Psst... We're working on the next generation of Chowhound! View >
HOME > Chowhound > Cookware >
Jan 6, 2010 08:46 AM

Using a PID for sous vide

For Christmas, I received the Sous Vide Magic PID controller - a comparatively cheap method of turning a slow cooker, rice cooker, or hot plate into a temperature-controlled water bath for sous-vide cooking.

I'd like to give my impressions, and also ask a few questions of anyone who has experience with using a PID to cook sous vide.

I've seen a few people mention these devices on other threads, but I don't know how many people actually had one. I was curious if many other chowhounds have one and how they like it - if they find any limitations, durability issues, or whatever.

For my part - the first thing I noticed was that it was not as simple or easy as I expected. The user interface was not at all intuitive - I foolishly tried to navigate it without instructions, managed to miscalibrate its thermometer, change it's base settings, and then lock myself out of the controls, all without any clue what I had done. After a thorough reading of the instruction manual, I repaired my own damage and started cooking.

I also found I did not actually know what a PID does before I got one. I thought I did, but I was wrong. Strictly speaking, you don't actually need to understand what a PID does in order to use it - this one has an auto self-tuning function you can use in lieu of adjusting its settings yourself. But that is not an option for me given the tendencies of my psyche. So... more reading and fiddling before getting started.

On the other hand, though the interface was intimidating and I needed a several-hour long crash course on what a PID is to even figure out how it works, once you get going, it's really easy. Once familiar with the menu, you can set controls in seconds, warm the bath, add the food, and not worry about it until it's done.

I seemed to be able to attain temp control of about +/- 0.6 degrees Fahrenheit with a simple setting and +/- 0.2 degrees with a more elaborate setting using a medium-large sized slow cooker. So precision was more than adequate.

Overall, though the PID was less user-friendly than expected, I can't argue with results. I've now tried several long sessions without any glitches. Its temperature control is impressive. And it's far cheaper than the Sous Vide Supreme or an immersion circulator, even a used one.

Questions for any of you that have a PID:

1. Do you find a circulator is necessary when not stuffing your water bath full? Do some types of heating elements negate the need for a circulator? I understand the theoretical problems with forgoing the circulator, but my experiences with large stockpot + thermometer seem to indicate that it's not a big deal when heating from the bottom. If I do need one, what do you recommend?

2. The manual warns of overheating the PID itself - would, say, a 1000 watt hotplate on 'high' be too much power to plug into a PID?

3. Can I switch from PD mode (integral time set to 0) to a PID setting once achieving a rough stabilization without destabilizing the water bath for a little while?

4. Is there any reason I'd want to set derivative time to zero? The manual pointed out that I could do this and use it as a PI controller. Cool. So... why would I want to do that?


  1. Click to Upload a photo (10 MB limit)
  1. I'm still in the process of deciding whether to get the SVM, and so appreciate your post. As to your questions, I've found a little info on (1) and (2) on the internet. If you've seen the same stuff and are just looking for personal experiences, ignore the following.

    (1) I've read of some people using aquarium pumps to circulate the water. But the consensus seems to be that it's unnecessary unless you're making dramatic temperature changes (eg, flash-chilling after cooking).

    (2) Depends on your PID's power rating - the old SVM was 1400w. So a 1000w hot plate would be no problem, while a 1750w commercial rice cooker wouldn't be such a good idea. I think they raised the power rating on the new model to 1800w, so now the question is whether you're going to blow your 15 amp circuit...

    1 Reply
    1. re: alanbarnes

      Thanks Alan,

      Hope I didn't scare you off a PID too much. Now that I'm familiar with it, I can set it up in just a few seconds and expect it to work well. It has made long sous vide preparations possible. If you think about it, having to read instructions and fiddle around for a couple hours is not bad for picking up a new technology (new to me, anyway) - it's just more than I've come to expect from a new kitchen appliance.

