Industrial digital input chips provide serialized data by default. However, in systems that require real time, low latency, or higher speed, it may be preferable to provide level-translated, real-time logic signals for industrial digital input channel.
So, some industrial digital inputs sample and serialize the state of eight 24-V current sinking inputs under SPI or pin-based (LATCH) timing control, allowing for readout of the eight states via SPI. A serial interface is used to minimize the number of logic signals requiring isolation, which is particularly beneficial in high channel count digital input modules.
Serialization of logic signals uses simultaneous sampling of the signals so that the signals become time quantized. This means that real-time information content is lost, which can be of concern in certain systems. Examples are applications where timing differences between switching signals are of concern, such as incremental encoders or counters.
These applications either necessitate the use of high-speed sampling with high-speed serial readout or the use of non-serialized parallel data, as provided by the MAX22195, an industrial digital input with parallel output. Using the MAX22190/MAX22199 industrial digital input devices with parallel operation provides the benefit of diagnostics and configurability.
This article delves into the characteristics, limitations, and design considerations regarding techniques for generating parallel logic outputs with industrial digital inputs.
Design details
The technique is based on repurposing the eight LED outputs to function as logic signals. LEDs serve to provide a visual indication of the digital input’s state—useful for installation, maintenance, and in service. The characteristics and specifications of industrial inputs are clearly defined in the IEC 61131-2 standard, with the output state being binary in nature: either on or off.
The MAX22190/MAX22199 chips feature energyless LED drivers that power the LEDs from the sensor/switch in the field, not drawing current/power from a power supply in the digital input module. These devices limit the input current to a level settable by the REFDI resistor. This is done to achieve the lowest power dissipation in the module.
For the common Type 1/Type 3 digital inputs, the input current is typically set to a level of ~2.3 mA (typ) to be larger than the 2.0 mA minimum required by the IEC standard. The ICs channel most of the ~2.3 mA field input (IN) current to the LED output pins, and only ~160 µA are consumed by the chip.
With the LED drivers being current outputs, not voltage, the current needs to be converted to voltage for interfacing with other logic devices like digital isolators and microcontrollers. Resistors are the simplest trans-resistance element for this purpose, as shown in Figure 1.
Figure 1 LED pins are used as voltage-based logic outputs. Source: Analog Devices Inc.
Using the LED output pins in this manner is not documented in the product datasheets. This article investigates the characteristics and possible limitations.
LED pin characteristics
When using ground-connected resistors on the LED pins to create voltage outputs, the following needs to be considered:
- What is the maximum voltage allowed on the LED pins?
- Is there interaction/feedback from the LED_ pin to the IN_ pin?
- Specifically, does voltage on the LED pins result in a change of the input current, as minimum current levels are mandated by the IEC standards?
- Do the LED output currents show undesired transient behavior, such as overshoots or slow rise/fall times?
- Are the LED outputs suitable for use as high-speed logic signals when the inputs switch at high rates?
- Are the LED outputs filtered (as programmable by SPI)?
The MAX22190/MAX22199 datasheets’ absolute maximum ratings specify the maximum allowed LED pin voltages as +6 V. This indicates that the LED pins are suitable for use as 5 V (and 3.3 V) logic outputs, with the caveat that the voltage may not be higher than 6 V.
The impact of the LED pin voltage on other critical characteristics needs to be evaluated. Of particular concern is the change of the input current with the presence of high LED pin voltages, as the current is specified by the standards. The critical case is with the field voltage close to the 11 V on-state threshold voltage, as defined for Type 3 digital inputs.
Figure 2 shows the measured field input current dependence on the LED pin voltage for three field input voltages close to the 11-V level: 9 V, 10 V, and 11 V. The 10-V and 9-V levels were chosen as these are within the transition region for Type 3 inputs, and their input currents have no defined minimum, while the minimum for the 11 V input case is 2 mA.
Figure 2 Field input current is dependent on the LED pin voltage. Source: Analog Devices Inc.
With the field voltage at the 11-V threshold, the blue curve shows that the input current starts decreasing when the LED voltage is higher than ~5.8 V. The current decrease is only 0.6% at 6 V. For cases of 9 V and 10 V, which are in the transition where the currents are not defined, the measurements show that the input current is still above 2 mA for up to 5.5-V inputs.
In conclusion, this shows that the MAX22190/MAX22199 will produce 5-V LED logic outputs (as well as lower voltage logic like 3.3 V) and still be compatible with Type 3 digital inputs. For Type 1 digital inputs, the case is trivial since the on-threshold is much higher at 15 V, meaning that the LED pins will also provide 5-V logic levels without any impact on the field input current.
Parallel operation example
Figure 3 shows a 10-kHz field input (yellow curve) with the resulting LED output voltage in blue. A 1.5-kΩ resistor was used on the LED output, which provides a 3.3 V logic signal. Glitch filtering was disabled (default bypass mode).
Figure 3 In 10-kHz switching, Channel 1 has field input and Channel 2 has LED output. Source: Analog Devices Inc.
Regarding the transient behavior of the LED output current under switching conditions, Figure 3 shows a case of 10-kHz switching. A 1.5-kΩ resistor was used to convert current to voltage. The scope shot illustrates that the LED outputs do not produce transient overshoots or undershoots that could damage logic input devices. The rise and fall times are fast and do not lead to signal distortion.
Using the SPI interface
The MAX22190/MAX22199 devices feature SPI-programmable filters to enable per-channel glitch/noise filtering. Eight filter time constants up to the 20-ms level are available as well as a filter bypass for high-speed applications. The selected noise filtering also applies to the LED outputs to make the visual representation consistent with the electrical signals.
