SPG accuracy

[3ricj]

Calibrated gauges are accurate for a little while but can become out of calibration by things like use and neglect- impact etc. Most divers I see have at some point had their gauges hit something.

Even more expensive sensors experience sensor drift. Some manufacturers provide a value for the expected drift or long term stability, but these numbers are still based on use in very stable environments, making them somewhat irrelevant in normal use.

Gauges used commercially often require calibration to verify accuracy. The local hydro facility requires it every 6 months.

Most scuba gauges are sealed units that do not allow for calibration. Of course if you could send it out for calibration at a cost near to the cost of a new one my guess is you would buy a new one.

So if you consider that your new gauge may have inaccuracies in its reading great. if you think yours is accurate think again.

There is a reason for promoting recreational divers to return with no less than 500psi. 10% of full scale should should allow some air to remain in the tank. Even if gauge is reading high by 7-8% of full scale the diver should notice the regulators performance decreasing as the pressure in the tank drops below the regulator intermediate pressure, but then they should be at the surface.

Here is what Scubapro lists for its pressure gauges.

50 bar: ± 5 bar
750 psi: ± 75 psi
100 bar, 200 bar: ±10 bar
1500 psi, 3000 psi: ±150 psi
300 bar: ±15 bar
4350 psi: ±225 psi

Hi Dean,

First of all, Thanks for your detailed explanation. This makes lots of sense regarding calibration and error.

Innovation in scuba diving equipment moves slowly. In part, because changes cause risk due to misuse and shortage of updated training can put people at risk. Things tend to change slowly in scuba diving, in part due to self preservation.

With that said, modern pressure transducers do not suffer from the typical failures of equipment from 5-10 years ago. Modern small digital pressure transducers have such small membranes, low mass, mems frontends, auto calibration and testing and thus are highly resilient to shock.

A few years ago I was testing a "high accuracy" honeywell transducer:

full scale range was +/- 1 psi. accuracy 0.1%, 14 bit data. Drift was also very low. It had full self test, auto cal. We tested a 6' drop,10,000 cycles while operating in -20 to +40c. The amount of error during and after testing was not measurable. This was the HSC series. The quantity one pricing was $30, much lower in bulk. While the HSC series doesn't go into the pressures we need, I suspect we could find a nice cheap sensor which gives us high accuracy. Modern sensors can also report when they need service (well, more specifically, they can report error due to mishandling).

It seems that most scuba gear is using old analog gauges and digital sensors which are highly sensitive to shock and abuse. Some of the scuba pressure gauges seem to have calibration screws inside of their fittings. I am unsure if this is only offset cal or what. I'm very happy to see published information from scubapro, thanks for passing it along.

Cheers,
-3ric

 

[TSandM]

I was definitely taught somewhere in my early training that pressure gauges were most inaccurate at the ends of their ranges -- and that inaccuracy at the high end wasn't critical, but inaccuracy at the low end was part of the reason why one never plans to surface with less than 500 psi.

I know gauges are inaccurate. We have a tank checker which reads 600 psi high at its high end, which is a little annoying. However, I'm not sure I'd be willing to pay much more for an SPG that was more accurate, and I can't imagine any kind of electronic gauge not costing significantly more than my little brass-and-glass SPG does. I start all my dives with more gas than I am going to need, and end most of them with a third or more of my tank remaining, so the inaccuracies at the end of the range don't really matter very much to me.

 

[tbone1004]

Eric, first thing is you can't calibrate a bourdon tube. Once it is out of calibration, there is literally nothing that can be done to fix it other than replace the gauge face with a different background, it's just a fact. The cost of doing this would be prohibitive.

For a digital gauge, you have the cost of the battery+the screen, all in a housing that is submersible to at minimum of 11ATA, preferably 16ata to satisfy deep technical diving. The screens have to be backlit for cave diving as well as night diving/deep ocean diving, the units can't be any larger or heavier than the current 2" SPG's, and they have to provide a compelling reason to go away from the analog spg's. 2" SPG's are the "standard" for backmount, 1.5" slimlines are becoming the standard for sidemount. Pricing is $50-$80 retail, and standard industry markup is 100%, so you have to produce them for $30-$40 ish to make it profitable for the dive shops. You also have to make the battery easy to change, design it so it is long lasting *must last minimum of 100 hours or so for most divers to consider it worthwhile and the display must also show battery readings so tech divers can change it out at somewhere around 50% battery life*.

If you think this is possible you might see divers transition, but unfortunately I don't think it is and the risk/reward is basically I risk going from an analog gauge that always works but may be inaccurate, to a very accurate gauge that might have an electrical fit and cause me to have to turn the dive and then pray I can get a replacement quickly.

Now, with that all being said, that is only applicable for digital SPG's only, at that point you might as well integrate it into a dive computer which is where you'd make money anyway and now you have a lot more room to play with price wise with the average AI dive computer, non-wireless, being around $500. You already have the pressure sensor, already have the display, already have the housing, might as well just add a clock, depth sensor, and small chip for an algorithm to keep track of everything, Mares Puck, Oceanic Veo 3.0 etc. Now you can actually make money because these computers are only about $100-$150 more than their non-AI counterparts, and when you have to spend $70-$80 on an SPG anyway, you may as well spend a little more and get the more accurate pressure gauge as well as the instant SAC rate etc.