      1) I was mostly curious as to whether no circulator would mean less efficient heat transfer - whether cooler water sort of encapsulated the food while cooking the same way air does in a non-convection oven, or whether there would still be enough convection in an uncrowded bottom-heated water bath to compensate. Since I don't use a temp probe in my sous vide bags while cooking, so far I have been cooking a while past pasteurization times in order to compensate just in case. Thanks for the tip about aquarium pumps.

      2) A quick look in the instruction manual confirms that my unit's rating is 1800w. Don't know how I missed it, but thanks for pointing it out. This of course makes me very happy. Lets see how big a water bath I can create. 15 amp circuit be damned.

    2. I have no clue about the specific controller. But I do some industrial controls for a living, and my smoker runs off a homemade PID program.

      With that in mind, you do want to set the derivative time to zero, if that essentially eliminates the derivative calculation. All the stuff I do, and I think hot water in a crockpot would be similar, doesn't need derivative control to meet/hold setpoint. It also tends to result in some funky responses.

      Integral, however, you do need, and it's basically going to be the time you want it to take to get back to setpoint when disturbed. A bigger value means it'll take more time but will be less likely to oscillate around the setpoint; smaller reacts quicker but is more likely to overshoot.

      Hope that helps. Here's a link explaining how a PID works. It's specific to Unitronics PLCs but the basics are there:

      14 Replies
      1. re: ted

        Thanks for the link, Ted.

        I'm still not sure I fully understand though. When cooking, I have tried setting my integral time to zero a few times. The system works - comes to temperature quicker than other settings I used and oscillates by about +/- 0.6 F. For this setting, my proportional band value has been low and my derivative time high.

        Meanwhile, a high(ish) integral time along with higher P and lower D values took a good deal longer to reach desired temperature. It seemed to stabilize at a lower temperature a couple times for a while before reaching desired temperature. I don't remember if it overshot or not, but once it stabilized there was very little oscillation.

        My understanding is that integral time controls how aggressively the device moves towards your desired temperature with respect to both how far off the current temperature is and how long the device has been at the wrong temperature. But the actual value I am adding for this is in seconds - is it the amount of time my PID will stay at one setting and accumulating error before readjusting/increasing output? Is my understanding wrong?

        I suspect I should try using an integral value - just a smaller one. Of course, there always is the self-tune function. Truthfully, just running in PD mode (integral time = 0) is convenient and accurate enough for my purposes. But its bothering me not to understand.

        1. re: cowboyardee

          I'm not 100% sure on how your controller is programmed. Looked for a copy of the manual online but didn't find one. And the ones at Auber Instruments don't have the instructions online either.

          My understanding of integral time is that it's how long you want the controller to take to get back to setpoint once disturbed. You're right in that it's impact on control is determined by how far away you are from setpoint and how long the measurment has been away from setpoint.

          Since the heating element is strictly on-off (I'm used to dealing more with variable-speed devices like pumps/fans), it's probably varying the amount of off time.

          The Wiki entry may be more useful than the Unitronics stuff:

          I would follow the Ziegler-Nichols tuning method. Probably will be a pain due to time investment and no datalogging (gotta love being able to walk away and later look at what's happening in Excel). Start with zero for I and D. Adjust P until you get oscillation, then add I based on oscillation frequency. Keep D at zero- trust me.

          1. re: ted

            Sorry to butt in...

            Does the pid vary the output voltage to the device or is it a binary on-off sort of thing?

            1. re: Paulustrious

              The "P" in PID stands for "proportional." In other words, the device makes varying changes to output based on error value.

              I'm sure others will correct me if I'm wrong, but it's my understanding that a PID device by definition must be able to deliver variable output.

              1. re: alanbarnes

                The reason I asked is that some people were talking about using the PID to control other devices. If they have electronic controls then how can these work when the output voltage drops to (say) 50v?

                1. re: Paulustrious

                  I imagine it depends on the device being controlled. But yeah, I've been wondering the same thing.

