A Tale of Phase Rotation
These days, we hear a lot about revival of retro technologies like the cassette and the film camera. Personally, as someone who is obsessed with analog power supplies, I am hoping that the retro trend will see the return of the “dropper-type” power supply.
This story involves something that happened to me very early in my career. At many offices, the end of the year means it’s time to have a big cleanout. However, when it gets time to do the cleanout at my work, I find myself reminded of a lesson I learned a long time ago about phase rotation.
Boosting the Efficiency of Series Dropper Power Supplies
My story takes place over 25 years ago. Back then, switching power supplies had yet to catch on and the market for DC power supplies for testing purposes was dominated by “series dropper” power supplies. One of my design assignments at the time was improving power supply efficiency at below rated output voltage. Transformers are efficient if you use them at their full output voltage, but when you attempt to drop that voltage, they become less efficient due to the large voltage difference that is applied to the control transistor and associated losses.
A phase-controlled circuit is not affected by fluctuations in output or input voltage, and is able to minimize losses by minimizing the voltage difference across the control transistor. However, because of the way these circuits lower output voltage by reducing their conduction angle, at lower voltages they tend to have poor power factors or emit high-frequency waves. In applications where there a significant power loss would arise across the control transistor, it is therefore better to use the transformer’s full output voltage without dropping it.
A Mode-Switching Three-Phase Full-Bridge Rectifier
In the old days, engineers would progressively improve a power supply’s power factor by switching taps so that the full output voltage of the transformer could be used, or progressively switch voltage by changing the method of rectification in accordance with the transformer’s output.
This fact inspired me to construct a phase-controlling circuit using a mode-switching, three phase, full-bridge rectifier. At low voltages, the circuit would rectify input power using phase voltage and the three-phase bridge rectifier, then, at higher voltages, the three-phase bridge rectifier would become fully conducting and the downstream thyristor (SCR) would turn on, transforming the circuit into a full-bridge rectifier that controlled line voltage. The design progressed well, and as shown in the graph of power consumption in Figure 1, I observed an improvement in efficiency at mid-range output voltages.
The Mystery of the Doubling Voltage
My approach was starting to bear fruit, and I was upbeat as work shut down for the year. During our annual end-of year cleanout, I unplugged my prototype so that I could clean my work area and my desk. With one last office party, the year was over.
After arriving back at work after the break and seeing off our first product shipment for the New Year, I wired up my prototype again and recommenced my experiment. My prototype, however, despite having performed well at the end of the previous year, was now showing double the output voltage. I had no idea why this was happening. My New Year’s bubble was burst already!
In fact, the problem was the result of my having disconnected the device during the big cleanout. The connections between the terminals on my prototype and the input phases (U, V, W) had remained unchanged as I had not reconfigured the phases─up until the time of the cleanup, that is.
The issue was “phase rotation”, a phenomenon whereby when the U phase voltage signal is used to drive the V phase input, a zero U voltage will drive a near full-voltage V phase on account of the V phase being 120° ahead of the U phase. In a power supply with three-phase input, when the input wiring is out of phase with the SCR (thyristor) that performs phase control and the gate pulse is driving the SCR, phase control becomes nonlinear, meaning that as soon as the power is switched on, the circuit’s output increases to full power, and when you attempt to obtain rated power, phase control works against you, resulting in an insufficient power output.
When building my prototype, I did not even consider the issue of phase rotation. My experience taught me the importance of carefully designing the gate pulse, which triggers the circuit to switch from phase voltage to line voltage. To overcome the problem, I needed to add a circuit that would detect phase voltage and turn off the output if it detected a phase error were incorrect. I was just grateful that the malfunction did not happen on the client site.
A Clean Out is Also a Chance to Reset
You may have had a similar experience when you had a cleanout. Sometimes, when you become too engrossed in a project, you can forget time, holed up in your “engineer’s cave”. A cleanout is a good opportunity to reset. You feel better having a tidy workspace, and the cleanout will sometimes helps you notice things that you had overlooked because you were so engrossed in your work. At the risk of laboring the point, I believe we should all take cleanouts seriously and think of them as a chance to “reset” our work as well.
