quarter of that, and my CO2 production rate (extrapolated from my oxygen consumption) is ~0.375 that. Under normal conditions, that brings the expected duration to something more useful for me.
- Stock 44” LP hose from the DS4 to the Needle.
Price comparison for a 2022 Sidewinder with dual monitors, as of 31 Dec 2022, to be “ready to dive” and excluding harness, regs, cylinder, and assuming one needs a MAV, and that Delrin is a less good insulator than Black Amalgon. The scrubber basket material matters for diving, and thus you need syntactic foam.

Note, a purchaser could do some other things to cheapen this, but I wished to use the KISS configurator to attempt to bring them together to something many view as safer. Blah, blah—always know your PO2. If two monitors saves one life, the quant in me says the math of everyone needing two monitors is positive expected value. Even better, an instructor can monitor student PPO2 on a Fathom HUD farther away than a Petrel (Even scarier: instructors that permit new students to only use a Shearwater Nerd).
What is still lacking on this hypothetical Sidewinder compared to a Gemini? The N@90 heads up display (HUD) is much worse than the Fathom version. You’d still be working with Molex cells and a head that isn’t potted. And you have a depth limit in the 70m/230 ft range. (*Mike has made and sent a few SMB style splitter boards, but they’re Molex in and SMB out, so you’re still drowning in six Molex connectors).
If we were attempting to make this phantom Sidewinder match the Gemini as closely as possible in function (higher depth limit, cap ADV, etc), we’d be out the needle valve ($499 for the SubGravity, which is cheapest publicly available on market, but only allows for oxygen in, as opposed to the Fathom or KISS needle valves with two inlets, and thus eliminating the MAV) as well as appropriate first stage, stiffer spring, blanking cap, and OPV+ADV caps. The price discrepancy would then increase further.
Can someone argue that the Sidewinder is designed to be simple and shouldn’t need splitters and dual monitoring and a needle, and that sump heads would solve some of these? Sure. It’s not how I use it, and it’s not how I think it should be taught. An uncertified user also isn’t “allowed” to order sump heads with a new unit. You could order the Sidewinder without dual monitoring, but you can’t order a Gemini without a HUD. I can’t see a NERD 2 in a true whiteout, but I can see the HUD, or I can feel a (vibrating) Petrel 3. I’d rather be able to stay on the loop and count via the HUD than stay on the unit and trust my—or most diver’s—ability to dick around counting breaths in semi-closed rebreather (SCR) mode.
The above information is all pre-purchase for me. I put down a deposit at DEMA and took delivery in early-January. Current instructors as of this writing are Kelvin Davidson (instructor trainer), Jon Kieren, and Giovanni Gastaldo at Third Dimension in Tulum, Mexico, and Jon Bernot (instructor trainer) at Cave Country Dive Shop in High Springs, Florida. After a hiccup with my work coverage interfering with a class with Bernot, I was able to find a mutual timeframe to sneak to Mexico for some time with Kelvin.
First Impressions
My unit arrived at my office on 4 January 2023 in a blue crate weighing about 12kg. My initial thoughts:

- Loop hoses are much shorter, but much more flexible.
- Removable scrubber baskets mean I won’t be able to use that dead space for packing (I formerly put two computers and my oxygen first stage inside of scrubber cans.

- Very tight tolerances on all connections. Almost “break an O-ring if you don’t lube them” tight.
- Weed Wacker cable seals the bottoms and tops of the cans to the lid. Coincidentally, this is also how the one atmosphere exo suits work.

- Not sure how I’m going to feel about the water dump placement. I debated for a long time drilling holes in the bottom of my sidewinder cans and installing Light Monkey tinkle valves. This way, I could dump water before it hits the lung. I elected not to proceed, as I worried it would harm resale value and, at the end of the day, rebreathers are a means to an end, and I don’t have emotional attachment to units.
- I’m pretty compact, the two button Fathom needle valve is not. (I’ve had one in the past, this wasn’t really a shock)
- No loop hose weights (I had them on the sidewinder and never experimented with removing them).
- Holes on the scrubber screens are a little larger than expected. Fractions of a millimeter but would wait to put sorb in them to find out whether this concern was warranted.
- The choice to flow gas from right to left means that for the oxygen valve to remain on the right side, the hose into the needle must pass along the bottom of the harness and up the left side of the back (as opposed to straight up the right side like the SW).
- Cell placement is an improvement. In a horizontal swimming orientation, cell faces are down toward the ground rather than backwards toward your fins. Unlike Molex cells, backs of JJ and Fathom cells are covered with the wire lead.

