The Celtic Engineer is a weekly newsletter produced by Celtic Engineering Solutions. We hope you enjoy it. If you have any suggestions for topics, would like to give feedback or want your email added to the distribution list please send an email to
Extremes and extremists
If you own a TV these days it seems we have lost the centrists; everyone appears to be an extremist. While that may get you to where you want to be in politics, in electronics, that is not the case. Things are just better in the middle. If you are working on low-power hand-held products, you probably know what I mean. Voltages are getting lower and lower. I cut my teeth on electronics when the voltage rails were +/-15VDC. Talk about extremes, that feels like a lot of power, and it was. Transistors and OP Amps run cooler and transition faster at 3.3V than they do at 30V. The problem though is that we don’t have a ground at 1.65V. And OP Amps are a reverenced device.
You could just define a 1.65 voltage reference, but our signals are referenced to ground. With today’s rail to rail OP Amps it is a temptation to power them at 0 and 3.3V. It will work, sort of, most of the time. The problem is that we expect an op amp to be a linear device, Vout = -(R2/R1) * Vin, or Vout = (1 + R2/R1) * Vin, that sort of thing. Those are linear equations. When you push them to the extreme, that is against the rails, they are not a linear as we might like. They do a much better job of acting the way we expect if we operate them away from the rails. If you design a circuit that operates away from the rail, great, but if zero signal means zero volts things get uncomfortable.
By decree I declare ground to be
So, one way to deal with this problem is do declare a reference point. In this case 1.65 is dead center. During the fall of the voltage rail, there was a time when this is exactly what you would do. You will find some of the older A/D converters expected you to have a zero above your ground.
Even today, if you have a voltage that can take on negative voltages, you will have to define a ground above the bottom of the voltage range as zero. What I mean is that if you can have a value that is +/- 1V, then zero must be at least 1V to get into your ADC, again 1.65V would probably be better.
It would be so much easier if there was a +3.3V and a -3.3V supply in the device. Especially for sensor boards.
You could make one of course, but that sounds like hard work. They PM’s will say its risky and upper
management will see all those parts and say it sounds expensive.
There must be a better way
There are now DC/DC converters that do not require external parts. Case in point is the IA0505S by XP Power.
This is a 1W device, about 100 mA per output. The input is between 4.5 and 5.5VDC and it gives you +/-
5VDC out. You could operate off of the +/-5, and make sure you never go above 3.3 when you hit the ADC,
but that is asking for problems. To make things simple, you then put two linear voltage regulators and you have
+/- 3.3VDC, see figure 1. This assumes your ADC is a 0-3.3V input. Some are 0-1.8V input, in which case
you might want to have a positive voltage regulator of 2.0V to limit your ability to kill the ADC.
Figure 1 Getting +/- 3.3V the easy way, see featured image.
This solves so many problems, I just don’t know where to start. Rarely do you need 100mA for sensors so the
power is not an issue. You might choose to connect a second positive regulator to Vraw and produce a 3.3V
power supply for the MCU or other noisy digital things. Having these supplies around 0V is convenient
because ground is now in the center of the OP Amps range and so there is no straining to make it linear. You
can user either an inverting or non-inverting OP Amp without problem.
You might ask why use linear regulators to get down to the 3.3V and not just a device that does that directly. In
many cases you might choose a converter that does that, but you will be subject to the converter noise. Plus,
you have 10V you can use for other parts of the circuit. Sometimes you need 5V for a device, such as displays.
Or you might use the 10V for lighting, LED’s. Either way there is a price to pay, TANSTAAFL – Heinlein, the
efficiency of this device for +/-5 is 74%. That is another reason to have a second 3.3V regulator for items like
the MCU.
Happy New Year! I hope this one is a great one for all of you.
Next week is the Consumer Electronics Show, will you be there? I wonder what new toys will be announced.
Final thoughts
This newsletter is sponsored by Celtic Engineering Solutions LLC, a design engineering firm based out of West
Jordan, Utah, which can be found on the web at: You can find the
newsletter on the company blog, LinkedIn or in your inbox by subscribing. Send your emails to The Celtic
Engineer at:, with the subject line SUBSCRIBE.