Sous Vide: Recipes, Techniques & Equipment (Part 5)

KennethT · April 16, 2009

I've actually been thinking about doing something similar, along the lines of the SV burgers.... someone PLEASE let me know if this is a little crazy, or just not practical...

Anyway, I've been thinking about doing a bolognese sauce and, separately, meatballs by SV at say 55 or 60C for long time (i.e. 24 hours, ,etc)... or what about either 67 or 83C for 8 hours or so?? I gather that a really good bolognese needs to simmer for a long time to get a really great mouthfeel.. so I figured why not do the same thing you'd do with tenderizing a chuck steak, but to ground meat?

Similarly with meatballs - I've had way too many meatballs made the conventional way where the meat was gray and really tight... either from simmering too hard, or not long enough... I figured, even if the connective tissue int eh meat is ground up, it's still there... if you cook it long enough to convert it to gelatin, wouldn't that make a melt-in-your-mouth product???

Any thoughts???

KennethT · April 16, 2009

My understanding of botulism was similar - that you need a long time in a zero oxygen enviroment for it to grow and produce toxins... the FDA shows in the food code how long you can keep pasteurized ROP products at different temperatures.... From what I understand, if you do cook-chill, you can keep stuff at refrigerator temps (34F) or lower for a max. of 30 days (according to 3-502-12D(2)(e)(i))

So, theoretically, you can cook your burger to 55C for a time long enough to pasteurize depending on thickness, and then chill them down and refrigerate for a month - then just take out and sear on the grill to bring back to temp whenever you're ready... If you want to cook at lower temp. (for a really rare burger), I don't see why you couldn't do it, just so long as your burger wasn't 4" thick so that it would sit in the danger zone for too long.... then eat right away - can't store in the refrig. since it's not pasteurized....

Am I reading this correctly? Someone please correct me if I'm misunderstanding this....

e_monster · April 16, 2009

As I understand it, 131F is the minimum temp for pasteurizing meat. So, you can't extrapolate from 131F to find a time to pasteurize at 127.5 F as 127.5 below the temp at which you can pasteurize. As I understand it, once you drop below 131F, you are potentially in the danger zone which makes it more dangerous the longer the food is there because some bugs can survive and reproduce below 131F.

Howsmatt, I can try to answer your four recent questions.
4. If you look at St. Douglas's tables again, 5:14 is enough for even a 70 mm thick piece of meat at 131F, vs. 4:26 at 136F. I cook my steaks at around 131 because my wife doesn't like really rare roast beef, and it isn't worth firing up two SVM systems (although I now have three -- a 1500A, 1500B, and 1500C,. and three rice cookers of varying sizes from 1.8 liters to 10 liters). So if you are talking about 127.5F, extrapolation would suggest that another 38 minutes (5:52) would be more than enough, even for the thickest steak or brisket you are likely to serve to an individual, and 40 hours is way over the top. (This assumes you aren't picking up exotic bacteria from undersea lava vents or the pools at Yellowstone, or from Mercury or Venus.)
←

MaxH · April 16, 2009

Just as a reminder, the C. botulinum occurs in three steps: spores, then bacterial growth in a friendly anaerobic environment, then toxin production. The spores are extremely heat-resistant (surviving several hours of boiling) so there is no concept of "pasteurization" safely preventing this organism (standard food-processing prevention requires higher than sea-level boiling temps., therefore elevated pressure). The bacteria (like other claustridia and further anaerobic species) have a range of growth temperatures and a time scale, though famous fatal botulinum poisonings have happened after two weeks of cool storage. The C. bot toxin, a side effect of the bacterial growth, is fragile enough that ordinary cooking methods are normally a safety measure; but conditions listed in standard references as sufficient to kill C. bot toxin in food entail higher temperatures than sous-vide uses:

