Pressure Gauge Errors

[Brooothru]

So, I've got a perplexing problem with pressure gauge accuracy that's been bugging me for a while, and it's finally come to a head. I'm hoping someone can offer an explanation and possible solution. I've got a pressure fermentation currently in progress in a Unitank using WLP-808 "Mythical Hammer" mixed blend lager yeast that calls for warm temperature (63F) @ 1 BAR pressure. I'm monitoring the pressure with the tank's original equipment pressure gauge similar to this:

I'm controlling the pressure with a Spike Brewing combination spunding valve:





The Spike valve is a great innovation that combines pressure readings with spunding capabilities, a Gas In port for CO2 additional pressurization, and (most importantly) a built-in large disc PRV that dumps whenever tank pressure exceeds about 16-17 psig., all connected to a single 1.5" TC port. I've used the Spike valve several times as a stand-alone device to free-up a couple of TC ports on my unitank as well as simplify the installation and clutter of the various components during fermentation. I also have a Spike manifold:

...which was essentially replaced by the Spike spunding combination device.

Here's the problem. The gauges on both Spike devices appear to have significant errors. I first noticed this on the manifold device when I used it during forced carbonations. The beers seemed flat when I'd initially tap a keg, but later, after a few days in the kegerator, they seemed fine. The cause never got chased down since the problem 'solved' itself. Later after purchasing the Spike spunding valve, I started noticing that the PRV was releasing prematurely at around 10-12 psig instead of the designed release point of 16~17 psig. I guessed that the internal spring mechanism was weak, thus allowing the PRV disk to become unseated at a lower pressure. Not a big deal, since I would normally shoot for around 12 psi @ 48F when spunding, but it was an annoyance since, when cold crashing post fermentation, the carbonation level would drop below the desired 12 psi for volumes of CO2 at the cold crash temperature (38F). Once again, a few days in the kegerator fixed the deficiency.

So during this current fermentation, I mounted both the OEM pressure gauge and the Spike spunding valve on the unitank. The OEM gauge started showing a modest pressure increase not long after the 12 hour mark following the yeast pitch. There was a small volume of bubble activity in the spunding cup, but no indication of pressure rise in the Spike gauge. Eventually the Spike gauge began to show some increase by Day 2, but lagged far behind the OEM gauge. By Day 3 the OEM gauge registered 15 psig/~1 BAR, but the Spike gauge only showed 10.5 psig. But which gauge was indicating correctly?

Intuitively, it would appear that the Spike gauge was in error, since the difference of ~4.5 psig was about the same pressure that the over-pressure PRV was opening 'early'. But intuition doesn't constitute proof, so I attached two different Gas In post mounted inexpensive pressure gauges. Both they and the OEM gauge were all in agreement +/- 0.5 psi. It looks like the Spike gauge is in error by 4~5 psi, which is certainly not an insignificant value in this application. In retrospect the gauge on the Spike manifold is probably off by 3-4 psig as well. Maybe just a coincidence, possibly a random chance that I happened to get a couple of bum gauges from a very reputable vendor (Spike Brewing). who make some very high quality and innovative products.

One thought did occur to me. The OEM gauge and the two Spike gauges are both liquid filled (glycerin) damped gauges. The hand-held Gas In port gauges are cheap ($10) NPT mount dry gauges. There is no provision for periodic recalibration of either type of gauge. The liquid filled gauges however all come with fill ports with neoprene caps that are removable. Somewhere in my distant past memory bank I seem to recall that when they are filled and the cap is replaced, you 'seal in' the local barometric pressure. If that pressure is either abnormally high or low (deviating significantly from Standard 1013.4 millibars at sea level), then there will be an inherent error in any subsequent readings. If for instance the Spike gauges were manufactured and assembled in Denver (5230') on a low pressure day, they might indicate in error on a high pressure day in Death Valley (-200'). Don't know if that error could be 4.5 psig, hypothetically. But being cheap by nature, if I can pop the cap on these gauges to equalize the ambient pressure, I could save myself a few 'Benjamins' on replacement gauges.

