Herms system.

[CanMan]

Basic control for my 3v Herms system. Two pumps on off latched switches. One or two elements on for both hlt and boil vessels. And a switch for temp feedback control for the hlt direct and mash return temp controlling the hlt vessel. I figured it's better to control the hlt directly to the return mash temp.

 

[Adr_0]

Good to see some Omron gear getting around.

What size element are you using in your HLT? I think you might struggle with response (overshoot, step change performance, and reduced ramp rates to minimise these) because your HERMS coil is extremely long and your HERMS vessel looks to be very big. Most people end up with a 1-2L HERMS vessel and a much shorter coil, as by the time the heat input gets to the wort (the start of the coil to the end of the coil) it is too late before the heat can back off it the temperature overshoots. And with the size of the HERMS vessel, it takes extra time to actually change temperature because of the large volume.

It sounds like a great idea but it's not the best from a process control point of view.

 

[CanMan]

Hmm interesting. Yes it's about 30 litres in that hlt and 15 mtrs of coil. Using two elements at 2200 watts each. Switchable for one or two depending on how things eventuate. I figured I would heat the hlt to mash temp then switch across and monitor the return of the sparge arm to control the hlt temperature. I was hoping that the omron would auto tune to hold the hlt water at the correct temp for my wort return.

 

[Adr_0]

What's your target batch size? Do you know how to calculate ramp rates and boiloff rate based on your element size? 2200W would be ok for a 23L batch size but pushing it for a 36L for e.g. but yeah 4400W would smash it. 30L is just big enough if you are looking at 36L batch size... but that's about the biggest you could do using rough numbers.

Your mash for 36L (28L water + 9kg grain) might theoretically ramp at 12min/10°C for each 2200W element (6min/10°C for both) but it would need the gain ramped right back to prevent a lot of overshoot and some serious tuning.

Possible but annoying... or are you thinking 23L?

 

[CanMan]

Ramp rates etc are new to me being an extract brewer up till now. I was only aiming at the usual 23 litre batches that I would normally punch out of a can. I've purely gone off copious hours of reading googling and researching others systems to get to where I am atm. Im sure there will be some comical blunders as I learn how to run all this.

 

[Adr_0]

I don't want to discourage you attend all from building a rig and making great beer, just trying to adjust your expectations a bit of you are expecting a fast response with no overshoot, the reality will be different and you maybe be unhappy.

23L is good and I would think that a 2200W in the HLT would be ok and the other 2200W in the kettle would work.

Whereabouts are you? It would definitely be worth getting some advice and a bit of an overview from a local Brewer with a HERMS system, or experience with one. RIMS and HERMS can offer awesome brewing performance and flexibility, but if there are any elements not done correctly you will likely be disappointed.

 

[CanMan]

All good. No discouragement at all. Any input is always good to hear. I actually have two 2200w elements in the hlt and two in the boil kettle. I have the two three position switches setup to either run one element alone or both together for each of the vessels. I do realise it's not instantaneous as far as response times also. I've made the temperature control of the hlt water switchable also so I can control it directly or indirectly from the wort return. (Well that's my plan anyhow ). I was going to do a bit of testing with water etc before biting the bullet and doing a mash. Im sure I'll be messing around with getting that part working spot on but have enough confidence that I'll get there. End of the day it may be a case of directly controlling the hlt water but at a degree or two above mash temp and use wort return purely as monitoring.

 

[TheWiggman]

Your concept will work but as Adr_0 has said it won't be a perfect performer.

You have a PID controller there (and a handy one at that) and they tend to be 'conservative' to avoid overshoot unless they are well programmed for your system. That is to say, it will slow down the rate of heating the closer you get to your set temp. Not great when you're already heating 25l of water as well as 18l of water in the MT with 5kg of grain with only 2200W of power.

On the contrary, the long coil will mean the 'coil out' temp will be probably be the same temp as the HLT. This will aid in preventing overshoot.

You already have the gear so you can certainly continue as planned. If I were you I would sparge through the coil because you will have quite a bit of wort in there due to the length of the coil.

I'd also put as much power as possible into the HLT to maximise ramp times. I reflect on the wise words of the prophet Tim 'The Tool Man' Taylor: "more power!".

 

[Adr_0]

I think regardless of control you will need to properly insulate the return line to the mash tun and ensure you have good distribution of wort over your bed to try to minimise low flow zones.

Regarding the coil length, actually it will increase overshoot if it is longer.

If you double the coil, you are doubling the time constant which is not a good thing. This directly affects overshoot as the time constant relates to capacitance of the system, or heat input before you see a result.

The maths is supported by the concept that for a given tube out temp, a longer tube will have a lower tube in temp. When you get close to your set point, your tube in temp is very close to your set point. Unfortunately the HLT will still be adding heat, so the tube out temp then overshoots and your system oscillates a bit.

A shorter tube is better, with a high flowrate (not ridiculous). Both the reduced tube volume and higher flow will reduce your time constant (a very good thing for process control) , and the higher flow should mean better temperature uniformity in your mash tun, assuming the liquid distribution and collection is spread nicely over the bed.