Merck Manual (2006) 30 minutes at 80 C

CRC Handbook of Food Toxicology (2002) 20 min 79 C or 5min 85 C

World Health Organization (2002) 5min "above 85 C" or "few min" at 100C

ETA: The three classic ways to prevent botulism poisoning in foods stored anaerobically without first pressure-cooking are: consume soon (before bacteria can develop much) -- recurrent phrase is "a few days" if refrigerated; stop bacterial growth chemically (preservatives or high acid); or stop the growth thermally (i.e., freeze it). The quirk about sous-vide is that a band of temperatures slightly below normal professional sous-vide cooking temps. actually greatly accelerates anaerobic bacertial growth, compressing its time scale. A respected biologist friend who pointed this out years ago uses the same equipment to cultivate bacteria in a laboratory that professional kitchens use for sous-vide cooking!

A looming side issue, by the way, concerns meat confits (non-sous-vide) which in modern recipes are lightly salted, not otherwise preserved, cooked at atmospheric pressure, then sometimes stored anaerobically under fat for months, even unrefrigerated. That doesn't meet existing guidelines for C. bot safety and the problem would come if such aged confit were served or even carelessly handled (or "sampled") without much further cooking.

Edited April 16, 2009 by MaxH (log)

howsmatt · April 16, 2009

I have a sous vide magic setup with a 15 cup rice cooker and 18 quart rotisserie oven. I could use 4 of these setups for the price of one industrial version but some items I would like to cook are big (think 16 English cut short ribs) and might not fit in the rotisserie (again I'll have to see). I also want to make sure there is sufficient space for water to move around the food--at home I'm not too worried but in a restaurant I don't want problems with food inspectors.

I did the brisket (not corned) for 40 hours. Still not great.

My understanding is that a burger heated to 131 within 4 hours (realistically 1-2 hours for safety) should be safe. Searing for flavour is important but shouldn't be required from a bacteria standpoint. My curiosity here would be focused on how long and longer cooking times might affect the texture. Guess it's time for more research.

DouglasBaldwin · April 16, 2009

Sous Vide Hamburgers

First off, there is no greater risk of botulism in hamburgers than other food products prepared sous vide. As stated up thread (and in my guide), to prevent the outgrowth and toxin formation of Clostridium spp. (e.g., C. botulinum and C. perfringens) either: (a) hold the food at or above 54.4C/130F until ready for service or (b) rapidly chill (in ice water), chill hold (at below 38F/3.3C for up to four weeks or freeze indefinitely), and rethermalize/reheat (at between130F/54.4C and the original cooking temperature) until ready for service.

Can you pasteurize ground beef sous vide? You betcha! In fact, most the studies which determine the pasteurization times for beef use ground beef (which has been inoculated with a pathogen) and placed in either a vacuum bag or a test tube which is placed in a water bath controlled with an immersion circulator for a set length of time! Thus, you can pasteurize your hamburger using the times in Table 5.8 of my guide.

You cannot, however, linearly extrapolate pasteurization times to temperatures lower than 131F/55C! From Table 5.8 in my guide, you can see that I use D_60^9.22 = 3.63 min; this means that at 55C/131F it takes 6*3.63*10^((60-55)/9.22) = 75.9 min or at 53C/127.5F 6*3.63*10^((50-53)/9.22) = 125.1 min. However, C. perfringens can grow and produce toxins at temperatures up to 127.5F/53C. Therefore, 54.4C/130F is the lowest practical temperature which meat can be safely held or pasteurized at.

Robert Jueneman · April 17, 2009

Douglas, thanks for responding to this thread!

I'm happy enough cooking and serving beef at 131F/55C, and I don't particularly want to wait 2 hours before serving it, either. (24 or 48 hours is fine, however.)

But I'm not a microbiologist, and so I don't know what C perfringens is, how hazardous it is, or (especially) how common it is and how it spreads. As I understand it, this should only be a problem if you cook-hold for more than four hours, correct?

Would this risk be significantly reduced if I were to grind my own burger meat, and freeze it promptly and quickly, so as to reduce the possibility of outside contamination? I suppose I could even pre-sear the outside before grinding it, if that would help significantly.

