CNC Transformation of T for Large-Scale Hose Plate Parts
A radial arm drill can be converted to CNC and will still drill hose plate holes. The question is whether the converted machine can hold position on a 4,000 mm plate. This page explains the mechanics, the boundary conditions and the point where a gantry mill becomes the cheaper route.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
Key takeaways
What the CNC transformation of t actually changes
A radial arm drill is a manual machine with a heavy cast column, a horizontal arm that slides up and down, and a head that travels along the arm. The operator positions the spindle by hand and feeds a quill into the work. A CNC transformation of t replaces the hand wheels with servo drives and adds a control that reads a program, but it does not change the load path. The column, arm and head are the same castings they were before.
That distinction matters on large hose plates. A converted machine can position the spindle to a program coordinate instead of a chalk mark or a layout line. It can repeat that position thousands of times without fatigue or reading error. What it cannot do is become stiffer. The arm still deflects under thrust, and the quill still extends over a changing overhang as the head moves along the arm.
So the first honest framing is this: a retrofit buys positioning and repeatability, not rigidity. If your hose plate tolerances are driven by hole location, a converted arm drill can do the job. If they are driven by hole roundness, bore straightness or surface finish in a deep hole, the retrofit will not get you there and no amount of programming will fix it.
Every retrofit project should start with a process capability check on the existing machine, not with a controls shopping list. Measure what the current spindle does in your material on your plate thickness. That number, not the spec sheet of the servo kit, sets the ceiling for the converted machine.
Why large hose plate parts are hard to drill
A hose plate, or tube sheet, is a thick plate carrying a dense array of holes that receive tubes, hoses or fittings. Diameters run from roughly 10 mm to 100 mm, plate thickness from 20 mm to 200 mm, and overall plate size can reach 4,000 mm across. The holes are not decorative. Each one has to be in the right place relative to every other hole, because the mating manifold or header was drilled to the same pattern.
That is what makes the job different from general plate drilling. On a small bracket you can tolerate a 0.1 mm position error and nobody notices. On a tube sheet, position errors accumulate across the pattern. If hole 1 is 0.05 mm high and hole 400 is 0.05 mm low, the total spread across the plate is 0.1 mm, and the tube bundle may not assemble.
Heat is the second problem. Drilling a 40 mm hole through 100 mm of 304 stainless generates a lot of heat in a local area. The plate grows around the hole, the tool rubs, and the next hole drilled nearby starts from a thermally shifted datum. On a manual machine the operator compensates by feel. On a converted machine the control has no idea the plate moved unless you give it a way to check.
Chip evacuation is the third. Deep holes in ductile material produce long stringy chips that wrap the tool and pack the flutes. A CNC retrofit with a peck cycle handles this far better than a manual operator turning a handle, which is one of the genuine wins of the conversion.
How positioning error builds up on a converted arm drill
Positioning error on a converted arm drill comes from four places, and they add up differently than on a machining center. The first is the scale or encoder feedback on each axis. If the arm axis uses a rack and pinion with a rotary encoder on the pinion, backlash and rack wear translate directly into position error that the control cannot see.
The second is thermal drift of the column and arm. A cast iron arm 2 m long grows about 0.022 mm per degree Celsius. Run the spindle hard for two hours and the arm can be 3 to 5 °C warmer than when you probed the fixture. That is 0.07 to 0.11 mm of drift at the far end of the arm, in the direction the arm points.
The third is spindle and quill deflection under cutting thrust. A typical radial arm drill quill extended 300 mm will bend measurably under a 5 kN thrust. The hole goes in slightly off-axis and the exit position differs from the entry position.
The fourth is workpiece movement. This is the one people underrate. A 4,000 mm plate weighing several hundred kilograms, held on a few clamps with jack stands in between, will flex and shift as material is removed and as clamps are repositioned. A converted machine with a probe can catch this. A converted machine without one cannot.
