Testing Pulse Transformers for Maxim Integrated Remote Sensor Technology

Abstract: This application note describes how to design, test, evaluate, and characterize pulse transformers to be used for the 71M654x family of electricity metering ICs in conjunction with the 71M6203, 71M6201, 71M6103, 71M6113, or 71M6601 remote sensor ICs. The transformer design and tests are also applicable to the combination of Maxim Integrated’s power monitoring ICs, the MAX78700 and the MAX78615+LMU.

This document is written for a highly specific application and contains information about proprietary technology, and is only available to customers whose requirements closely match the application. To request the full document, please complete the form below.

Please enter information in English only.



Remote sensor interface ICs (the 71M6103, 71M6113, 71M6201, 71M6203, 71M6601, 71M6603) are normally con-nected to host ICs (the 71M6541D/F/G, 71M6541DT/FT, 71M6542F/G, 71M6542FT, 71M6543F/G/GH/H, and 71M6545/H) via small, low-cost pulse transformers. A similar connection using small, low-cost pulse transformers is used between the power monitoring ICs MAX78700 and the MAX78615+LMU. This application note explains how to design, test, and evaluate pulse transformers to be used with Maxim Integrated’s remote sensors in electricity meters.
Just like their analog counterparts (current transformers, or CTs), pulse transformers can be influenced by strong external magnetic fields. In the worst case, the transformers saturate and therefore lose their inductivity, which turn the transformers into low-ohmic loads to the drivers in the electricity metering device. This shorts out the differential drivers and consequently, causes the supply voltage to collapse. Even before this happens, data integrity, and with it, meter accuracy, are lost.
Consequently, a metering system consisting of a 71M654x electricity metering IC and one or more 71M6xxx remote sensor ICs that are connected to current shunt sensors is not immune to magnetic fields. Meter manufacturers have the option to use magnetic shielding, or better, to specify pulse transformers with high-saturation cores when magnetic resistance is required. Such transformers can be built by virtually any experienced manufacturer of magnetic components; many meter manufacturers will want to use products made by their preferred vendors of magnetic components.

Transformer Design and Test

Note that protection from magnetic effects cannot be implemented with a one-size-fits-all solution. Each and every case is different, depending on whether AC or DC fields are present, whether the meter is immersed in a magnetic field or just exposed to a magnet placed on the outside of the device, or whether the meter enclosure provides enough space to move the transformers away from the field source. Further product variation arises due to the required isolation voltage, which can be up to 10kV in some markets.
These considerations imply that a large variety of pulse transformers will be needed. This variety cannot typically be provided by a single vendor. In addition, meter manufacturers may want second sources for pulse transformers.
The question then arises how to design and test a pulse transformer that is compatible with the 71M654x family of metering ICs. This application note seeks to answer that question.
Below, the basic specifications for a pulse transformer are introduced. Note that building a transformer to these exact specifications alone does not guarantee proper functioning. Tests must follow that involve the 71M654x metering IC and a 71M6xxx remote sensor. Test equipment and test procedures that allow the designer to evaluate the transformer to ensure proper functioning are described.

Basic Specifications for Pulse Transformers

The basic specifications are listed below:
  • Winding ratio: 1.08:1 to 1.1:1 primary/secondary, no taps
  • DC resistance: < 1.2Ω
  • Current capacity: > 20mA
  • Differential operating voltage: 3.3V
  • Operating temperature range: -40°C to +85°C (can be narrower for indoor applications)
  • Inductance: > 60µH at 10kHz (primary)
  • Operating frequency range: > 16MHz
  • Isolation: 7kV, or higher, depending on the application, for one minute
  • Winding-to-winding capacitance (CWW) should be minimized to limit RF emissions. Emissions typically couple through the CWW into the shunt sensor and from there, into the grid wiring. Realistic values for CWW are 5pF to 10pF, which can be achieved by separating the primary and secondary windings physically (with no wire overlap).
  • Saturation flux density: > 700mT (higher values are needed for challenging requirements)
  • Packaging: SMT plastic package, or a flat header with:
    • 4 pins, as narrow as possible for the given core size
    • Primary to secondary pins > 15mm apart (for clearance and creepage)
    • Wires and core in non-conductive molding material

The Remote Sensor Interface

The Driver Circuit

Figure 1 shows the driver circuit in a 71M654x host with a transformer and a 71M6xxx remote sensor IC. The differential driver basically consists of two switches with four positions each. Selecting position A will connect the transformer to V3P3SYS (3.3V, typical) and GND. This setting produces a power pulse that is rectified by the power supply in the remote sensor IC to generate its DC supply. Other switch settings enable the driver to generate softer data pulses (C), remove the flyback voltage from the transformer (B), or float so that the driver can receive the data pulse from the remote sensor IC (D).
Figure 1. Driver circuit with transformer and remote sensor IC.
Figure 1. Driver circuit with transformer and remote sensor IC.