To change the subject slightly, when cooking lean burgers (12% fat, according to Costco) they are coming out rather dry, yet only a little liquid is left in the bag. Would letting the burgs rest for 10 minutes or so help to reabsorb the juice, or do I need to switch to ground chuck or something that is a little less lean?

Paul McMichael · April 17, 2009

What is the maximum size pot I could use with an immersion circulator? It would be for a restaurant so getting up to temp time is not important--I just need to keep it steady all day/night. Obviously the more prep I can do at once the better.
Matt
←

It is all a matter of insulation. I have a 1000 watt circulator ($30 on ebay) that normally sits on a 10 liter tank. Last week, to do some brisket, I clamped it on to a 40 qt colman cooler. works like a charm.

Paul

Robert Jueneman · April 17, 2009

Sous vide problems.

Today was not a good day!

First, I had problems yesterday with my FoodSaver suddenly not pulling much of a vacuum. It turns out the lower gasket was torn somehow. Solution -- order two replacement gaskets ($3.00 each from FoodSaver, plus a upgraded FoodSaver (what appears to be a very nice unit -- the Professional III, with five different vacuum settings and other goodies, all at a very nice 25% discount), plus five boxes of quart bags (also discounted), plus second day air shipment -- total $345.66. Ouch!

Lesson learned #1 -- order spare gaskets and other spare parts ahead of time -- they are rather fragile.

Second, two different probes failed on my Sous Vide Magic 1500A and 1500C, and unfortunately they are not interchangeable. Although the probe wires are thin enough to allow my rice cooker to close on them, apparently they have somehow gotten kinked or broken. One doesn't work at all, and the other is erratic, with temperatures vacillating plus or minus 20 degrees F. Damn!

I'm going to order two replacement probes tomorrow, and maybe a spare or two, and then I'm going to encase the probe wires in poly tubing to protect them. If the lid won't close completely, that's too bad, but I'll live with it

Lesson learned 2 -- protect any external probes from damage, and keep a spare or two on hand if you have gotten used to preparing all of your meals sous vide, or if you are running a restaurant.

A couple of days ago, the power went out (just on our block), and so at 8:30PM we gave up and went to the Elephant Bar in Fremont, CA. I ordered lamb shanks, but was very disappointed, as they were too well done, and covered with a too salty sauce that tasted like an old pot roast. "I can do better than that," I said.

Wrong! And thank you, darling, for not being so unkind as to remind me of my brave words.

Cooking two lamb shanks at 55.5C/132F for 24 hours left the lamb moderately tender, but not quite not melt-in-your-mouth, and probably too rare, even for my taste, with barely enough meat to make it worth the effort. Scratch that idea.

Cooking tiny carrots sous vide for 30 minutes with a little orange marmalade and butter revealed the fact that the carrots had probably suffered from freezer-burn, and tasted a little off. Mediocre. Note to self -- need bigger, fresher carrots.

The tiny "Tipsy" cocktail onions soaked in vermouth, SV with butter and a pinch of salt and later glazed on the stove were better -- perhaps the hit of the meal.

Not quite a disaster -- so far we are still alive -- but certainly not up to par.

Sigh.

Edited April 17, 2009 by Robert Jueneman (log)

tomdarch · April 18, 2009

I did a small test chunk of dry-rub (salt/sugar/smoked paprika) pork belly this week. I cooked it at 65C for roughly 8 hours, then seared it in a cast iron skillet.

It wasn't amazing. The flavor was great, but the meat was a bit dry and/or a bit tough.

I'm trying to figure out if the temp was too high, if the cooking time wasn't long enough, or if the salt rub just pulled too much liquid out. Any ideas?

Also, the pork bellies have their skin on. With the test chunk I did, I didn't realize that it was there, so I went ahead and seared it. I pretty much liked it. Some recipe I've seen comments on removing it. Of the two chunks I'm using, the one I didn't take the test from has a thicker, more noticeable skin. Now that I'm thinking about it, I'd rather get the skin off, so I can really sear/crisp the fat. They are in ziplocks covered in rub currently. Would I be better off removing the skin after rinsing off the rub, or removing it when they come out of the water bath?

