Sous Vide: Recipes, Techniques & Equipment (Part 6)

[jk1002]

This is Monkfish done at 60.5c for an hour, poached in olive oil. Then finished off under a broiler for 1 or 2 minutes to get some color on it.

I did brine for 20 minutes in 10%, I am guessing this wasn't enough since the fish leaked a lot of protein

I know monkfish can be tough, when done right it should have a texture like lobster. This one here was not tough at all, more flaky and not at all what Lobster is like. For a normal flaky fish I thought the thinner parts were slightly overdone, the thicker pieces were perfect - I was just expecting the Lobster texture which wasn't there. In fact, texture wise it was similar to the salmon I did few days back.

I think for fish with no skin the torch is definitely the way to go to get some color on, I don't think I would do the broiler again just to have more control.

I know fish at an hour is long but I am extra careful and follow the Baldwin tables.

P.s. Note I know Monkfish is on the Monterey Aquarium list, Wholefoods was selling it though and my local Monger had some as well so I picked it up.

Edited January 31, 2010 by jk1002 (log)

[Robert Jueneman]

In preparation for a Sous-Vide class I taught last July, I calibrated 12 of the 1500C Sous Vide Magic controllers at 100°F, 130°F, 160°F, and 190°F, using a NIST-certified thermometer certified accurate to 0.1°F. All twelve were within 1°F at temperatures between 100°F and 160°F (where it really matters). Nine were perfect throughout the range (within the 1°F precision of the 1500C), and three were 1°F high at 190°F. But if you are cooking veggies at 190°F, I doubt that 1 degree makes that much difference!

I am presently negotiating with Frank Hsu of Fresh Meals Solutions to be the US distributor of the Sous Vide Magic, mainly to decrease the shipping time from Canada. If that comes to pass, I would be willing to offer one-point and four-point calibration for new 1500D units, prior to shipment.

In any case, I am willing to offer such a calibration service for eGullet subscribers for after-market units, including previous versions of the SVM and other PID controllers, as well as for various thermometers. This will include a signed calibration sheet, and a calibration sticker for the unit.

To date, we do not have enough long-term experience to say anything definitively about the extent or cause of calibration drift. In some of the beta testing we have done, we have seen some drift that was apparently caused by the sensor, and may have been the result of continuous immersion and/or other problems, but it is at least possible that the reference circuitry in the controller could drift. But in either case, if you are cooking sous vide at temperatures around 55°F/131°F for more than four hours, you really want to be certain that your thermometer/sensor is accurate -- at least within 1°F. Funeral expense cost a lot more than a thermometer calibration!

My recommendation, therefore, would be to calibrate any PID controller and associated sensor as a matched set, and not attempt to mix and match units, or calibrate only the sensor. Until we have more data points, I would recommend a yearly recalibration. (Note to anyone who buys the Sous Vide Supreme -- save the box! And I'd be curious to know what the instruction manual says about calibration, and whether the PID parameters can be tuned.)

If anyone is interested in this service, contact me by PM with the details of what you need, and I'll try to figure out a reasonable price and turn-around time.

Just a couple of other points. I have NOT found the digital fever thermometers to be acceptably accurate, either for use as a secondary thermometer, or even as a fever thermometer. I would recommend the non-mercury analog basal thermometer made by Geratherm in Germany. It used to be offered by Longs Drug Store, but since they were acquired by CVS, I don't know whether they still distribute it.

The Germans and the Swiss seem to know how to make accurate thermometers -- the Chinese and the US manufacturers don't seem to have adequate quality control. You get what you pay for, but then only if you are lucky.

Here are my current recommendations, but with the caveat that just because one unit worked well for me (or didn't), a different unit might behave differently:

Brand & Model Number Error at 131°F Comments

Control Company Traceable 4000 ±0.05°C at 60.002°C My primary reference thermometer, traceable to NIST standards. Precision 0.001°, accurate within ±0.05°C from 0°C to 100°C. Recalibration due 7/31/2010. Very Highly Recommended.

Geratherm basal thermometer ±0.00°F at 100°F My secondary reference. Liquid, non-mercury. Highly Recommended.

