A thread from a press plant where the return-line filter was the monthly scapegoat until somebody broke open the cartridge and read the delta-P log. Names trimmed, ISO codes left real.
From: F. Costa, Maintenance — To: R. Almeida, Design; cc hydraulic technician
Subject: HPU-03 return filter alarm again, 4th cartridge this quarter
Ricardo, the 160-litre hydraulic power unit on the double-acting press keeps setting the return-filter differential switch, four cartridges in three months on a circuit that used to eat one a year. The element is our standard 10-micron return filter with a bypass, the switch is set at 1.5 bar delta-P, and it alarms every three weeks or so, the gauge freezes right under the switch setting when the press is cycling hard and resets when the unit idles. Oil tests show the ISO 4406 code has crept from the commissioning 16/14/11 to 18/16/13, the lab flags water at 450 ppm, and the reservoir breather looks clean enough that I stopped suspecting it. The line manager wants me to fit an oversized return filter or throw finer elements at it, and I want to know whether I am treating a symptom, because a return filter that fills monthly is usually telling you the tank has a particle source, not a weak cartridge.
From: R. Almeida, Design — To: M. Berg, Filter Supplier Applications; cc F. Costa
Subject: RE: HPU-03 — particle ingress investigation before we resize the filter
Monica, I have a hydraulic unit clogging its 10-micron return filter on a three-week cadence and a lab code that moved two classes in five months, and I want the contamination-source logic, not a bigger element. The suspects the shop has already cleared: reservoir breather looks clean, cylinders were rebuilt last year so no shaft scoring visible, and the filter housings are clean at change. The two things nobody has done: count the particles that come back with each piston on a cold start, and check the charge, because the makeup pump breathes through the tank and if the fill point is open during filling, that is a classic way to pull 60 steel and silica particles per millilitre straight into the reservoir on every fluid top-up. I would like your standard worksheet: where particles enter a 160-litre unit at 210 bar, in order of likelihood, and the sequence of checks that either clears or convicts each one, because the shop will only do the checks in the order you write them, and if you hand them a lecture they will burn it.
From: M. Berg, Filter Supplier Applications — To: R. Almeida; cc F. Costa
Subject: RE: HPU-03 — the worksheet, in the order you run it
Ricardo, here is the order, and there is method in it, you rule out the sources that dump a lot of particles fast before you chase the ones that dribble. On a 160-litre unit at 210 bar these are, in likelihood order, and I am folding your cleared suspects back in because cleared-on-looking is not cleared-on-measuring.
[1. Make-up and filling practice] Measure the fill point first. If the tank is filled through an open port or a drum pump without a desiccant filter, every top-up is a contamination event, 20 microns and up, steel and dust, straight in. The lab count at 18/16/13 with a hump in the larger sizes reads exactly like fill event, not pump wear, pump wear shows a hump at the small end. Question for the shop: how is the fluid topped up, and on what schedule.
[2. Reservoir breather] Your shop cleared it by eye, but a breather that looks clean can still be wet-loaded, if it has a desiccant cartridge check the colour, and measure the vacuum in the tank above the pump: 20 millibar vacuum at start draws air and whatever the breather lets past. Replace breather at the element interval, not when it looks bad.
[3. Cold-start ram and cylinder rod wipers] On a cold start the rods wipe down the outside dirt, if the wiper is worn the return stroke carries that dirt straight back into the tank, no filter sees it until the return strainer, and the return filter eats the rest. Check wiper lip for cracks and rod surfaces for pick-up.
[4. The filter itself and its beta] Your element is 10 micron nominal, which usually means beta x = 200 at 10 micron, meaning 1 particle in 200 passes. That is fine for a return filter once the source is gone, but it will never rescue a tank that keeps being fed, so do not level a cartridge at an open fill port.
Do these four checks in exactly this order, one morning, and bring me the delta-P log from the switch and the fill log from the machine logbook, because the switch reading over a cold start tells me whether the element is loading from a source that acts every cycle or only at fill, and the two tell different stories.
From: F. Costa, Maintenance — To: R. Almeida, M. Berg; cc hydraulic technician
Subject: RE: checks done — the fill port is the criminal, and the vacuum is the witness
Ran the four checks in your order. Fill practice: convicted immediately, the drum pump sits on the fill flange without a filter and the port is open to the floor dust, and the logbook shows a fill event every two to three weeks, exactly the cadence of the alarm, so the filter has been eating the same dirt over and over. Breather: measured 35 millibar vacuum at the pump inlet on a cold start, so it is drawing air past the seal, and the element on the breather is older than the machine record, replaced it today. Cold start delta-P log: the switch shows a slow ramp, not a step, meaning the load builds across the shift, consistent with a trickle of dirt, not a burst. Wipers: all four rods clean, no pick-up. So the answer is the fill port, and I will say it out loud so it is on the record: we have been topping up a 160-litre unit through an open drum and blaming a 10-micron filter for doing its job.
