Sous Vide: Recipes, Techniques & Equipment, 2011

[rotuts]

is that 'Good Goaty' or ' goaty'

and the pics .....

just a joke!

too bad it seems that only in TX you can get <1 Y.O. goat.

thats said to be very very tasteeee.

and the Tamarind sauce?

[jmolinari]

It was, to me and my daughter, good goaty, to my wife, not so good goaty:) But then again she doesn't care much for lamb either.

The tamarind glaze was good, not great.

[rotuts]

Sooooooo many thanks for the pic.

many people would like lamb is there was no fat. its the lamb fat that bothers many.

me somewhat.

id be very interested in your thoughts on the tamarind not being perfect:

tamarind aside, what elements do you think that are not in that glaze would make it Mighty Fine?

if you do this again, Id think Lime.

[jmolinari]

I think the glaze needed more heat (but then my 2.5 year old might have had an issue), as well as more sour. So more Sriracha and some lime would have been good.

[nickrey]

The problem with many souring agents is that they are delicate and their effect is destroyed by heat. for this reason, vinegar or lime juice are best added at the end of cooking, I'd thus be tempted to give it a lift at the end (perhaps tamarind in the marinade/glaze and some lime squeezed over for serving). I always find it handy to keep the Asian food mantra of Sweet Sour Salty Hot in mind. If you balance these, you are less likely to see deficiencies in the finished dish.

[Mikels]

When I looked into SV, one of the requirements was that I could cook for up to 20 people without breaking the bank - the answer was to put together my own. For three years I have been using a 32gal system with a immersion heater and a PID controller. Not only does it cook more food, but when you put the food in, there is almost no drop in temperature when food is placed in it. But, it takes up space. I have not had any problem with either the 1000 watt immersion heater or the PID controller.

A while ago, there was a discussion of the food safety of plastic bags. I am under the impression that many think thicker is better. I don't think this is correct. The thicker the bag, the more chemical chemical can leach out; if I am correct the increased chemical is lineally proportional to the thickness. More importantly, the chemicals used to make the bag and its construction are the key factors.

[vengroff]

Couple more things i would like:

1) Would love to see a "pasteurization" option step. This could be a switch that would add a pasteurization time once the core temp is reached. Time would be based on final item temperature. I realize this might be a bit complicated to add for beef, chicken and fish since the times are difference, but starting with chicken would be great, since that's the one that we all pasteurize and the one that really NEEDS to be cooked to pasteurization. without that option i'm stuck having to remember to add it to the end of the timer and look up how long pasteurization is at each temp.

...

Love the app!

I'm in the process of adding this, and in light of recent discussions in this thread I'd like to ask for some feedback. Specifically, would you be interested in the time required to pasteurize the surface, the core, or an option for one or the other? I've put together the basic code for both, and the results are pretty interesting, especially for large cuts of meat.

For example, here is a scenario where we are cooking a 50mm steak to 54.4°C in 55.4°C water. This is basically what the app does today.

If we look at what it takes to pasteurize the surface, we see that the cooking time does not change at all. The surface has a log reduction of E. Coli of 5 (99.999% reduction) before the core even reaches the final temperature. This is shown be the dashed red line. The other two pathogens have been reduced off the chart by this time. Note that up means more reduction, i.e. less of the pathogen still alive.

If we want to pasteurize all the way to the core because we have reason to believe there are internal pathogens, then we haven't done the job. According to the solid red line, we've only got an E. Coli log reduction of 1.77, or about 98%. If we want to pasteurize all the way to the core, we need more time, as shown below.

We finally got a log reduction of 5 for E. Coli (the solid red line) over an hour after we had it at the surface. The other pathogens have log reductions off the chart by this time.

So back to the original question. Is this the level of detail and the kind of options you would like to see in the app? Is the surface-only option useful--i.e. do you trust your rancher and butcher that much? Or is core pasteurization what most people want?

