The night shift runs unattended at a six-machine medical-component shop in the upper Midwest, and those lights-out burns are supposed to be the cheapest hours on the calendar. Day shift parts gauge clean. The same program, started at ten at night, comes off pitted and inconsistent, so the crew re-tunes the machine, swaps a wire guide, and argues about flushing pressure. Nobody opens the water tank. That dielectric water depends on EDM resin with a finite service life, and once the bed nears exhaustion the loop starts conducting current it is supposed to hold back. The machine is usually not the problem. The water loop nobody logs usually is.
Night Shift Scrap Usually Starts In The Tank
Deionized water in an EDM does two jobs at once, insulating the spark gap and flushing debris out of it, and conductivity governs both. A tank sitting at 5 microsiemens per centimeter at the start of first shift can drift past 20 by the end of a twelve-hour unattended burn, because the bed is stripping ions the entire time and nothing on the control panel announces that it has stopped keeping up. Day shift rarely feels it. Somebody is standing there, the bed is fresher, and a marginal part gets caught before the next twenty go the same way. At two in the morning the machine simply keeps cutting, and the scrap bin does the reporting.
Calendar Based Media Swaps Miss The Real Signal
Most shops swap media on a calendar, twice a year, because that is what the last guy did. The case we see most often is two beds in the same building living completely different lives, one coasting through eleven months of light graphite work while the other is finished in fourteen weeks after a titanium run and a leaking chiller push hardness into the loop. EDM resin exhausts on duty cycle and inlet quality, not on a date. A cheap conductivity meter, read at the start of every shift and written on the same clipboard as the spindle hours, turns an invisible variable into a trend line anybody on the floor can follow. When that number climbs three shifts running, the bed is on fumes and the parts will say so before the maintenance schedule does.
Servicing A Live Machine Is A Costly Shortcut
Changing media on a machine that is still energized, to save fifteen minutes, is the one shortcut worth naming outright. Shut the machine down, de-energize it, and lock it out before anyone opens, drains, or swaps media in the dielectric loop, every time, with no exception for a quick look. Water and live electrical equipment do not negotiate. There is a money argument stacked on top of the safety one, because a bed swapped hot with the pump still cycling tends to channel and push fines downstream into the filters, and the shop pays for that twice, once in filter elements and once in the next burn.
Regeneration Math Beats Buying New Media Outright
Ten years ago the reflex in most shops was to buy virgin media every cycle and never think about it again, because the line item was small enough to disappear inside the maintenance budget. That reflex has not changed much, but the pricing underneath it has. Federal Reserve Economic Data, publishing the Bureau of Labor Statistics series, put the producer price index for plastics material and resin manufacturing at 333.855 in June 2026, well above the low 300s it tracked through the previous winter. Resin is the input cost sitting under ion exchange media, so a purchasing habit formed when that number was flat has quietly turned into an expensive one.
Run the arithmetic for a six-machine shop with an annual media budget under $9,000, using illustrative numbers rather than a quote. Say each changeout of virgin media runs $1,150 and the shop does eight of them a year, which comes to $9,200 before freight is added. Send the exhausted beds out for regeneration instead at $640 a changeout and those same eight swaps total $5,120, and even after $360 in round-trip freight the year lands at $5,480 all in. That is roughly $3,700 back inside a budget that had already blown its own ceiling, and the exhausted media goes back into service rather than into a drum behind the building.
Regeneration is not automatically the right call for every bed. Media fouled with tramp oil, or hit with a metals load past what the resin can release, will not come back to spec, and a supplier who promises otherwise is selling a second disappointment. What usually decides it is inlet quality and whether the shop caught exhaustion before the bed went hard. The shop that logs conductivity gets the cheaper option; the shop that runs beds into the ground buys new and calls it the cost of doing business.
Water Quality Deserves The Same Discipline As Spindles
Nobody would run a spindle for a year without checking bearing temperature, yet the water loop feeding a quarter-million-dollar sinker gets looked at only when something breaks. OSHA’s flood response guidance calls for turning off power at the main breaker or fuse on the service panel before anyone works around water, and states plainly that nobody enters a flooded area or touches electrical equipment while the ground is wet, which is the same discipline a resin changeout deserves on an ordinary Tuesday. Put conductivity on the shift log, set the number that triggers a swap, and hand the exhausted media to somebody who regenerates it. The night shift stops being a gamble. Scrap that comes off at two in the morning was never really a machine problem, and the shops that work that out stop paying for the same lesson twice a year.