Turning, 5-axis milling, a lower turret and a second spindle. One setup, and one program to coordinate them.
The Mazak INTEGREX i-350H ST is a horizontal multi-tasking machine that combines turning with 5-axis milling, and adds a second spindle and a lower turret. A part can be finished complete in a single setup, with several tools working on it at the same time.
That is the value of the machine. It is also the reason a valid set of toolpaths is not yet a safe or efficient program.
What the i-350H ST gives the programmer to coordinate
The features that shape the NC program are:
- A main spindle and a second spindle, for transfer and back-side machining
- A milling spindle with B axis for 5-axis contouring, angled drilling and multi-surface machining
- A lower turret, optionally with rotary tools
- Balance cutting with the milling spindle and the lower turret, machining with two tools at the same time
- Automatic jaw changing on both spindles, with several sets of chuck jaws
- A MAZATROL SmoothAi control, with programs in conversational or ISO/EIA form
Each of these is an opportunity to shorten the cycle. Each is also a dependency between parts of the machine that share the same working area.
Valid code that runs wrong
On a machine with a milling spindle, a lower turret and a second spindle, the typical failure is not a syntax error. The NC program loads, the control accepts it, and the result is still wrong:
- the milling spindle and the lower turret reach the same zone at the same moment
- the B axis swings the milling head while the lower turret is still close to the part
- the second spindle picks up the part before the cut-off is complete
- a balance cut starts on one tool before the other is in position
- one side sits waiting for the other far longer than necessary
- a hand edit made at the machine never returns to the CAM model, so the next revision reintroduces the original problem
None of these produce an alarm at load time. They show up as a crash, a damaged part, or a cycle time nobody can explain.
Why the usual workflow misses it
A conventional workflow for a machine like the i-350H ST looks like this:
CAM > individual operations > postprocessor > manual synchronization > manual edits > prove-out on the machine
Three weaknesses stand out.
The postprocessor is not a programming layer. It converts operations into NC code. It does not, by itself, decide how work should be divided between the milling spindle, the lower turret and the second spindle, where each should wait for the others, or how the part should be handed over.
Manual refinement becomes part of the process. Synchronization and handover logic are added by hand after posting. That work is repeated for every part and depends on the individual programmer.
Simulation usually runs before the postprocessor. CAM simulation shows what the CAM system intended to produce. It does not see waiting codes, macros, offsets or hand edits added afterwards. Toolpath simulation tells you what the CAM system intended to produce. G-code simulation tells you what the CNC machine has actually been instructed to do.
Where Eureka NC Coder fits
Eureka NC Coder is the layer between your existing CAM and the final NC program. It is not a replacement for CAM, and it does not require you to change CAM platform.
The division of work is straightforward:
- CAM creates the machining operations, in the system your team already uses.
- NC Coder assigns operations to the machine’s channels, manages synchronization and part transfer, and applies the machine-specific postprocessing logic.
The result is a multi-channel NC program in which waiting points, handovers and channel assignments are generated in a controlled way rather than typed in afterwards. Because NC Coder works from the operations rather than from a single postprocessor, it applies to the CAM systems already used for multi-channel work, including Fusion 360, Mastercam, Siemens NX, PTC Creo and hyperMILL.
Where Eureka G-Code fits
NC Coder produces the program. Eureka G-Code checks the program the machine will actually run.
Eureka G-Code simulates the final, post-processed NC program on a digital twin of the i-350H ST and its control. It interprets the real program, the machine kinematics, the controller logic, macros and variables, work and tool offsets and machine limits, and runs all channels in their real relative timing. That is what exposes the problems that only exist when the parts of the machine interact:
- collisions and near-misses between the milling head, the lower turret, the spindles and the chucks
- a B-axis move that swings the milling head into the lower turret zone
- a second spindle closing on the part at the wrong moment
- axis limit violations
- material removal errors, checked by comparing the simulated part with the design model
Balance cutting and 5-axis milling: powerful, and unforgiving
Balance cutting with the milling spindle and the lower turret puts two tools on the part at once, which removes material faster and reduces deflection. 5-axis milling with the B axis removes secondary setups.
Both remove the safety margin that a sequential program enjoys. In balance cutting, two tools depend on each other’s timing. With a tilting milling head, the tool position depends on the kinematics of the entire head, not only on the programmed coordinates. A small error is not a wasted second, it is a collision on a machine that costs a great deal to repair.
