Echoes and Challenges of Technological Limitations
How did engineers in the past manage to measure electrical power without modern microchips and DSPs? This article explores the Energomera CE6806P, a device created in 2006 for verifying electricity meters, yet built using 1980s-era technology.
We’ll take a closer look at its design, principles of operation, and how discrete-analog solutions were used to achieve high accuracy. The Energomera is a fascinating example of engineering and ingenuity, giving us a unique perspective on the evolution of electrical measurement devices.
Modern Energy Metering Technologies
Today, measuring electrical energy is a straightforward task, thanks to specialized metering chips. This is the same approach we use in our WB-MAP devices, which rely on Microchip’s ATM90E32AS and ATM90E36A. There are also many other manufacturers of metering ICs, including Western, Chinese, Russian, and even South African companies.
Nearly all of these chips operate in a similar way. Voltage signals (from a calibrated voltage divider) and current signals (from a current transformer or shunt) are fed into 24-bit sigma-delta ADCs with a very high dynamic range. The sampled values are then processed by a dedicated digital signal processor (DSP).
Typically, the DSP:
Downsamples voltage and current readings to 2 kHz
Multiplies these values to obtain instantaneous power every 0.5 ms
Integrates the results to determine active power consumption
Additionally, the DSP calculates other useful parameters:
Network frequency
RMS voltage and current
Reactive power
Phase angles
But these microchips have only been in widespread use for about 30 years. So how was this done before?
Let’s Dive into History
In the earliest DC power networks, engineers even used electrochemical meters that worked by transferring mercury between two electrodes through electrolysis. Incredibly, similar devices were still used in Soviet measuring instruments to track operational time until the 1980s.
https://creativecommons.org/licenses/by-sa/4.0, via Wikimedia Commons (by Geni, CC BY-SA 4.0)
Here’s how they worked:
The voltage coil induces eddy currents in the disk, interacting with the field of the current coil to create a rotational force.
The current coil does the same, producing its own rotational force.
These forces sum together, resulting in a rotation proportional to instantaneous power.
A magnetic brake slows the disk, balancing the force.
Rotation is transferred via a worm gear to a mechanical drum counter, which accumulates and displays total energy consumption.
https://creativecommons.org/licenses/by-sa/3.0, via Wikimedia Commons (by Wefo at de.wikipedia, CC BY-SA 3.0 )
Amazingly, despite being over a century old, this mechanism performs the exact same power calculations as modern metering chips. The disk’s speed is directly proportional to active power consumption, even accounting for phase angles and waveform distortions. The drum counter is simply a mechanical integrator that accumulates energy over time.
In three-phase meters, this system is duplicated two or three times, with additional aluminum disks. The rotational moments are summed, just like in modern three-phase power calculations.
And just look at the insane number of washers, wires, and tiny screws used for calibration — all adjusted by hand! Can you really compare this to an electronic meter, which an automated system calibrates in minutes?
Transitioning to Electronic Power Calculations
By the 1970s-1990s, engineers attempted to perform power calculations electronically, but high-resolution ADCs and DSPs were still unavailable. Instead, they used hybrid discrete-analog methods, combining:
CMOS 4000-series logic
Analog MOS switches
Operational amplifiers
Recently, we got our hands on an interesting relic. Although manufactured in 2006, its technology harkens back to the 1980s. Unlike standard energy meters, the Energomera CE6806P wasn’t designed for metering but for verifying and calibrating electricity meters. As such, it was meant to have higher accuracy—though, after analyzing its construction, we’re not so sure about that! 😊
Energomera
Energomera is one of the largest post-Soviet manufacturers of electricity meters, producing both household and industrial solutions. While best known for its mass-produced utility meters, the CE6806P belongs to a different category — a specialized calibration tool rooted in 1980s engineering.
Wiren Board specializes in industrial, facility, and home automation, designing modular, Linux-powered controllers for diverse monitoring and control applications. With production facilities in Astana and Yerevan, we serve both home appliance solutions and industrial automation needs. For us the Energomera CE6806P reflects a key transition in metering technology — bridging analog precision with early digital logic — an evolution that aligns with our commitment to advancing automation solutions.
Energomera, the company behind this device, is one of the largest post-soviet manufacturers of electricity meters, producing both household and industrial energy metering solutions. Since the 1990s, it has developed a wide range of devices, from mechanical induction meters to modern digital smart meters with automated data collection. While the company is best known for its mass-produced utility meters, the CE6806P represents a different category—a specialized calibration tool designed to test and verify other meters. And unlike the company’s mainstream products, this device carries the distinct flavor of 1980s engineering.
