Are Milltronic Partner CNC Machines Any Good?

The Partner series was built as a bed mill with a cast iron base and a box-way or linear-guide column, depending on the year. Early units used heavier castings than their price class suggested, which is why many are still cutting today. The weak point is rarely the frame. It is the spindle bearings and the way surfaces after years of service.

On a machine this age, check spindle runout with a test bar at 300 mm from the nose. Under 0.010 mm is workable for general milling. Above 0.020 mm and you are looking at a bearing or taper regrind before the machine is worth much. Push the table by hand at the extremes of travel. Any visible step or tight spot means the ways need scraping.

The Partner handles aluminum and brass all day. Cast iron and mild steel are fine at moderate depths of cut. Where it falls short is high-feed roughing in hardened or gummy alloys, because the spindle taper and drawbar were not designed for the side loads a modern high-torque spindle takes.

A light finishing pass on a 300 mm part is a different problem than a 25 mm face mill in 4140. Match the cut to the machine, not to the brochure.

• 1
  Good fitPrismatic parts, plates, housings, fixtures, one-off tooling.
• 2
  MarginalDeep pockets in 17-4PH or Inconel, long-reach tools.
• 3
  Walk awayVisible scoring on ways, spindle runout over 0.020 mm, cracked castings.

A well-maintained Partner in good thermal conditions holds ±0.025 mm on position and about ±0.013 mm on repeatability for a single setup. That is a general machining figure. It is not a precision grinding figure, and no amount of calibration turns it into one.

What kills the number is heat. Run the spindle hard for an hour and the column grows. Operators who hold tight tolerances on these machines warm them up for 20 to 30 minutes, then take a skim cut and adjust the offset. That single habit decides whether you meet a ±0.025 mm callout or fight it all afternoon.

If your drawing calls for ±0.005 mm, the Partner is the wrong tool. At GreatLight we hold ±0.005 mm (±0.0002 in) and Ra 0.2–0.8 μm on our 5-axis and mill-turn centers, and that is the class of machine that tolerance belongs on. Buying a Partner to chase it is a losing trade.

For a job shop, the useful question is not 'what is the best it can do' but 'what does it hold on a Tuesday afternoon in August'. The honest answer is ±0.05 mm unless you control the room.

• 1
  PositioningAround ±0.025 mm on a tight, well-kept machine.
• 2
  RepeatabilityRoughly ±0.013 mm within one thermal cycle.
• 3
  Needs a warm-up20–30 minutes of spindle run before the first tight cut.
• 4
  Not designed for±0.005 mm fits, mirror finishes, hard-turned profiles.

The Partner's conversational control is the part most people either love or replace. It lets an operator build a pocket, a bolt circle or a face pass at the panel without CAM. For repair work and simple fixtures, that is genuinely fast. An experienced operator can go from print to first cut in under ten minutes.

The friction shows up when you feed it modern CAM output. Old controls have small memory, slow drip-feed and limits on file size. A 3D surfacing toolpath from a current CAM package may not fit, or it feeds so slowly that the cycle time doubles. Check the memory and the communication ports before you buy. RS-232 and floppy drives are not a workflow, they are a workaround.

The practical middle path is to keep the conversational side for simple jobs and post-process complex ones for a modern controller if the machine has been retrofitted. Many Partners now run a PC-based control or a modern front end. A retrofit changes the value calculation entirely. A clean mechanical machine with a fresh control is a better buy than a low-hour machine with a dead CRT and no memory.

Ask what software version the control runs, whether the ladder logic is documented, and whether the seller can supply parameters. Without a parameter backup, one dead battery can turn the machine into a project.

• 1
  Strong atManual-style programming, simple pockets, bolt circles, facing.
• 2
  Weak atLarge 3D toolpaths, high-speed surfacing, file transfer.
• 3
  Before buyingConfirm memory size, ports, and a full parameter backup.

Mild steel, 4140 and 4340 cut fine on a Partner with the right speeds, feeds and coolant. Titanium is possible but slow, and the machine is not where the limit sits. The limit is the spindle taper and the rigidity at the tool tip. A Ti-6Al-4V pocket with a long 8 mm end mill will chatter on a Partner long before a modern 5-axis center does.

If your titanium work is small, shallow and toleranced loosely, the Partner can do it. If it is deep, thin-walled or needs Ra 0.8 μm, run it on a machine built for the heat and the load. We machine TA1, TA2, TC4 (Ti-6Al-4V) and Inconel daily on 16 simultaneous 5-axis centers and 16 mill-turn centers, and the difference in tool life alone pays for the machine class.

The automatic tool changer is the most common failure point on older Partners. Early hydraulic units need steady pressure, clean fluid and fresh seals. Leaks, slow arm movement and dropped tools usually trace back to a tired pump or a clogged filter. Pneumatic and umbrella changers are simpler and more forgiving.

Before you commit, run the changer through 30 to 50 full cycles under power. Watch for arm hesitation, listen for the air blast, and check that every pocket holds the taper cleanly. A changer that drops one tool a week will cost more in scrapped parts than the machine saves.

• 1
  Steel1018, 1045, 4130, 4140, 4340, A36 all workable.
• 2
  TitaniumOnly shallow, open, loosely toleranced features.
• 3
  ATC check30–50 powered cycles, then inspect every pocket.

A used Partner makes sense when your parts are simple, your tolerance is loose, and you have an operator who knows the control. Around 15,000 to 30,000 USD for a clean unit is the typical range, and the machine pays back on repair work, fixtures and low-volume brackets. That is a real market. It is not a market for precision contract work.

The cost of ownership is where the math turns. Spindle rebuilds, way scraping, control batteries, and the hours you spend finding obsolete boards add up. If the machine is down a week a quarter, the effective rate is worse than outsourcing to a shop with 127 CNC machines and spare capacity.

For a prototype shop, the better play is often to keep one old mill for soft material and simple features, and send the tight, contoured, or hard-material work out. You get the flexibility without carrying the maintenance and the thermal drift.

If you do buy, get the maintenance history, a parameter backup, and a spindle runout report before money changes hands. No history, no deal.

• 1
  Buy it ifSimple prismatic parts, in-house repairs, loose tolerance.
• 2
  Skip it if±0.005 mm fits, Ra 0.2–0.8 μm, deep titanium pockets.
• 3
  Always verifySpindle runout, way condition, control backup, ATC cycles.