                  1. re: cowboyardee

                    AFAIK the only things you can control with the sous vide controller are devices that lack electronic controls. So it works well with a "dumb" slow cookers or rice pots, but not with a temperature-controllable crock pot or a neuro-fuzzy-logic rice cooker.

                    1. re: alanbarnes

                      I understand that to be true as well. But what I was wondering was whether even "dumb" electronics could have minimum power input needs in order to work at all. This does not seem all that likely with most basic heating elements, from my understanding. But that's the problem - I'm woefully undereducated with respect to how electronics actually work.

                      1. re: cowboyardee

                        There's no such thing as truly "dumb" electronics. By definition, anything electronic controls the flow of electrons, usually through the use of semiconductors.

                        "Dumb" electrical appliances, on the other hand, just complete one or more circuits using household power. When you put an incandescent bulb into a socket, all that happens is that 110v power flows through the filament, causing it to glow. Same thing with a basic rice cooker or crock pot, although the current flows through elements that generate heat rather than light. There are no processors or controls on the heating circuit, so the PID works great.

                        1. re: alanbarnes

                          Clearly you have not met my DVR. After TIVO, you know what dumb is!

                          1. re: firecooked

                            Tivo is smart, but deranged. There's a difference.

                        2. re: cowboyardee

                          So I'm guessing that an electric kettle would also be a good starting point, one with the element 'out of sight'.

                          Then you can try a larger size using one of those (non-instant) hot water dispensers.

                          Finally you can finish off with a household hot water tank with the thermostat disabled.

                          Just rambling, but I'm searching for suitable devices. Currently I'm imagining a small (20 gallon) hot water tank with a low level heating element with the tank sawn off horizontally at about 18 inches. You can pick up a damaged one fairly cheaply. They also have a drain plug which could be helpful.

              2. re: ted

                Ted, if you're interested in the instruction manual, check out the link to the sous vide magic...
                In the bottom left corner of the page is a link to download the user manual as a word file. My model is the 1500D.

                The slow cooker I use is strictly on/off (actually 3 dial settings - low, med, and high- I set mine to high), but the PID is varying the output to the slow cooker as part of its algorithm. That should in turn vary the output of the slow cooker, right?

                Thanks for mentioning the Ziegler Nichols method. Hadn't heard of it. Will try.

                The manual states that derivative time should help to minimize overshoot (when figured correctly) and also to help the PID respond quickly when the temperature in the system is disturbed - like when food is added to the water bath. Both very useful in my application if that's actually what it's doing. So... what is it actually doing?

                1. re: cowboyardee

                  They definitely hid that manual away on the page. Will take a look.

                  For alanbarnes and Palustrious, PID is used with on-off devices. It's actually a concept that took a while for me to get, also. Rather than modulating the output to the device, the controller is modulating the balance between run time and off time. That allows the controller to be simpler- it's just switching a relay on-off. Outputting variable voltage or a pulse-width-modulated signal is a more complicated proposition.

                  Folks are using similar controllers for refrigerators for fermenting beer and in homebrew systems. In those cases (fridge compressors, gas burners), there really isn't an option for an economical variable-output device. The BBQ Guru and Stoker fan systems for smokers are also running on similar principles.

          2. Let me try to help, as I've been using the SousVideMagic for about a year now (1500A, 1500B, 1500C, and 1500D), and have experimented rather extensively.

            1. At least when using a rice cooker (which heats from the bottom), a circulator is generally unnecessary. Of course, if you overstuff a smallish container, anything might help. The laboratory circulators are more often used with broad, flat containers, where circulation might be more of a problem. Likewise, a CrockPot, which heats from the sides, might be more problematic. If you do feel the need to stir things, i would suggest an inexpensive garden fountain pump (around $20). Just don't use it above 160F. An aquarium pump will also work, but some of the air stones also have problems at higher temps.