As an automation hardware designer and manufacturer with production facilities in Astana and Yerevan, we at Wiren Board feel both a historical connection to this device and a deep engineering curiosity about its design. The CE6806P embodies a transitional era in metering technology, blending analog precision with early digital logic — a fascinating approach that resonates with our own interest in the evolution of industrial automation. Let’s take a closer look at how it works.
Design of the Energomera CE6806P
Housed in a rugged plastic case (similar to a Peli Case), the front panel features:
Terminals for voltage and current channel connections
Clamp meter ports for external current transformers
A specialized sensor for detecting the rotation of a meter’s disk
Key components include:
Microcontroller board (top-right)
Current channel amplifiers (top-left)
Rotary switches for voltage and current channel settings
LCD display module (center)
Power transformer and protection components (bottom section)
Here we see three high-quality voltage transformers—no expense was spared on materials. This is necessary to ensure minimal voltage loss due to winding resistance and, consequently, high accuracy. We can also spot three current transformers hidden under the terminal blocks, along with the calculator board itself.
Computing Board: A Mix of Technologies from the Perestroika Era
The Perestroika era was a time of transition, where the Soviet electronics industry, once isolated and reliant on domestic components, began integrating imported technologies amid economic and supply chain disruptions. This led to a peculiar mix of old Soviet engineering traditions with newly accessible modern components, creating devices that combined legacy circuit design with whatever parts were available. The result was often an eclectic blend of cutting-edge ideas and outdated manufacturing constraints, as engineers adapted to a rapidly changing technological landscape. The computing board in this device is a true Perestroika-era mix of components. It includes both branded modern microchips from TI, Philips, and Onsemi, as well as the good old Soviet K561LN2 logic chips.
The core of the board is 4000-series CMOS logic, complemented by:
OP07 operational amplifiers — old but fairly precise,
LM211 comparators,
Soviet JFET operational amplifiers K544UD1,
Hybrid SES4 microcircuits — laser-trimmed resistor arrays forming a 6-bit R-2R ladder network.
Hybrid microchips with the unusual name SES4 are laser-trimmed resistor arrays in the form of a 6-bit R-2R matrix. In the Russian internet, you can find photos of these chips with their casings removed.
Typically, R-2R matrices are associated with DACs, but here they serve as precise fixed dividers in reference voltage circuits. Unfortunately, the capacitors used in this design are standard Soviet polyester K73-17, which have never been known for outstanding quality. Surprisingly, they are even used in precision circuits. For calibration, the device includes a large number of multi-turn trimmer resistors.
Processor Board
At the heart of the board is an AT89C52 microcontroller, a classic Atmel chip from the enhanced Intel 8051 family — the dominant architecture before the rise of AVR (Arduino, etc.).
One can only wonder how much life is left in the electrolytic capacitors from 2006.
Display Board
The designers resisted the trend of the time to use character LCD displays with a Hitachi controller and parallel interface. Instead, they went their own way, opting for a passive 7-segment LCD display.
Typically, such displays require a dedicated driver chip with exclusive OR gates on the outputs, designed to generate AC signals for LCD operation. But here, instead of a standard solution, they built a custom driver using six cascaded MC14094 chips — an 8-bit shift register, essentially an older cousin of the now-popular 74HC595.
A critical point: LCDs must be driven by AC signals with balanced half-cycles. If driven by DC or unbalanced AC, electrolysis occurs, rapidly destroying the display.
Dedicated LCD controllers handle this automatically, but in this custom circuit, the microcontroller must frequently update the display, ensuring proper waveform balancing.
We checked with an oscilloscope — and yes, that’s exactly how it works. Let’s just hope the firmware properly maintains waveform symmetry.
Current Channel Amplifier
This amplifier sits between the current transformers and the computing board.
An interesting discovery: sloppily soldered K73-17 capacitors — were they factory-installed, or is this a repair job? These capacitors correct parasitic phase shifts in the current channels.
And once again — clusters of multi-turn trimmer resistors (SP5-2) for calibration across different current ranges.
Some of these trimmers are classic aluminum-bodied ones with a bakelite base, while others are cheaper plastic versions from the Perestroika era.