Where this breaks down is when you look at the way the industry is trending right now. Cave/Technical diving aside which have no use for these types of computers due to gas switching *most of those style computers computers allow decompression, but backgas only, the Veo 3.0 does allow gas switching, but doesn't monitor the other tank pressures, so you still have analog gauges on your other bottles*. The big thing right now is sidemount diving, you have divers with two tanks that are being alternated frequently, you can go to a wireless computer which brings all of the discussion points above to the party, but now you have to have two very expensive wireless transmitters and have to hope and pray the link stays connected and you're now into very expensive parts, very expensive computer modules, and there just isn't a good way to do it where the tech divers will buy into it. Not trying to burst your bubble, just giving you the reality of why this technology hasn't taken off and isn't really being looked into. The ROI just doesn't work for the R&D involved in developing a module, combined with the marketing problems of getting divers to buy into it.

 

[Nemrod]

I just look at about where the needle is and the rate it is falling over time. Beyond that, I care not at all.

N

 

[Colliam7]

My core question remains: Why is accuracy information not published? Why are SPGs not tested for accuracy on a schedule such as regs and valves and tanks?

The two questions are reasonable, and well worth considering.

From a manufacturing perspective I cannot answer the first. Several have suggested that such information may reflect unfavorably on a particular product, and that sounds as good as anything I can come up with. Or, it adds a step to the production process (although I would presume that manufacturing quality control already includes sampling testing of each production batch), which adds to production cost. Would the availability of such data influence consumer purchases of SPGs? I don't know, am just asking the question. It would seem to me that a SPG line which supplied accuracy data would have at least some degree of marketing edge. But, maybe not.

The second question is interesting from a dive shop perspective. Let's assume that I want to test all SPGs attached to regulators brought to my shop for periodic service. Purchase of the calibrated, digital gauge will mean an investment of perhaps $400-500. That isn't a huge amount, but I want to be sure that I recoup my costs. And, let's say I want to do that over 12 calendar months, as I plan to periodically replace the gauge. If I service 100 regs / year, I am going to add $4-5 to each service bill, which is also not a lot. A bigger question is, what will I - and my customer - do with the data? If I give them data that show the divergence between my NIST calibrated gauge, and their SPG, at several pressures (measured with my calibrated gauge), say 500 psi, 1500 psi, and 3000 psi, and the divergence of their gauge is 100 psi above at 500 psi, 182 psi below at 1500 psi, and 397 psi above at 3000 psi, what would I recommend they do? I am asking questions not to refute or deflect the premise of your post, but to 'think out loud' as much as anything.

 

[Stoo]

Ok. Let's unroll that a bit.

Let's be clear... Accuracy and precision have no place in diving... ;-)

 

[RJP]

It would seem to me that a SPG line which supplied accuracy data would have at least some degree of marketing edge. But, maybe not.

Absent standards or comparative data... one branding having accuracy data would mean very little. Unless it showed 100%.

Practically however, what does "accuracy" really mean when we're talking about a volume that changes significantly relative to things like temperature. What good is knowing that my hot fill is exactly 3750psi when it's going to be 3400 when it cools, 3200 when it sits in my car overnight, and 3100 when I get in the water. When I get out it will say 1250. After sitting in the sun during the surface interval it will then be 1,420. Until I jump in the water when it will drop back to 1250.

Then there's the whole "precision" thing. I love when you ask someone their pressure and in a blur of flying fingers they let you know that they have 1,874psi of gas left.

As long as I know my SPG isn't horribly off I'm pretty much OK with an understanding that the resolution of such a device includes roughly four amounts of gas:
- Starting pressure
- About half of starting pressure
- About a quarter of starting pressure
- Empty

Relying on greater precision than that is an exercise in folly. (Especially with 49yr old eyes in low-light, low-vis settings...)

 

[raftingtigger]

Maybe I'm being really simple minded. I look at my pressure when I set up my gear, then take a breath or 2 from each regulator, fill my wing, and then close the tank valve and breathe the air out and make sure the SPG goes to 'near' zero.

This does several things:
1. Checks all my regulators
2. Checks my Wing and inflator
3. Checks my SPG at the top end and the bottom end
4. Keeps from accidentally emptying my tank in the time (often considerable) between set-up and diving.
5. Makes it obvious the air is off.

Seems to me if the gauge goes to zero as the last of the air hisses out I know that at least the bottom end is somewhat informative.

 

[tbone1004]

1. yes
2. yes, though you should be orally inflating your wing for positive pressure checks and then tap the inflator a couple times to verify the OPV is working
3. no it doesn't, it just makes sure that it doesn't get hung up at any point in its cycle
4. yes, though you should be leaving your gear pressurized with the valve turned off during setup to make sure nothing gets into it
5. yes, which is the negative pressure check you're supposed to do with the dust cover on, but off is off on the tank valve

 

[Colliam7]

Absent standards or comparative data... one branding having accuracy data would mean very little. Unless it showed 100%.

And, that's why I added the 'maybe not'. However, perception (of greater accuracy) not reality (of greater accuracy), may be the key. A wise philosopher recently wrote, 'that's the very essence of marketing! The perception of my shoes is far more important that the reality of my shoes.'