compressor for sailing

[Bob3]

There is a lot to be said for a decent hookah setup; the ones we use for commercial harvesting can keep 2 divers down all day on a half gallon of gas.
You may not have the same range of "freedom" as with tanks, but it's pretty darn hard for someone to get lost on the end of a 150' hose.

 

[Genesis]

I own one and use it with a Hyperfilter to fill PP-mixed Nitrox tanks.

The cfm rating on ALL of these compressors is "free air displacement", which is not what you can expect over a REAL fill. Second, all fills are "hot" and thus when cooled will not be to spec, which also inflates the numbers.

So, what's the REAL output on a compressor?

First, there is no such thing as "free air" when you're filling a tank. Sorry. The idea is to make the tank fill with compressed air, and as you do, the output goes down some. The "floating" third stage design is an attempt to get constant displacement over pressure, but the first and second stages don't float, and as such displacement is NOT a constant.

Second, any air spaces you use as buffers (e.g. filters and coalescers!) consume some air, because (1) there is no free lunch, and (2) they have to be pressurized AND DUMPED for condensate. This counts too.

So, to compensate for all this and get ACCURATE figures, we must
norm the output to what the tank WILL BE when cooled, and we must "discard" the air tossed when we dump the condensate.

You have to do this if you're mixing Nitrox anyway, since your PP computations are based on a tank at the final temperature and pressure, not the "hot fill" temp. Otherwise, your FO2 will be off!

Here, on the Florida coast, with its typical near 100% humidity you have to dump condensate about every 10 minutes. Less often than that and the filter stack begins to accumulate excessive amounts of water, which is very bad (and can trash your filter if allowed to get too far.)

What this means is that you typically "overfill" to accomodate for the heat of compression; when the tank cools, it will be spot-on in final pressure. This is perfectly acceptable as the "rated pressures" are all at 70F, and the manufacturer allows for this pressure gradient (as does the burst disc) as you might have the tank somewhere on a 100F day.

When you're filling an AL80, this typically means you fill to right near 3200 psi (hot); it will cool to 3000. For an HP100, you might fill to 3750 (rated pressure is 3500), and again, it will cool right to the rated over the next hour or so. Again, "rated" pressure means pressure at 70F; you can't fill a tank at 32F to "rated", as when it warms up (if it does) it will be significantly overfilled.

When you measure ACCURATELY, for ACTUAL output, you find that these compressors produce about 2.5-2.8cfm of ACTUAL output across the tank, assuming you do NOT start completely empty, and that includes the dumping of the condensate (which hurts you significantly.) Remember, starting completely empty gives the compressor a leg up; this is doubly true if there is no backpressure maintaining valve on the dryer! Also, when you dump the condensate you release pressure in both the condensator on the second state (inlet to third) and the filter cannister, and both of those take time to recover pressure before air delivery resumes. This is particularly true as pressure rises in the tank(since more delta is released.)

I have BOTH a backpressure-maintaining valve on the dryer (you want one, as the dryer doesn't do crap worth of a job without it until the tank pressure comes up) and I account for the cooling of the gas; in addition, I'm filling through a hyperfilter which has fairly significant internal gas volume.

My statistically-valid average ACTUAL delivery on HP100s, accounting for all of the above and making allowance for none, leaving the H/F stack pressurized between both fills and when the system is idle, for the Alkin W31 is approximately 2.4cfm.

This is from my fill log spreadsheet which records the starting pressure of the tank, O2 added, final pressure after cooling, and both start and finish times on the hobbs meter of the compressor to the nearest hundreth of an hour, all over more than 40 actual operating hours.

NONE of the other small compressors do any better, and some do a lot worse! BTW, cylinder head temps never go over about 150F on this unit - that's quite cool, and very nice. Air output from the filter stack to the hyperfilter is just a few degrees above ambient; the cooling tubes do a very nice job. All of this has been measured with an IR thermometer.

BTW, the W31 220V 1ph compressor should be able to be started by most 8kw boat Gensets. Mine can start it - I've tried - if I turn off all the AC systems. The issue is surge current, not running current (I have plenty of running current) on startup.

Its a VERY quiet compressor compared to the others I've been around on the market. The 1100 rpm speed has a lot to do with that. You can easily stand right next to it while its operating; it doesn't make you reach for the hearing protection.