Component Design Northwest Q2-450 ±0.0°F Handheld probe with lanyard, ice-bath calibration option. Highly Recommended.

Sur La Table TSP572 -0.1°F Dishwasher safe probe, timer, ice-bath calibration capability. made by Component Design Northwest. First probe had to be replaced because of inaccuracy. Recommended after calibration.

All-Clad +0.2°F Meat preference alarm, ice-bath calibration capability. Recommended.

Taylor Tru-Temp 3518 +0.6°F Digital “ice-pick” pocket probe. Acceptable.

Taylor 9840 +0.6°F° Digital “ice-pick” pocket probe. Acceptable.

Engle +0.6°F Probe,

+1.2°F Ambient, at 75°F Refrigerator thermometer with 10-foot cord. Displays minimum, maximum, current, and ambient temp. Acceptable.

CVS KD192 -0.4°F at 99.88°F Digital Fever Thermometer. Not Recommended as a secondary standard.

Taylor 1478-21 oven thermometer +1°F Only reads to 1°F. First unit was off by 5°F. Not Recommended.

No name “Oil and Candy Thermometer” +2.3F Long clip-on probe, settings for candy and frying. Not Recommended.

Bob

[e_monster]

Robert,

When you say "I have NOT found the digital fever thermometers to be acceptably accurate", what do you mean by "acceptable". What sort of disparity (in quantifiable terms) did you find? I'd be curious to know the range of error you found -- and to know more about how many brands you tried.

What was the standard that you were measuring against?

[Robert Jueneman]

I tried two. One (listed in the chart) was off by 0.4F, the other (whose name I don't recall, was worse. I haven't tried dozen of different digital fever thermometers, obviously, either one brand or multiples. I'll leave that up to Cooks magazine, or Consumer Reports. I would NOT levee it up to NSF, as their standards are too loose, IMHO.

Now, 0.4F would be acceptable for sous vide, but not for a secondary standard, and not for a fever thermometer, I believe, especially since it cost a dollar or two more than the Geratherm, which was accurate to within 0.00F, i.e., as close as I could read it while interpolating between 0.1F markings. This was checked against my NIST-certified Traceable 4000 reference thermometer.

[e_monster]

While that amount of error obviously isn't optimal, it seems reasonable for our purposes. An error of .4F seems reasonable to me for setting the offset for sous-vide cooking. Even if the error had been .6F that would be reasonable though obviously smaller error would be nice

So, the blanket statement that digital fever thermometers are not acceptable for setting the offset does not seem warranted -- unless someone is cooking things were .6F is going to make an important difference. A sample size of two is hardly sufficient to condemn all digital fever thermometers -- especially when we don't know the brand or reputation of the one whose error was great than .4F.

Glass fever or basal thermometers are probably generally more accurate and less error prone than inexpensive digital fever or basal thermometers, but I think all but the poorest quality will be within .5F -- and the more reliable digital fever thermometers will be more on the order of .2 to .4F.

The more important question is whether our PIDs have a predictable or unpredictable variation regarding offset in relation to absolute temperature.

If the variation is predictable, it would be fairly simple to come up with a a formula or graph that would allow one to map from the displayed temperature to the actual temperature within a reasonable margin of error. This would obviate the need to send out the unit to someone to have it calibrated or re-calibrated.

[PedroG]

I agree with e_monster that digital ovulation thermometers are not generally unacceptable as a secondary standard. I have a "Cyclotest Lady" by UEBE Medical GmbH in Wertheim, Germany. It has a resolution of 0.01°C and is specified to be accurate to ±0.10°C. In fact, its offset against my ISO-calibrated Greisinger GMH3710 high precision thermometer was -0.06°C at 33°C and +0.01°C at 42°C, see http://www.mydrive.ch/download/50592073/Calibration_GMH3710_SVM1500A_testo_Mastrad_GourmetCheck.xls

By extrapolating from 33-42°C to 55°C, we might risk to be 0.5°C off, but this is still better than no calibration at all and just trusting any junk thermometer, and regular checking of the SVM sensor against an ovulation thermometer may indicate a beginning shift in the sensor (some sensors do not withstand permanent immersion for many months or years).