From: R. Almeida, Design — To: F. Costa, M. Berg; cc plant engineer
Subject: RE: fix, and the cleanliness target we should hold
Agreed on the fix, and let us make it complete so we are not back here in June. Three actions, in order. First, stop the open fill today, the drum pump gets a 3-micron absolute fill filter with a desiccant pre-stage, and the fill log goes into the machine logbook with a timestamp, one fill event per line. Second, replace the return cartridge now, this one started contaminated and it is 80% loaded by the look of the gauge, and set the differential switch at 1.2 bar so we hear trouble before the bypass cracks. Third, set the oil cleanliness objective on paper: ISO 4406 16/14/11 at commissioning, and I want 17/15/12 as the running limit after the breather and fill filter land, with a lab check at the next oil change to prove the tank has settled. Monica, confirm the beta rating on the fill filter element reads 200 at 3 micron or better, because if the fill filter passes 10-micron dirt on the way in, we have simply moved the problem ten minutes upstream.
From: M. Berg, Filter Supplier — To: R. Almeida, F. Costa; cc purchasing
Subject: RE: fill filter confirmed — beta 200 at 3 micron, drawing attached
Confirmed, the fill filter media is beta 200 at 3 micron measured to ISO 16889, desiccant breather matches the 160-litre flow, and I am adding a coarse strainer on the drum pump inlet because drums spend their life on wet floors and the bottom third carries settled grit that would otherwise wear the filter out ahead of schedule. Part numbers attached. One number to hold the whole story together: you were feeding the return filter about 65 milligrammes of dirt per fill event by the look of your lab count drift, and a 10-micron beta 200 return element holds roughly 80 grammes before bypass, so one drum pump blow-in a month outran the filter by a factor of about five, which is why the cadence was three weeks and not six months. Resize nothing; fix the fill and the filter keeps its original duty.
From: R. Almeida, Design — To: F. Costa; cc plant engineer, M. Berg
Subject: RE: closing out HPU-03 — one table for the training folder
Closing HPU-03, and here is the table I want in the annual maintenance refresher, because the same argument comes around every machine. It lists where contamination gets in before it ever sees a filter, in the order this unit was caught.
| Source | Tell | Check |
|---|---|---|
| Open fill port / drum pump | 2-3 week clog cadence, hump at large sizes | fill log, fill filter installed |
| Wet breather | vacuum over 20 mbar at start | desiccant colour, element age |
| Rod wipers | rod pick-up, grainy return oil | wiper lip, rod surface |
| Pump/cylinder wear | small-particle hump, steady rise | wear debris analysis |
For HPU-03 specifically: fill filter beta 200 at 3 micron fitted, breather replaced, cartridge renewed, switch moved to 1.2 bar, running limit 17/15/12, lab proof at next oil change. The filter was never the problem, and it was never supposed to be the problem; a return filter is the last gate, not the first line of defence.
Reference data — 160 L unit, 210 bar; return element 10 µm, beta 200, bypass; switch set 1.5 then 1.2 bar; oil ISO 4406 16/14/11 commission, 18/16/13 degraded, running target 17/15/12; fill filter beta 200 @ 3 µm; breather vacuum measured 35 mbar; clog cadence ~3 weeks vs historical 1 year.
Glossary — ISO 4406: the three-number cleanliness code abbreviating particle counts per mL at 4, 6 and 14 micron; beta ratio: the ratio of upstream to downstream particles at a given size, beta 200 meaning 1 in 200 pass the element; differential pressure switch: the device that alarms when the element loads to the set pressure drop; bypass valve: the spring valve that cracks to route oil past a clogged element; desiccant breather: a breather that dries intake air and filters dust in one cartridge.
Field note — the filter that got blamed for the oil
A return filter that fills on a rhythm is evidence of a source, not a symptom to be resized. The lab code, the vacuum reading and the fill log convict the real culprit, an open drum pump dumping the first floor grit into a 160-litre tank every three weeks, and no cartridge, however big, outruns a fill event done on a schedule. Order the checks by particle volume: fill practice first, breather second, rod wipers third, wear fourth, and only then talk about element grade. Clean the source, hold the cleanliness class on paper, and the original filter finishes its original job, which you will know because the alarm moves from a three-week rhythm to a yearly one, and the oil test stops moving.
From: F. Costa, Maintenance — To: R. Almeida; cc M. Berg
Subject: RE: fill filter installed and first fill event logged
Fill filter and desiccant breather are in, the drum pump strainer is on, and the first top-up since the change went through the new filter with the timestamp in the logbook. The differential gauge has sat under 0.4 bar all week through full press cycles, which is where it used to sit when the unit was new, and I have told the night shift that an open fill port is now a logged near-miss rather than a habit. The lab will draw a sample at the next oil change and we will hold HPU-03 to 17/15/12 on paper, as agreed.
Endpoint of the thread: the machine logbook now carries a one-line fill record per event, and the return filter is back on its annual cadence. The 10-micron cartridge that everyone blamed for the oil had never once been the problem.