Curve Shapes and Timing

Figure 2 shows curve shapes and timing. The power pulse is shown higher in amplitude, due to lower driver impedance, in comparison to the data pulses. Both pulse width and amplitude are decoded by the remote sensor IC. Data pulses must have a certain width to be recognized. This implies that the transformer must truthfully reproduce the timing of the pulses, which is only possible if the spectral content of the pulses is maintained (estimated up to the 10th harmonics).
Figure 2 shows the idealized signal shapes. An example for real signal shapes obtained with a differential probe is shown in Figure 3 (green traces).
Figure 2. Signal shape between the 71M654x host and the 71M6xxx remote sensor IC.
Figure 2. Signal shape between the 71M654x host and the 71M6xxx remote sensor IC.
Figure 3. Scope picture of the signal shape between the 71M654x host and the 71M6xxx remote sensor IC.
Figure 3. Scope picture of the signal shape between the 71M654x host and the 71M6xxx remote sensor IC.

Testing the Transformer

Test Tools and Methods

A variety of test tools are required to test and verify the transformer functionality (see Table 1). All tests should involve a demo board (71M6541F-DB, 71M6543F-DB/71M6543F-DB-CT), for which one or two of the three built-in standard transformers (Wurth Electronics Midcom’s 750110056) can be exchanged for the custom transformer prototype (DUT). Leaving one 750110056 on the board allows for convenient comparison testing.
The demo board comes with demo code in the flash memory of the 71M654x IC. The demo code provides a command line interface (CLI) that can be accessed by connecting the board to a PC equipped with HyperTerminal® or any other serial communication program. For details on connections, operation, and the command set, see the relevant user manual for the particular demo board.1 The demo code provides commands that initiate communication with the 71M6xxx remote sensor IC as well as some commands that exercise the communication link in a loop.
An oscilloscope with a differential probe provides a good tool for visual comparison of the signal shapes. A simple DMM can be used to verify the local supply voltage of the 71M6xxx remote sensor IC.
A hipot tester should be used to check the isolation properties of the transformer.
Finally, a permanent magnet can be used to test the magnetic immunity of the transformer.

Table 1. Test Tools
Test Tool Test Coverage
DMM Verify supply voltage at 71M6xxx remote sensors (pin 1 to pin 4)
Oscilloscope with differential probe Verify curve shape, eye test
71M654x demo board with 71M6xxx remote sensors Test function of interface
Hipot tester Isolation
Magnet and flux-meter Resistance to magnetic fields

Setting Up the Test with the 71M6543F-DB Demo Board

Let’s take a look at how to test with the 71M6543F-DB demo board.
The 71M6543F-DB demo board contains three Wurth Electronics Midcom 750100056 transformers, one for each phase. To monitor the signal shapes with the standard transformers, a differential oscilloscope probe can be applied to the pins IAP/IAN, IBP/IBN, or ICP/ICN. To synchronize an oscilloscope to the signal, one can use the following CLI commands to generate a synchronization pulse on the TMUX pin (TP21) of the 71M6543F:
  • RI2502=20 (for the 71M6x0x remote sensor IC on phase A)
  • RI2502=21 (for the 71M6x0x remote sensor IC on phase B)
  • RI2502=22 (for the 71M6x0x remote sensor IC on phase C)
A simple test of the communication link can be done by issuing the commands 6R1.20 (phase A), 6R2.20 (phase B), or 6R3.20 (phase C). These commands return the reading from the temperature sensor (STEMP) of the 71M6xxx remote sensor interface in a two-byte hexadecimal format (e.g., 0xFFDF). Negative readings are signaled by the MSB being 1. The temperature reported by the remote sensor IC is:
T = 22°C + (STEMP × 0.33 - (STEMP2) × 0.00003)°C
For example, if STEMP = 0xFFDF, the decimal equivalent is -32. The temperature calculates to 22°C – 10.59°C = 11.4°C.
If the temperature is displayed reliably and consistently with the 6Rn.20 command, i.e., without parity errors, this can be interpreted to mean that the communication link works in both directions. Temperature variations of several degrees are normal.