BTW in the ominous 54.4°C safety limit, there is already a safety margin of at least 1.4°C, see http://www.hi-tm.com/RFA/food-path-summ.pdf

So longtime cooking at nominal 55°C should be safe even with a 1.5°C deviation of the sensor.

Pedro

[Robert Jueneman]

I'm not trying to make this into a high-precision physics experiment, and I agree that a certain amount of error is probably tolerable. My comment about digital fever thermometers was perhaps not very artfully worded, and reflected my disappointment. I should not have condemned the entire industry, nor specific countries, without testing more products. Nonetheless, when I buy a clinical thermometer for my own health, I expect 0.1F accuracy.

Since these thermometers only cost about $20, it would be reasonable to buy two different brands from two different sources. If they agree, they are probably correct. If not, buy a third one, and either average them if they are all reasonably close, or return one of them.

What we really want to do is catch any significant drift away from a calibration point, as that would indicate an incipient failure.

A completely different issue is the linearity of the probe/controller, and that is something that is not going to be easily measurable by the average user without a laboratory grade reference. An accurate ice-bath calibration isn't all that easy to do, and a boiling point calibration can be terribly inaccurate, depending on the weather, elevation, etc. So an inexpensive ovulation thermometer is about the only option, and excellent starting point.

Now, if you have or buy a more general purpose kitchen thermometer, AND it agrees quite closely with your secondary reference at say 100F, then you can use that as a working thermometer, and double-check the linearity of the PID controller. Again, if they match, then presumably all is well, and if not, well, buy another one!

As we gain more experience in the characteristics of particular sensors, it ought to be a simple matter to incorporate a firmware correction for any first-order nonlinearity, if any is found.

[PedroG]

[quote name=PedroG' date='31 January 2010 - 12:37 AM' timestamp='1264894657'

Interesting article. Thanks for the link. If I read the graph correctly, at temps above 55C the maximum error is less than 1C and doesn't happen until 90C. Yes, and fortunately all sensors had minimal deviation at 55°C, as you can see in my original spreadsheet http://www.mydrive.ch/download/50434411/1500D_sensor%20calibration_on-the-fly_25-90%C2%B0C.xls

[PedroG]

I saw a couple of posts regarding this upthread, but has anyone else noticed that when you SV beef for 72 hours @ ~132F with no other flavourings, it gets quite an odd smell/taste?

Can you describe this "odd smell/taste" more precisely? Was it still pink? Was it falling-apart?

[dougal]

Can I remark that these are somewhat extreme conditions.

The lack of response indicates that this is not a usual situation.

72 hours is a "very long time" even for sv.

It is possible to overcook with sv!

132F is 55.5C, which is rather close to the edge of safety.

And that in turn raises the question of whether this meat has really had 132F, or whether that is merely what an uncalibrated controller is reporting, from a different part of the waterbath, with an incompletely evacuated pouch floating horizontally ... you get the picture?

If bits of the meat have had less than 132F, and/or there was a significant amount of air in the pouch (I'll leave it for others to define 'significant amount'), then there is definite potential for some undesirable microbiology.

Now put the two things together and the longer time you give it, the more that microbiology will proceed. It might be merely spoilage, but it could well be much worse.

So, I'd simply suggest "Don't go there" -- especially as you mention approximately 132F ("~132F").

There is an edge.

The closer to that edge that you want to go, the better equipped you need to be -- see the calibration discussion above. Thinking about that "~" approximately -- it makes sense not to go nearer the edge than a pessimistic (ie safe) estimation of the errors in your temperature control system.

But, anyway, I don't think you should need to go 72 hours ...

If there is ONE THING that non-tech readers of this thread should take away from the foregoing calibration discussion, it is that just because the digital readout says 132F, you cannot assume that your food is getting 132.000°F. Its like oven thermostats (or anything else really) - there's a margin of error.

And its important to understand what it might be -- ideally what it actually is -- for your own equipment.

