sous vide thread


 
Ok I get to ask the first question :)

Why can't you just use a crock pot hooked up to the SSR and then use the LM to control the temp?

I saw from Evan's equipment list a special heater and pump. Is that necessary? I could see that circulating water would be advantageous as it would reduce localized temperature spots. Is that the idea?

I've got a crock pot, and a LM with a SSR, maybe I'll give an egg a shot?

dave
 
Dave - that theoretically should work. Rice cookers, Food Warmers have all been used with success.
As far as some increased form of circulation it may or may not be required deprnding on the device. Easiest (and cheapest) is a small aquariam air pump with an airstone.
If you do try the egg, make sure it's not sitting on the bottom of the pot.

You might want to check out these post for some ideas and inspiration:
Momofuku-48-hr-Short-Ribs
SV-Rib-Roast
Sous Vide Pork
 
Thanks for the links, I'll check them out.

I considered using my crock pot to test with. It will probably work... I used to use one at work for melting wax that even had a thermistor with setpoint, etc... Walmart sold it.

Let us know how it goes! Circulating the water would be ideal, but just stir it around a bit and it should be fine, and try to keep the meat off the sides so it cooks evenly.

I have heatermeter parts in the mail, so I'll rig up an SSR to my crock pot and see how well it holds temp.

My other plan is to use my "kill-a-watt" to see what various PWM duty cycles do to with respect to percent of heater power, to make sure it works well over the full range. Using the Slow fan PWM mode might work better if you set it to 100% (though I'm not sure that has been tested in the software by Bryan, should work though).

EDIT: I just recalled that I have an electric turkey roaster. I'll try heating it up in the next few weeks and see how well it can hold temp with a heatermeter.

EDIT 2: I also just realized that some crock pots require you to push the "high" or "low" button when you plug them in for them to start the heater. You'll probably need the cheaper, "dumber" crock pots, or you can jumper in the heater, which is not recommended :p Luckily the turkey roaster ($20!) only has a knob.
 
Last edited:
I have an older style crock pot that just has high, med, low, simmer.

I hooked it up and cooked 3 eggs. They turned out just like they were supposed to.

Today I'm doing a pork loin. I rubbed it with bacon fat, sprinkled on my favorite pork rubbed and it will go in at 140F for 4-5 hours. I'll let everyone know how it turns out.

I did measure ~2 degree difference between the top of the water and the bottom. The top of the water being hotter. I think a little air stone wouldn't hurt. I'll have to go dig through my aquarium equipment drawer and try it out sometime.

I have a turkey roaster as well and think that will work just fine as well. I use it to poach my sausage after taking them from the smoker. I never thought about controlling the temp with the LM, but I'll definitely do that next time as you don't want the water to get too hot and render the fat.

dave
 
Turned out AWESOME!

I didn't get any pictures as it disappeared as fast as I could slice it. Here is a picture of the cook. I started out with hot water so you can see it cooling off at the beginning of the cook, but after that it was dead on 140F.

sousvideloin.png


dave
 
Well done Dave! It looks like the control is working well.

Did you set the "fan" to slow mode, or just leave the settings as default? Either way, it looks like you were well below 10%, so you were probably in the "slow" pwm range anyway. Did you notice the LED on the SSR turning on for small amounts of time?

I got my work bench all set up to start building my LM/HM :)

Geir, your cooks look great! the only thing they are missing is internet monitoring and control :p
 
I did not change the fan mode, just ran it per normal.

What I saw was at say 3%, the SSR light would turn on for 2 seconds, then run off for 10 seconds or so, then turn back on. It wasn't what you would think of just running the fan very slowly at 3% of max speed continually. Hopefully that makes sense.

dave
 
That makes sense.

Just for clarity.... some of this information is duplicated in another thread, but here goes:

PWM:

If you run above 10% (by default), the PWM is in "normal" mode, with a frequency something like 500Hz or whatever the Arduino puts out on the PWM pins. This is similar to a true "analog" value, which is why arduino calls it an "analogWrite()". So, at 50% power requested, the output pin is a square wave with a period of .002 seconds, where it is on .001 second and off .001 second.

SRTP:

Once the "requested" fan speed (or in this case heater power) goes under 10% (by default), the PWM of the pin gets set at 100% duty cycle, but the pin is then cycled on and off much slower than before. The ratio of the on time to the off time is still equivalent to the duty cycle, say 5%, but it is slowed down to be on for 0.5 seconds of every 10 seconds, so the average is 5% on time. The code is such that we only have 6 "steps" of resolution on the output, but hopefully we can increase that with a small code change. For the fan it is just fine, but for the heater we seem to be running in the < 10% mode, so increased time resolution here might smooth the line out. I have been calling this mode SRTP, for "Split Range Time Proportional", which is a term I've become familiar with in the PLC world... It makes this topic easier to discuss in writing if we have some nomenclature defined.

Another tidbit about SSR (Solid State Relays)...

SSRs do not provide proportional control, at least not the way we are using them here. If you look inside a standard 3-35vdc triggered SSR, you'll find a bit of circuitry, an opto-isolator, and an SCR. The SCR is our "problem" here, since they will only switch in phase with the 60Hz sine wave of the mains voltage you apply it (or 50Hz in some places). So, the ability to define our PWM duty cycle in increments less than 1/60 second does little to no good. It will make for a nice flicker for Halloween lights, but that is not so great for powering a heating element. I imagine it makes a lot of excess heat as well. Just to be thorough, you can provide proportional control, but that involves syncing to the 0 crossing point of the sine wave and offsetting by a phase angle. There are chips that provide the timing in a microcontroller friendly way, but this is very much overkill for our setup. If someone finds a reason to control lights with the heatermeter, that is a different story all together.... Anyway, that is why we need the "SRTP" mode, so we get predictable and efficient control of the heater.

