How to move CNC chips to the auger: washdown vs manual clearing

Clear the floor by hand first: a soft squeegee kept at the door moves two-thirds of a PVC chip load to the auger in about 15 seconds, once every two or three cycles. Add a coolant-fed wash only where chips have an open route to the trough, and only after you check one balance — pump delivery against coolant return. A wash manifold feeding 60 gal/min [230 L/min] into an enclosure whose screens pass 10 gal/min [38 L/min] back drains the tank in about 10 seconds and leaves your pump running dry. Air blast scatters stringy PVC into the way covers instead of the trough, so it stays a spot tool, not a system.

What moves PVC chips and what does not

PVC cuts into strings, the strings knot into mats, and liquid runs underneath a mat instead of carrying it. That habit decides which method works where: momentum tools win on open surfaces and light chips, and your hands win everywhere else. Aluminum chips and firm plastics such as acetal and ABS move on a directed jet; a knotted PVC pile stays put and grows.

At least one production vertical machining centre shipped from the factory with a small coolant manifold at the rear casting, above the Y way covers: whenever flood ran, it sprayed left and right along the cover tops and walked acetal and ABS chips from back to front. That is the pattern worth copying — modest flow, along a surface with an open end. For chips that fly rather than pile, small CNC enclosure design for chip control is the better read.

The balance that decides every washdown

Tank level moves by one line of arithmetic: change per minute = return flow − pump delivery, with both flows in the same units. In the first seconds of a wash the coolant is in transit — pooling on the floor, filling the trough, crossing the screen — so the tank pays the full delivery and the return is still zero.

Work it for a typical small vertical. Your tank holds 40 gal [150 L], with 8 gal [30 L] of usable volume above the pump intake. The wash manifold feeds 60 gal/min [230 L/min]; trough, screen and baffle pass 10 gal/min [38 L/min] back once the floor is flooded. Net loss = 60 − 10 = 50 gal/min [190 L/min]. Time to a dry intake = 8 gal ÷ 50 gal/min = 0.16 min, about 10 seconds. That is the machine that drained its tank in seconds and left the floor under an inch of coolant, every time the wash ran.

The cure is arithmetic, not hardware. Throttle the wash branch until tank level holds flat for 30 seconds; that flow is what your return path can carry, and everything above it buys you a flood. When the balancing flow is too small to move chips, the wash is the wrong tool for that spot and the squeegee is faster.

Three symptoms separate the three failures:

• Chips stay on the floor with tank level steady — a transport problem. The jet has no clear path or the pile is matted; move the nozzle or use your hands.
• Tank level falls while the floor pools — a return problem. Clean the screen, clear the trough and throttle the delivery down.
• Chips reach the auger but the chute packs — a downstream problem. Jog the auger in short bursts so the chute drains between them.

The manual routine, step by step

You need one tool for the baseline: a 12 in [300 mm] soft squeegee on a hook by the door, plus the machine's normal access procedure. Follow it — spindle stopped, motion stopped, hazardous energy isolated where entry calls for it — before your arms go past the door.

1. Park with the door closed: retract the tool, then run Y forward. On most verticals that opens a clean reach over the way covers.
2. First pass, rear to front: reach over the covers with the squeegee and drag the heavy pile forward under the table. Two pulls move most of a PVC load.
3. Retract Y and finish: bring the table back, then sweep the front pile into the trough.
4. Jog the auger two turns, so the chips in the chute drain their coolant instead of packing into a plug.
5. Set your interval from what you see: start at every two cycles, stretch it while the trough pile stays under about 1 in [25 mm], and shorten it the first day the chute bridges.

That is the whole baseline: one squeegee, no plumbing, no tuning. If reaching over the covers fights the machine, the clearances are the real problem, and machine enclosure layout for service and expansion covers what to change.

Adding a wash nozzle without draining the tank

Copy the arrangement that earns its keep: a small manifold above the rear way covers, teed into the flood line behind its own ball valve. It sprays left and right along the cover tops whenever flood runs, needs no control logic, and stops when flood stops. Set it up like this:

• Tee off after the pump and give the wash its own valve, so you can throttle it without starving the tool.
• Aim fan nozzles along an open route to the trough; a jet pointed at a ledge spends its momentum there and rebuilds the pile.
• Trial at quarter flow for 10–15 s, one person on the valve and one calling tank level, and stop the moment the level moves.
• Add flow one step at a time, only after the return has proved it carries what you already feed it.

The branch works when the floor stays drained, the tool keeps full flood, and the pile you used to drag walks to the trough on its own. The extra volume tests everything around it, too — containing machine-tool coolant at enclosure gaps covers keeping that spray off the panel seams and the electronics.

Where air blast fits

Air has a place, and it is small. A 1/8 in [3 mm] nozzle at 80 psi [5.5 bar] consumes on the order of 15 scfm [425 L/min] of compressed air, transfers little momentum to a wet stringy mat, and puts fines in the way covers, the cable trays and the shop air. Use it to dry an exposed ledge or to chase the last handful after a coolant pass. Leave the main load to methods with a return path.

The Y-axis wiper that clears while you cut

A flexible blade hung on the axis — riding over debris on the forward stroke, digging and pushing on the return — is the only approach here that needs no operator and no coolant. It is a machine modification rather than an accessory: it must survive the full Y travel without snagging, and the drag it adds on the digging stroke lands straight on the servo. Before you design one, read how KV factor varies 0.6 to 3.7 because axis loop tuning differs. A wiper that digs on one stroke changes the load your axis loop was tuned against, and that shows up as following error long before it shows up as a clean floor.

Checks that prove it worked

Judge the change on chips leaving the machine, not on how vigorous the wash looks:

• Chips reach the auger and discharge from the chute — watch the chute end, not the trough.
• Tank level recovers within a minute of the wash stopping, then holds flat at full wash flow.
• The screen still passes after five consecutive clearing cycles; one clean trial hides a screen that loads up cycle by cycle.
• Your manual time drops from a full clearing pass to spot checks under 5 s — otherwise the wash is decoration.

When the enclosure itself is the fix

If the chips keep winning, change the floor. Slope the enclosure floor pans 5° to 10° (about 1 to 2 in per foot [85 to 170 mm/m]) so coolant and chips walk to the trough under gravity, pull the trough under the table's drip line, and take out the ledges that trap PVC strings. That kind of one-off geometry is custom enclosure work: you set the material, the width, height and depth and every cutout, and it is delivered in 10 business days from your approval of the drawing.

The wash you add lands on everything inside the machine, the control box included. Keep that box rated NEMA 4X / IP66 by the manufacturer — ours are not UL listed — with glands that seal against spray, and pick it by environment in the enclosure selector.

Enclosures for this job

Catalog sizes that suit the job this article answers, picked from its subject; open one for its full specifications.

Questions