Diagnostics are provided via SPI, like low power supply voltage alarms, overtemperature warnings, short-circuit detection on the REFDI and REFWB pins, and as wire-break detection of the field inputs.
The power-up default state of the register bits is:
- All eight inputs are enabled
- All input filters are bypassed
- Wire-break detection is disabled
- Short-circuit detection of the REFDI and REFWB (only MAX22199) pins is disabled
Hence, the SPI interface does not need to be used in applications that do not require glitch filtering (for example, for high-speed signals) and diagnostics. In cases where the per-channel selectable glitch/noise filtering is needed or diagnostic detection is wanted, SPI can be used.
The LED output waveform does not show overshoots or other undesired irregularities such as varying voltage in the on-state. This illustrates that the LED outputs can be used as voltage outputs. Its characteristics and limitations are investigated.
Glitch filtering
The MAX22190 and MAX22199 devices provide per-channel selectable glitch filtering. The following content demonstrates the effect of the glitch filters on the LED outputs by example of a 200-Hz switching signal with filter time set to 800 µs. Defined glitch widths were emulated by changing the duty cycle. Both positive and negative glitches were investigated.
Figure 4 shows an example of 750-µs positive pulses being filtered out by the 800-µs glitch filter. So, positive glitch filtering works both for the LED outputs as well as the SPI data.
Figure 4 Here is an example of positive glitch filtering. Source: Analog Devices Inc.
Negative glitches are, however, not filtered out at the LED outputs, as shown in Figure 5, where a 750-µs falling pulse propagates to the LED output. This differs from using the SPI readout, for which both positive and negative glitches are successfully filtered.
Figure 5 This image shows negative glitch filtering. Source: Analog Devices Inc.
Figure 6 shows the LED output signal with an 800-µs glitch filter enabled and input switching with a 50% duty cycle. The rising edges are delayed by ~770 µs while the falling edges show no delay. This illustrates that the filters do not work properly with the LED outputs.
Figure 6 This image highlights the filtering effect on LED output. Source: Analog Devices Inc.
High frequency switching
For applications with high switching frequencies, low propagation, or low skew requirements, glitch filtering would be disabled. In bypass mode (glitch filters) and 100-kHz input, the LED output results in the waveforms shown in Figure 7.
Figure 7 The 100-kHz input switching is shown with filter bypass. Source: Analog Devices Inc.
While the falling edges show low propagation delay of ~60 ns, the rising edges have significant propagation delay as well as jitter. The rising edge jitter is in the range of ±0.5 µs with an average propagation delay of ~1 µs. The rising delay and jitter are due to the ~1 MHz sampling documented in the datasheet. Sampling does not occur on the falling edges, hence the fast response.
This illustrates that the LED outputs have rise time/fall time skews of up to ~1.5 µs with jitter. Channel-to-channel skew is low on the falling edges but much higher on the rising edges. This could limit the use of the LED outputs in some applications.
Design considerations
This section discusses some considerations required when using the LED output pins as voltage outputs.
Ensure that the MAX22190/MAX22199 current-drive LED outputs are voltage limited to not exceed the safe levels of the logic inputs that they drive. While the REFDI resistor sets the field input current to a typical current level, the actual input current has a tolerance of ±10.6%, as specified in the datasheets. Thus, the voltage across the resistor will be in the ±10.6% range.
Logic inputs typically have tightly specified absolute maximum ratings, like VL + 0.3 V, where VL is the logic supply voltage. When interfacing two logic signals, a common VL supply is often used to ensure matching as standard logic outputs have push-pull or open-drain outputs whose maximum output voltage is defined/limited by a logic supply, VL.
One can make the typical LED pin’s output voltage lower to ensure that absolute maximum ratings are not exceeded for the input. Alternatively, one can consider that the LED pin’s ~2.3 mA output current will not damage a logic input, as these are commonly specified for tolerating much higher latch-up currents, in the 50 mA to 100 mA range. This needs to be verified for the device under consideration. The third, less attractive, option is to limit the voltage by clamping.
Standard logic outputs are push-pull and thus low impedance, providing high flexibility in driving logic inputs. In contrast, the LED outputs are open-drain outputs where the pull-down resistor with parasitic capacitance determines the switching speeds.
Without additional capacitors, switching rates of 100 kHz and higher are feasible.
The MAX22190/MAX22199 industrial digital inputs can be used as an octal input having eight parallel outputs, despite being documented for serialized data operation. To this purpose, the LED drivers, originally intended for visual state indication, are repurposed as voltage-based or current-based logic outputs. When using parallel operation in this manner, the use of the SPI interface is optional and provides all the diagnostics as well as device configurability with some limitations.
Wei Shi is an applications engineer manager in the Industrial Automation business unit of Analog Devices based in San Jose, California. She joined Maxim Integrated (now part of Analog Devices) in 2012 as an applications engineer.
Reinhardt Wagner was a distinguished engineer with Analog Devices in Munich, Germany. His 21-year tenure primarily involved the product definition of new industrial chips in the areas of communication and input/output devices.
Editor’s Note
This article was written in cooperation with Chin Chia Leong, senior staff engineer for hardware at Rockwell Automation.
Related Content
- Analog outputs for industrial applications
- Isolating analog signals using a digital isolator
- Isolated PLC digital inputs for industrial control
- Audio chip moves machine learning from digital to analog
- Using Smart I/Os to build pin-level digital logic functionality and reduce CPU loading
The post Design digital input modules with parallel interface using industrial digital inputs appeared first on EDN.