            The SVM will handle 15 amps, so a 1000 watt/120V hotplate should be fine. if necessary, you can dial down the power output using one of the menu options.

            Several of us have been working on a PID tutorial, which will hopefully be available on the Fresh Meals Solutions web site shortly. Until then, here is a radically condensed version.

            You can try the auto-tune function, but you may be able to get better results by simply following the recommended settings for various devices in the user manual.

            If you want to try it manually, start by setting I=0 and D=0, then pick a value for P that is relatively high (100 or more for Celsius, or 180 for F). Observe how long it take to come up to temperature, and whether there is any overshoot. If there is significant overshoot (more than 2F), try increasing P by 50%, and try again, then back it back down as appropriate. Note that if you start with cold water, overshoot is more likely, so filling the tub with water that is close to the desired temp is easier.

            Once you finally get it to the point where it is heating up "fast enough" and not overshooting "too much," you will probably find that it settled down to equilibrium at a little bit under the desired set point. Then it is time to start playing with the I value. Start high -- like I-500 (note that I=0 is effectively I = infinity), and then reduce it and see what happens. Eventually, it will progress from a steady-state undershoot to a long-term overshoot, and then you can zero in on perfection -- at least for that particular cooker.

            The D value is of some value if you tend to start the water coming up to temp, e.g., by starting a timer while you are still at work, and then dropping a bunch of frozen food in the pot. Increasing the D value can reduce the time the bath requires to come back to steady-state, but I can also lead to overshoot, or even instability. A rough rule of thumb calls for D to be about 25% of the I value, but even that may be too much for your particular style of cooking. if you leave it at 0, life will go on.

            Note that a little overshoot in the bath temperature may not be all that harmful, as the meat or whatever still has to come up to temperature internally. But if the meat is already thawed (close to room temperature), you don't want much of an overshoot.

            As far as controlling "smart" devices, understand that the SVM works by controlling the output power. Whether it does that by cycling at full voltage, with a different amount of time on vs. off, or whether it reduces the power via an analog circuit, I don't know, but if you have anything more than an on/off switch on the device it is probably going to become very confused. Also, the device is intended to control a resistive load -- I wouldn't use an induction heater.

            Hope that helps.

            3 Replies
            1. re: Robert Jueneman

              Bob, thanks for your very concise condensation of our PID-tutorial.
              @ aquarium bubbler: use the air stone as a weight to pull the tube to the bottom of the bath, but cut a sideward hole in the tube just above the air stone; the much larger bubbles will produce less cooling by evaporation than the fine bubbles from the stone.
              @auto-tuning: in my experience, results in °C mode are more reliable than in °F mode, maybe because a hysteresis range of 1°F is too small.
              @controlling output power: the SVM controls output power by simple on/off in cycles of 2 seconds. Some cookers do the same, e.g. my stockpot has an analog wheel, controlling temperature by varying the proportion of on-time/off-time, so to control it by the SVM, the wheel on the stockpot must be turned to maximum temperature to give 100% on-time.

              1. re: PedroG

                Air stirring: I did a compromise thing: I weighted a food grade silicone hose with two 1/2" nuts. I folded the hose and pushed it through the nut when doubled, this created a plug, then I used my diagonal cutters to make a series of holes that were shaped like the edge of a crescent moon. This gave me bubbles in a line that worked as sort of a stirring curtain.

                I am doing this in a roaster, in which I've measured a 6F difference between the top and bottom of the water during heating.

                Someone suggested that the primary negative regarding bubble stirring is evaporation, but this post suggests loss of heat is the issue? What is it? It has to be tons more reliable than other methods of manual stirring on the cheap, my bubble stirrer is top end (food grade silicon tubing, small air pump that has a volume controller built in) and my investment was $25 with most going to the hose. I've seen those small via pumps I guess that people say fail because of the heat - they are cheaper to start, a single pump will cost $12 or so. I have also seen a small pump online that says it can pump liquids up to 105, and works on 6v. It is fine for the applications listed but when you compare the 15 lpm volume of commercial circulators to 1 lpm, well, it seems inadequate. So I'm still left with no real good, obvious way to stir.