Eye Diagrams

The timing of the signals on the remote sensor interface is extremely important for proper functioning. The data pulse from the remote sensor IC is sampled by the 71M654x 355ns after the rise of the power pulse (see Figure 4). Immediately before this has concluded, the data pulse is integrated over a period of 35ns. If the signal is clearly defined as positive or negative during the integration time, i.e., if it does not cross zero during this time, it can be easily detected by the 71M654x. If the data pulse appears too early or too late, or changes its timing (jitter), detection is difficult and will result in data errors.
Figure 4. Ideal timing for the data pulse.
Figure 4. Ideal timing for the data pulse.
A simple scope picture similar to Figure 4 will not tell the whole story. It is best to generate an eye diagram with an oscilloscope capable of infinite persistence and watch the signal for a long time. If the data pulse changes its timing over time or temperature, this will appear on the scope as a closing eye.
The eye diagram test shows jitter over time and temperature and provides a good test criteria for the transformer.
The recommended test procedure is as follows:
  1. Insert the DUT (transformer prototype) into the position of the original standard transformer on the demo board.
  2. Power up the demo board with the 5VDC supply that was supplied with the demo kit.
  3. Establish a synchronizing pulse on TP21 (TMUXOUT) by issuing one of the RI2502 commands described earlier.
  4. Use an oscilloscope with a differential probe to monitor the signals on the 71M654x side of the transformer. Set the oscilloscope to infinite persistence mode and select a time resolution of 20ns to 25ns/division.
  5. Set the trigger settings of the scope to the synchronizing pulse.
  6. Watch the eye developing underneath the synchronization pulse, which is formed by positive and negative data pulses. Over time and over temperature, the following should apply:
    • The eye should never close. This condition would signal too much jitter or a change of the timing with temperature.
    • The eye should be 40ns wide or wider.
    • The trigger point where the data pulse is sampled is at 355ns after the rise of the power pulse. The polarity of the data pulse should be clearly defined during the 35ns preceding the 355ns trigger point.
A slower flyback from the transformer can shift the response from the remote sensor IC out in time, which may cause the data pulse to appear too late. In that case, it can be helpful to set the RFLY_DIS bit in the 71M654x. This register is bit 3 of I/O RAM 0x210C. With RFLY_DIS = 1, the transformer is not floated during the fly-back phase, but actively driven by the 71M654x, resulting in shorter fall times.
Figure 5 shows the scope capture of positive and negative data pulses on a 71M6543F-DB demo board. The vertical dotted marker by the red arrow is adjusted to be 355ns after the rise of the power pulse. The time slice marked by the two yellow arrows is 35ns wide and marks the integration period for the data pulse. The eye of the data pulse is 47ns wide. Note that the data pulse is significantly lower in amplitude than the power pulse and is also clearly defined as either positive or negative during the integration time.
Figure 5. Timing of a data pulse obtained with an oscilloscope.
Figure 5. Timing of a data pulse obtained with an oscilloscope.

Loop Read Test

Now back to our example of the 71M6543F-DB demo board.
The standard demo code for the 71M6543F-DB demo board supports read commands that are executed in a loop while statistics on data and parity errors are kept and displayed. An example for the command 6Ta.b is shown below:
resp:0xD65F ,6100 ,d:0 ,p:0 ,t:0
In this example, an internal fuse of the remote sensor IC in phase A was read 6100 times and reported no data (d) or parity (p) errors. The fuse value was 0xD65F.
resp:0x0000 ,29300 ,d:59 ,p:59 ,t:0
resp:0x0000 ,29400 ,d:159 ,p:158 ,t:0
resp:0x0000 ,29500 ,d:259 ,p:258 ,t:0
resp:0x0000 ,29600 ,d:359 ,p:358 ,t:0
resp:0xFFFF ,29700 ,d:459 ,p:458 ,t:0
resp:0xFFFF ,29800 ,d:559 ,p:558 ,t:0
resp:0xD65F ,29900 ,d:595 ,p:594 ,t:0
resp:0xD65F ,69300 ,d:595 ,p:594 ,t:0
In the example above, in the first 6 lines, the fuse was not read properly, returning values of 0x0000 and 0xFFFF.
resp:0x00B5 ,16400 ,d:0 ,p:0 ,t:0
In this example, the band-gap trim fuse was read 16,400 times. Zero data (d) and zero parity (p) errors were encountered. The remote sensor IC used in this example was the 71M6103. Other remote sensor ICs, such as the 71M6601 and the 71M6603 will return 0x0000, which is not a good test case. Tests should be performed with the 71M6103.

Test Sequence

The recommended test sequence is as follows:
  1. Insert the DUT (transformer prototype) into the position of the original standard transformer on the demo board.
  2. Power up the demo board with the 5VDC supply that was supplied with the demo kit.
  3. Measure the supply voltage that develops between pins 1 and 4 of the 71M6xxx remote sensor IC. It should be above 2.7VDC, preferably above 3.0VDC.
  4. Establish a synchronizing pulse on TP21 by issuing one of the RI2502 commands described earlier.
  5. Use an oscilloscope with a differential probe to monitor the signals on the 71M654x side of the transformer. The signals should resemble those shown in Figure 3, with the slopes steep and the power pulses higher in amplitude as compared to the data pulses.
  6. Issue the M15 command for the CLI, causing current display on the LCD. With no current applied to the shunt sensors, the display should be close to zero on all channels.
  7. Apply various currents between 1A and 100A to the channel equipped with the DUT and observe the display for that channel. The display must accurately reflect the injected current (within ±0.5%).
  8. Obtain the eye diagram with a storage oscilloscope as described earlier. Make sure the eye diagram of the first data pulse is acceptable over operating temperature and board supply voltages.
  9. Perform a loop read test over several minutes.
  10. Repeat the loop read test at various temperatures, depending on the temperature range of the meter, and at various board supply voltages.


  1. Application note 5469, “User Manual for 71M6543F-DB and 71M6543-DB-CT.”

APPLICATION NOTE 5680,AN5680, AN 5680, APP5680, Appnote5680, Appnote 5680