Finally, I'm currently going with 5 LR (100,000 to 1) for E. Coli, 6 (1,000,000 to 1) for Listeria, and 6.48 (3,000,000 to 1) for Salmonella. These are the same numbers Baldwin uses, but I'd appreciate any feedback on them.

Thanks in advance.

[mgaretz]

Couple more things i would like:

1) Would love to see a "pasteurization" option step. This could be a switch that would add a pasteurization time once the core temp is reached. Time would be based on final item temperature. I realize this might be a bit complicated to add for beef, chicken and fish since the times are difference, but starting with chicken would be great, since that's the one that we all pasteurize and the one that really NEEDS to be cooked to pasteurization. without that option i'm stuck having to remember to add it to the end of the timer and look up how long pasteurization is at each temp.

...

Love the app!

I'm in the process of adding this, and in light of recent discussions in this thread I'd like to ask for some feedback. Specifically, would you be interested in the time required to pasteurize the surface, the core, or an option for one or the other? I've put together the basic code for both, and the results are pretty interesting, especially for large cuts of meat.

For example, here is a scenario where we are cooking a 50mm steak to 54.4°C in 55.4°C water. This is basically what the app does today.

w

So back to the original question. Is this the level of detail and the kind of options you would like to see in the app? Is the surface-only option useful--i.e. do you trust your rancher and butcher that much? Or is core pasteurization what most people want?

Finally, I'm currently going with 5 LR (100,000 to 1) for E. Coli, 6 (1,000,000 to 1) for Listeria, and 6.48 (3,000,000 to 1) for Salmonella. These are the same numbers Baldwin uses, but I'd appreciate any feedback on them.

Thanks in advance.

Both, please. But what I would really like is recommended cooking times for cooking effect. My roast may be done and pasteurized in 4:36, but would it better cooked for 24-48? I know this is no longer mathematics, but if I'm going to have a reference, then I shouldn't have to look somewhere else to find out how long I should cook something vs. how long I must cook it.

Edited December 3, 2011 by mgaretz (log)

[pep.]

So back to the original question. Is this the level of detail and the kind of options you would like to see in the app? Is the surface-only option useful--i.e. do you trust your rancher and butcher that much? Or is core pasteurization what most people want?

Definitely both, please. And while you are at it, could you please add temperatures below 0 °C to the starting temperature selection? I like to cook stuff directly from the freezer and while the temperature curve doesn't change much, it is a concern with relatively short term cooking. I think going down to -30 °C should cover most non-laboratory freezers.

Anyway, thanks for such a great SV tool - it is the first SV app that is really useful! Great work!

[jmolinari]

I'd say at LEAST core, if not both. But i would also suggest making the pasteurization time a separate counter that starts once the main counter reaches its end. That way i can see how long to get to core and then how long to pasturize. I'd keep them almost as independent timers. Am I explaining that clearly?

[vengroff]

I'd say at LEAST core, if not both. But i would also suggest making the pasteurization time a separate counter that starts once the main counter reaches its end. That way i can see how long to get to core and then how long to pasturize. I'd keep them almost as independent timers. Am I explaining that clearly?

I think I understand. Basically you would like the app to tell you something like, "the core temperature goal will be reached in 45 min, then pasteurization will start and it will be complete one hour after that." So you'd like to see a 45 min timer start, then when it reached zero a 60 min pasteurization timer would start, and when the second timer reached zero the food would be ready.

Assuming this is correct the issue I have with it is that pasteurization starts before you reach the target temperature. It is slower, but it is already happening even before core temperature is reached. You can see this in the pasteurization graphs I posted above. 2-1/2 hours into the cooking process, we are still an hour short of reaching core temperature, but we are at a high enough temperature that pathogens are being reduced. This is shown by the positive slope (up and to the right) of the various colored log reduction curves. As we get closer to and eventually reach our target temperature, the slope gets even steeper, meaning pathogens are dying off even more rapidly, but the process was already underway long before this.

