What Packing Does That a Mechanical Seal Cannot

Why this matters

Packed pumps are routinely destroyed by someone trying to help. Somebody sees a drip, tightens the gland until the drip stops, and walks away pleased. That drip was the only liquid cooling and lubricating the packing, and stopping it turns a consumable into a heater that scores the shaft. Packing is not an older, worse version of a mechanical seal; it is a component with a different bargain, and the bargain is that it leaks on purpose in exchange for surviving six conditions a face film cannot. Knowing which of those six you have is what stops a shop from converting a pump that was correctly packed and repacking a pump that should have been converted.

Before you touch a gland

Adjusting a gland on a running pump is the normal and correct method, and it puts a wrench within reach of a rotating shaft. The coupling guard stays on. The only things reached are the two gland nuts, with a wrench, from outside the guard. Sleeves buttoned, no rings, no loose or oversized gloves, nothing that can be drawn in. Anything beyond those two nuts means stopping the pump and locking and tagging the disconnecting means under 29 CFR 1910.147 rather than switching it off.

The drip is at system temperature. On hot service it is a scald source and it must be routed to a drain rather than allowed to run down the base. Cracking a gland that has been over-tightened releases whatever was trapped behind it: stand out of the plane of the gland, wear a face shield rather than glasses, and expect a surge rather than a drip.

Never judge gland temperature by hand. Read it with an infrared thermometer from arm's length, outside the guard.

Repacking is a stopped, locked-out, depressurized, drained job. Old packing on any service other than clean water is contaminated by whatever the pump handled and is handled per that fluid's SDS; if the rings have dried out and are going to be picked at with a hook, the exposure route is inhalation and the control is respiratory under 29 CFR 1910.134, not a glove.

What packing is and how it seals

Rings of braided material sit in a bore around the shaft called the stuffing box. A gland follower squeezes them along the shaft axis, and because the material has nowhere else to go, that axial squeeze becomes radial pressure against the shaft or against a replaceable sleeve fitted over it.

That is the whole mechanism, and the important consequence is that packing seals by restriction, not by barrier. It does not stop flow; it makes flow difficult. There is always a path from the box to atmosphere, and the only question is how much goes through it.

The pressure gradient decides which way that path runs, which matters more than it sounds. On a pump with positive box pressure, liquid comes out. On a suction-lift pump the box can sit below atmospheric, and then the same path draws air in, silently, with the gland bone dry. That is why a packed box on a lift needs a lantern ring: a spacer ring partway down the stack, fed by a flush at a pressure above box pressure, so the flow across the inner rings is always outward. The same arrangement on an abrasive service is what keeps grit out of the packing, for the same reason.

The leak is the design

Two jobs are done by the liquid moving through the packing, and only one of them is obvious.

It lubricates the interface between the packing and the shaft, which is a sliding contact under real radial pressure. And it carries away the friction heat generated at that contact, which is the job nobody credits it with and the one that fails first.

Stop the drip and both jobs stop. The interface runs dry, the temperature at the shaft climbs, the packing glazes and then chars, and the shaft or sleeve scores. The next repack lands on a grooved surface and leaks from the first hour, which is how a shop ends up believing packing does not work.

A common field target after break-in is a countable, steady drip, on the order of tens of drops a minute rather than a stream and never nothing at all, with more on hot or high-speed service. The pump or packing manufacturer's own figure governs, and it is worth asking for, because the number varies with shaft diameter, speed and pressure. For scale: at 45 drops a minute, with a drop on the order of a twentieth of a millilitre, that is a bit under a gallon a day. A bucket a week, not a puddle.

The six conditions it survives that a face film does not

Abrasives. A seal's film is thinner than the particles a suction line routinely carries, so grit that reaches the faces scores them permanently and the seal is finished. Packing wears against grit and keeps sealing, and with a lantern ring flushed above box pressure the flow across the inner rings is outward, so the grit never arrives.

Dry starts and momentary loss of liquid. A face film that flashes heat-checks the hard face in seconds, and heat checking does not heal. Packing run dry gets hot, squeals and glazes. It is damaged rather than destroyed, and it announces itself audibly before it is.

Shaft deflection and runout. A face pair has to track a face that is moving, and there is a limit, commonly a couple of thousandths of an inch of total indicated runout at the seal location, with the seal manufacturer's figure governing. Packing is compliant material and follows a shaft a face pair would reject.