One thing to be aware of - like all compressors, the filter life ratings are rather, uh, "optimistic", as they are rated for 70F inlet conditions at an "average" humidity level. Quite possible if you live in Michigan. Ain't gonna happen in Florida. You'd be wise to install an "eyeball" indicator for humidity if you get one of these, and change the filter when the 60% RH indicator just starts to turn color or before. If you let that second band turn pink, you have waited too long.

If the filter gets saturated with water, it doesn't filter any more.

Beware, as this is one way you can get oil in your output! If you have a hyperfilter stack it will be rapidly polluted if you allow the inlet air to go outside of Grade "E" specs.

I've found that the 3000 cfm "hot" rating is fairly close using this metric.

BTW, I point a pedestal fan at my compressor when its running in an attempt to keep things a bit cooler, and allow that fan to run for several minutes after I shut it down - all part of an attempt to keep the valves from getting gunked up. Also, I change oil FAR more frequently than specs require (the alkin only needs ~250ml of Chemlube 751, so an oil change is VERY cheap - you can buy a gallon of the stuff for about $30.)

 

[pescador775]

I don't understand your statement about floating pistons:

First, there is no such thing as "free air" when you're filling a tank. Sorry. The idea is
to make the tank fill with compressed air, and as you do, the output goes down some.
The "floating" third stage design is an attempt to get constant displacement over
pressure, but the first and second stages don't float, and as such displacement is
NOT a constant.

The three 'floating' plungers in the Kidde 4 stage compressor are driven by a cam. The internal pressure in the second stage is fixed at 450 psi, third stage is fixed at 1100 psi and that in the fourth stage is variable. However, the displacement of #4, as in the second and third stages, is determined by a cam and by head clearance. Although the pressure of the fourth stage is variable, the mass air charge it receives is identical each stroke. As long as the head clearances are correct the charge cannot vary, either independently or in relation to pressures generated by the other stages; that is, except for variance in blowby which does increase as delivery pressure rises. The variance is approx 0.1 cfm over the full pressure range unless the unit is faulty in some way.

Is this how the Alkin works?

Incidently, I believe readers may get confused over your descriptions of measurements of a fill cycle. I hope this is a simpler explanation; run the compressor to final delivery pressure and then open the tank valve. At this point begin timing the fill. Either water cool or, if hot filling, overfill 7-10% depending on tank metal and air temp or, if hot filling to rated pressure, add 7-10% to the fill time for purposes of calculating the compressor output. This method will give the actual output of the compressor neglecting the effects of filters and condensators which will, of course, increase fill times when filling the first tank of the day, as will venting air from the filter set.

 

[Genesis]

I have put interstage gauges on my Alkin; they're not standard, but they're not difficult to add (just go get some gauges and fittings rated for the pressures involved.)

The operating interstage pressures are NOT constant over final delivery pressures.

The Alkin is a three-stage compressor. The first stage runs from 40-50 psi typically (lower for lower output pressure, higher for higher), the second from 300-440 psi, and the final is of course the output pressure. The higher range of these figures is at roughly 3600-3700 PSI (my "final pressure" for HP tanks), the former is when the priority valve (set to 1600 psi on my unit) opens and actual output air begins to flow.

The mass air charge moved cannot possibly be the same. If you take in the air at 0 psig and discharge it at 40 psig for a given stroke of the piston, the mass of air moved is not the same as if you discharge it at 50 psig.

The manual for the unit shows a range of "normal" values at full output pressure which match well with what I observe in actual use.

The Alkin has two mechanically-coupled pistons (the piston that compresses the air is physically connected to the crankshaft with a connecting rod) and one (the last stage) is a "floating" piston; the actual piston that compresses the air is not mechanically connected to the one driven by the crankshaft (it rides "on" it, but is not pinned or otherwise connected to it.)

You can only get an equal-volume output across final delivery pressures if the intake pressures at each stage are the same; otherwise, the mass of air taken into the cylinder for each stroke changes with the intake pressure.

The Alkin's interstage pressures are NOT fixed and definitely vary some (enough to be significant) with the final output pressure.

Yes, the procedure you give works for timing fill rate, but you have to dump the condensators frequently enough in Florida (every 10-15 minutes) that it WILL affect your ratings significantly. I do essentially the same thing with the Hyperfilter stack; I run the unit until the pressure in the stack is at the final pressure, then open the fill valve. The values I gave for my "actual" delivery account for both that (e.g. the hobbs time is clocked when the valve is opened, not when the unit is turned on) and also for the dumps made during filling (which can't be avoided.)