Because pasteurization is a continuous process that starts slowly but well before we reach target temperature I think it might be a little misleading to have the two timers. That would suggest that pasteurization does not begin until we reach the target temperature, which is not the case.

[jmolinari]

Understood. Isn't that math already accounted for in the tables for pasteurization available in MC and Douglas' book and the FDA tables? I thought the way you were going to implement it was much simpler. Just a switch, and a lookup table for the pasteurization time required for each temperature. That time would be added as a secondary timer after the 1st one is complete.

At least, that's how we do it now, is it not? We look at the item thickness, calculate time to core temp, look up pasturization time at that temp, and add it to the "time to core temp".

At least, that's how i've been doing it:)

[Robert Jueneman]

So back to the original question. Is this the level of detail and the kind of options you would like to see in the app? Is the surface-only option useful--i.e. do you trust your rancher and butcher that much? Or is core pasteurization what most people want?

Definitely both, please. And while you are at it, could you please add temperatures below 0 °C to the starting temperature selection? I like to cook stuff directly from the freezer and while the temperature curve doesn't change much, it is a concern with relatively short term cooking. I think going down to -30 °C should cover most non-laboratory freezers.

Anyway, thanks for such a great SV tool - it is the first SV app that is really useful! Great work!

This seems like such a great tool --- any chance of porting to a Droid? Pretty please?

From my perspective, I think it depends on what you are cooking. When I cook a rib-eye, I cook it at 51C, so pasteurization isn't an option. But when I cook a chuck steak (24 hours) or brisket (72 hours) at 55C, pasteurization is guaranteed, so it doesn't matter. Similarly, fish is never going to be cooked to pasteurization, at least by most people.

So that leaves poultry, and there pasteurization definitely is an issue. Now, I don't know of anyone Jaccarding poultry, but the process of injecting brine seems somewhat suspect, and many chicken parts have crevices that would suggest that the core temperature is what matters.

I certainly agree with adding the low temperature starting point -- almost of the time I cook or rewarm from frozen, i.e., -10 to -20C.

But now, how about adding another very important feature -- chill time?

Now this is an area where the surface vs. core temperature is very much an issue, and still somewhat uncertain, at least in my mind.

Is it sufficient to chill the surface to say 4C, and then put it (bagged) in the refrigerator, even though the core would still be warmer? Assuming you aren't chilling a 100mm thick roast, how much warmer would the core be, and then how much longer would it take it in air to chill down the rest of the way? According to the FDA, it takes DAYS for food pathogens to grow to a dangerous level at 4.4C, so it doesn't seem likely that the time required for the core to be chilled vs. the surface is likely to be that significant.

This is an issue that PedroG and I have been discussing offline, but with another twist.

I have been doing something I haven't seen anyone else mention. I keep a couple of 1.5L bottles of cheap 80 proof vodka in the freezer at -20C. Then when I want to chill something quickly I pour the chilled vodka into a metal pan of a suitable size, put it on top of my Anti-Griddle (which eventually gets down to -36C), and submerge the bagged food in the vodka. I don't have to fuss with a container with lots of ice (a large bag of ice takes up more room in the freezer than the vodka does), and the vodka is reusable, unlike the ice.

In addition to using this for meats that are prepared cook/chill, I also use to save any leftovers, in case my wife or I don't eat all of a nice steak. (In that case, since someone likely stuck a fork into the meat that had just been in their mouth, pasteurization certainly seems like a good idea when reheating it, although I doubt anyone normally does that.)

I tried this technique on a 50mm potato slab, and starting at around 22C with the vodka at -16.5C, the core temperature dropped to 5C in 20 minutes, but it took a full hour to reach 0C. I'm going to try it again, and this time monitor the temperature just below the surface as well.

Now, Pedro has questioned whether I really want to freeze the surface in order to cool the core to a safe level, and he fears that freezing and then thawing slowly (in the refrigerator) might degrade the quality of the meat. I can see that it MIGHT, but I haven't done the blind tests necessary to convince me one way or the other. And there is still the issue of whether it is sufficient to merely cool the surface to say 4C, or even 0C, before putting it in the refrigerator.

