Hello DigiKey,
I have chosen ScioSense ENS130 as the sensor of my next project. I have designed the first draft of my circuit. However, the major components are made by different factories.
So, may I submit a Design Check about my circuit built to use the ENS130 sensor to DigiKey?
Yours faithfully,
Timmy
The best we can offer is to post everything you got here, so we can take a look.
Hi HKPhysicist,
The only difference I see between your circuit and that provided in the datasheet for the ENS130 is that yours is missing the buffer op amp provided by the second op-amp in the TP5552 package.
Example circuit from page 10 of the ENS130 datasheet:
Since the TP5552 has very low input bias current (in the low hundreds of pico-amps) one could argue that the buffer circuit for Vref is unnecessary, but since they use the second op amp as a buffer in their reference design, they probably thought it would, at a minimum, improve performance, and it may be necessary for even “good” performance.
Regarding supply voltage for the TP5552, since it is a rail-to-rail input and output device, the inputs can be right at the supply rail values, so one can use 3.3V for the positive supply rail and 0V (ground) for the negative supply rail and the circuit should behave as expected.
Note that in order to get accurate results, you will need to do calibration and compensate the output for temperature variation, as described in the ENS130 datasheet.
Answer: My mistake. I misread your schematic previously. 
You have it right. Amp “A” in your schematic IS the buffer and you have it connected correctly.
HI @HKPhysicist ,
For a better stability, I’d connect R5 to a precision reference instead of +3.3V. Also a good practice is to shunt the R3 with a, say 1uF capacitor, but here the inverting amp’s low-pass cut-off frequency may be so low, that capacitor should be much larger to have any practical effect.
In addition, place a 1uF…10uF capacitor between the +3.3V supply and GND as close to the op-amp as possible.
Do you take that +3.3V from a lab supply or regulate locally on-board?
Cheers, heke
Hello heke,
I get the +3.3V from a MCU standard +3.3V output. The MCU circuit is powered by a USB 5V; and also a 3.7V Li+ battery as the backup whenever the USB 5V is down caused by AC power off.
Thanks for the clarification.
I am going to layout PCB for prototyping/production if there is no more improvement suggestion. 
Hi HKPhysicist,
No problem. By the way @heke makes some very good points.
The externals to your circuit are implied, but make sure you place a bypass cap on the supply pins for your op amp, just as you should for your microcontroller and every other IC. In the datasheet for the TP5552, they use 0.1uF, which is pretty much the universal standard bypass cap value for IC’s. Using that in parallel with a higher value, with the 0.1uF part placed nearest the op amp pins is common practice for high precision analog circuits. Make sure you use an X7R type of ceramic for the 0.1uF cap. Do not use Y5V type ceramic. It has terrible performance over temperature variations.
Regarding your power supply, I’m not clear from your description as to the source of your power. Is it 3.3V coming from a voltage regulator or an output pin of a microcontroller? If the latter, I would have significant concerns. It really should come from a regulator.
Assuming that your supply is coming from a voltage regulator, then the next question would be, does your ADC also use the 3.3V supply as it’s reference voltage, or does it use a different voltage reference (in many MCU’s and external ADCs, one can select whether one uses the supply voltage or some other reference voltage for it’s ADC’s reference voltage)?
If it uses the 3.3V supply as it’s reference, then I would keep the 3.3V supply as the input to the simple voltage divider you have (R5 and R3) as the reference for the buffer circuit, as any fluctuations in supply voltage will cause both the buffer reference and the ADC reference to track together, maintaining relative accuracy.
However, using a precision reference rather than the 3.3V supply for your reference voltage for both the buffer circuit and the reference for the ADC will give the best accuracy. The key is that if you use the 3.3V supply as the reference on the buffer circuit, then use it for both.
Hello David,
I am facing a new problem. I choose a retail ESP32 board to partner this ENS130 circuit.
Then, I find from the official ESP32 document that its ADC default reference voltage is 1.1V as shown.
Then, I add a voltage regulator to output a 1.1V voltage to the ENS130 reference voltage as shown:
Now, my question is:
What should be the Vs supply voltage to the precision op-amp TP5552 now? Can I still use the 3.3V regulated voltage supply from my ESP32 board?
Yours faithfully.
Hi HKPhysicist,
Yes, you can still use 3.3V for the TP5552 +Vs supply. It’s generally preferable to have extra headroom in an op amp’s supply relative to the expected output range, so that’s actually a good thing.
The one issue there is that if that supply were to fluctuate during operation, typically caused by fluctuations in current draw (i.e. load variations) from various components on the board, then this can cause fluctuations in the supply voltage of the op amp, which in turn can cause output variations from the op amp.
That can be reduced by capacitance placed between +Vs and -Vs (-Vs being tied to the ground plane in this case), so following @heke’s advice is important. An optimal arrangement is to place a 0.1uF cap between +Vs and the ground plane as close as possible to +Vs and the ground plane to minimize parasitic inductance, and then putting a 10uF cap next to, and in parallel with the 0.1uF cap. This helps keep that supply voltage more stable. The reason for using both caps is that the smaller one placed close is more effective at shunting high frequency noise to ground while the larger value holds more charge and does the primary work of stabilizing the voltage supply.
I don’t think the separate 1.1V regulator is necessary, and if you did use it, you would probably need to change the resistor values. The purpose of the reference voltage on op amp “A” is to set a minimum output voltage when the sensor measures zero CO. Their circuit sets this voltage to 0.148V. The reason they do this rather than leave it at zero volts is because the op amp is incapable of outputting 0V when -Vs is tied to ground (0V). According to the datasheet for the TP5552, the lowest guaranteed output is 30mV, and it will perform more accurately when staying well above this low voltage.
You could actually leave that circuit as you had it with the 3.3V input to the divider circuit if you don’t expect to read concentration levels somewhere in excess of at least 600ppm. Using the example values given in the datasheet on pages 11 and 12 for “I0” and “S”, I calculate that you can read up to 773ppm using 0.148V as a reference voltage without exceeding your 1.1V ADC reference limit.
Nothing obvious that I can see.
Make sure you carefully read through the calibration info in the datasheet for the ENS130, as you won’t get good results if you don’t follow their recommendations.
Good luck to you!
