By: AD Lionel Wallace, Field Application Engineer | Jason Fischer, Application Engineer | Ben Douts, Field Application Engineer


Is there an easy way to create a high voltage supply suitable for sensor biasing applications?


Of course, just use ICs with integrated precision feedback resistors.


Adjustable high-voltage power supplies that provide high-precision output are difficult to build. Drift due to time, temperature, and differences in the production process often contributes to errors. Resistive networks traditionally used for feedback are a common source of error. This paper presents a novel design that exploits the feedback path of an integrated circuit (IC). This circuit is used in sensor biasing applications and offers higher accuracy, lower drift, more flexibility, and even cost savings than designs that utilize a resistor network to provide feedback.

Figure 1 shows the traditional approach to building an adjustable high voltage bias circuit. The DAC is used to generate the control voltage and the op amp is used to provide the gain. The circuit in Figure 1 provides an output of ~0 V to 110 V with a control voltage range of 0 V to 5 V.

Because high-voltage sensors are often quite capacitive, a resistor (R2) is typically used to isolate the op amp output from the load, avoiding potential stability problems.

Figure 1. Traditional Approach for High Voltage Adjustable Bias Circuits

In some cases, these circuits work very well. Implementing feedback with an IC is beneficial when greater accuracy or consistent long-term performance is required.

IC feedback implementation

The circuit shown in Figure 2 is configured with the following design goals in mind:

  • Control voltage: 0 V to 5 V
  • Output Voltage Adjustable Range: ~0 V to 110 V
  • Output current > 10 mA
  • Initial accuracy: ±0.1% (typ.)
  • No need for external precision resistors

The circuit in Figure 2 mainly consists of three parts: the control voltage, the integrator, and the feedback path. As mentioned above, the feedback is provided by an integrated circuit rather than a resistor network.

The control voltage input range is 0 V to 5 V. A circuit gain of 22 provides output bias voltages from ~0V (0 V×22) to 110 V (5 V×22). To generate the control voltage, the AD5683R was chosen. The AD5683R is a 16-bit nanoDAC® with an internal 2 ppm/°C reference. The 5 V output range was chosen to allow the circuit to provide a bias voltage from ~0 V to 110 V in ~1.68 mV steps.

The integrator chooses the LTC6090. The LTC6090 is a high voltage operational amplifier capable of providing rail-to-rail output and picoamp input bias current. Low input bias current is critical to achieve the required high accuracy. In addition, the open-loop gain provided by the LTC6090 is typically greater than 140 dB, so system errors due to finite loop gain are greatly reduced.

The LTC6090 compares the feedback voltage to the control voltage and integrates the difference, or error, to adjust the output (VBIAS) to the desired setpoint. The time constant formed by R1 and C1 sets the integration time, which does not affect amplifier accuracy, so precision components are not required. For testing, the load was modeled as an 11 kΩ resistor in parallel with a 2.2 μF capacitor.

Figure 2. LTspice® Schematic for ~0 V to 110 V Bias

Figure 3. Screenshot of LT1997-2 Design Tool, Attenuation = 22

The LT1997-2 difference amplifier provides 22 times (gain = 0.4545…) attenuation to the feedback loop. The connections required to achieve 22x attenuation can be easily determined with the LTC1997-2 online calculator. A screenshot of the tool is shown in Figure 3.

The LT1997-2 is very flexible and supports a wide range of gain/attenuation combinations. Examples are provided in the data sheet, and the evaluation board supports many gain combinations with jumper selectable settings.