Its a (common but fundamental) mistake to confuse display precision with accuracy. A reading of 132 on your display does not (by itself) guarantee that the temperature at the probe (let alone the whole bath) is truly between 131.5 and 132.5.

Regarding calibration standards, I note that the current Thermapen offering is supplied with a traceable-calibration certificate for that specific instrument complete with its probe. Resolution is 0.1F while the accuracy is specified as better than ±0.7F over its entire range.

Given that it can be field-recalibrated at one temperature (if required), I think it sounds like a reasonable, and highly convenient, secondary (working) standard for the very serious/pro, and a decent primary standard (and all round kitchen workhorse) for 'the rest of us'. http://www.thermoworks.com/products/thermapen/splashproof_thermapen.html

Another reason I'm going to HAVE to get one, dammit. (The current model can now do C or F, is even faster and splashproof too. Dammit.)

[Robert Jueneman]

I had that problem once (or was it twice?) with a flat iron steak, which I think I cooked of 48 hours.

I've done short ribs and brisket for 72 hours with minimal seasoning (dry rub only), with no off taste.

What kind of beef was it, where did you get it, has it happened more than once, yada, yada, yada.

[FoodMan]

In Heston Blumenthal's Fat Duck book, he devotes a few pages talking about meat cookery and Sous Vide specifically. He recommends dunking the vac packed meat in an 80 or 85 C water pot for about 30 seconds before cooking sous vide at a much lower temp (say 55 C). This he recommends is done to kill off any surface bacteria, specifically if memory serves Lactobacillus type, that could cause off flavor and odor. I've been doing that and have not had a problem with any long cooked piece of meat. So, the odd smell could be due to a surface bacteria that was present.

Edited August 14, 2010 by LindaK (log)

[Robert Jueneman]

That makes good sense.

Long before the 72 hours was up, all of the bacteria should have been killed, but that doesn't necessarily mean that in the first several hours they couldn't have been working, and perhaps producing some lactic acid. The lactic acid, in turn (which presumably WON'T go away), could then be breaking down the meat over that extended amount of time. Biologically, the meat is presumably safe to eat, but it might not be very pleasant.

Now, having to have another pot full of water at 80 or 85C is a bit of a nuisance, and if you want to go from taking a package out of the freezer and thawing it in the sous vide bath as I do, it might not get hot enough.

In reading some preliminary drafts of Douglas Baldwin's book I see that he is sometimes pre-searing beef before packaging it, and then post-searing -- either with a torch or a hot skillet. I haven't tried that method, but presumably it would accomplish the same thing -- killing any surface bugs.

You could try that and see if it helps, but then again, the next piece of meat you buy might not have any bacteria on it, so it might not be conclusive. Isn't science fun?

[FoodMan]

Exactly. The bacteria might still be active for the first hour or more at 132 and causing lactic acid. It really is not much of a nuisance to heat a small pot of water to dunk the meat in though. Honestly, it's less of a hassle than pan searing IMO. Although in Douglas' method, I assumed the pan searing is also to contribute some flavor not just for safety or killing off bacteria.

As far as frozen meat, I would think that meat would be bacteria free anyways, so the need to "clean the surface" is not that much of a necessity. I would do it with fresh-just bought or maybe meat that has been thawed already.

[e_monster]

Freezing does not make meat pathogen-free. Bacteria and mold spore are in suspended animation for the most part. Some parasites are killed (which is why when you cook salmon at low temperature you want to use salmon that was frozen and thawed).

You should consider frozen meat to be just as likely to be contaminated as unfrozen meat.