Contents:
- Lung (4.5L)
- Scrubbers
- Scrubber baskets
- Gas addition head
- Electronics head with HUD + 6” female 4-pin (Charlie will make it longer if you ask/need)
- 40” Inflator Hose DS4 w button gauge and beefy Apeks OPV
- Needle + hose to head
- QC6 male + hose to first stage
- Flow meter
- DSV + nipples
- Three bolt snaps (can tops, bungee across front)
- Bungee
- Paracord
- Fathom hat and sticker
Taking The Class
I built the Gemini to copy over the rigging from my Sidewinder but did not get it wet before class. I made it to Tulum, Mexico, and took care of my crossover. Kelvin was using a Razor 4, Lanny an XDeep Rec, and myself a Katana2. Takeaways are listed below in positive, neutral, and negative order.
The Positives
+ SMB cells. Larger surface area, fewer small connections. Molex are trash, and I’m happy to fight anyone that says they prefer them.
+ Proper threading on the head to loop connection. The LM ones aren’t super well machined, and most people have a chip on their first thread because of it. The KISS ones are…well KISS ones. Caveat: Dobbykins and I had the first retail set of LM towers, so I don’t know if this has been improved in the past two years.

+ Connection from the cans to the CL do not have a circ clip that could break or be lost.
+ Dewatering ability. I have not tested it. There’s enough alkalinity in the freshwater in the cenotes in Mexico that I was worried about frying the cells. Next time I have a bit of sorb time left and the ability to pull my cells out, I’ll hop back in the water and totally flood the thing a couple times to see what happens. Between Edd and one or two others’ experiences having flooded one side, I’m not super concerned with some moisture, assuming the inhale side remains unflooded.

+ Easy BOV integration. I’ve danced both sides of the BOV DSV debate. On a Sidewinder, the integration with a BOV was either very messy, or needed Fathom parts, or would add an additional hose somewhere. I did the Jason Richards thing for a while but grew to hate the ADV and all the additional hardware necessary. On a Gemini, going from DSV to BOV requires a single 30” LP hose and an elbow (plus the BOV). One could ditch the necklaced backup and thus ease gearing up. On all CCRs, I firmly believe you need either a BOV or a necklaced backup that is always breathable.

+ The two button needle (and the other Fathom needle and other Fathom MAV) are upstream. I’ll use a quote from Charlie here to avoid confusion: “The manual addition button flows gas opposite of normal buoyancy compensator inflators. We did it to prevent the leaks that the SW MAV is plagued with. Increased IP pushed it open in the normal downstream direction but assists in keeping it shut with upstream flow. It eliminates leaks and free flowing MAV. That also means that a HP seat failure on the oxygen 1st stage won’t result in a boom scenario.” I actually didn’t know this, and I don’t currently know why it isn’t advertised more heavily.