The original Soviet SP5-2 trimmers were quite decent, but modernized versions look disappointingly cheap.
Why use these at all? By 2006, Bourns trimmers were readily available.
How It Works
Now, let’s break down how the power computing unit operates.
By 2006, analog multipliers existed, but their accuracy and dynamic range were far from ideal. Moreover, they rely on PN-junction properties, making them highly temperature-dependent. Compensating for these effects is complex, expensive, and unstable, making it difficult to achieve the ±0.1% precision required.
The Solution: A Discrete-Analog Computing Method
Instead of purely analog multiplication, this device employs a hybrid discrete-analog approach. Below is a highly simplified block diagram—many secondary components are omitted!
Core Idea:
Voltage signals are converted into Pulse-Width Modulated (PWM) digital signals.
Using analog MOS switches, they build bridge circuits, where:
a. One diagonal receives the current signal.
b. The other diagonal outputs a current proportional to active power.
To generate PWM signals, a triangle wave generator runs at ~550 Hz — not synchronized with the power grid (which is actually preferable).
The triangle waveform is produced using a standard integrator + comparator circuit, with amplitude stability ensured by a precision Zener diode and symmetry maintained by the SES4 resistor array.
The comparator output controls the MOS switch bridge, modulating the current signal. The average output current is proportional to the instantaneous power — just like in modern DSP-based meters.
This method is cheaper and more precise than using analog multipliers.
Digitization and Frequency Conversion
Since we need high accuracy and a wide dynamic range, a simple ADC wouldn’t be enough.
Instead, the designers used a current-to-frequency converter:
The integrator accumulates charge from the current signal.
When the voltage exceeds a threshold, the comparator generates an output pulse.
The charge resets, and the process repeats.
The output pulse frequency is proportional to input current, achieving high accuracy.
Finally, the microcontroller simply counts pulses over a fixed time interval — producing a value proportional to total power across all three phases.
Are There Any Issues?
Almost perfect — but not quite. This design introduces a fundamental signal processing flaw:
It samples voltage and current without filtering out high-frequency components.
This violates the Nyquist-Kotelnikov theorem, leading to aliasing effects.
A proper low-pass filter would solve this—but at 550 Hz modulation, designing one is impractical.
A sharp roll-off filter would be complex, expensive, and would distort phase angles, which is critical for accurate metering.
Instead, the designers relied on statistical averaging — assuming aliasing errors cancel out over time.
This is fine for a calibration device in a controlled lab environment — but would be problematic in real-world usage.
Oscilloscope Readings
The waveform traces below were captured directly from a real meter, at points marked with colored circles in the circuit diagram:
Blue: Voltage signal.
Yellow: Triangle wave modulator signal.
Purple: Comparator output (drives the multiplier bridge).
Green: Output of the voltage-to-frequency converter—the pulses counted by the microcontroller.
The current signal at the bridge output likely has a complex waveform, but measuring it without cutting PCB traces would be tricky — so we left it untouched.
Reactive Power Measurement
To measure reactive power, the current signal must be multiplied by a 90-degree phase-shifted voltage.
Modern DSP-based meters handle this digitally, using FIFO buffers or precomputed lookup tables.
But how does Energomera CE6806P do it?
It employs an old trick from electromechanical meters:
Instead of shifting voltage digitally, it uses the difference between two phase voltages.
A-phase current is multiplied by (B-phase voltage – C-phase voltage).
This naturally produces a 90-degree phase shift, allowing reactive power computation.
Limitations of This Approach
Phase imbalance completely ruins accuracy.
Harmonics distort the calculation.
Even IEC standards struggle to define reactive power under real-world conditions with unbalanced loads.
The Giant Rotary Switch
Reactive power was calculated by multiplying the phase current with the voltage difference between opposite phases, naturally creating a 90-degree phase shift. Switching to reactive power mode required manually rotating a large rotary switch, which physically reconfigured transformer windings and signaled the microcontroller to adjust the scaling factor accordingly.
Conclusion
We initially hoped to use the Energomera CE6806P as a reference instrument for using it on a testing setup-stand for developing devices. But, as usual, expectations were not met.
Main Concerns:
Too many trimmer resistors of questionable quality — one was faulty and had to be replaced.
Sloppy soldering in some areas.
Ultimately, this quirky relic of the past is heading to our museum of interesting engineering solutions — a fascinating look into pre-DSP era metering techniques.