Is this possible, or am I once again experiencing age-related misremembering induced hallucinations?

 

[Deadalus]

I just replaced a pressure gauge on a Buon Vino Filter and I bought a glycerin filled replacement since it has a pump and maybe would be vibrating some. There were instructions to pop the plug off and then put back in (which I did). I don't remember however why it said to do that, but if I find the instructions I can possibly corroborate. (Offhand I believe you are correct as to the reason for doing it, but it could be group psychosis!)

I also have a Spike spunding valve. Haven't used it yet but I will keep an eye on it and report back if I notice any differences. I have another spunding valve and then I put a pressure gauge on a QD so I also have alternatives but as you noted, it may be hard to tell which one is right!

 

[Brooothru]

I just replaced a pressure gauge on a Buon Vino Filter and I bought a glycerin filled replacement since it has a pump and maybe would be vibrating some. There were instructions to pop the plug off and then put back in (which I did). I don't remember however why it said to do that, but if I find the instructions I can possibly corroborate. (Offhand I believe you are correct as to the reason for doing it, but it could be group psychosis!)

I also have a Spike spunding valve. Haven't used it yet but I will keep an eye on it and report back if I notice any differences. I have another spunding valve and then I put a pressure gauge on a QD so I also have alternatives but as you noted, it may be hard to tell which one is right!

Thanks for the reply. I couldn’t find the instructions or quick start guide for the Spike spunding valve. Guess I should just go to their website since it’s probably posted there. Duh! Why didn’t I think of that?

I really like the concept and design of the Spike unit. There are a few minor gripes I have that I would like to see though. First, I’d make the cup deeper. If you depress the PRV release, the sudden pressure dump will empty not only the tank pressure, but also all the liquid in the cup, mostly on yourself and the surrounding area! I’d make the sides of the cylinder taller.

Likewise, the drain pipe for the cup is awkward to use when the valve is installed on your fermenter, almost to the point of being nonfunctional. A flexible drain line might be better, so the discharge could actually be collected rather than spilling on the fermenter.

I’d also increase the PRV release pressure slightly to 16~17 psi. I’m assuming that Spike set it at ~15 psig since that’s the normal rated operating maximum pressure (not maximum) of their Flex fermenters, but 15+ psi is pretty much the nominal pressure setting for forced carbonation and pressure fermentations. Having 15 psi as a limiting pressure doesn’t give much ‘maneuvering room’ for normal tasks like resetting spunding pressures.

With those small tweaks it would be pretty much the perfect add-on TC device. I really like mine, and all its versatility.

 

[whoaru99]

The liquid filled gauges however all come with fill ports with neoprene caps that are removable. Somewhere in my distant past memory bank I seem to recall that when they are filled and the cap is replaced, you 'seal in' the local barometric pressure. If that pressure is either abnormally high or low (deviating significantly from Standard 1013.4 millibars at sea level), then there will be an inherent error in any subsequent readings. If for instance the Spike gauges were manufactured and assembled in Denver (5230') on a low pressure day, they might indicate in error on a high pressure day in Death Valley (-200'). Don't know if that error could be 4.5 psig, hypothetically.

On the presumption it actually works that way, and not sure either way, 4.5 PSI would be a stretch given the difference between the highest and lowest ever recorded barometric pressures is ~3 PSI.

Considering a range of more normal min/max barometer, it's around 1 PSI.

 

[Brooothru]

On the presumption it actually works that way, and not sure either way, 4.5 PSI would be a stretch given the difference between the highest and lowest ever recorded barometric pressures is ~3 PSI.

Considering a range of more normal min/max barometer, it's around 1 PSI.

No. Barometric pressures are corrected to average mean sea level (29.92” mercury /1013 mb). The lowest barometric pressure recorded atop Mt. Everest is 9.3”/315 mb, so a difference of about 20” Hg/800 mb.

