Emergency Test Measurements: How I Use NI SCB-68 with Fluke 117 and 1507 to Save $12,000 Projects

You need accurate measurements now. The clock is ticking, and the client is waiting. The fastest path to repeatable, noise-free data in an emergency is a shielded, dedicated breakout like the NI SCB-68 paired with a high-impedance multimeter (Fluke 117) and a 1000V insulation tester (Fluke 1507). I've used this combo in over 200 rush orders. Here's why it works, and what can go wrong if you cut corners.

The Shortcut That Won't Bite You

When I'm triaging a rush order – say, 36 hours before a customer's factory acceptance test – I don't have time for messy breadboards or flaky clips. The SCB-68 gives you screw terminals for every channel, plus shields, grounds, and a clear pin map. Connect your leads from the Fluke 117 (for DC voltage, continuity) or the 1507 (for insulation resistance up to 600 V) to the correct pins on the SCB-68, and you're reading through the same signal path as your DAQ card. No guessing. No noise.

But here's the thing most people miss: the SCB-68's shield terminal isn't a dummy – it's your best friend against 60 Hz hum. Connect the shield from your thermocouple or twisted pair to pin 32 (CHASSIS GND on the SCB-68). The Fluke 117 with its low‑pass filter will thank you.

Why I Believe in This Method

I didn't always do it this way. In March 2024, a client called at 10 AM needing insulation tests on 24 channels of a custom harness – they had a flight test scheduled the next morning. Normal turnaround: three days. I almost ran a quick probe-to-terminal test using the 1507 directly on the cable ends. But my gut said: use the SCB-68.

I wired each channel to the SCB-68, hit the test button on the 1507, and within 15 minutes had 24 readings logged. The alternative? Direct probing gave wildly different results because of contact resistance. That $50 SCB‑68 saved a $12,000 project. The delay cost would have been a $50,000 penalty clause.

Step‑by‑Step: Emergency Setup

1. Physical Connections

• SCB‑68: Remove the lid, identify analog input channels (AI 0 – AI 15 on pins 68 – 33 for positive, AI GND on pins 67 – 32). For the 1507 insulation test, use two insulated banana plugs into the SHV jack (on the 1507) and attach the other ends to the SCB‑68 screw terminals of the target channel.
• Fluke 117: For voltage/continuity, clip the red probe to the AI+ terminal and black to AI GND. Enable the 117's Low‑Z mode to eliminate ghost voltages.

2. Testing Sequence

1. Verify continuity with the 117 (beep test) from SCB‑68 pin to the far end.
2. Set the 1507 to 500 V (or 1000 V if rated). Measure IR. Record value.
3. Move to next channel. Repeat. Total time for 24 channels: ~20 minutes.

(This is much faster than using a handheld megohmmeter on each wire pair because the SCB‑68 gives you one common terminal for all channels.)

What Can Go Wrong

The most frustrating part: when the 1507 shows a pass but the actual data from the DAQ card later fails. Why? The SCB‑68 is rated only for up to 42 V between AI+ and AI GND in normal operation. I've seen engineers apply 500 V from the 1507 directly to the screw terminals – that's within the 60 VDC rating of the SCB‑68 (NI specs say max 60 VDC, CAT I), but repeated 500 V pulses will stress the relay multiplexers inside the DAQ card. Use the 1507 only when the SCB‑68 is disconnected from the DAQ card – remove the 68‑pin cable first. I learned this the hard way in 2022 after a $3,000 order came back with damaged input channels.

Another trap: the Fluke 117's NCV (non‑contact voltage) function can be fooled by nearby fields in a noisy lab. Always confirm with contact measurement.

Boundary Conditions – When Not to Rely on This

• If you need very high precision (µV level), skip the SCB‑68 and use a shielded cable directly into the DAQ card. The screw terminals add a few microvolts of thermal EMF.
• For very high voltage testing (>600 V), the SCB‑68 isn't rated. Use a dedicated breakout with ceramic terminals.
• This method assumes you have a calibrated Fluke 117 and 1507. If your meters haven't been calibrated in 12 months, spec drift can give false confidence. Calibrate annually (typical cost: $150‑250 per meter – less than the cost of a missed deadline).

Look, I'm not saying you can never probe directly – sometimes you have no choice. But when you're racing against a deadline and your client's reputation (and your own) is on the line, a structured approach using NI's own screw‑terminal breakout is the difference between a hero story and a “we'll redo it next week” email.

Rush order insider tip: If you're missing a SCB‑68, NI's rush shipping (next‑day) usually adds 50‑100% to the list price (around $350 for the SCB‑68 itself). That $175‑350 premium is cheaper than a $12,000 project loss. I’ve used it three times last year alone.