Verifying these moves on a digital twin that reproduces the real machine kinematics, before the program runs, is what allows them to be used with confidence.
It is also about cycle time
Once the whole sequence is visible, the question changes from “Will the program work?” to “Is this the best way to run the machine?” You can see:
- the lower turret or second spindle sitting idle while the milling spindle finishes
- waiting points that could be moved earlier
- operations that could be overlapped or balanced instead of run in sequence
- a part-transfer sequence that spends more time waiting than cutting
On a machine of this class, minutes per part are at stake. They can only be found when the full multi-channel sequence can be seen and measured before it runs.
The M32 workflow, end to end
CAM (Fusion 360, Mastercam, NX, Creo, hyperMILL)
Machining operations
↓
Eureka NC Coder
Channel assignment
Synchronization
Part transfer
Machine logic
Postprocessing
↓
Final multi-channel NC program
↓
Eureka G-Code
Digital twin of the i-350H ST and its control
Collision detection
Limits
Material removal
Complete multi-channel sequence
↓
Mazak INTEGREX i-350H ST
The program that is generated is the program that is simulated, and the program that is sent to the machine.
Running a Mazak INTEGREX or another multi-channel mill-turn machine? Request a demonstration on a digital twin of your own machine and controller, and see how Eureka NC Coder and Eureka G-Code connect your CAM, synchronization, postprocessing and G-code simulation in one workflow.
FAQ
Can I program a Mazak INTEGREX i-350H ST from my existing CAM system?
Yes. Eureka NC Coder sits between the CAM system and the final NC program, so you keep the CAM you already use. It takes the machining operations, assigns them to the machine’s channels, manages synchronization and part transfer, and produces the machine-specific multi-channel NC program.
Why is the i-350H ST harder to program than a conventional lathe or machining center?
It combines a milling spindle with B axis, a lower turret and a second spindle in one working area. The difficulty is not any single toolpath but coordinating the parts of the machine: waiting points, part handover, balance cutting and clearance between the milling head and the lower turret.
Is a postprocessor enough for multi-tasking machine programming?
A postprocessor converts operations into NC code, but it does not normally distribute operations across channels or design the synchronization between them. That work is often done by hand after posting, which is where errors and lost cycle time come from. NC Coder is designed to handle that step.
What does Eureka G-Code check on the i-350H ST?
It simulates the real post-processed NC program on a digital twin of the machine and its control: kinematics including the B axis, controller logic, macros and variables, work and tool offsets, machine limits, collisions between the milling head, lower turret, spindles and chucks, timing across all channels, and material removal.
Why simulate the NC program instead of the CAM toolpaths?
CAM simulation runs before the postprocessor and before any manual edits. It shows what the CAM system intended. Simulating the final NC program shows what the machine has actually been told to do, including synchronization codes and hand modifications.
Can simulation help with balance cutting?
Yes. Balance cutting depends on precise timing between the two tools. Running all channels together on the digital twin shows whether both tools are in position when the cut starts, and whether they clear each other throughout.
Can simulation help reduce cycle time on the i-350H ST?
Yes. With all channels visible in one simulation, idle time, late waiting points and inefficient part-transfer sequences become easy to spot. The program can be rebalanced offline before it runs on the machine.
Do I need to change my CAM system to use NC Coder?
No. NC Coder works with the CAM systems already used for multi-channel work, including Fusion 360, Mastercam, Siemens NX, PTC Creo and hyperMILL.
Next step
Eureka G-Code — request a demonstration on a digital twin of your own machine and controller.
Verify the real ISO program — on a digital twin of your sliding-headstock machine, before the first bar.
Related Articles
Ask any experienced machinist ‘what’s in your safe start?’ and you’ll get a line like G17 G40 G49 G80 G90 — the block that forces the control into a known state so your program doesn’t inherit leftover modal settings from whatever ran before.
On a Citizen Cincom L20, the pains are multi-system ($1/$2/$3) synchronization, the G600-series machining-mode calls that govern gang, front, back and superimposed work, wait codes that must appear in every system, and Z-shift and guide-bushing traps. Eureka G-Code runs the real ISO across every system on a digital twin of the machine. Request a demo on a twin of your machine.
Master Bumotec s191 process orchestration across turning, milling, and grinding. Discover how digital twin simulation prevents costly first-cut errors.
A Swiss-type lathe flips the geometry, packs the tool zone, and lets one tool do many jobs — so the mistakes that scrap parts and crash gang tools are ones a conventional-lathe programmer has never had to think about.