+ Swappable canisters. Credit is due to rEvo designer Paul Raymakers for conceiving of the split or dual scrubber canisters—the ability to swap scrubbers does present interesting possibilities. Say I fly down to Florida or Mexico for a week. Arrive by midday Saturday and want to get a shakedown dive in before doing much larger dives over the next few days. Now, I can get a two hour dive in, swap the old inhale into the new exhale, repack the new inhale side, and be happy putting in a very long dive (Caveat: you’ll all need to figure this out by yourself, and I’m not going to tell you my rules for sorb use. If you’re pushing max scrubber durations, you should be in a big boy or big girl state of mind and be able to evaluate these risks.)
+ You’re more likely to hurt yourself trying to disassemble the unit topside than you are to hurt it under water.
+ No battery box on the loop hoses
+ WOB. I saved this one for last, as I don’t think humans can measure these things objectively by feel or with much precision, and I don’t want to lead readers astray. I feel that the Gemini breathes slightly better than the Sidewinder. 10%? 15%? And then I had to ask myself why (and question as well whether I was still in the honeymoon phase with the unit). I think, very simply, it has fewer bends for the gas path and slightly less turbulent flow through the hoses. The towers run straight out of the lid, there is less dead space in the heads, and the entry at the bottom of the scrubber is smoother.
Hopefully xDeep/KISS will publish the results of the testing they did, Fathom will do the same, and we will have some data. The important thing to remember is the manner the cans and lungs are rigged will have the highest impact.
Neutral
- SW users are split running their MAV and GAV across their chest vs over the shoulder. I will need to get used to this being on my left shoulder.
- Packing sorb is going to take longer. I’ll trade this for the ability to pull scrubbers out to let cells dry without needing third party caps or to be able to toss the scrubber baskets into a drybag.
- Haven’t squeezed into anything that makes me prefer the straight routing vs the angled-in routing of the KISS and LM towers. If the straight piping did improve build quality and WOB, I’m all for it, and I do believe the straight plumbing puts the incompressible parts lower on the body
- Flow right to left
- Fathom advertises Black Amalgon as 400x better insulation than aluminum. Aluminum 6061 is in the ~167 W/m-K range, Acetal (Delrin) is 0.23 W/m-K range, and epoxy coated fibrous synthetic materials seem to be in the ~0.04 W/m-K range. I don’t have the equipment to test whether the dual design of Amalgon + inner scrubber basket material is superior to the newer Sidewinder Delrin cans + syntactic foam coating.
The Negatives
— HUD uses a 2032 battery. Time will tell how long this lasts me, and it’s another thing to keep in the kit. Takes about three minutes to carefully change.

— My finger fits and has the dexterity to remove the heads via fingering. If one doesn’t, you’ll need a guitar pick or a spudger or a relatively high amount of grip strength to remove the heads when you need to disassemble the unit. I think Charlie should include a plastic bike tire lever, as I promise you someone is going to use a non-coated tool and either mark the cans or the head or chip the head loop connection. There is a trick to it, but GFL if you didn’t lube the o-rings sufficiently.
— The fishing crimp on the end of the weedwacker line could mar the canister. The Gemini is a tool I’m going to scrape through rock, and the crimp is not in a place to cut my suit. That said, I wrapped it in heat shrink so I could grab it better with wet hands.
— I’d like to see a notch machined into the head and canister to ensure future users always line the heads up properly. It’s minor, and matters less than on the Sidewinder because of the straight piping. The argument against doing this is that the bottom of the cans can also be removed (but seems to be more for maintenance and manufacturing ease than a need). Fathom would need to mark tops and bottoms or one could use a sharpie and be done with it. There’s an unfortunately high amount of variation among SW instructors on correct counterlung bung to head tower angle.
— The mesh screen size is ever-so-slightly too big. I’m going to use some JJ scrim material (filter paper) on the cans to ensure less dust and no granules poking out. I popped through 5? 10? granules per time I packed the sorb. At very least, there needs to be a scrim on the side of the basket nearest the cells and one nearest the water dump on the lung. UPDATE: scrims are now included.
— I don’t think the spring on the scrubber basket adds much value. As more classes come through, I think a standard will end up developing. I’m kind of worried a muppet will bend the screen via the bolt, and the bottom screen (top as packed) isn’t of the same machining tolerance as every other piece of the unit.
— No water trap. I’ve experimented with different ideas on the Sidewinder (including a very small t piece in the middle of a loop hose) to absorb the sickening amount of spit I drool into the cans), but as of now, drool path is the same. Sorb works decently whilst wet, but eh. Still a complaint people will have.
— I have to suck it up and return to using an Apeks product.
— I dislike the stock GG mouthpieces. I’ve switched to a Divex (JJ) which I like better than the soft Scubapro one, which I prefer over the Comfobite for CCR use. On OC, my needs are slightly different. I’ll inevitably end up with a gag strap should I move back to BOV use.
Conclusions
The Gemini works about as I expected it to. It’s not wildly different from the Sidewinder, but it is better, and I trust it more. Any issues that I have with the Gemini are minor or are inherent in the design of split-backmount-not-really-a-sidemount-unit units. If you believe (as I do) that dual monitoring is a need, then I do not believe there is a reason to buy a new Sidewinder as of Jan 2023.
If you do not believe in dual monitoring but do believe in reducing the number of things that can go wrong on your unit via plugs, or sump heads, or LM towers, or a new first stage, or not using Omni-Swivel QDs, or dislike circlips, etc., then the question is a little more difficult. If demand outweighs availability of the small instructor pool for the Gemini, there’s an inevitable bottleneck. Currently, there is also a three to 12-month lead time for a new Sidewinder, depending on who you ask.
On an ending note, my motivation for putting such effort into this review comes from the fact that I have one of the first couple non-prototype units, and wanted to share what I learned.
Dive Deeper
InDEPTH: Not All mCCRs are Created Equal: The Case for the Needle Valve by Charles Roberson
InDEPTH: Keep It Simple Sidewinder By Jake Bulman and Skanda Coffield
InDEPTH: InDepth’s Holiday Rebreather Guide: 2022 Update