Now, if I cook a 50mm steak from frozen, Baldwin's latest tables predict that it will take 5:15 to come up to the core temperature. (His previous tables suggested 3:45, so this is something that needs to be confirmed experimentally -- 5:15 seems pretty long, to me.)

By comparison, the heating time from 5C is only 3:30, so it is apparently taking 1:45 to thaw the frozen meat from -18C to 5C.

Now, 3:30 is within the four hour rule for non pasteurized meat above 5C. However, the chill time for the same 50mm steak in an ice bath (from 55-80C down to 5C -- presumably the core temperature) is 2:45. But 3:30 (cook time) plus 2:45 (chill time) is 6:15, and that exceeds the generally accepted standard of 6 hours for cook/chill.

On the other hand, using the chilled vodka technique, the chill time to a core temperature of 5C would only be 20 minutes, or less than 4 hours total above 5C.

In the case of cook/eat/chill leftovers, assuming it takes less than 30 minutes to eat dinner, and maybe another 30 minutes to clean up and set up the vodka, I'm still less than five hours above 5C, compared to the seven plus hours using an ice bath.

Does this make sense to people?

What is the general consensus -- would freezing the surface of the meat in order to chill the core to 5C be likely to cause an appreciable degradation in quality?

[Anna N]

Devastated to discover it will not work on my i-Touch due to an older OS.

[Robert Jueneman]

Tonight I tried the Anti-Griddle/vodka chilling again, and did some measurements.

I had cooked nearly 3 kg of a chuck roast, cut into six steaks. I took three of those steaks at 55C, totaling 1.4 kg, and put them in a roasting pan along with 3L of vodka at -16C. After 30 minutes, the vodka was up to about -4C. Obviously I should have used more vodka, but I was limited in the size roasting pan I had. And the bottom of the pan wasn't entirely flat, so it was only touching the Anti-Griddle surface in one small spot. And I don't have a Styrofoam cooler large enough to cover the pan, so some heat was lost to the air. Sob!

After 30 minutes, I removed the steaks, and cut open one bag to measure the temperature. The subcutaneous surface temperature was 5.4C, and surprisingly, the core temperature was only 5.7C. That steak was about 40mm thick.

I figured that was good enough, so I put the steaks in the freezer.

I still need to chill the remaining steaks, but I'll get out another couple of cold vodka bottles for that.

Overall, I'm quite pleased with this technique.

Bob

Edited December 4, 2011 by Robert Jueneman (log)

[dcarch]

"---The tamarind glaze was good, not great."

Tamarind + 1/4 oyster sauce = great!

dcarch

[vengroff]

Understood. Isn't that math already accounted for in the tables for pasteurization available in MC and Douglas' book and the FDA tables?

By my reading, that is exactly what Baldwin does. And for the cases I've checked my numbers match his to within the 15 minute precision he provides. Douglas, if you are on this thread I'd love to hear you chime in with a confirmation.

I thought the way you were going to implement it was much simpler. Just a switch, and a lookup table for the pasteurization time required for each temperature. That time would be added as a secondary timer after the 1st one is complete.

At least, that's how we do it now, is it not? We look at the item thickness, calculate time to core temp, look up pasturization time at that temp, and add it to the "time to core temp".

In Baldwin's pasteurization tables (specifically the ones at http://www.douglasbaldwin.com/sous-vide.html) he uses the heading "Pasteurization Time from 41°F (5°C)." So the time he is giving is not the time you need to add once you reach your target temperature, but the whole time, including the time to get from 5°C to the target. So if you add this time to the time to reach temperature, then you are cooking longer than you really have to in order to pasteurize.

The reason it isn't ideal to simply add a pasteurization time once you reach temperature is that you can't know how close you are to pasteurized when you reach the target temperature without knowing what kind of partial pasteurization occurred on your way to the target temperature. That depends on the temperature curve you followed to get to the target temperature, which in turn depends on your starting temperature, among other things.