Edited April 18, 2009 by tomdarch (log)

tomdarch · April 18, 2009

Separate question: has anyone done parsnips sous vide? Their similarities to both carrots and potatoes makes me think they may be winners! (a little butter, salt, maybe some herb, then in the bath at 85C for 30 to 60 minutes?)

jackal10 · April 18, 2009

I did a small test chunk of dry-rub (salt/sugar/smoked paprika) pork belly this week. I cooked it at 65C for roughly 8 hours, then seared it in a cast iron skillet.
←

8 hours is not long enough. Try 24 to 48

Cook the skin seperately, 80C for 12 hours. Scrpe off the fat etc for he inside. SEson well with salt, pepper, crushed garlic. Place the skin between two sheets of silicone paper, then place in between two flat oven trays. Weight to flatten, then place into the oven 200C to bake for 10-12 minutes. Remove from the oven and immediately place the crackling onto a board and cut into strips before it has a chance to cool. Return the strips to the oven and continue to roast for a further five minutes, or until the crackling is crisp and golden

e_monster · April 18, 2009

Doug Baldwin's great sous-vide pages mention that eggs can be pasteurized at 135F by cooking for 1 hr 15 min or more. Anyone have any idea what the time to pasteurize would be at 131F and 133F? I tried to find some good data on the web but found that a lot of the pages (even from the food industry) wave their hands around and act like 140F is the minimum temp at which salmonella is killed.

I found one paper that confirmed that 131F was sufficient but I wasn't able to determine from the paper the time required to pasteurize against salmonella in eggs at that temp. (Apparently the substrate makes a difference in how long it takes to kill the salmonella).

I'd love to be able to pasteurize eggs when I am cooking at 131F and 133F.

Robert Jueneman · April 19, 2009

e-monster, I can understand that might be convenient to throw some eggs into the water bath while you are SV'ing some steaks.

But whether you use 135 or 140F wouldn't seem to matter much -- at the end of the evening, while you are cleaning up, etc., or at the beginning of the day before the guests start to arrive at your restaurant, jack up the temperature by a few degrees, cook 5, 10, or 20 dozen eggs -- whatever number you think you will need for a day or two -- for 1:15, then chill them and save them for use throughout the day.

I am neither a microbiologist nor an attorney, but if I could offer some strictly layman's advice, if something doesn't make a whole lot of difference to you, you are far better off following "conventional wisdom" in case a law suit ever materializes.

Now, if this is just for own personal use, you can consider the odds that an egg is infected (1 in 10,00 to 20,000) vs. the likelihood you will be struck by lightening, or a bus, or by an enraged deer, or a jealous husband, or die of a heart attack or stroke, and decide for yourself whether it is worth worrying about.

Obviously if you have AIDS or TB, are on chemotherapy, are pregnant or elderly, or otherwise might be immune compromised and more vulnerable than most, you should adjust your behavior accordingly.

Bob

DouglasBaldwin · April 20, 2009

Doug Baldwin's great sous-vide pages mention that eggs can be pasteurized at 135F by cooking for 1 hr 15 min or more. Anyone have any idea what the time to pasteurize would be at 131F and 133F? I tried to find some good data on the web but found that a lot of the pages (even from the food industry) wave their hands around and act like 140F is the minimum temp at which salmonella is killed.
I found one paper that confirmed that 131F was sufficient but I wasn't able to determine from the paper the time required to pasteurize against salmonella in eggs at that temp. (Apparently the substrate makes a difference in how long it takes to kill the salmonella).
I'd love to be able to pasteurize eggs when I am cooking at 131F and 133F.
←

That is an interesting question, I'm not sure why most studies do not consider anything less than 135F/57C in eggs or poultry; . For an educated guess, we can use the D-values for Salmonella enteritidis of 4.5 min at 58C and 6.0 min at 57C in [J Appl Microbio 83 (1997) 438--444]. These D-values give a z-value of 8C (1/log_10(6.0/4.5)). Therefore, at 131F/55C it would take 6.5*6.0*10^((57-55)/8) + 35 = 105 min and at 133F/56C it would take 6.5*6.0*10^((57-56)/8) + 35 = 87 min (where the study found that the egg took 24--35 min to come up to temp).