Grant Tobin is a derivatives trader and risk analyst based out of Chicago, IL. A native Floridian, he began diving in 2006 and cave diving in 2009. His last several years have been focused on dives and projects ranging from participating with Karst Underwater Research (KUR) in Florida, to caves in Missouri and Mexico, to the wrecks of Bikini, Scapa, and the Great Lakes with MWUE. Outside of diving, he can be found racing long distance triathlon and rock climbing.
Cave
N=1: The Inside Story of the First-Ever Hydrogen CCR dive
This Valentine’s Day, Dr. Richard Harris, aka ‘Dr. Harry,’ and the Wetmules made the first reported hydrogen (H2) rebreather dive to a depth of 230m/751 ft, in The Pearse Resurgence, New Zealand. The 13 hour dive, which was nearly two years in planning, was a field test to determine the efficacy of using hydrogen to improve safety and performance on über-deep tech dives. Harris’s dive was the deepest “bounce” dive in approximately 54 experimental H2 dives—the majority SAT dives—that have been conducted over the last 80 years by military, commercial and, yes, a group of technical divers. Now in this first published account, InDEPTH editor Ashley Stewart details the inside story behind the dive, a dive that will arguably be remembered 100 years from now!
By Ashley Stewart. Images courtesy of Simon Mitchell unless noted.

On March 11, a little more than three weeks after completing what is believed to be the first-ever rebreather dive with hydrogen as a diluent gas, Dr. Richard “Harry” Harris convened the group of scientists and researchers who had spent years helping to plan the attempt.
He started with an apology. “All of you had the sense that you were party to this crime, either knowingly or suspecting that you were complicit in this criminal activity,” Harris, an Australian anesthesiologist and diver known for his role in the Tham Luang Cave rescue, told the group.
The apology came because the dive was dangerous—not just to Harris who was risking his life, but for the people who supported him were risking a hit to their reputations and worried their friend may not return home. Harris and his team put it all on the line to develop a new technology to enable exploration at greater depths.
A significant challenge to deep diving is an increased work of breathing and CO2 buildup as breathing gas becomes more dense at greater depths. This can not only culminate in fatal respiratory failure but also increases the risk of practically everything else divers want to avoid, like inert gas narcosis and oxygen toxicity. For this reason, helium is favored by divers for its low density and non-narcotic effect. However, at such great depths, helium increases the risk of tremors and seizures from High Pressure Nervous Syndrome (HPNS). This can be ameliorated by keeping a small amount of narcotic nitrogen in the mix. The problem is that even small amounts of nitrogen makes the mix too dense past 250 meters.
Harris’s experiment would determine if divers can turn to an even lighter gas: Hydrogen, the lightest in the universe. Hydrogen is about half the density of helium. It’s also slightly narcotic and hence thought to ameliorate HPNS, thus allowing elimination of nitrogen from the mix.

