Our sous vide is accurate to 0.2 °C and yours can be too!
‘Of the many questions regarding a sous vide setup, one is always accuracy. There are two ways a measurement can deviate from the actual value of what is being measured: systematic error and random error. We will find that the first can be corrected quite easily, and the second is small enough not to worry about, on the order of 0.2 °C.
UPDATE: Aspects of PID tuning here may not be correct. See our revised article for better PID values.
TheType K thermocouple we suggested has an official accuracy of 1.5 to 2.5 °C. Much of this is due to manufacturing error, and should lead to systematic error (rather than random error). We need an absolutely calibrated thermometer to determine this error.
After setting the SV (set value) to 60 °C and waiting for the bath to reach an equilibrium, we measured the temperature every minute for twenty minutes (Fig 1). The measurements ranged from 61.0 to 61.2 °C. With this data, we can expect the sous vide to reach temperatures about 1.1 °C above the SV.
We can correct this within the PID by changing the “PSb” parameter, a temperature offset. To do this press SET, enter the value: 0089, press SET, scroll to “PSb”, and change the value to “-001″ (the first digit can be “0″, “1″, “-”, or “-1″). A button-by-button description is given for changing “Inty” and “outy” in the same way in our post about how to make a $50 dollar sous vide.
With our original PID parameters, when the heater is first turned on the SV overshoots by four degrees, then 30 minutes later it undershoots by two, and goes up and down a few times by a degree or so before settling down (Fig 2). While the temperature eventually reaches the equilibrium, putting in your cold food can set it off again.
The PID is a “Proportional, Integral, Derivative” feedback system to heat the water to the right temperature. There are three main values that determine how the PID respons, named P, I and d. Our PID has an added in overshoot suppression and control period (time between intervals turning the heater on).
The PID contains an automatic algorithm to figure out values for P, I and D. To set this up, all you need to do is turn the PID on, set a value for the SV, press and hold the “>” arrow, and wait. The “AT” (auto-tune) light will light up as the tuning starts, and the heater will turn on to reach the SV. When the AT light turns off, the PID will work as usual to keep the temperature at the SV. More about this process is covered in the JLD 612 manual.
Value: P / I / D / SouF / ot
Before: 5.0 / 100 / 20 / 0.2 / 2
With AT: 1.0 / 266 / 66 / 0.2 / 2
With the auto-tuned values, the approach to equilibrium was somewhat smoother (Fig. 3). However there is still some undershoot, which may be corrected by decreasing SouF. To do this we press SET, enter “0036″, and then scroll to “SouF,” press SET, and change the value.
Finally, to reduce wear on our relay electrodes, we will increase “ot” in the same menu to 10.
A final source of error is variation of temperature in the sous vide vessel. With nothing in the container, local variations of up to 0.2 °C were measured. The variation was primarily across the container rather than vertical, which is a sign that the mixing is working but the container is too large. The error might be increased if a bag prevents water from mixing in a certain region. Make sure your bubblers are working!