From Table 4 of the survey article [J Food Sci 71 (2006) R23--R30] (which includes many studies at 55C, 57.5C, 58C, ..., 70C for Salmonella spp in poultry), I compute a D-value of 4.8 min at 60C with a z-value of 6.46C. Thus, a 6.5D destruction of Salmonella spp is: 6.5*10^(9.97- 0.1548*55) + 35 = 220 min at 131F/55C and 6.5*10^(9.97- 0.1548*56) + 35 = 165 min at 133F/56C. However, since Salmonella enteritidis (which is the strain found in eggs) is a less thermally resistant strain than the Salmonella senftenberg containing `cocktail' used in most the studies, the times in the previous paragraph should be sufficient.

Edit: Added extra detail.

Edited April 21, 2009 by DouglasBaldwin (log)

blackp · April 22, 2009

What should my first SV meal be?

This is my first post.

I've got a chamber vacuum machine (Henkelman - previously only used for packaging for the freezer). A commercial rice-cooker (Tiger) and I'm waiting for my Sous Vide Magic PID controller to get from Toronto to Sydney - hopefully it'll be here later this week.

I've spent the last several days reading this entire thread and have learned heaps. Now I'm raring to go!

If I am going to impress the sox off the good lady of the house what should I try first?

I've made Tetsuya's Confit of Ocean Trout in the oven under oil @60C and that was a great success (According to one friend it was as good as the original, but I still think that Tets does it better).

That got me thinking of fish for the first SV meal - or should I get a normally tough cut of meat and tenderise that over a long time - or should I just try a steak?

What do those more experienced SV users recommend?

Thanks in advance,

Peter.

e_monster · April 22, 2009

Here are my thoughts based on my personal experience and my reading of others' experiences.

Fish sous-vide is a bit hit and miss -- some people love it, some are lukewarm to it and some downright don't like it (including Tom Colicchio). Personally, I love salmon at 113 F but my wife is lukewarm. She liked it but doesn't ooh and ahh the way she does for some dishes.

A nice thick rib-eye cooked sous-vide at 129F for an hour or two (just to bring it up to temp) followed by a quick (30 seconds to a minute per side) sear always gets raves when I serve. The nice thing about this one is that you don't have to wait a day or two. This is mostly something for those that love steak medium rare with a thin well-browned crust. The fat softens nicely in a way that it doesn't seem to when prepared traditionally.

Short-ribs (well-trimmed ahead of time) and cooked at 133F to 135F for 48 hrs and then subject to a quick sear consistently get raves (even from people that don't think of themselves as big meat eaters). I serve this with sour cream and horseradish as if it were roast beef.

Pork tenderloin cooked for a few hours at 140F (I usually put a marinade of equal parts cider vinegar, soy sauce and sugar in the bag along with a cap of liquid smoke) MOSTLY gets raves -- but there have been one or two outliers that were lukewarm because they like their pork well-done. For a couple of friends, this is their favorite.

Flat-iron steak at 131F is another that gets raves. I cook it for 24 hours and jaccard it per Douglas Baldwin's instructions.

Another big winner for me has been tri-tip (bottom sirloin) jaccarded and cooked at 133F for 5 to 8 hours with marinade in the bag -- and then flamed under the broiler or with a blow torch to brown the outside. I like a marinade that is 2 parts soy sauce, 1 part vegetable oil, 1 part cider vinegar and 1 part sugar. A capful or half capful of liquid smoke adds nicely.

Other meals with a quick pay off are chicken breast at 135F to 140F (be sure to cook long enough to pasteurize) and duck at 131F. If you love chicken, the texture of sous-vide chicken breast is great very very tender -- use the best quality chicken you can find. If you love medium-rare duck breast, there is no better way to prepare it than 131F followed by a quick sear. (Remove the skin before cooking and cook it separately in the oven so that you get lovely duck breast topped by crisp skin -- I have not had great luck getting the skin to crisp if I leave the skin in the bag.