A worn shaft or sleeve. Packing conforms to a surface a seal will not accept, within reason, which is what keeps older pumps in service.

Intermittent, seasonal or unattended duty. A seal that has sat dry through a shutdown may come back with a set elastomer and faces stuck together. Packing that has sat idle is just compressed material: it takes one adjustment and runs.

Field repair with hand tools. Rings are cut, wound and set with a packing hook and a pair of hands. A cartridge seal has a setting length, often a setting device, and a way of being installed wrong that looks perfect from outside.

What the bargain costs

Continuous leakage, which disqualifies packing outright on some services. Anything toxic, flammable, regulated for discharge, or costly to lose is not a packing application at any drip rate, and that judgement is made before any of the six conditions above are considered.

Wear on the shaft. The shaft or sleeve is the wear partner, which is why a sleeve is fitted: so the wear lands on a replaceable part rather than on the shaft itself.

Shaft power. Gland friction is real work and over-tightening is where it comes from. A gland cranked down to stop a drip is charging the motor for the privilege.

Attention on a cadence. Packing is a maintained component. A shop that cannot get to it, or will not, has chosen a technology it is not going to support.

Worked artifact: one break-in log

A general-service pump, water at 75 F, ambient 78 F, freshly repacked. Gland surface temperature read with an infrared thermometer at the same marked spot every time, from arm's length, guard on. Drips counted for a full minute each time.

Read the series literally. There were three adjustments, at 15, 30 and 45 minutes. The drip fell at each of the three and then kept easing on its own after the last one, from 55 to 48 to 45. The gland surface rose at each of the three and then stopped climbing within thirty minutes of the last one, moving only one degree between 60 and 90 minutes.

The correction that has to be visible. The number that matters is not 107 F, it is the rise over the pumped liquid, because the packing can only ever be cooled to something above what is flowing through it. That is 107 minus 75, or 32 F. Reading it against the 78 F ambient instead would have given 29 F, which is the wrong denominator on a hot service and badly wrong on a chilled one.

The second correction. An infrared reading of a painted or bare metal surface carries an emissivity error, which is systematic for a given instrument on a given surface. Because every reading in the log came from the same instrument at the same marked spot, that error is common to all of them and largely cancels in the rise. It would not cancel between two different spots or two different finishes, which is exactly why the log names one spot rather than "the gland."

The stopping rule the log demonstrates. Adjustment stopped at 45 minutes, with a countable 55 drops a minute, because the temperature had begun to level and the drip had entered the band. It did not stop when the drip stopped, because the drip was never going to be allowed to stop.

What would have made this a failure. If at 90 minutes the surface were still climbing, the correct action is to back the nuts off and let the drip come back up, not to wait and see. A gland still gaining temperature half an hour after the last adjustment is generating more heat than the leakage can carry, and every additional minute is taken out of the shaft.

And the failure mode in the field: somebody arrives at 60 minutes, sees 48 drops a minute, decides that is untidy, and takes another flat and a half. The drip goes to nothing, the log stops being kept because there is nothing to count, and the next person to look at this pump finds a scored sleeve and a charred stack and concludes packing was a bad idea.

How to verify you got this right

Come back at four to eight running hours and count again. Packing beds in, so a stack that settled at 45 drops a minute will usually have eased down on its own, and the correct response to a rate that has drifted low is to back off, not to leave it. That second visit is the one shops skip and it is the one that decides how long the stack lasts.

Then pull the shaft record. If the sleeve was measured before repacking and gets measured again at the next repack, the difference over that interval is the honest score for how the gland was run. A sleeve that lost measurable diameter in one packing life was run too tight regardless of what anyone remembers about the drip.

Finally, write the drip target on the pump itself, on a tag, in drops per minute. The gland is adjusted by whoever walks past it, and a target that lives in a manual is a target that loses to the first person who thinks a drip is a defect.

References

• 29 CFR 1910.147, control of hazardous energy, for lock and tag plus pressure relief before repacking or any work beyond the gland nuts
• 29 CFR 1910.134, respiratory protection, where old packing is disturbed on a service whose SDS names an inhalation route
• Pump and packing manufacturer documentation for the drip rate, the box temperature limit, the ring count and the correct packing material for the service
• See related: What a Mechanical Seal Is Actually Sealing Against; How to Decide Between Repacking and Converting to a Seal; What Air Entrainment Does That Cavitation Does Not