800 mb is equivalent to 11.6 pounds force per square inch. 4.5#^2 (my observed psi discrepancy between gauges) is equivalent to about 310 mb. Average baro pressure in Death Valley is 1026 mb/14.7” Hg. Average baro in Denver is 615 mb/25.3” Hg, a difference of 411 mb.

That difference equates to 5.96 psi, so the plausibility of my example may be at the extreme margins, but the possibility is well within normative values. Does it ‘prove’ my hypothesis? No. Though it does suggest a causative approach to explaining my gauge errors.

 

[whoaru99]

No. Barometric pressures are corrected to average mean sea level (29.92” mercury /1013 mb). The lowest barometric pressure recorded atop Mt. Everest is 9.3”/315 mb, so a difference of about 20” Hg/800 mb.

800 mb is equivalent to 11.6 pounds force per square inch. 4.5#^2 (my observed psi discrepancy between gauges) is equivalent to about 310 mb. Average baro pressure in Death Valley is 1026 mb/14.7” Hg. Average baro in Denver is 615 mb/25.3” Hg, a difference of 411 mb.

That difference equates to 5.96 psi, so the plausibility of my example may be at the extreme margins, but the possibility is well within normative values. Does it ‘prove’ my hypothesis? No. Though it does suggest a causative approach to explaining my gauge errors.

OK, well, I was going with generally plausble situations since I presume you're not really trying to imply your gauge was built on top of Everest and you're using it in Death Valley type of deal.

 

[FunkedOut]

Pressure fermenting in a Spike conical Unitank.
Using the same Spike manifold with gauge, safety PRV and gas port.
Using a different spunding valve that is adjustable.

I noticed I could not get the pressure much above 10psi, and after testing a bit, I found the safety PRV on the manifold (fixed at 15psi) was releasing early.
Called Spike and they provided a new PRV. New one releases at 15psi.

Also, I see up to 3psi difference between gauges across the fermenter, spunding vlave, CO2 tank, kegerator. It's close enough for government work. I'm OK with the beers bumping up that last bit of carbonation during the first few days of being tapped.

 

[Brooothru]

Pressure fermenting in a Spike conical Unitank.
Using the same Spike manifold with gauge, safety PRV and gas port.
Using a different spunding valve that is adjustable.

I noticed I could not get the pressure much above 10psi, and after testing a bit, I found the safety PRV on the manifold (fixed at 15psi) was releasing early.
Called Spike and they provided a new PRV. New one releases at 15psi.

Also, I see up to 3psi difference between gauges across the fermenter, spunding vlave, CO2 tank, kegerator. It's close enough for government work. I'm OK with the beers bumping up that last bit of carbonation during the first few days of being tapped.

It's reassuring to know that I'm not the only one with pressure issues, re: the Spike spunding valve combo system. In my case I believe that the PRV is probably releasing at or near the proper (for this device) point of ~15 psi, but that the pressure gauge is in error by about 4.5 psi on the low side. I.e., the gauge reads 10.5 psig (incorrectly) and the PRV is releasing (correctly) at its 15 psi set point. I'm basing this after measuring the Spike gauge combo value against three other independent pressure gauges (one, the expensive oil dampened OEM pressure gauge, and the two separate cheap handheld gas port mounted gauges).

The three non-Spike gauges agree within +/- 0.5 psig. The difference between the average of these three gauges and the Spike gauge is 4.5 psig. For a home-brew application, +/- 0.5 psig would be well within "government use" error parameters, but -4.5 psig on the low side is around a 35% error towards unsafe operating pressure levels. Most mechanical/structural operating limits follow general protocols of Normal Stress Limit (100%) is just what it says: "normal" maximum limit, for everyday use; Overstress (125% of Normal Stress) with possible weakening, hidden deformation and structural compromise; and Limit Load (150% of Normal Stress Limit) where one or more structural elements will likely fail possibly causing the entire structure to fail catastrophically. In the case of a pressurized unitank, if you are pressurizing to 15 psi(g) but the gauge you are relying on is reading 4.5 psi low, you're already at 19.5 psi unless the PRV is working properly, and well into an over stress situation at risk of structural compromise.