Static checks miss parametric logic errors. Execute Macro B variables, IF/WHILE loops, and family-of-parts math on a digital twin to catch crashes off-line.
CAM simulation checks the toolpath it generated. The machine runs the posted G-code — with constructor cycles, macros, variables, subprograms and probing the CAM never modeled. On multi-channel and sliding-headstock machines, that gap is where the crash lives. Eureka G-Code simulates the true G-code on a digital twin of your machine. Request a demo on a twin of your machine.
Cut cycle time, extend tool life and save energy by re-modulating feedrates based on real tool engagement. Optimize G-code programs offline
Avoid two-turret crashes on Nakamura-Tome machines. Learn to verify pinch turning, superimpose modes, and waiting M-codes across channels.
Unattended bar work turns a crash from one scrapped part into a stopped run and a damaged machine. Here’s the checklist of what has to be verified before you walk away — and how Eureka Cloud runs those simulations overnight, in queue, so the whole night’s programs are proven before the lights go off. Request a demo on a twin of your machine
Most scrap isn’t a crash—it’s a valid program making an out-of-spec part. Compare real ISO simulation to CAD models to catch shifted offsets offline
Discover common multichannel sync errors on the Okuma Multus U4000. Prevent crashes and deadlocks by simulating real ISO code on a digital twin.
A digital twin of your sliding-headstock (Swiss-type) machine that reads the real ISO code and proves it out before the bar runs. Any controller, any kinematics
G32 cuts a thread one pass per block — every pass, infeed, chamfer and pull-out is on the programmer, and the feed must match the pitch in sync with the spindle. It’s valid code that scraps the thread or crashes the tool when a value is off. Eureka G-Code runs the real G32/G92/G76 passes on a digital twin of your machine
Verify automatic head changes, kinematic transforms, and large-envelope ISO code on PAMA Speedram and Speedmat machines with Eureka G-Code
Master B-axis signs, sub-spindle logic, and Fanuc cycles on the SMX 2100ST. Verify actual ISO code with Eureka G-Code digital twins
Feedrate optimization isn’t a lab trick. Across real production — hardened-steel molds, aluminum castings, precision components, large gantry work and continuous 5-axis parts like impellers — Eureka Chronos has cut cycle time by roughly 6–26% while holding surface quality and dimensional accuracy, with tool wear reduced or unchanged. And it needs no material or tool libraries to do it. Request a demo.
Why G83 peck depths, G98/G99 retracts, and control-specific behaviors cause crashes and scrap. Simulate real canned cycles on a digital twin offline.
Aero-engine blisks and casings push machining to the edge: tough titanium and superalloys, thin deformable parts, and airflow passages between blades so narrow the tool can interfere on a five-axis move
Deep-hole drilling for injection and die-casting molds (IMSA, CHETO) isn’t just a toolpath: depth vs diameter, chip evacuation, guide bushings, special cycles and process monitoring all matter
Single-block on a multitasking machine is defusing a bomb with the timer running. Step through the real G-code off-machine on a digital twin — watch where a variable actually drives the tool, follow subprogram flow across channels — and find the bug at a desk. Eureka G-Code runs and edits the true G-code on a twin of your machine. Request a demo on a twin of your machine.
Eureka Cloud turns every PC on your network into a simulation agent, so a whole shift’s worth of NC programs is verified overnight, in parallel — cleared or flagged before they reach the machine. And despite the name, it runs on your own servers: a private cloud, so your NC programs, geometry and IP never leave your network. Highly customizable, it integrates with your CRM, MES or other apps through APIs and scripting, and can automate post-processing, simulation, and feedrate optimization end to end. Request a demo
Prevent slow, costly collisions on large vertical lathes like the Pietro Carnaghi AP80. Verify part zero, attachments, and G-code before production.
Master horizontal spindle kinematics, A’/B’ swivel table logic, and undercut operations on the GROB G350 with Eureka G-Code digital twins.
Program the Citizen Cincom M32 from your existing CAM. Synchronize channels and part transfer with NC Coder, then verify the final G-code on a digital twin.