The addition of hydrogen into a breathing gas, however, comes with one small technical uncertainty—the extremely explosive nature of hydrogen. History confirmed this reality with the 1937 Hindenburg disaster in which the hydrogen-filled dirigible airship burst into flames. As Harris tells it, he set out to dive hydrogen in his diluent gas while avoiding the nickname “Hindenburg Harry.”

Hydrogen in the Mix
Why would anyone attempt to breathe hydrogen? Harris and his colleagues have spent more than a decade and a half exploring the Pearse Resurgence cave system in New Zealand. This extremely challenging, cold water cave system (water temperature is 6ºC/43ºF) has been explored by Harris and his team, who call themselves the Wetmules, to a maximum depth of 245 meters/803 feet in 2020. Their gas density at depth was 7.2 g/l, significantly above the recommended hard ceiling of less than 6.2 g/l.

Diving past this point introduces increased risks, not only of CO2 buildup, but narcosis, decompression sickness, HPNS, cold breathing gas, having adequate gas supply or bailout, and isobaric counter diffusion (ICD) in which different gasses diffuse into and out of tissues after a gas switch causing bubble formation and related symptoms, cold breathing gas, and having adequate gas supply or bailout.

Divers have been examining hydrogen as a breathing gas for decades. The Swedish Navy was the first to experiment with hydrogen as a possible deep diving gas during World War II. The U.S. Navy in a 1965 paper proposed replacing helium with hydrogen due to projected helium scarcity. Later, beginning in 1991, researchers at the Naval Medical Research Institute (NMRI) in Bethesda, Maryland spent a decade studying hydrogen’s potential physiological impacts and biochemical decompression. French commercial diving contractor Comex (Compagnie maritime d’expertises) launched its hydrogen program in 1982, and the Undersea Hyperbaric Medical Society (UHMS) held a workshop “Hydrogen as a Diving Gas,” in 1987.
Even technical divers considered hydrogen. Legendary cave explorer Sheck Exley considered hydrogen in the early 1990s to mitigate HPNS symptoms, which are ultimately believed to have contributed to Exley’s death at Zacatón in 1994. Nearly all of the experimental hydrogen work up until this point used surface-supplied systems and saturation diving versus self-contained diving, and none of it, as far as we know, has been done with a rebreather.

The primary objective of Harris’ hydrogen experiment was to address the issue of increased work of breathing. Harris’s team had previously encountered CO2 incidents at the Pearse Resurgence. In one incident, while at 194 meters/636 feet, explorer Craig Challen—Harris’s primary dive buddy since 2006—lost buoyancy but was unable to find his buoyancy compensating button quickly. He kicked up a couple of times to stop his descent and immediately got a CO2 hit. Challen was able to grab the wall, calm down, slow his breathing, and survive. Based on such incidents, it’s clear to the team that they have reached the limits of the gas. “I feel we are on the knife edge all the time,” Harris said, in terms of physiology and equipment.
While hydrogen in the diluent breathing mix was expected to address increased work of breathing, the rest of the issues associated with deep diving were “major unknowns,” and some (such as respiratory heat loss) were potentially even made worse by hydrogen.
“At what depth do the risks of introducing this new technology outweigh the risks of carrying on with trimix?” Harris said. “That’s a very difficult question to answer. At some point we are going to have to consider different technologies and, at this point, hydrogen is perhaps the only one available to us.”
H2 Working Group
In 2021, the year after Harris completed his deepest dive at the Pearse Resurgence, InDepth editor-in-chief Michael Menduno was taking a technical diving class and reading about the government looking at hydrogen as a diving gas again. “Technical divers should be at the table,” Menduno said he thought to himself at the time, “our divers are as good as anybody’s.” He called John Clarke, who had spent 27 years as scientific director of the U.S. Navy Experimental Diving Unit (NEDU), and discussed setting up a working group. Menduno’s next call was to Harris, who had shared his troubles with gas density at the Pearse Resurgence. Harris had also, separately, been thinking about hydrogen.
The so-called H2 working group met for the first time in May 2021 and included many of the top minds in diving medicine and research, including Clarke, NEDU’s David Doolette and Greg Murphy, research physiologist Susan Kayar who headed up the US Navy’s hydrogen research at the Naval Medical Research Institute (NAMRI), along with her former graduate student Andreas Fahlman. There was diving engineer Åke Larsson who had hydrogen diving experience, deep-diving legend Nuno Gomes, decompression engineer JP Imbert who had been involved in COMEX’s Hydrogen diving program, and anesthesiologist and diving physician Simon Mitchell. The group was later joined by Vince Ferris, a diving hardware specialist from the U.S. Navy, and explorer and engineer Dr. Bill Stone, founder of Stone Aerospace.
The working group met regularly with the goal of figuring out how one might possibly operationalize hydrogen for a deep technical dive using the Resurgence as an example. During one of their meetings, Clark used a breathing system simulator built for the Navy to predict how hydrogen would affect gas density in a closed circuit rebreather at depths to 300 meters/984 feet.
To Doolette, who has known Harris for decades and supervised his Diploma of Diving Medicine project in 2001, it was immediately clear this was not a hypothetical discussion. “Unlike some of the scientists, I was under no illusion that the question before the working group was fiction, I knew that Harry was likely to try a H2 technical dive in the Pearse Resurgence,” said Doolette, a cave explorer in his own right, who has laid line in the Resurgence.