**nickrey** · April 22, 2009

What should my first SV meal be?
This is my first post.
I've got a chamber vacuum machine (Henkelman - previously only used for packaging for the freezer). A commercial rice-cooker (Tiger) and I'm waiting for my Sous Vide Magic PID controller to get from Toronto to Sydney - hopefully it'll be here later this week.
I've spent the last several days reading this entire thread and have learned heaps. Now I'm raring to go!
If I am going to impress the sox off the good lady of the house what should I try first?
I've made Tetsuya's Confit of Ocean Trout in the oven under oil @60C and that was a great success (According to one friend it was as good as the original, but I still think that Tets does it better).
That got me thinking of fish for the first SV meal - or should I get a normally tough cut of meat and tenderise that over a long time - or should I just try a steak?
What do those more experienced SV users recommend?
Thanks in advance,
Peter.
←

Welcome to eGullet! You have found an invaluable resource.

From my experience on eGullet, there seem to be a lot of national differences in what people like and recommend. It seems a lot of US-based contributors (not all) are lukewarm about dishes that we in Australia find very appetising and appealing (and vice versa).

If you like Testuya's ocean trout dishes and your good lady does as well, by all means try sous vide fish.

Make sure, however, that you prepare her and others for it being noticeably cooler than pan fried fish.

Moreover, it is still a part of a total: if you don't have an appropriate sauce, accompaniments and presentation, even the best cooked sous vide product will not be appealing.

For more Australian cuts of meat, topside done at 56C comes out perfectly and any leftovers make a wonderful sandwich filling.

Feel free to PM me if you need help setting up the controller with the Tiger cooker, it took me a while to get the settings correct.

howsmatt · April 22, 2009

Souvlaki using lamb shoulder cooked 131 for 48 hours then seared is amazing (oregano and mint before searing). Short rib is great but really requires the right sauce. Chicken breast at 141 then seared is really tender and anyone can immediately appreciate the difference. Fish is good but I've found that every fish needs a different temp. and if you're over (as always with fish) it's no good--perhaps not the best choice for a first try.

Good luck and welcome.

As for the burgers: You can get a nice compressed burger with no filler cooked medium. The downside is that it seemed the burger released a lot of flavour into the liquid. Perhaps a panada would be required. Still not sure I can get results better than classic cooking.

Matt

Edited April 22, 2009 by howsmatt (log)

blackp · April 23, 2009

Thanks for the suggestions.

The SVM arrived this morning, and I'm trying to calibrate it for the Tiger (Thanks for your assistance nickrey). At the moment it overshoots by 3.5C which is a bit much for accuracy. In the last hour it has come back down to it's set temp and so far seems to be holding fairly stable. Perhaps opening the lid when it comes up to temp will be one way of counteracting the overshoot faster.

Impatience will get the better of me for sure and the Rib-Eye steaks in the freezer are going to be tonight's meal - calibration fixed or not. In any case if I'm only SVing them until the internal temp comes up to target I figure I won't have well done steaks.

The souvlaki suggestion from howsmatt sounds great too - I will give that a go when I have a piece of lamb.

Cheers,

Peter.

e_monster · April 23, 2009

That is a lot of overshoot.

Did the unit come with a guide to adjusting the PID settings?

Does the temp stabilize if left long enough? Until I got my roaster setup calibrated, I would get a few degrees F of overshoot when first coming up to temp and it would then stabilize rock solid after an hour. (Until I got the settings right, I would wait until the temp stabilized to add food -- nowadays, I might get 1 degree F of overshoot initially and then it quickly stabilizes).

Did you start with hot or cold water?

I would wait to get your calibration before doing any experiment you care about -- otherwise you might be disappointed by the results. 3.5C is a huge overshoot.

