Haas VF-Series Mid-Program Freeze: Troubleshooting LVPS Failure

Problem Overview

A Haas VF-series machining center (VF-0E through VF-6, approximately 1997–2006 vector-drive/brushless-motor era) that intermittently halts mid-program — pausing as if in single-block mode without any active single-block flag — and resumes normally on Cycle Start is exhibiting a classic symptom cluster tied to marginal DC bus voltages on the control electronics. The fault escalates from "once a month" to "every 30–60 seconds" as the power supply degrades. Secondary symptoms include freezes during tool changes, Running Wait or Running Reset status on the lower display with no operator input, and eventually the machine hanging on the Haas Automation boot splash screen.

Root Cause Analysis

The primary cause on machines of this vintage is a failing Low Voltage Power Supply (LVPS) — the Condor SP1730 (also labeled SL Power / Ault SP1730, Haas PN 32-0211). As electrolytic capacitors age, output ripple increases and regulation degrades. The control CPU, MOCON board, and I/O PCB all share the same regulated rails; even brief under-voltage transients on the +5 V rail cause the processor to assert a wait state or trigger a watchdog-style pause that the operator sees as a freeze. Because the supply may still read within spec under no-load or light-load conditions, the fault is intermittent and worsens as thermal load builds during a running program.

Secondary contributors that should be ruled out in parallel:

• Dead or weak SRAM backup battery — causes memory corruption that manifests as erratic program execution or parameter loss on power cycle.
• Failed Real-Time Clock (RTC) crystal or battery — on some boards the RTC losing time triggers supervisor-level interrupts that stall the motion planner.
• Corrupted SRAM chip — confirmed by the Battery Backup Unit (BBU) self-test; a failing memory IC can inject garbage into the look-ahead buffer.
• PC/DNC serial communication contention — a host PC holding the RS-232 line active while the control tries to run can force a Waiting state (lower display reads WAITING); disconnect DNC and retest before replacing hardware.

Diagnostic Procedure

1. Isolate DNC/serial — Unplug any RS-232 or Ethernet DNC cable. If the freeze disappears, the fault is software/communication, not hardware.
2. Check Single Block & Feed Hold state — Press DIAGN, navigate to Inputs page; confirm input 0108 (Feed Hold) = 0 and Parameter 57 bit 0 = 0.
3. Measure LVPS output voltages under load — With machine powered and spindle/axes idle, probe the Molex connectors at each major board (MOCON, IOPCB, video board, processor board). Use a true-RMS DMM. Acceptable ranges and Condor SP1730 specs:

4. Run BBU Memory Test — Power down, remove the processor board battery (see Battery Replacement section below), power up, and allow the BBU self-test to run to completion. Any FAIL indication identifies the defective SRAM chip (U-designator reported on screen). A clean PASS confirms memory integrity; a fail here requires board-level SRAM replacement or processor board exchange.
5. Inspect for Alarm History — Press ALARM > F4 (History). Alarms 102 (Servo Drive Fault), 119 (Encoder Loss), or 900-range servo faults appearing in the log just before freeze events indicate axis drive issues rather than LVPS.

Solution: LVPS Replacement (Condor SP1730)

The Condor SP1730 / SL Power SP1730 is the correct replacement for Haas PN 32-0211. New-old-stock units are available from industrial surplus suppliers; verify the output trim pots are accessible and confirm the connector pinout matches your machine's harness before installation.

1. Power down the machine and lock out / tag out per your site LOTO procedure. Confirm capacitor discharge (wait minimum 5 minutes after main contactor opens).
2. Photograph all wiring harness routing before disconnecting.
3. Disconnect the AC input leads (typically 115 VAC or 230 VAC single-phase — verify your machine's input voltage label on the LVPS chassis).
4. Disconnect all DC output Molex connectors; label each connector with destination board.
5. Remove the four mounting screws and extract the LVPS chassis.
6. Mount the replacement unit; reconnect DC harnesses and AC input in reverse order.
7. Before full power-up, use a current-limited bench supply or measure DC output at the LVPS terminals (load disconnected) to verify output voltages are within spec and trim pots are not misadjusted.
8. Power up the machine and verify all axes home successfully and a complete program runs for > 10 minutes without freeze.

Preventive Maintenance: Battery, RTC, and Memory

On Haas controls from the 1997–2006 era, perform these steps any time a machine is exhibiting erratic behavior, or proactively every 3–5 years:

Battery Replacement

The processor board uses one or two Panasonic BR2032 (3 V lithium, 190 mAh) or equivalent coin cells for SRAM retention. Machines with larger memory boards (64K+ SRAM) typically have two batteries in parallel. Replacement procedure:

1. With machine powered ON (to maintain SRAM contents), locate the battery holder(s) on the processor PCB — typically near the SRAM ICs at board center.
2. Swap batteries one at a time if two are present; do not remove both simultaneously or SRAM contents will be lost.
3. If the machine must be powered down first, record all parameters (Settings, Parameters, and Tool/Work Offsets) to a DNC backup before removing batteries.
4. After replacement, confirm the date/time is correct and re-run the BBU test.

Real-Time Clock Replacement

The RTC module is typically a Dallas/Maxim DS1287 or DS12887 (DIP-18 package) socketed on the processor board. These contain an internal lithium cell rated for ~10 years. When the clock loses time between power cycles despite a good backup battery, replace the entire RTC module — the internal cell is not user-serviceable. Haas P/N varies by board revision; cross-reference the DS12887 directly from Analog Devices (formerly Maxim) DS12887 product page.

Post-Maintenance Verification Checklist

FAQ