How G90/G91 modal errors, lathe U/W mix-ups, and subprogram mode shifts cause wild moves and crashes. Verify real ISO reference frames on a digital twin
Why safe Z heights set below clamps or raised features cause rapid collisions between operations. Verify real ISO rapids over actual fixtures on a twin
After a coordinate rotation or a tilted working plane, ‘up’ isn’t machine-up anymore — so a retract programmed in the rotated frame, or with the wrong axis order, drives the head straight into the part instead of clearing it. It’s a crash during what should be a safe disengage. Eureka G-Code runs the real ISO — rotation and kinematics included — on a digital twin. Request a demo on a twin of your machine.
A digital twin of your multi-channel turn-mill that reads the real ISO of every channel — synchronization, multiple spindles, sub-spindle transfer, constructor cycles and hand edits included — and proves it out before the turrets move. Any controller, any kinematics, no axis limit. Request a demo on a twin of your machine.
Moving legacy CNC programs to new horizontals risks costly crashes. Learn what breaks in G-code, offsets, and M-codes, and how to verify them
Prevent sub-spindle handoff crashes, front/back side errors, and XB B-axis mismatches on Nomura NN-20J3 Swiss machines with Eureka G-Code.
Why rewrite the same helical interpolation code for every thread size when a macro can generate it from a few parameters? Eureka G-Code simulates the real G-code with macro expansion and helical interpolation, before you cut
Prevent head order errors, turret mirror mismatches, and EIA/ISO sequence crashes on the Mazak i-300 ST with Eureka G-Code digital twins
CAM simulations assume correct offsets and miss wrong H numbers or missing G43 codes. Verify runtime control offsets and G49 states on a digital twin
How shortest-path rollover, degrees/min feed errors, and C-axis wind-up cause 5-axis crashes. Verify multi-axis kinematics and TCP cycles on a digital twin.
On paper the DVF 5000 Gen 2 gives you a big envelope and full 5-face access. In real setups it’s tighter: at tilt the reach shrinks, the fixture and trunnion interfere, the tool setter gets in the way, and the table sits off-center from the spindle. That gap between brochure envelope and usable envelope is where setups crash.
How wrong call numbers, missing M99 returns, and carried modal states cause program flow crashes. Verify nested M98/M99 call structures on a digital twin
On the DMG MORI NTX 1000, the recurring pains cluster around tool offsets that rotate with the B-axis or sub-spindle, wrong origins between the main and counter spindle, a post that doesn’t match the real kinematics, and one-setup programs that assume a machine state that isn’t there. Eureka G-Code runs the real posted ISO on a digital twin of the actual kinematics. Request a demo on a twin of your machine.
Program the Nomura NN-20J3 from your existing CAM. Synchronize front and back machining, part transfer and both spindles, then verify the final G-code.
The main-to-sub-spindle handoff is a timed dance — advance, grip, match, release before cut-off — and getting it wrong drops the part or crashes the spindles. It’s timing, not geometry, so a listing and a toolpath sim both miss it. Eureka G-Code executes the real transfer on a digital twin of your machine. Request a demo on a twin of your machine.
On a multi-channel turn-mill, each channel’s program can be flawless and the machine still crashes — because the collision lives in the timing between channels, not in any one line. Eureka G-Code simulates the real ISO of every channel, synchronization commands included, on a digital twin of your machine, so wait-code mistakes surface before the turrets meet. Request a demo on a twin of your machine.
Why wrong plane selection causes arcs to cut into parts and canned cycles to plunge off-axis. Verify modal G17/G18/G19 plane state on a digital twin.
CAM cycle-time math is optimistic on any machine — and on parallel channels with synchronization waits it’s worst of all, because the waits and the interleaving aren’t in the toolpath. Eureka G-Code computes cycle time from the real ISO on a twin of your machine, and Eureka Chronos optimizes it. Request a demo on a twin of your machine
A move that’s geometrically fine can still drive a rotary or B-axis into its travel limit — and because it isn’t a geometry error, a toolpath simulation doesn’t catch it. Eureka G-Code checks overtravel and near-miss against your real machine’s travels on a digital twin. Request a demo on a twin of your machine.
Meta description: Program the Okuma MULTUS U4000 2SW from your existing CAM. Synchronize the B-axis head, lower turret and opposing spindles, then verify the final G-code.
A probing cycle sets offsets at runtime and changes what every later move does — yet it’s the part of the program almost nobody simulates, because a toolpath render doesn’t run it. Eureka G-Code executes probing cycles on a digital twin of your machine, and checks pre-holes for tapping too. Request a demo on a twin of your machine.
Standardize program verification across plants with automated, on-premises batch simulation. Eliminate bottlenecks and ensure no G-code runs unverified