The real fundamental tenant of SousVide Dash is to try not to rely on any pre-computation but instead to do as detailed a simulation as possible for exactly the scenario the user has chosen. The app tries to simulate everything that happens temperature and pathogen-wise second by second from the moment the food enters the water bath. I went this way because you can't easily pre-compute every scenario anyone will ever want to try and store it in a table, even if the table is pretty dense and you have decent interpolation.

The desire to fully simulate the cooking process is also why I haven't yet done the cooking from frozen feature that some of you are asking for. There is a fair amount of added complexity that comes into play as ice in food melts. It absorbs a great deal of heat in doing so, in a very different way than water that in simply becoming warmer water does. Thus accurately keeping track of what is going on in a partially frozen, partially thawed piece of food gets a little more complicated.

[vengroff]

This seems like such a great tool --- any chance of porting to a Droid? Pretty please?

The number one request I get. I'm gonna have to break down and buy a droid of one form or another. Can anyone recommend a good cheap plan-free option that will run on WiFi even without a SIM? That's all I probably really need. In any case, it's a tentative 2012 project.

I certainly agree with adding the low temperature starting point -- almost of the time I cook or rewarm from frozen, i.e., -10 to -20C.

See my last post above.

But now, how about adding another very important feature -- chill time?

Wow. Now that could almost be a whole new app, especially with you fancy vodka technique. I may have to give that a try.

[DouglasBaldwin]

Understood. Isn't that math already accounted for in the tables for pasteurization available in MC and Douglas' book and the FDA tables? I thought the way you were going to implement it was much simpler. Just a switch, and a lookup table for the pasteurization time required for each temperature. That time would be added as a secondary timer after the 1st one is complete.

At least, that's how we do it now, is it not? We look at the item thickness, calculate time to core temp, look up pasturization time at that temp, and add it to the "time to core temp".

At least, that's how i've been doing it:)

vengroff is doing what I do. We compute the destruction of pathogens based on the temperature throughout the heating process. Indeed, this is why some of the heating times in Table 2.2 are longer than the pasteurization times in Table 4.1. (Though, my tables always assume that the core is being pasteurized.) For healthy people, a 3D reduction of the Salmonella species is generally considered sufficient and that might be a nice option.

MC takes the simpler approach that you just mentioned. The tables from the 2009 Food Code also don't take this into account except at the highest temperatures. Nonetheless, what we do is science-based and well justified in the literature.

[DouglasBaldwin]

...

The desire to fully simulate the cooking process is also why I haven't yet done the cooking from frozen feature that some of you are asking for. There is a fair amount of added complexity that comes into play as ice in food melts. It absorbs a great deal of heat in doing so, in a very different way than water that in simply becoming warmer water does. Thus accurately keeping track of what is going on in a partially frozen, partially thawed piece of food gets a little more complicated.

Doing calculations from frozen is much more complicated. I'm the first to admit that my heating-from-frozen tables aren't what I'd like them to be; that's why I don't have pasteurization-from-frozen tables! I coded up some numerical methods back in 2009 and they seemed like they worked great until I actually compared them with experiments. I think if I had about a hundred hours I could improve my numerics to a point that I'd trust them for pasteurization, but don't count on me having that kind of time anytime in 2012.

[Kerry Beal]

Vengroff - out of curiosity - if I'm using a SVM which equipment should I chose - as I notice that there seems to be different timing depending on the equipment chosen and SVM is not included.

[blackp]

Vengroff - out of curiosity - if I'm using a SVM which equipment should I chose - as I notice that there seems to be different timing depending on the equipment chosen and SVM is not included.

