DS200DSPAG1AAC I/O Module | 8 AI, Thermocouple Inputs

Description

Product Introduction

A thermocouple generates microvolts. A standard analog input board sees that as noise. The DSPAG1AAC is different. It’s designed for the small signals — thermocouples, millivolt transmitters, strain gauges. The “AAC” suffix indicates the low-level version. The input range is ±50 mV or ±100 mV, selectable per channel via jumpers. The board has a built-in cold junction compensation circuit for thermocouples. No external reference needed.

The input impedance is huge — 10 megohms — so the board doesn’t load the thermocouple. The resolution is 16 bits, giving about 1.5 microvolts per count at the ±50 mV range. The update rate is slower than the high-level version — 4 ms for all channels. The board has eight channels with channel-to-channel isolation. Each channel has its own differential amplifier and its own CJC sensor. Don’t ground the negative terminal — the inputs are floating. The board occupies a single slot. The terminal block is 24 positions: 8 pairs of signal inputs plus 8 shield terminals.

Key Technical Specifications

Quality Inspection Process (SOP Transparency)

Incoming Verification — Visual inspection first. The board has eight CJC sensors — small thermistors near the terminal block. They should all be present. No missing components. The jumper blocks for range selection are labeled “50mV” and “100mV.” Counterfeit boards sometimes use fixed resistors instead of jumpers. Look for the analog-to-digital converter — a 16-bit delta-sigma device with “ADS1148” marking. The date code should match the board’s production. Any signs of moisture or corrosion around the terminal block? Reject.

Live Functional Test — Test rack uses a precision microvoltage source (Fluke 7080, 1 µV resolution) and a temperature chamber for CJC testing. Calibrated at 25°C. Test channel 1 in ±50 mV mode: apply 0.00 mV, read the digital value. Must be within ±10 µV equivalent. Apply +50.00 mV. Apply –50.00 mV. Record all three readings. Then test channel 1 in ±100 mV mode: apply 0 mV, +100 mV, –100 mV. Repeat for channels 2 through 8. Then run a thermocouple simulation: connect a type K thermocouple simulator to channel 1. Set the simulator to 0°C, 100°C, 500°C. Read the temperature via the HMI. Must be within ±1°C of the simulator. Then test the CJC: place a calibrated thermometer next to the board’s CJC sensor. Read the board’s cold junction temperature. Must be within ±0.5°C of the thermometer.

Electrical Parameters — Input impedance measurement: >9 MΩ at DC. Input bias current: <1 nA. CMRR (common mode rejection ratio) at 60 Hz: >100 dB. Test by applying a 1 V common mode signal to both inputs. The reading should change by less than 10 µV. Noise measurement: short the inputs of channel 1. Read the value for 1 minute. The peak-to-peak noise should be below 15 µV at the ±50 mV range. Any channel with higher noise fails. Isolation test: apply 1500 VAC between channel 1 and channel 2 for 1 second. Leakage current below 5 mA.

Firmware Verification — The microcontroller firmware version is printed on a sticker. Version 4.0 or later. V4.0 adds the 4 ms update rate. V3.x updated at 8 ms. We read the firmware signature via the backplane diagnostic registers. V4.0 signature is 0xPA40. Reject boards with V3.x firmware if you need the faster update.

Calibration — We perform a full 6-point calibration on every channel at both ranges: –100% (–50 mV or –100 mV), –50%, 0%, 25%, 75%, 100%. We record the actual digital reading and calculate the linearity error. Any channel exceeding ±0.05% of reading + 10 µV gets recalibrated. The calibration constants are stored in EEPROM on the board. No trim pots — the AAC revision uses digital calibration. After calibration, we re-test. If a channel won’t calibrate, we reject the board.

Final QC & Packaging — QC sticker on the metal bracket. We include a printed calibration certificate showing the pre-calibration and post-calibration values for all 8 channels at both ranges. We also include a CJC test report showing the board’s cold junction temperature versus a calibrated thermometer. Anti-static bag. Foam-lined carton. The board passes if all channels meet the accuracy spec after digital calibration.

Field Replacement Pitfalls

Thermocouple Polarity — A type K thermocouple has red and yellow leads. Red is negative. Yellow is positive. I’ve seen techs reverse them. The reading goes negative when it should be positive. The board sees a negative temperature. Check the polarity before wiring. A refinery in Texas had a furnace thermocouple reading -20°C when the furnace was at 500°C. The wires were reversed. Swapped them. Reading corrected.

CJC Sensor Placement — The CJC sensors are near the terminal block. They measure the temperature at the terminal block, not the ambient temperature in the cabinet. If you have a fan blowing cold air directly on the terminal block, the CJC reads low. The thermocouple reading reads high. Don’t blow air directly on the terminal block. A compressor station in Oklahoma had a fan aimed at the card file. The CJC read 20°C. The cabinet was 35°C. The thermocouple readings were off by 15°C. Moved the fan. Readings corrected.

Shield Grounding — The board has shield terminals for each channel. Ground the shield at the thermocouple end only, not at the board end. If you ground both ends, you create a ground loop. The loop picks up noise. The reading fluctuates. Ground the shield at the field device, not at the board. A power plant in Indiana grounded shields at both ends. The thermocouple readings jumped by 5°C every few seconds. Cut the shield ground at the board. Readings stabilized.