The practical consequence: on a converted arm drill, hole-to-hole accuracy across a large plate is usually limited by thermal drift and workpiece movement, not by the servo resolution. Buying a finer encoder does not help if the plate is moving.
Where the retrofit stops working
There is a clear crossover point. Below it, a CNC transformation of t is the economical answer. Above it, you are paying to fight the machine's geometry.
The retrofit works well when hole pitch is generous relative to hole diameter, when plate thickness is under about 100 mm, when the material is mild steel, aluminium or a free-machining stainless, and when position tolerance is in the ±0.1 mm to ±0.2 mm band. It also works well when the plate can be drilled in one setup with the clamps outside the pattern.
The retrofit struggles when hole pitch is tight enough that the webs between holes are thin. Thin webs mean low stiffness, and the thrust from drilling one hole can deflect the material next to the neighbouring hole. It also struggles on deep holes in gummy stainless or on Inconel, where the required thrust is high and the arm deflection shows up directly in the hole.
Interrupted cuts and angled entry faces are another boundary. If the hose plate has a curved or stepped surface, a converted arm drill without a controlled Z reference will wander on entry. A machining center with a rigid spindle and a probe handles that far better.
And if the part needs bored, reamed or threaded holes with a true position under ±0.05 mm, the retrofit is the wrong tool. That is not a controls problem. It is a stiffness problem.
Fixture and clamping rules that decide the outcome
Most failed hose plate jobs are lost before the first hole. The plate is not supported evenly, or it is clamped in a way that springs when the clamps come off, or the datum is set off a flame-cut edge that is not square to anything.
Support the plate over its full area, not at four points. A 4,000 mm plate on four jack stands will sag under its own weight, and the sag changes as you move the drilling head and as the operator steps on it. Use a grid of supports at roughly 500 mm spacing, shimmed so the top face reads flat within 0.05 mm across the diagonal before you start.
Set the datum from a machined reference, not a cut edge. Ideally the plate is machined on two adjacent edges on the same machine that will do the drilling, so the datums are true to each other. If that is not possible, probe the reference features and let the control establish the rotation offset.
Clamp next to the supports, not between them. A clamp over an unsupported span bends the plate down locally, drills fine, and springs back when released. The hole is then out of position in the released state.
Leave the clamps alone during the cycle if you can. Re-clamping mid-job changes the stress state of the plate. If the pattern forces a clamp move, probe the datum again after the move rather than trusting the original setup.
Finally, break the edges of drilled holes before inspection. A raised burr will sit proud of the plate and throw off any measurement taken against the surface.
What we watch on converted machines in production
In our Dongguan plant we run a mix of converted and purpose-built machines, and the converted ones teach the same lessons every time. Warm up the spindle before probing the fixture. A cold machine and a warm machine will not agree, and the difference shows up as a drifting pattern.
Use a spot drill on every hole, even the small ones. It costs a few seconds and it stops the twist drill from walking on the plate surface. On a converted arm drill with some residual play in the arm axis, the spot also gives the control a chance to correct position before the full diameter engages.
Peck deep holes with a retract that clears the flutes fully. On 304 stainless at 100 mm depth, a 3 × diameter peck with full retract keeps the tool cool and the hole straight. Shallow pecks that never clear the chip just rub the material.
Check the first hole, then check a hole at the far corner of the pattern. If the two agree within tolerance, the setup is stable. If they differ, the plate or the machine moved, and drilling the rest of the pattern will only confirm the error.
Log the plate temperature if the job runs more than an hour. A simple contact thermometer on the plate surface tells you whether thermal growth is worth compensating for in the program.
And measure after the plate has cooled to room temperature. Measuring hot gives a number that means nothing the next morning.