Hi Kerry,

The behaviour of your cooking rig will be more related to the kind of device you are controlling with the SVM rather than any behaviour of the SVM itself. If for instance you are using a Fresh Meals Magic with air bubble circulation or are using a pump to circulate the water it will behave in a very similar way to an immersion circulator. If however you use it to control a large rice-cooker (as I do) it's behaviour will be closer to the Sous Vide Supreme - or even better as a cylindrical tank with heating underneath has a better chance of even heat distribution by convection currents than a square or oblong shaped tank does. Also most rice cookers are better insulated than the SVS so that too could impact the behaviour.

Of course Vengroff may have also taken the wattage of the heating element into account. Given that SVM users use it to control a myriad of heating devices it would be good if we could select the parameters which are important to the calculation in Sous Vide Dash. Maybe if we could select SVM from the main list we could then select "circulated or not", "insulated or not" and "wattage" assuming that these are the key variables. (Vengroff can you advise?).

Regards,

Peter. (Edited to fix some typos)

Edited December 4, 2011 by blackp (log)

[Kerry Beal]

Vengroff - out of curiosity - if I'm using a SVM which equipment should I chose - as I notice that there seems to be different timing depending on the equipment chosen and SVM is not included.

Hi Kerry,

The behaviour of your cooking rig will be more related to the kind of device you are controlling with the SVM rather than any behaviour of the SVM itself. If for instance you are using a Fresh Meals Magic with air bubble circulation or are using a pump to circulate the water it will behave in a very similar way to an immersion circulator. If however you use it to control a large rice-cooker (as I do) it's behaviour will be closer to the Sous Vide Supreme - or even better as a cylindrical tank with heating underneath has a better chance of even heat distribution by convection currents than a square or oblong shaped tank does. Also most rice cookers are better insulated than the SVS so that too could impact the behaviour.

Of course Vengroff may have also taken the wattage of the heating element into account. Given that SVM users use it to control a myriad of heating devices it would be good if we could select the parameters which are important to the calculation in Sous Vide Dash. Maybe if we could select SVM from the main list we could then select "circulated or not", "insulated or not" and "wattage" assuming that these are the key variables. (Vengroff can you advise?).

Regards,

Peter. (Edited to fix some typos)

Ah yes - should have been more specific. FMM with 1800W element and air bubble circulator. I assumed I should choose Polyscience Professional but don't really know the wattage etc on those.

[vengroff]

blackp is correct that circulation vs. non-circulation is what really makes the difference. Technically, what I am changing from one machine to the next is something called the surface heat transfer coefficient, also known simply as h. This constant models how heat is transferred from the water into the food. For non-circulators, this constant is lower than for circulators because the circulators do a better job of bringing warm water to the food surface to replace the ever so slightly cooler water that has lost some heat into the food. Non-ciculators rely on the small changes in density of the cooler water to cause it to sink and be replaced by warmer rising water to create natural circulation paths.

Wattage is really only a factor when you start talking about large amounts of water and/or food. You need enough wattage to keep adding heat to the water at the rate it is losing heat so that it does not drop in temperature. The water mainly loses heat into the food, via evaporation at the surface, and through the sides of the bath. Which or these is dominant depends on all kinds of geometric, temperature, and material factors. When the heater doesn't have enough wattage to do a good job of keeping up with these losses you are in trouble.

I don't try to model the effect of underpowered heaters. Instead I assume that people are using appropriately designed and appropriately sized vessels for their heaters. You best bet here is to make sure you stick within the guidelines the manufacturer recommends. If they say they can handle a 50 liter bath and you use 200 liters, you are probably in for a bad experience.

[rotuts]

heat loss to the container and room is one thing anybody can control and it makes every system water bath response much better no matter the watts.

I stumbled on coleman beer coolers of three sizes and am very happy with them

the smaller ones are powered by two to three inexpensive immersion heaters that are about 120 - 150 watts each.

good luck useing these to get your bath to temp. i use hot water from the tap and take addtional hot tap water and micowave it in a large pyrex measuring container: why? uses less energy and has a nice glass handle still cool to add to the bath.

i do not use these smaller elements to bring the water to temp. the SVM goes only in the largest cooler and I still add hot tap water to start out.

Coooooooooooooool App !