[MartinH]

In Heston Blumenthal's Fat Duck book, he devotes a few pages talking about meat cookery and Sous Vide specifically. He recommends dunking the vac packed meat in an 80 or 85 C water pot for about 30 seconds before cooking sous vide at a much lower temp (say 55 C). This he recommends is done to kill off any surface bacteria

I wouldn't trust a 30-second dunking to pasteurize the meat. Not long enough. Actually, HB says something a bit different. What HB says is:

"Sometimes with red meat, when it's going to be cooked at 56C, it's necessary to sear or grill the meat before vacuum-packing it, and plunge the bag in water above 80C for three to ten minutes, according to thickness, to kill lactobacteria on the surface, which will otherwise affect the flavour." (The Fat Duck Cookbook p. 429)

But I'm doubtful of HB's suggestion. First, if the aim is to kill surface bacteria, there's no need to vary the time from 3-10 mins according to thickness. Thickness has no bearing on the matter of killing stuff on the surface. Second, to first sear/grill the meat and then give it a 3-10 minute hot bath seems like overkill. Has anyone tried this double surface-heating method? Surely one or the other would suffice. Third, HB doesn't say exactly when the "sometimes" are that one needs to do this. How is one supposed to know when to do this and when not to? Fourth, I'm not sure how much or what flavours lactobacteria can leave behind during a straight 56C SV with no pre-searing. Fifth, why just red meat and why not say chicken too?

[FoodMan]

Sorry, I did not have the book in front of me and I must've misspoke as to timing. not sure where I got the 30 seconds from! The goal here as I understand it is not to pasteurize the meat, but to just kill surface bacteria. 3-10 minutes should be more than enough. As to the difference in timing, It might depend on the temperature of the meat now, wouldn't it?

Chicken cooks at a much higher temp than red meat, so the bacteria issue might not be as prevalent, but a dunk might not hurt either. I usually go for the shorter dunk not the 10 minutes. The practice makes sense to me and figured it might help the original poster with the unexplainable weird smell.

[Markdmd]

Actually cooking pasta sous vide would be a bad idea, because the pasta will continue to absorb water so long as there is water in the bag.

←

Ya, thats why I want to measure the amount of water it absorbs when cooked right, then only put that much in the bag. Instead of the normal bloating of the bag from escaped juices, I imagine the pasta would swell to fill it. Probably have to lay the pasta flat to try and keep the water absorption even. If it works, it would mean I could cook pasta in more expensive liquids than I would otherwise (not wanting to turn a couple gallons of duck-stock or Sauternes into pasta water.)

←

100% semoline dough absorbs about 80%-85% water to obtain an al-dente texture.

When you take into account the regular ratio of dough/filling in a raviolo, the absorption is about 35%-40% (YMMV).

However, you would have to check what the minimum temperature is for the semoline starch to gel, which I would guess is in the 60-70 C ballpark. Given that the temperature you're looking for the yolk is closer to the lower bound of this ballpark figure, I'm not sure this is the best approach for what you're trying to obtain.

You should also take into account that the pasta will stick together, that the yolk can break when you seal the bag, etc.

Freezing the yolk and then filling the raviolo might be a simpler low-tech solution with better results, IMHO.

Is there a followup somewhere on this topic? Specifically cooking pasta sous vide, the temperature required and ratio of liquid for al dente from dry pasta? What if the pasta is cooked directly in sauce rather than water? What if I filled my water oven with sauce rather than water then cooked the pasta directly in the sauce without vacuum? I'm thinking for a really light fluffy ricotta gnocchi/gnudi dough?

[KennethT]

The guys at Cooking Issues at the FCI are putting together a sous vide/low temp primer... you can find it here:

http://cookingissues.wordpress.com/2010/02/02/coming-soon-sous-vide-and-low-temp-primer/

Looks pretty good and informative so far!

[casquette]

Bruno Gousseault one of the fathers of sous vide cooking recommend browing before for a couple of reasons; Browning on high heat will help eliminate surface bacteria, and in the other hand, because sous vide keeps and amplifies flavors your food will be infuse with the flavors created by the Maillard reaction, which is pretty cool. Just make sure that when browning the core temperature doesn't go above 30 degrees Celcuis. After cooking you sear it again for service.

Cheers

n.

[e_monster]

I realize that conventional wisdom is that pre-searing before bagging amplifies these flavors. But my experience has been that pre-searing has minimal effect if you post-sear. I did a number of blind-tastings of short-ribs and steak where some samples were pre-seared and some were pre and post seared and some only post-seared. I repeated these a few times and consistently post-seared only was preferred.