Figure 4. LT1997-2 Evaluation Board (Gain Set via Jumpers and Additional Wires)

Test setup

The circuit is modeled in LTspice and conforms to the design goals. Use the following evaluation boards to help with hardware testing:

  • EVAL-AD5683R: AD5683R DAC Evaluation Board
  • DC1979A: LTC6090 140 V Rail-to-Rail Output Operational Amplifier Evaluation Board (modified for testing)
  • DC2551A-B: LT1997 Configurable Precision Amplifier Demonstration Board (modified for testing)
  • DC2275A: LT8331 Booster Demo Board, 10 V ≤ VIN ≤ 48 V, 120 VOUT, up to 80 mA
  • DC2354A: LTC7149 Buck Demo Board, Configured for Negative VOUT; 3.5 V ≤ VIN ≤ 55 V; VOUT = –3.3 V/–5 V/adjustable to -56 V, up to 4 A

generate control voltage

Use the AD5683R evaluation board to set the control voltage for the circuit. The board is connected via a USB port to a laptop running Analog Devices' ACE (Analysis, Control, Evaluation) software. ACE provides a simple GUI to configure the AD5683R and set the DAC output voltage. The output voltage provides the setpoint for the high voltage bias output.

Figure 5. Test Configuration Block Diagram

Figure 6. Screenshot of the ACE interface of the AD5683R evaluation board

DC accuracy

The measurements in Table 1 and Figure 7 were performed using a Keysight 34460A DMM at 24°C ambient temperature. The output of the AD5683R evaluation board is calibrated to four decimal places and controlled through Analog Devices' ACE software. These results are from a set of boards and do not represent min/max specs.

Table 1. Measured vs. Expected Output Voltage

Figure 7. Output Voltage Error vs. Bias Voltage

Note that below the ~40 V output, the error is dominated by the amplifier offset within the circuit. At low bias voltages, the magnitude of the offset is larger than the gain error. At higher bias voltages, the offset contributes a smaller percentage of the error and the gain error dominates. Error analysis and more detailed information are provided later in this article.

AC response

Apply a step function to control inputs of different voltages. Measure the output and feedback voltages (see Figures 8 to 10). Note that the bias voltage ramps smoothly to the desired value.

Figure 8. Step Response (0 V to 1 V Control Input)


Figure 9. Step Response (0 V to 2.5 V Control Input)

Figure 10. Step Response (0 V to 5 V Control Input)

start waveform

Observe the startup waveforms of the power and signal. This is to ensure that high voltages are not accidentally applied to the biased output. The AD5683R provides a control voltage starting from 0 V. Small glitches of ~3V were observed at the bias output as the supply voltage increased. Given the high voltage nature of the bias output, this is acceptable for testing purposes.
If this circuit is to be used in a production system, it is recommended to control the power supply sequence so that the control voltage is applied first and then the high voltage power supply is activated. This power-up sequence will avoid the possibility of high voltage spike pairs on the bias voltage output during start-up. A simple sequencer such as the ADM1186 is sufficient for this function.

Figure 11. Startup Waveforms – Power

Figure 12. Startup Waveform – Signal

Test setup photo

The LTC6090 evaluation board is mounted on the bottom of the LT1997-2 evaluation board. The test setup only requires modification of these evaluation boards. The DAC and power supply evaluation boards are used in stock configurations and are not shown for simplicity.

Figure 13. LT1997-2 Evaluation Board and Bottom Mounted LTC6090 Evaluation Board

Error Analysis

We performed an error analysis. The main error sources in the circuit and their typical and maximum values ​​are shown in Table 2.

The maximum error is calculated to be 0.0382% or 42 mV at 110 V biased output, including all errors due to device variation and variation over the full temperature range (-40°C to +125°C). The typical error at 110 V bias output is calculated to be 0.00839%, which matches the measured result (0.008% or 9 mV).

A note on power

The hardware used during testing was powered by ±5 V, 24 V, and 120 V supplies. Here are some additional notes on how to choose these power rails:

The AD5683R DAC requires a 5 V supply.

  • In order to achieve the 5 V output of the DAC, the supply voltage may have to be slightly higher than 5 V. Even small loads may limit the maximum output value. See Figure 38 on page 15 of the AD5683R data sheet for additional information.

