What is a servo press?

A servo press uses servo motors for precise force and position control during forming operations.

How do I calculate servo press tonnage?

Use our free tonnage calculator at servopress.club/tonnage-calculator.html

Yaskawa SGM7J parameter setup - anyone have a baseline?

mchen_servo

Just installed a Yaskawa SGM7J motor on our press and the default parameters are... not great. Anyone have a good baseline setup for press applications?

**Initial Parameters
**
Motor parameters (from nameplate):
- Pn170: Rated current (A)
- Pn171: Rated speed (rpm)
- Pn172: Rated torque (Nm)

Position control:
- Pn205/206: Electronic gear ratio
- Pn207: Position loop gain (start at 40)

Safety:
- Pn520: Position deviation alarm (3-5mm in pulses)
- Pn522: Torque limit (300% for press)

**Electronic Gear Ratio Example
**
SGM7J 23-bit encoder = 8,388,608 pulses/rev
10:1 gearbox, 20mm ball screw pitch, 0.001mm resolution:

Pulses/mm = 1000
Pulses/rev at motor = 20,000 x 10 = 200,000
Ratio = 8,388,608 / 200,000 = 41.94

Set Pn205=4194, Pn206=100

**Tuning for Press
**
High inertia + variable load = auto-tuning often fails.

Manual procedure:
1. Pn100=50Hz, Pn101=20ms, Pn102=40/s
2. Run 10 strokes at 50% speed
3. If stable, increase Pn100 by 10Hz
4. Stop when oscillation appears, reduce 20%

Typical result for 160T press: Pn100=80-120Hz, Pn102=60-80/s

anna_plc

The electronic gear ratio (Pn205/206) is the most important parameter to get right. It converts encoder pulses to physical distance. Get it wrong and your position commands won't match actual slide movement. Calculate it from: encoder resolution ? gear ratio ? ball screw lead. Double-check by commanding a known distance and measuring with a dial indicator.

jim_apps

For press applications, I always set the position loop gain (Pn102) lower than the default. Presses have high inertia and compliance in the mechanical train. High position gain causes oscillation. Start at 50% of default and increase until you see the first hint of vibration, then back off 20%. That's your sweet spot.

carlos_mx

Pn102 advice above is solid. I'd add ??watch Pn110 (speed loop gain) too. On our 200T with SGM7J, the default speed gain caused a nasty 120Hz resonance that showed up as a whine at mid-stroke. Dropped Pn110 from 400 to 280 and added a notch filter at Pn409/Pn40A. Completely gone.

For the torque limit parameters (Pn402/403), don't set them too tight initially. We had a press that would fault on E.741 (excessive torque) during cold starts in winter because the grease was stiff. Set forward torque limit to 300% for the first 5 minutes of operation, then a PLC routine drops it to your normal 150-180% once the oil temp sensor reads above 15C.

One more ??if you're using the absolute encoder option, make sure Pn002.1 is set correctly for your battery backup. Lost our home position twice before we figured out the battery voltage threshold was set too aggressive and it was resetting during weekend power blips.

pressroom_tom

One thing nobody mentioned yet — before you even touch gains, verify your inertia ratio. Run the Yaskawa auto-measurement (Fn00B) with the die out and slide at mid-stroke. If the measured inertia ratio (Un00D) is above 15:1, you're going to fight tuning forever. Our 300T was showing 22:1 because the coupling between motor and ball screw had worn — 0.08mm backlash that the encoder couldn't see.

Also be careful with Yaskawa's auto-tuning (Fn002). It works great on simple linear axes but on a press it tends to set the speed loop way too aggressive because it tests with no load. I've seen it set Pn110 at 500+ which is asking for trouble once you hit material. My approach: run auto-tune, then immediately cut Pn110 by 40% and Pn102 by 30%. That's your real starting point.

The other gotcha with SGM7J specifically — check Pn00B (motor code). If you replaced a motor and the drive still has the old motor code, the current loop parameters will be wrong and no amount of speed/position tuning will fix the vibration. Ask me how I know... cost us two days on a weekend changeover.