Your mileage may vary or tastes be different, but I have noticed that other people that did blind comparisons found the same thing.

[casquette]

I'll do the test tomorrow and post my result. I have to say that because it needs to be chilled back after searing before vacuum packing I always post sear. But now I'm interested to find out, I'll write tomorrow after my test.

N.

[jk1002]

For the pasta SV question - I recall that Alan Ducasse is / was using an ancient method where the pasta would be cooked in little water which later would become its sauce. Alessi has a pot for that ....

Alessi

There is a also a Bittman article video about this, essentially it is cooking pasta Risotto style.

Bittman

I would think with SV you may loose some of the creaminess, it may become unpleasant if you run it too dry as I would think the pasta still leaches some starch. I would also think you do need boiling water, I had pasta once that was "soaked" in hot but not boiling water - It had a funny texture.

[howsmatt]

I think the key point to take away from Heston is that you should "pasteurize" the bags as well as the meat.

[PedroG]

"The time has come," the Walrus said,

"To talk of many things:

Of germs - and molds - and botulism -

Of meat and fish - and pasteurizings -

And why the searing is so hot -

And what the PID-controller brings."

Most of the posts from #3210 to #3250 deal with food safety aspects, temperature/time combinations, and accuracy of temperature measurement and stability. So it seems justified to compile a few data.

They may be found in http://www.hi-tm.com/homeprep/Home-2006-2col-forpdf.pdf and http://www.hi-tm.com/RFA/food-path-summ.pdf and http://amath.colorado.edu/~baldwind/sous-vide.html

Food safety and spoilage:

Meat and fish is safe below 29.3°F/–1.5°C (Listeria monocytogenes is the limiting pathogenic microorganism).

However, spoilage bacteria begin to multiply at 23°F/–5°C. So with improper refrigeration, spoilage will begin before pathogen multiplication.

Food can be safely held at or above 126.1°F/52.3°C (Clostridium perfringens is the limiting pathogenic microorganism). The upper limit of the “danger zone” has arbitrarily been set to 130°F/54.4°C in the FDA Food Code. After cooking, the food must be rapidly chilled in ice water (crushed ice/water 1:1) to prevent spores of Clostridium perfringens from becoming active bacteria and producing toxins.

If the food is safe to eat raw and you are not pasteurizing it, then the FDA Food Code requires that it is between 40°F/ 4.4°C and 130°F/54.4°C for less than 4 hours.

Surface pasteurization by dunking or pre-searing:

A 6.5D reduction of Salmonella takes 85 seconds at 149°F/65°C, 11 seconds at 158°F/70°C, 1.4 seconds at 167°F/75°C, and 0.17 seconds at 176°F/80°C. So when dunking, add 30 seconds for heat penetration through the bag and a 1mm surface layer, and you are on the safe side.

Pre-searing would theoretically sterilize the surface, but handling during cooling and bagging before vacuum-sealing may introduce new germs.

Importance of accurate temperature control:

Pasteurization (a 6.5D reduction of Salmonella) takes 110 min at 54.5°C or 89 min at 55.0°C, so when relying on the pasteurization times looked up in Douglas Baldwin’s tables (http://amath.colorado.edu/~baldwind/sous-vide.html ), a temperature variation of ±0.25°C makes a significant difference, although this would rather not influence the organoleptic outcome. However, in Douglas’ tables there is already a safety margin.

If a 0.1°C error of the SousVideMagic’s sensor adds up with a ±0.1°C oscillation of water bath temperature, this may already result in a 0.2°C error. When using a SousVideSupreme with a specified “sensitivity” of ±0.5°C (accuracy is not specified), safety margins have to be accordingly larger.

Whatever your SV equipment is, a calibrated reference thermometer is highly recommendable, and quoting Bob Jueneman (post #3232 ), funeral expense cost a lot more than a thermometer calibration!

Thanks a lot to Douglas Baldwin for proofreading this post and making corrections in the “food safety and spoilage” section.

We are looking forward to further clarification by Douglas Baldwin.

Pedro