-5 V is to allow the LTC6090 and LT1997-2 to operate with a control voltage input close to 0V.

  • The input common-mode range of the LTC6090 is limited to 3 V above V-.
  • For convenience, use the LTC7149 demo board to generate the -5 V rail.
  • The LTC7149 evaluation board is capable of delivering up to 4 A output.
  • The circuit requires less than 25 mA at -5 V and a simple charge pump inverter will suffice. As an example, consider the ADP5600.

120 V is used for V+ of the LTC6090.

  • While the LTC6090 provides rail-to-rail output, V+ requires additional headroom under heavy loads.

24 V is used as the positive supply for the LT1997-2.

  • This voltage was chosen to avoid Over-The-Top® operation. Certain characteristics of the LT1997-2 degrade in the Over-The-Top region. See page 14 of the LT1997-2 data sheet for additional information.

Table 2. Output Voltage Error Analysis

* Includes device variation and full temperature range
** at 25°C

Comparison of IC Feedback and Traditional Resistor Network Feedback

Let's compare several design metrics for the traditional method shown in Figure 1 and the IC feedback method shown in Figure 2. For this comparison, the LT1997-2 (see Figure 14) was chosen as the IC for the feedback network. Note that highly matched precision resistors are embedded in the LT1997-2.

Figure 14. LT1997-2 functional block diagram

Table 3. Comparison of LT1997-2 with two 1206 discrete precision resistors (Note: 1206 was chosen because it operates at 200 V)

Table 5. Comparison of LT1997-2 and silicon-based precision resistors

*MAX5490VA10000+, 1,000-piece price from Maxim's website as of December 2020

LT1997-2IDF#PBF, 1,000-piece price from ADI website December 2020

While the LT1997-2 is significantly more expensive than two chip resistors, it performs much better. Compared to metal film resistor networks, the LT1997-2 offers size and cost advantages. Compared to silicon-based resistor networks, the LT1997-2 has advantages in accuracy and operating voltage. In addition, the integration of different resistor values ​​within the LT1997-2 is an advantage over all competing solutions, providing gain flexibility with external jumpers when needed.

There is another, perhaps not obvious, advantage of using ICs with integrated precision resistors. The amplifier's summing junction is buried within the device and not exposed to the PCB. Therefore, these sensitive nodes are protected from disturbing inputs. Also, in many gain configurations, internal resistors are externally connected to ground or the output, avoiding leakage paths that can affect circuit accuracy. Leakage paths are a common source of error in higher voltage circuits. For more information on this topic, see page 14 of the LTC6090 data sheet.

in conclusion

Adjustable high voltage bias circuits have traditionally used op amps to generate precision outputs through a resistive feedback network. While this approach is easy to understand, achieving precise, repeatable performance is difficult. Using an IC instead of a resistor network to provide feedback can provide more accurate and consistent results.

About the Author

Lionel Wallace joined Analog Devices in 2009. During his tenure at Analog Devices, he held various engineering and sales roles. Lionel currently works in Alabama as a Field Applications Engineer. Lionel holds a bachelor's degree in electrical engineering from Auburn University and a master's degree in electrical engineering from the University of Alabama at Huntsville.

Jason Fischer is an applications engineer assisting the Americas East sales team at Analog Devices. He is responsible for supporting multiple products with a focus on prototyping and evaluation of switch-mode power supplies for industrial, telecom, medical, and military applications. His previous work experience includes production management, circuit design, test system development and RF regulatory testing. Jason received his BS in Electrical Engineering from Bloomsburg University, Pennsylvania, in 2014.

Ben Douts is a Field Applications Engineer at Analog Devices in South Carolina. He has worked in a variety of areas, including test engineering and integrated circuit design, with a focus on precision analog circuits and power management. Ben received his bachelor's degree in electrical engineering from the Massachusetts Institute of Technology in 1998.

Leave a Reply

Your email address will not be published.