If the SVM machine has auto-tune mode, use it. If the SVM machine is essentially the same as one of the Auber Instruments PIDs that doesn't have auto-tune, you will probably have to tweak the settings a little bit -- differences in insulation can make a big difference in the required settings -- and out-of-the-box it is probably set to something that responds pretty quickly.

You should email their support people for help -- the people at Auber were very helpful and hopefully the sous-vide magic people will have good support, too. Suyi at Auber mentioned that it is worth exploring settings that use just PD when one has a setup that overshoots a lot (especially true of setups that use a lot of water or have a lot of 'inertia')

blackp · April 23, 2009

Report on meal #1 using SV with Tiger rice cooker and SVM PID Controller.

When I went to the freezer I found that I had veal cutlets instead of the rib-eye steaks that I thought were there - oh well.. keep going anyhow.

My observations are as follows:

The temp overrun is relatively short lived, although I will experiment with the PID settings to attempt to fix it.

I set the temperature to 54C. When it came to the target temperature I put the room temperature packages (about 20C) into the water. The temp dropped a couple of degrees then overshot to about 57C before stabilising at 54C after 35 minutes. The temp then remained at 54C consistently until I removed the packages from the water after a total time of 90 minutes.

I then seared the cutlets on a very hot cast-iron pan briefly.

The result was a more tender piece of meat than I've been used to and a slightly different flavour - I guess due to the less Mailiard effect as a result of the shorter time on the high heat.

At the same time I put some ripe tomatoes in a bag with a couple of sprigs of fresh thyme from the garden and some S&P along with some EVOO. They were vacuumed to 90%. I would not do it that way again - they needed to be weighed down with a cake rack to keep them submerged. I need to figure out how much vacuum will work to remove more air without squashing the tomatoes. They were very good, but there was no noticeable addition of flavour from the thyme or the EVOO.

As to e_monster's suggestion re tuning the PID - I did get a return email from Frank @ SVM, but have yet to try his advice. I'll report back once I've had time to test it and check the results.

PS

If I knew how to post photographic evidence on this forum I would do so.

Cheers,

PB

Robert Jueneman · April 24, 2009

blackp, I am working with Frank Hsu at Fresh Meals Solutions (source of the Sous Vide Magic) to put together a guide to tuning the SVM. I was all set to fire up my data logging thermometer tonight, as a matter of fact, but I realized I first have to find a USB to serial port adapter -- I no longer have a computer with an old-fashioned serial port.

Here are my initial suggestions, but they are primarily based on my experience with the large 10 liter rice cooker, instead of my 1.8 or 1.2 liter Tiger and Ikeda units.

First, forget the Autotune function -- it seems to come up with some rather bizarre settings, at least for the large pot, even after running all night.

Instead, first determine an approximately optimum P (proportional) setting, with I=0 and D=0. I would try the settings in the instruction book first, and then experiment from there.

At least initially, start at ambient temperature, and record and plot the temperature of the bath (preferably using an external thermometer with 0.1 degree resolution) every 30 seconds for say 20 minutes.

If the overshoot is too great, the P value is probably too low. As a result, the SVM is putting out full power until the unit is close to the set point temperature, and at that point the trapped heat makes the device go into overshoot, even after the power is essentially off. Then, because rice cookers are well insulated, it takes a long time to drop back down.

On the other hand, if there is relatively little overshoot, but it takes forever to come up to temperature, the P value is probably too high. I don't know that I would cut it in half -- that might be too much -- but cut it substantially. It should then go into overshoot mode, and by increasing or decreasing by 1/2 of the difference each time, you should be able to optimize the P value fairly quickly.

(Of course, one good way to simplify this entire process is to add preheated water of approximately the right temperature to the pot to start with!)

If the final result, once everything has stabilized, is an undershoot, then you may need to adjust the I value. So far, with the large pot, I haven't found the need to do this yet, but if you check out "PID controller tuning" on Wikipedia, you will find some suggestions.