By fall of 2022, it was clear to many in the group that Harris was going to attempt the dive. The group had mixed feelings ranging from cautious optimism to comments like, “My friend is going to die.”
Doolette was concerned Harris and Challen would not survive the dive due to either ignition of hydrogen—in the worst case, inside the rebreather at depth—or a serious adverse response to respiratory heat loss (the latter was especially if Harris attempted diving beyond 245 meters/803 feet as he had originally planned) he said. “I have known Harry for longer than most in the group. I encouraged him to take up cave diving, so I felt a personal responsibility toward him,” Doolette said. “I have a lot of experience in operationalizing new diving technology. My goal was, if unable to discourage him, to force him to focus on the important issues.”
Leading up to the dive, Menduno scheduled Harris to give the banquet talk about the expedition at the Rebreather Forum 4 industry meeting in April. The outcome of the dive, of course, was uncertain, and the two had to make an alternate plan in the event that Harris did not return. “We had to say we were going to talk about your dive one way or another,” Menduno said. “If you don’t make it back, Simon Mitchell is going to have to give a presentation about what went wrong. Harry made some typical Harry joke like, ‘Well, as long as you don’t stop talking about me.’” Harris’s lighthearted tone betrays how seriously he took the dive and its preparation, people close to him said.
While no one involved was taking as big a risk as Harris and Challen, they were risking a hit to their professional reputations by being associated with a controversial dive, especially in the event of a tragic outcome.
“At heart, I’m an explorer, and that was pure exploration,” Mitchell, who was the diving supervisor on Harry’s dive, said when asked why he would take such a risk. “Exploration in the sense that we were pioneering a technique that hadn’t been used for quite some time and never in technical diving, not deep technical diving.” He also emphatically added, “I was more worried about my mate dying than about my professional reputation.”
Later, in planning Harris’s trip to the RF4 event, Menduno had occasion to speak to Harris’s wife, Fiona who brought up the dive.

“She said to me ‘I hope Harry is going to be OK’,” Menduno said. “I had no idea how much Harry told her, what she knew and didn’t know. All I could say was he’s got the best people in the world on his team, and if anybody can do it, he can.”
“We all held our breath and waited,” Menduno said.
‘Hydrogen Trials’ at Harry’s House
Ahead of the dive, Harris was preparing at home. The first thing Harris said he had to get his head around was—no surprise—the risk of explosion, and how to manage the gas to mitigate that risk. The potential source of explosion that Harry was most concerned with was static ignition within the CCR itself, plus other potential ignition sources like electronics, the solenoid, and adiabatic heating. Industrial literature—or “sober reading” as Harris calls it—suggested that the tiny amount of static necessary to initiate a spark to ignite hydrogen is .017 mJ, 400 times less than the smallest static spark you can feel with your fingertips and several hundred times less than required to ignite gasoline. “It ain’t much, in other words,” Harris said, noting that counterlung fabric rubbing against itself could generate just such a spark.