The kicker is, the hidden damage of these overstress situations is cumulative over multiple cycles. That's why there are service life limitations on many structural components (cycles) and on the entire structure itself. Every time you unknowingly wander into the range of Overstress, you unwittingly remove some of the stucture's (or it's components) ability to withstand future stresses, and may also cause weakening that can permit catastrophic failure before even reaching Limit Load. Think: the Hawaiian Airlines 737 that blew off about 25' of its roof flying between Kona and Honolulu. In its service life, it had seldom come anywhere close to Overstress or Limit Load pressurizations, since it seldom climbed above 15,000' cruising altitude between the islands. But it had hundreds if not thousands of pressurization/depressurization cycles of its fuselage due to its very short (15~:20 minute) stage length flights. Those cumulative cyclical stresses caused hidden metal fatigue that resulted in catastrophic failure at a point far below its Limit Load.

It's easy to become complacent with ubiquitous items containing fluids under pressure, like our friendly unitanks. But they can and do pose significant danger if proper safety precautions and practices are not followed. If you pump up a 15 psi rated tank of fermenting wort and your PRV fails/clogs/didn't get installed, pretty soon you'll exceed its 21 psi Limit Load, and then all bets are off. But two things will be true: 1) you'd better evacuate promptly; because, 2) it's guaranteed to leave a mark.

Anyone know of a good gauge supplier?

 

[Brooothru]

OK, well, I was going with generally plausble situations since I presume you're not really trying to imply your gauge was built on top of Everest and you're using it in Death Valley type of deal.

Not Everest, per se, but there is a lot of manufacturing in Denver, and most of the coastal and nearby inland population centers in the U.S. are within 300' of the floor of Death Valley, as well as the Dead Sea. So is it really such a stretch to think that a pressure sensing device might be assembled in a region of higher elevation (Salt Lake, Albuquerque, Mexico City, Calgary, etc) and get used by a customer in Houston, Tampa, New York or Los Angeles?

Actually, I just came across a You Tube video produced by a supplier of pressure gauges which explains how and why to reset a liquid filled pressure gauge. It does recommend venting (equalizing) the internal pressure of the gauge by temporarily unseating the fill cap to obtain ambient pressure inside the gauge that matches local conditions. This makes perfect sense since the gauge is measuring a displaced applied force pressure with its own internal pressure. If its internal pressure is significantly different from the ambient pressure outside the gauge, the result will be inaccurate readings.

The example (gauge) in the video showed an 'unloaded' pressure of around 4 psi. The narrator described the error as being caused by significant variations of temperature and/or pressure that occurred during shipping. Once again, this makes perfect sense and does not require a leap of faith (like my Denver-Death Valley scenario) to explain the anomaly. All that is needed is having the gauge shipped by air delivery from Spike to the consumer in a cold or unpressurized cargo hold for the ambient pressure inside the gauge to be altered.

The corrective cure is to simply vent and reseal the glycerin fill cap. Interestingly this condition only occurs with sealed liquid filled pressure gauges since the non-liquid filled variety of pressure gauges are not sealed and both sides of the pressure diaphragm are always exposed to whatever ambient pressure is present. Ya' learn something new every day, even when it's something you probably shoulda' known all along. Oh well. Hopefully this will solve the discrepancy.

link: https://www.youtube.com/@DirectMaterialcomIrving

 

[whoaru99]

Didn't watch the videos due to location presently with limited bandwidth.

But, a dry gauge would address the issue and I'm aware that Wika for sure, and likely others, makes liquid filled gauges with open/closed valves on them to facilitate equalization.

Of course, that type of venting valve is generally only on gauges where the span and accuracy class dictate its necessity to keep the gauge meeting specifications.