Larger condensators would permit less frequent dumping, which in turn might allow a closer-to-rated-delivery to be actually realized. I am considering adding larger ones, along with a home-made auto-dump system (not difficult to do with air-controlled pilot valves) as I'm thinking of putting in a bank here, and dumping by hand is a pain in the tush if you're filling a bank.

 

[pipedope]

Hey Bob3,

Why do commercial divers wear a harness with a tether?
ans. The standby diver is too lazy to look for the body.


Seriously, hooka rigs are great for any diving were you don't *need* to be free swimming.

 

[pescador775]

Genesis, that's interesting. Can you tell me the actual output of the Alkin, exclusive of the routine variables like condensate dumping?

The pressure variations that you are seeing are probably a function of heating and do not affect the mass of air moved by an individual piston stroke. If there were large variations I would suspect leaky valves. The mass of air being moved is a function of ambient air pressure and, more importantly, by the size of the first stage. Everything taken in by this stage MUST be moved somewhere and that 'somewhere' is out the final stage exhaust valve. Of course, if rotational speed changes when pumping at varying pressures that must be taken into account.

I wonder why Alkin appears to be intentionally avoiding a competitive advantage by not boosting compressor speed. The diameter of the pulley on the gas engine could probably be increased by, say, 1/2" to bring the speed up to 1450 rpm. This is still dead slow but would bring the advertised output closer to 4 cfm. The electric model, which is sometimes used to fill HP banks, should probably stay at 1200 rpm. These kinds of speed differentials are common in the industry, anyway.I have noticed though, that Bauer's Oceanus, rated 4 cfm, has a temperature switch.

 

[Genesis]

I will try though.... but its not trivial to do, as I don't dare not dump every 10-15 minutes, as the buildup of water gets excessive.

I could fill the pony (I usually just top it) but that's a small enough amount of air (19cf) that even very small errors in either start pressure or time would throw off the results.

I don't know about the temp being responsible... if it was, then it should get hot and stay hot. It doesn't - if I fill 3 tanks in a row, when I go from the first to the second the pressures come back down as the H/F stack bleeds into the new (empty) tank.

It would not be hard to crank up the speed, simply by changing the drive pully diameter. Of course that might require a more powerful motor, and it also might cause longevity issues - the slow RPM is actually quite nice, as it keeps it nice and cool and reasonably quiet.

 

[pescador775]

One week after an incident in June, a friend told me the story about the exploding condensator. It blew up with the power of a hand grenade. Since the compressor was located in the engine room of his 52' Hatteras no one was injured. The shards were taken to a retired engineer who was prominent at American Bristol. The finding was that there was a fire inside the condenser/filter and that the aluminum shell had failed due to heat. The cause of the fire was neat water entering the filter section causing the Vaporshell 13X to heat up. I told him of my missgivings concerning the condenser/filter combo and for that reason had installed, in my Kidde compressor, a separate condenser followed by a separate filter canister. The condenser can hold 7.5 ounces of water and easily go for an hour between drain intervals, even under high Rel humidity.

The condenser/filter was a modification of a Kidde dessicant container and contained a large quantity of 13X. The mods were done by my friend who has been making custom compressor installations for 30 years. The compressor itself is very interesting. It consists of TWO Kidde 4 cfm pumps operating in tandem and basically runs all day when the boat is at sea.

 

[Genesis]

Yeah, that's one reason I'm anally careful about dumping my condensators....

I also don't like the idea of having H/F cartridges ruined, as they're $100 a crack.....

 

[pescador775]

You must be talking about the 'hyper filter' when you say '100 dollars'. I don't use hyper air filters when doing partial pressure filling. The condenser/filter that is installed on the Kidde frame seems adequate for occasional topping off. The Capitano in my diving locker has 3 condensers, a dessicant can and a final triplex type filter with BP valve which seems to do the job. The Kidde, as mentioned, has a separate condenser with back pressure valve. From there the air goes into a filter canister which has its own air stone (shrouded in a bell housing) and chemical stack. I've yet to see a drop of condensate emit when blowing down the filter can. Both compressor filters can be hand packed for about 3-4 bucks ea. Generally, if one observes water in a SCUBA tank there is also a minute amount of oil. They are fellow travelers when forming an emulsion under high pressure. If not, I don't worry about oil either way. I've filled non SCUBA tanks without the final filter and the air expelled from the fill tank smells faintly like compressor oil. I would say that if you don't smell oil there probably isn't any. However, some types of oil may not have a noticeable smell. I think mineral oil is non odiferous.