Why?

For those who do not already know, the SRTP mode was created by Bryan to handle the fact that most of these blower motors are not brushed DC motors, they are 'brushless DC' (BLDC) motors... They have an MCU on board that does some magic to make the fan go, and they need a minimum voltage to run. The range they do not like to run in, we avoid via the SRTP. Just going out on a limb, I think the heater stuff will be best kept in SRTP mode at all times, especially if the resolution is increased.

Did anyone actually read all that? I added 'topic headers' for easy skimming :)
 
Hello,

I was curious what temp probe you are using for SV? It's my understanding the maverick probes are not capable of being used as a immersion probe.

Thanks.
Pete
 
I used the ET-72 Maverick probes. Any pointer to why they are not accurate enough?

dave

Hello,

I was curious what temp probe you are using for SV? It's my understanding the maverick probes are not capable of being used as a immersion probe.

Thanks.
Pete
 
Sorry,it not the accuracy but the longevity of the probe use as an immersion probe, below is from the FAQ on themaverick site

"What can i do to prolong or protect my probe?*

Wrapping your probe wire in aluminum foil can help protect the probe from liquids and flare ups during cooking.*"
 
That makes sense. Where my probe wire meets the probe is outside the pot to keep it away from the moist environment.

dave

Sorry,it not the accuracy but the longevity of the probe use as an immersion probe, below is from the FAQ on themaverick site

"What can i do to prolong or protect my probe?*

Wrapping your probe wire in aluminum foil can help protect the probe from liquids and flare ups during cooking.*"
 
Okay, I got a chance to rig up some stuff tonight.

I pulled out an SSR relay from the junk box, as well as a 486 heat sink for good measure. I found an "old work" construction box complete with a plug in my junk box as well (I have a lot of these junk boxes). They fit very nicely in the box:

sous1.JPG


sous2.JPG


sous3.JPG


The only problem with this setup is that my relay is only rated for 10A, as well as the IEC cable I cut up to feed into the box. I set my turkey cooker on the kill-a-watt and it pulls 12A. The newer SSRs are really cheap ($10 or so) and can handle 25A without a problem, and it should be easy to find a cable that will handle the 15A that the female plug is rated for.

Will it work? Possibly. I do not trust it unattended though, so watch your currents if you build something like this.

As for using it with the heatermeter... As expected, plugging in a stock heatermeter will not get "optimum" results. The 490Hz (or whatever) PWM is fast enough that any PWM duty cycle greater than 0 will allow the SSR to see a high signal at least once per AC cycle, and turn on the current until the next 0 crossing point. This results in 100% power at any duty cycle >= 10%. EDIT: THIS HAS BEEN FIXED!

Ideally we need to improve the resolution of the slow fan mode so that it can take advantage of smaller time slices of the 10 second time period. Feeding any "PWM" cycle to the SSR does no good, so only the special slow mode is functional. As it sits, it has 5 levels of power possible:
0-1% = 0% power (off)
2-3% = 20% power
4-5% = 40% power
6-7% = 60% power
8-9% = 80% power
10+% = 100% power

As discussed earlier, this should work fine, but the PID might hold tighter tolerances if it has a higher resolution possible for its control variable.EDIT: THIS HAS BEEN FIXED!

When I get some more time I'll poke at the avr code and try some more things.
 
Last edited:
Ideally we need to improve the resolution of the slow fan mode so that it can take advantage of smaller time slices of the 10 second time period. Feeding any "PWM" cycle to the SSR does no good, so only the special slow mode is functional. As it sits, it has 5 levels of power possible:
0-1% = 0% power (off)
2-3% = 20% power
4-5% = 40% power
6-7% = 60% power
8-9% = 80% power
10+% = 100% power
Aww yeah now this project is kickin' off. Actually the default code is even worse because in the SRTP mode, it doesn't run 100% for n/10 seconds, it runs 10% for n/10 seconds, so that's still PWMin'. You'd need to set the MINIMUM_FAN_SPEED in the code to 100% to get it to work. However, I just noticed that the code's math is actually incorrect and only arrives at the proper number for 10%. I'll have to fix that up.

Also I'm almost certainly going to take the PID loop back down to 1s and just ignore every other update in linkmeterd to prevent errors in the rrd.
 
Bryan,

10% PWM is still 12v pulses to the SSR (SCR), so it turns on no matter what the duty cycle is. The fact that it is PWM'ing the SRTP at the "low" number (10% by default) makes the fan work correctly, so I don't think any code needs changed for that reason specifically.

If you set MINIMUM_FAN_SPEED to 100%, the problem will be solved anyway, though I'm not sure how the SRTP will act under those conditions.
 
Yeah it is still 12V but I'm saying that when it is at, say 2%, it is running at 10% 490Hz PWM for 2 seconds out of every 10. The PWM is too fast for the SSR because it is still PWMing right?

In any case I've just pushed up a bunch of new code to github that adds the ability to set the min/max fan speeds from the web interface. I've also added the polarity inversion code but there's no checkbox on the web page for it yet.

There's a bunch of other fixes and changes too like being able to set the backlight from the button menu and a lower baud rate. You will lose your HeaterMeter config when you update because the configs are incompatible.
 

 

Back
Top