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

Tonnage monitoring without a dedicated system - DIY options?

steve_maint

We don't have budget for a dedicated tonnage monitoring system. Anyone rigged up a DIY solution that actually works?

**Using Drive Torque as Tonnage Proxy
**
The servo drive reports motor torque in real time. Torque is proportional to forming force (with some losses from the mechanical transmission).

**How to access torque data:**
- Siemens SINAMICS: r0080 (torque actual value, in Nm)
- Yaskawa: Un000 (torque reference monitor)
- Mitsubishi MR-J4: monitor parameter Md.Trq (torque)
- ABB ACS880: parameter 01.10 (motor torque %)

**Setting Up Monitoring
**
1. **Establish baseline**: run 50 good parts and log the peak torque value for each stroke. Calculate the average and standard deviation.

2. **Set limits**:
- High limit (overload): average + 3 standard deviations (typically 110-120% of average)
- Low limit (broken punch): average - 3 standard deviations (typically 70-80% of average)

3. **Configure drive fault output**: most drives can output a digital signal when torque exceeds a threshold. Wire this to the press safety circuit to stop on overload.

**Limitations vs. Dedicated Tonnage System
**
- Cannot distinguish between different punches in a progressive die
- Torque-to-force conversion changes with wear and lubrication
- Response time is 1-5ms vs. 0.1ms for strain gauges
- Cannot detect gradual punch wear (only sudden changes)

**Low-Cost Hardware Option**

A load cell under one of the press columns connected to a simple PLC: 50-ton load cell costs $200-$500, basic PLC with analog input costs $300-$800. Total under $2,000 for a functional tonnage monitor.

anna_plc

Yes, you can use the servo drive's current data as a proxy for tonnage. Motor current is proportional to torque, and torque is proportional to force (through the mechanical linkage). On Yaskawa, monitor Un000 (torque reference). On Siemens, monitor r0079. It's not as accurate as strain gauge monitoring but it catches broken punches and gross overloads.

lisaQ_metrology

We did exactly this on our 300T press. Set up the drive to output an analog signal proportional to torque, fed it into our data acquisition system, and built force envelopes for each die. Catches about 85% of what a dedicated tonnage monitor would catch. Not perfect but way better than nothing for $500 in parts.

mchen_servo

We did the drive torque monitoring route on three presses and it works, but there are some gotchas nobody mentions.

First — the torque-to-force relationship is not linear through the full stroke on a crank-driven servo. It depends on the crank angle. At BDC the mechanical advantage is highest so a small torque change means a big force change. At mid-stroke it is the opposite. You need a lookup table or a simple trig correction factor, not just a straight multiplier. Without that correction our readings were off by 30% at mid-stroke.

Second — drive torque includes acceleration forces. If you are monitoring during a fast approach, the torque spike from accelerating the slide mass will look like a forming force. Filter it out by only sampling during the actual forming portion of the stroke, say last 10-15mm before BDC. On Siemens S120 you can set up a position-triggered trace that only captures data in that window.

Third — recalibrate every time you change the drive parameters. We updated the jerk limit on one press and the torque baseline shifted 8%. Took us two days to figure out why parts were getting flagged as out-of-spec when nothing had changed on the tooling side.

carlos_mx

Following up on my earlier post about the crank angle correction — here's what we actually implemented if anyone wants to replicate it.

The math: Force = Torque / (r * sin(theta + phi)), where r is crank radius, theta is crank angle from TDC, and phi accounts for the connecting rod geometry. Near BDC (theta approaching 180), sin approaches zero and the force amplification goes to infinity theoretically. In practice this means your torque-based measurement gets very noisy near BDC because tiny torque fluctuations look like huge force changes.

Our workaround: we only use the torque-based measurement between 30 and 150 degrees of crank rotation. Outside that window, the signal-to-noise ratio is garbage. For most stamping operations, the actual forming work happens in that window anyway.

The $500 setup lisaQ_metrology mentioned is realistic. Here's our BOM:
- Analog output card for the drive (most drives have this built in, just need the connector): $0-150
- 24-bit USB DAQ module (we use a Measurement Computing USB-2408): ~$300
- Shielded cable and connectors: ~$50
- Python script running on a Raspberry Pi for logging and envelope comparison: $0 (plus a $50 Pi)

Total: about $400-550 depending on what you already have.

The limitation vs real strain gauge monitoring: you can't detect off-center loading. The drive sees total torque regardless of where the force is applied on the slide. A dedicated 4-corner tonnage monitor catches a broken pilot pin loading one corner — the drive torque method won't. For die protection (broken punch, double-blank, short feed), the drive method works great. For precision process monitoring, you eventually want strain gauges.

We ran the DIY system for 2 years, proved the value, then got budget approved for a proper Helm system. The DIY data actually helped justify the purchase because we could show management exactly how many bad parts we caught.