Ultimately, Harris came across research that suggested that static decreases with humidity. “I started to feel like there was no source of ignition inside a rebreather, but then again I said to myself, ‘Harry you only need to be wrong once’.”
The other concern was whether he could actually fill hydrogen safely while decanting, or filling one tank from another at the same pressure, and boosting the gas to reach higher pressures.
“I decided there is only one way to actually resolve this and that is to retire to the shed, order a sneaky bottle of hydrogen, and without telling my wife what was going on down the back of the house, start to actually have a bit of a play with this,” Harris said.
First Harris had to make his own DIN fitting (though not out of the ordinary for the anesthesiologist who built and tested his own rebreather before buying a commercial one in 2002) to decant the gas. Next he took his dual Megalodon rebreather with 100% hydrogen in one diluent cylinder and 100% oxygen in the other to the “test bed” in his backyard—his pool—and started to introduce hydrogen into his rebreather.
“Putting an explosive device into water was perhaps not the most logical approach because it becomes more like a depth charge than a bomb, but I thought, ‘Well, at least it might contain the blast somehow into the pool.’ I knew if I broke the back windows in the house or worse, my life wouldn’t be at risk just from the hydrogen. There would be bigger trouble afoot,” Harris said. “I left the lid of the rebreather unclipped in the vain hope it would spare me and the pool and the dog, who was helping with this experiment.”

He pressed the button of the Automatic Diluent Valve (ADV) on his rebreather, introducing hydrogen to the loop, and finally activated the solenoid before he started breathing from it. The first breaths were pleasant, he said. “It did feel very light and very slippery, and the hydrogen voice is even sillier than the helium voice, as you would expect,” he said. “I don’t want people to rush away thinking this is a safe and sensible thing to do. I’m under no illusions I’ve produced any evidence for you to see, but this is an honest account of the hydrogen trials at my house.”
The unit had not exploded with a fill of oxygen from zero to 70%, and very low humidity. “Harry, dog, and CCR survive,” as Harry wrote in his report of the trials. “Nothing bad had happened, so it was reasonable to move to the next step,” he said.


The Expedition
Harris, Challen, and other members of the Wetmules, arrived at the site of the Pearse Resurgence on New Zealand’s south island in February 2023. The cave system is so remote they needed around 10 helicopter trips to transport the team and all of its equipment. Mitchell, the diving physician, ran surface operations with “mixed feelings,” as Harris put it.
The group stayed for two weeks at a campsite, complete with a gas-mixing station, an electronics shelter for charging gear, and a “big green army tent where we meet and drink a lot of coffee and try and put off going back into the water each day,” Harris said.

The expedition was plagued with an unheard of number of problems, Harris said, “Every time we got in the water, something popped or blew up or failed.” The campsite is where Harris boosted hydrogen for the first time, from 100 to 150 bar. He flushed the booster and all the whips with hydrogen prior to boosting to make sure no oxygen was left in the system, but it was an anxious moment.
On dive day, Harris and Challen set out on what would be a 13 hour dive to 230 meters/754 feet—a “comfortable depth,” as Harris put it. Due to some problems during the expedition, it was decided that Harry would dive hydrogen, while Craig would dive trimix. At 200 meters/656 feet depth, Harris pivoted the switch block to introduce hydrogen into the loop. “The first cautious sip of hydrogen just to activate the ADV was satisfying,” he said. Gas density was not subjectively improved, but Harris noticed an obvious benefit—the HPNS-induced hand tremors he typically experienced after 180 meters/590 feet disappeared. Harris kept his setpoint at .7 during the descent and working portion of the dive, careful not to reach a fraction of oxygen above 4% which would make the mix explosive, and proceeded to the 230-meter test depth.