Converted arm drill vs gantry mill for hose plate work
Match the part to the machine before you commit to a retrofit budget.
| Factor | Converted arm drill | Gantry or bridge mill |
|---|---|---|
| Plate size | Up to roughly 4,000 mm with support | 4,000 mm and beyond, fully supported table |
| Position tolerance | ±0.1 to ±0.2 mm realistic | ±0.005 mm achievable |
| Hole pitch vs diameter | Pitch at least 1.5 × diameter | Down to 1.2 × diameter on thin webs |
| Plate thickness | Under about 100 mm | 100 mm to 200 mm and above |
| Material | Mild steel, aluminium, free-machining stainless | Stainless 316L, 17-4PH, Inconel, titanium |
| Typical hole count | A few hundred per plate | Thousands, dense arrays |
| Setup changes | One or two, clamps outside the pattern | Multiple, with probing between |
| Best fit for | Low volume, thick plate, moderate pattern | High volume, tight pattern, tight tolerance |
Which route to take
If your hose plate is under 100 mm thick, drilled in one setup, and toleranced at ±0.1 mm or looser, a CNC transformation of t is the cheaper and faster route. If the pattern is dense, the webs are thin, the material is stainless or Inconel, or true position must be ±0.05 mm or better, skip the retrofit and put the plate on a gantry mill.
Questions engineers ask before a retrofit
Can a converted arm drill hold ±0.05 mm true position?
Rarely, and not repeatably. The limit is mechanical, not electronic. Arm deflection under thrust, thermal growth of a 2 m casting, and backlash in the rack drive will each contribute more than 0.05 mm on a large plate.
If the drawing calls for ±0.05 mm true position, plan for a gantry or bridge mill with a rigid spindle and a fully supported table. A retrofit will produce parts that pass the first check and fail the third.
Does the control choice matter more than the mechanics?
No. A good control makes a converted machine easier to set up and easier to probe, but it does not add stiffness. A mid-range control on a well-scraped machine will out-drill a premium control on a worn one.
Spend the budget on re-scraping the arm slides, replacing the rack, and adding a scale on the arm axis before you spend it on a bigger control.
How do we hold position across a 4,000 mm plate?
Support the plate on a grid at roughly 500 mm spacing, machine two datum edges, and probe the datum at the start of the cycle. Do not rely on a flame-cut edge.
If the job runs long enough for the plate to warm up, probe again at the halfway point and let the control shift the work offset. That single step fixes most of the drift people blame on the machine.
What coolant and feed works on thick stainless hose plates?
Flood coolant at high volume, directed at the hole entry, and a peck cycle that fully clears the flutes. For 304 stainless, a surface speed around 20 to 25 m/min with a feed of 0.08 to 0.12 mm per revolution per tooth is a workable starting point on HSS-Co or carbide.
Watch the chip color. Light straw chips mean the speed is right. Blue or black chips mean you are burning the edge and the hole will come out undersized.
Can the converted machine also ream or thread the holes?
Reaming yes, at moderate depth and with a floating holder. Threading with a tap is possible in mild steel and aluminium, but rigid tapping on an arm drill with residual play is a risk.
For critical threads, drill and ream on the converted machine, then tap on a rigid machine or use a thread mill where the control supports helical interpolation.
How long does a conversion take and what does it cost?
We do not quote a fixed figure, because the scope depends on the condition of the machine. A conversion that keeps the existing spindle, gearbox and arm, and adds drives, scales and a control, is a different project from one that also re-scrapes the slides and rebuilds the head.
Send us the machine drawings, the plate drawing and the tolerance callout. We will review the geometry and tell you honestly whether the retrofit can hold the tolerance or whether the part belongs on a gantry mill.
Send us the plate drawing and the machine details
We review the hole pattern, thickness and tolerance callout, then tell you whether a retrofit holds it or whether the part should run on a 4,000 mm gantry mill. Quotation and free DFM analysis within 12 hours.
12-hour quote100% inspectionNo minimum order quantity
More machining notes
We publish setup notes, tooling trials and inspection data from the factory floor.