I THINK that the best way to do this is to apply a massive increase to the I value, in which case the system probably ought to start oscillating -- swinging back and forth slowly from a minimum to a maximum and back. Once that happens, record the period, i.e., the amount of time in seconds between two successive minima or maxima, then divide by two, set the I value to that., and experiment from there.

I'm not even close to determining an optimal way to set the D value, mainly because the amount of food I add to the water bath is small compared to the 10 liter mass of water, and things stabilize fairly quickly. But my plan is to take a set of "Blue Ice" bricks from the freezer one at a time, drop them in the water, plot the response, and tweak accordingly.

If Douglas Baldwin or other mathematician or process control engineer can suggest improvements to this procedure, I would certainly welcome their insights.

Once we get some better data for a number of different cookers, Frank has promised to post the results on his web site.

BTW, I assume, with no personal experience to back me up, that using a circulator, an electric turkey fryer, or similar device that has the heating element directly in the water along with a circulator would simplify this process, except perhaps for optimizing the response if you drop in some colder food all at once.

Note added: I've just been informed that the topic re sous vide thermometer accuracy has been relegated to the Kitchen Equipment section, at http://forums.egullet.org/index.php?showtopic=123336. (Why, I don't know.) In addition to tuning your PID controller, you really need to confirm the accuracy of the controller, so please reference that thread.

Bob

Edited April 24, 2009 by Robert Jueneman (log)

blackp · April 25, 2009

Thanks for your advice Bob.

BTW if this should be on the other thread about "Sous Vide Controller Accuracy" somebody let me know - I'm not sure of the subtlety of the rules yet.

For the first time in many years I now wish I'd listened more carefully in high school physics. I decided to "understand" the workings of a PID controller before moving forward and have managed to confuse myself. When checking on Wikipedia for information I read that increasing the "P" value will increase overshoot! Clearly this is at odds with your advice, Frank Hsu's advice and some help I've had from Nickrey.

Maybe it's a "scientific definition" versus an "implementation definition" problem complicated by my lack of understanding at this point. I guess the quickest way to resolve the issue is experimentation. Unfortunately the data-logger I've ordered is out of stock in Australia until mid May so the tool which would take the subjectivity out of any experiment is not available to me for a few weeks. (BTW Bob the logger I've ordered is USB - I don't have any serial port equipped computers any more either. It can keep 32,000 observations - it uses a K type thermocouple).

What I understand about the behaviour of the SVM is that increasing the "P" value widens the temperature range in which less-than-full power is applied. So setting "P" to 100 means that at less than 10C below the target (T) the unit will give full power and between Tminus10 and T the amount of power applied will be proportionately less. If I made the "P" value (say) 200 the full power would be attenuated at Tminus20 and proportionately decreased to zero at the target (T).

My assumption is that starting the temp reduction earlier would have 2 outcomes:

1. Time to reach target would be longer

2. Overrun would be less.

Maybe the scientists among us can fill me in - I'm guessing that because the SVM/RiceCooker is only applying heat (or not) the whole PID theory doesn't apply. We are only heating or not - we are not cooling to counteract overrun unlike other systems which I imagine can use PID controllers to correct in both directions.

I am determined to get to a reasonable understanding of this - and when I do I will post it in simple terms - or at least in terms which I can understand.

Cheers,

Peter.

PS - Next SV experiment is beef short ribs done for 48 hours - not sure if they are the same as what is called beef short ribs in the US - our butchers don't only cut things up differently but call the resulting products by different names too. As if it wasn't confusing enough!

PB

dougal · April 25, 2009

Actually, most PID '3-term' controllers only control from one side -- like only applying heat, not cooling as well.

Don't know whether you've done this, but autotune stands more chance of getting things close if the temperature is already pretty stable, and pretty close to target.

So, fill your bath with 2/3 boiling water, 1/3 cold and leave it for 10 minutes to warm the bath, (cooling the water) and even out (just stir it occasionally) - and then try the autotune with a set point of very close to the actual temperature at that time.

If you want to fiddle with the P, I and D values, then it should make sense to start from the values that the 'almost there' autotune gives you.