After completing their time at 230 meters, the team began their ascent. Harry shut off the hydrogen feed to the active loop of his dual Megalodon rebreather back at 200 meters, and then conducted a diluent flush every 10 meters/33 feet to remove the hydrogen from the loop until reaching 150 meters/492 feet. At that point, Harris boosted his PO2 to 1.3 from his set point of 0.7 (Challen remained at 1.3 throughout the dive), and they continued their ascent decompressing on a trimix (O2, He, N2) schedule, treating hydrogen as if it were helium. The complete technical details of the dive will be published in a forthcoming paper in the Diving and Hyperbaric Medicine Journal.

As soon as the team were helicoptered back to civilization, Harry called Michael from the road. “Michael, we did it!,” Harris said.
“Harry, you’re alive!,” Menduno responded.
N=1
At that March meeting with the H2 working group, Harris presented his findings from the dive. “I’m not sure what to conclude to a highly scientific, analytical, and evidence-based audience like yourselves,” he told the group. “Conclusions: N=1,” meaning it had been successful one time.
Doolette, who had been the most vocal in the group about his concerns, suggested Harris could add to his conclusions “the probability of survival is greater than zero.” Doolette, whom Mitchell contacted as soon as they reached civilization, said he “was relieved to hear that Harry survived this test dive” but remains disappointed with some aspects of the experiment, and concerned about possible future attempts. “For instance, I imagine among the engineers he consulted would have been someone with the ability and resources to do a computational fluid dynamic analysis of the Megalodon rebreather to establish the ignition risk, but instead Harry filled his rebreather up with hydrogen in his backyard.”
Overall, Harris said his findings are that hydrogen can be handled and boosted, hydrogen and CCR diving are compatible, a strategy to introduce hydrogen on descent was successful, a decompression dive was successful, a low setpoint at depth did not practically affect total dive time, strategy to reintroduce a high PO2 on ascent was successful, and HPNS and narcotic impacts were subjectively favorable.
“In introducing hydrogen we have addressed the issue of gas density, but we certainly have not established it is safe to use in terms of explosion risk, decompression of the thermal hazards,” Harris said.
Among his conclusions, Harris pointed out that he also managed to evade the nickname “Hindenburg Harry.” “Fortunately that was avoided,” he said, “but remains an ever-present risk.”
The Future of H2
Harris warns not to read too much into what his team achieved—a single data point that should in no way encourage others to repeat the dive. “David Doolette’s comment should be heeded,” Harris said. “All we have shown is that we got away with it on one occasion.”
Provided it can be safely proven and built upon, Harris said he thinks of his hydrogen dive as a window into the future that would enable tech divers to continue exploring into the 250 to 350 meter/820 to 1148 feet range. “Imagine the wrecks and caves that lay unvisited around the planet,” Harris said.
DIVE DEEPER
YouTube: Wetmules 245m Cave Dive in the Pearse Resurgence, New Zealand (2020)
InDEPTH: Hydrogen, At Last by Michael Menduno
InDEPTH: Density Discords: Understanding and Applying Gas Density Research by Reilly Fogarty
InDEPTH: Playing with Fire: Hydrogen as a Diving Gas by Reilly Fogarty
InDEPTH: High Pressure Problems on Über-Deep Dives: Dealing with HPNS by Reilly Fogarty
InDEPTH: The Case for Biochemical Decompression by Susan Kayar
John Clarke Online: Hydrogen Diving: The Good, The Bad, the Ugly (2021)
InDEPTH: Diving Beyond 250 Meters: The Deepest Cave Dives Today Compared to the Nineties by Michael Menduno and Nuno Gomes.
Undersea Hyperbaric Medical Society: Hydrogen as a Diving Gas: Proceedings of the 33rd UHMS Workshop Wilmington, North Carolina USA (February 1987)

InDepth Managing Editor Ashley Stewart is a Seattle-based journalist and tech diver. Ashley started diving with Global Underwater Explorers and writing for InDepth in 2021. She is a GUE Tech 2 and CCR1 diver and on her way to becoming an instructor. In her day job, Ashley is an investigative journalist reporting on technology companies. She can be reached at: ashley@gue.com.