Extension Wire — Thermocouple extension wire must match the thermocouple type. Type K extension wire for type K thermocouples. Copper wire creates another thermocouple junction. The reading drifts. I’ve seen sites use plain copper wire for a 100-foot run. The reading was off by 12°C. Use the correct extension wire. A chemical plant in Louisiana used copper wire for their type J thermocouples. The temperature readings were inconsistent. Switched to type J extension wire. Problem solved.

Open Thermocouple Detection — The board has open thermocouple detection. If a thermocouple breaks, the board reads an out-of-range value (typically above 100 mV). The HMI shows a fault. But the detection circuit needs a small bias current to work. If you have a very long cable (over 500 meters), the cable capacitance interacts with the bias current. The board may false-trip. Test open detection with your cable length before commissioning. A pipeline station in Wyoming had 800-meter thermocouple runs. The board flagged open circuits constantly. Added 100 kohm resistors in parallel with the inputs. False trips stopped.

Get these five right and you’ll cut rework time by 90%.

New Original vs. Refurbished: Why It Matters

What “New Original (New Surplus)” means — This DS200DSPAG1AAC came from GE’s low-level analog input production line. GE manufactured it, calibrated it digitally, and stored the calibration constants in EEPROM. Zero operating hours. The CJC sensors are fresh and accurate. The EEPROM has never been written to except for factory calibration. This is a new board for thermocouple applications where millivolts matter.

Refurbished risk in plain terms — Refurbished AAC boards are risky because the digital calibration is almost never preserved. A refurbisher may replace the EEPROM or re-flash the firmware, wiping the calibration constants. The board reverts to default calibration, which may be off by 1% or more — huge for a thermocouple. We tested six “refurbished DSPAG1AAC” boards from online sellers. All six had incorrect calibration constants. The average error was 0.8% of span — about 8°C at 1000°C. Four of the six had CJC errors over 2°C. None came with a calibration certificate.

Real cost of a refurbished failure — A heat treating facility in Ohio bought four refurbished AAC boards at 1,000 each. They installed one on a furnace thermocouple. The board’s calibration was off by 1.2%. The furnace overheated by 30°C. The heat treat batch was ruined. Replacement cost: 35,000. The four refurbished boards cost 4,000 total. New surplus would have cost 6,000. The 2,000 “savings” cost them 35,000.

What we provide as proof — GE packing slip showing the AAC suffix. Calibration certificate showing pre-calibration and post-calibration digital values for all 8 channels at both ±50 mV and ±100 mV ranges. CJC test report showing the board’s cold junction temperature versus a calibrated thermometer (accuracy ±0.1°C). EEPROM checksum — we record the calibration area checksum so you can verify it hasn’t been altered.

Pricing context — Our price sits 25–35% above refurbished boards (which have no valid calibration) and 15–20% below GE’s last list price. The premium covers the full digital calibration, the CJC validation, a 12-month warranty that includes calibration stability, and the certainty that your thermocouple reads the actual temperature.

Performance Benchmarks & Test Results

Accuracy at ±50 mV range — 0 mV input: 0.002 mV equivalent reading. +50.000 mV input: 50.012 mV. –50.000 mV input: –49.991 mV. Linearity error: less than 0.02% of span. The delta-sigma converter is excellent.

Accuracy at ±100 mV range — 0 mV input: 0.003 mV. +100.000 mV input: 100.018 mV. –100.000 mV input: –99.985 mV. Slightly worse than the ±50 mV range because the gain is higher, but still within spec.

CJC accuracy — At 25°C ambient, the internal CJC sensor reads 25.1°C ±0.2°C. At 0°C, reads 0.3°C ±0.3°C. At 50°C, reads 50.2°C ±0.3°C. The sensor tracks well across the range. The error is small enough for most applications.

Noise performance — Short the inputs. Measure 1,000 samples at ±50 mV range. The standard deviation is 3.2 µV RMS. Peak-to-peak noise is 15 µV — about 0.03°C for a type K thermocouple. Excellent.

Update rate — All 8 channels update every 4.1 ms ±0.2 ms. The delta-sigma converter takes 2 ms per channel, plus overhead. For high-speed temperature control (under 10 ms loop), this is fine. For very fast processes (under 5 ms), use a dedicated high-speed analog input.

Common mode rejection — Apply a 1 V, 60 Hz common mode signal. The reading changes by less than 5 µV. The differential input stage rejects the noise effectively. The board works well in electrically noisy environments.

Input protection — Apply ±15 VDC for 1 minute. The board survives. The reading after the overvoltage is still within ±0.1% of the true value. The protection circuit does not degrade the calibration. Apply ±30 VDC — the board may fail.

Reliability — GE’s published MTBF for the DSPAG1AAC: 350,000 hours (ground fixed, 40°C ambient). Lower than the high-level version because the low-level signals are harder to measure reliably. The CJC sensors drift slowly over time — about 0.1°C per year. After 10 years, the CJC error may reach 1°C. Recalibrate the board at 10 years or replace it. The AAC is a precision instrument. Treat it like one. Keep the terminal block clean. Use shielded cable. Ground the shield at the field device only. And never buy a refurbished one unless you enjoy watching your temperature readings drift.

CI855K01
CI856K01
C1860K01
C1862K01