              2. re: Robert Jueneman

                Thank you Robert (and Pedro). That was incredibly helpful.

              3. I've been using the SVM for a year or so and am extremely pleased with its performance. I use a Hamilton Beach Slow cooker (about $30) from Target and use it on the high setting. I religiously follow the manufacturers recommended operating procedure and have not had any trouble.

                I use a water pump for circulation that is made for an aquarium and attaches to the ceramic bowl of the slow cooker with a suction cup. The manufacturer is Hydor Koralia and here is a link to one of many suppliers


                I should warn you that the manufacturer of the pump does not recommend its use in liquids hotter than 95F, but I have been using it often for a year at temperatures up to 190F and it is still working fine.

                1. To those who have a PID - have any of you tried it with a hotplate? Any reason this would cause me problems? You would want to turn the hotplate to its high setting so that your on/off cycling is controlled by the PID only, correct? And if so, would a powerful hotplate necessitate a pretty large pot with a lot of water or could it effectively control a smaller bath as well with the right PID settings?

                  6 Replies
                  1. re: cowboyardee

                    I haven't personally tried it, but it was one of the methods I was going to try when I first got the controller.

                    However, the first safety caution in the owners manual states "This controller is designed only to be used with devices that have limited power and their own thermal cutoff protection, such as a thermostat or thermal fuse in case of controller failure. Do not use it with a cook top or hot plate."

                    1. re: trouttr

                      okay after just spending the last FOUR hours trying to figure out why my CrockPot nor my Rice Cooker wouldn't work...for sous vide...I came across the fact that they are too "smart"...just found a hot plate and hooked that up to my PID....IT'S WORKING! I know you aren't supposed to do this but I will let you know my results ....that is before I go out and buy yet another rice/slow cooker

                      1. re: heathermac

                        "Too smart" is rarely a problem around my house...

                      2. re: trouttr

                        Thanks for pointing out that warning.

                        *Disclaimer* - People should of course follow all warnings on their appliances for safety's sake - what I'm about to say is speculative and as I have no expertise in the matter, I could easily be wrong. *End of disclaimer*

                        Sounds like that warning is primarily concerned about risk for fires - if your PID malfunctions or your thermometer comes out of the bath in a rice cooker or crock pot, and you walk away or go to work/bed, you're a lot better off than if the same thing happens with a hotplate. Fire hazard.

                        If that is indeed the hazard, I would think a hot plate might be usable for any supervised cooking you do with it - the faster sous vide preparations.

                        I'm not gonna go telling others to try this. But if anyone can think of any other reasons not to use a moderately powered hotplate, please let me know. Because I'm buying one.

                        1. re: cowboyardee

                          I wouldn't. I used an "induction" hot plate, no PID, I regulated the temperature manually.

                          I measured around in the pot and noticed significant differences in temperature. I believe the problem is not that great with a rice cooker or hot pot since these are insulated which seems to help a bit.

                          If your sold on SV I would make that investment, otherwise start out on your normal stove and try it out.

                          Note also that fresh meal solution is coming out in the coming weeks with a "bubbler",



                          1. re: jk1002

                            I've already got a big slow cooker that works great with the PID.

                            Problem is I cook in competitions in other peoples homes. So I was looking for something more easily portable than the slow cooker for (relatively) short applications.

                            I also spent a good while cooking sous vide on the stove top with an instant read thermometer, with some success. For anyone interested, I found that using a very large stock pot of water went a long way in helping stabilize the temperature, for reasons that are obvious in retrospect. Still, I always had to give the pot an occasional stir to help ensure even temp within the pot.

                            Thanks for the link to the "bubbler" - I hadn't seen that.