 

[TheWiggman]

You sure? Picturing in my head two scenarios as I often do with these things...

Infinitely long tube:
The Tout of the coil will be equal to the temp inside the HLT. Once set temp is reached, heat turns off. In temp is irrelevant in this case.

Short tube:
The difference between Tin and Tout will be minimal and a higher temp in the HLT is required 'early on' to ramp the temp of the mash liquor.

Wouldn't the time constant be related to the volume in the HLT, or input energy?

 

[Adr_0]

If you consider your infinitely long tube, how long does it take to get to that'd higher temperature? Do you really think that an element turning off will stop overshoot that is already happening?

Dynamic process control is all about bang for buck. The long tube will transfer more energy (temperature) but at the expense of residence time, which increases the time constant. So although you might get more temperature for a given pass, you will overshoot more (all other things even).

The problem is, does a 2" long tube get the temperature gain we want? What is the correct answer here?

Good response should be a fast rate of change with no overshoot. A few concepts to consider:
-the element, sensing element and mash should be as close to each other as possible
-insulate your lines
-high pump flow increases process gain and decreases time constant. High pump flow is a very good thing, to a certain point
-rules of thumb dictate that what you control (the element) should have a direct impact on the thing you are measuring (wort). Is putting a big volume of water (a big buffer) between the element and wort improving response or is it hurting response?

As Wiggman has said, your setup will work. A long tube or big HLT volume will work, but it will overshoot. As Wiggman has said, PID control (derivative control) can reduce this overshoot, but at the expense of ramp rate.

Good system setup can mean all the difference. Those is why you see guys with a small HERMS tank, shortish coils and high flow pump and the system responds well. My RTD is right after my element, in a 500mL tube. This is why I don't even need the D in PID, so just use PI control and can use all of my power.

You can tune a system to respond well with an auto tuned PID controller but if you change your volume or flowrate you need to retune to get the same response.

 

[Camo6]

 

[TheWiggman]

And open your textbooks Camo.
I won't nitpick any further but what's 100% agreeable is that a high volume of water that is being heated will result in low ramp times. Consider CanMan that you can spend a few extra dollars now and have a better system now, rather than continuing and thinking about upgrading later.

My defeatist opinion is you should use the coil for a chiller. If you want to go down the HERMS route, then get a HERM-IT coil and a 2400-3600W element in a small vessel like the other HERMS folk do. You could grab this kit and build your own vessel with element and it'll probably set you back $250. Your control box will basically be a strap on unit that will work from the word go. You still need all the same fittings and thermowells that you would if you went down your current route but for the extra $75 + PVC and element, I know what I'd prefer.

 

[Adr_0]

Finally found this thread again...

I have to apologise to TheWiggman and CanMan for the bit about short coils possibly being better, where I thought the time to change the temperature would offset the temperature gained. Wrong.

Interestingly in all of my modelling (using feedback from about 5-6 guys and also my own RIMS testing) there is a definite 'trouble' spot with HERMS if your HEX/HLT (wherever the coil is) starts to get too big and your coil is too short. This 'trouble' is that the system will inherently overshoot. Although you can tune this out with PD control, you sacrifice a lot of ramp time in the process. A bigger element helps the ramp time but likely makes the overshoot worse, which needs more D and a sacrifice in ramp time to prevent the overshoot.

If your system is inherently stable/no overshoot, then you can change elements, tuning, pump circulation and the system should respond well regardless. You can probably almost do it just on P, but probably need a bit of D control as well.

Seems like if your HEX is more than 1/3 of the volume of your mash and you have a short coil (e.g. 2m of copper) you will run into some overshoot problems. Your ramp rate will also be pretty crappy.

And having said all that, since you have a 30L HLT/HEX and 15m of stainless with 2 x 2200W elements, you will be fine. Theoretically 5-8m of stainless might have some problems but 15m is great.

Happy brewing...

 

[CanMan]

I went ahead and temporarily plumbed the hex with the hlt for my own experimental benefits. I've ended up leaving it as planned. Meaning it worked as good as I could have hoped. It holds the temperature within +/- .3c. I'm happy with that result so never bothered altering the length or the hlt size etc. I have two elements (2 x 2.2kw ) in the hlt and can turn one off. Two works great for ramp times and seeing as I'm there messing about I turn one off around SP. I did read all your posts and models and found it interesting. You put a crap load of work into all that. End of the day my system worked to my expectations and I'm happy.
Cheers for the input. I'll shoot a few pics of my finalised system and post them. I've pushed out two brews so far after some wet tests and tuning. One was a slightly modified smurtos and I couldn't have asked for a better start to AG brewing.

 

[CanMan]

Couple of quick pics.

 

[Adr_0]

Mate that is unreal. Very glad to hear... The only important thing is that it meets your expectations and is good to brew on, so that's awesome.

And congrats on coming to the dark side... Your world will never be the same.

 

[CanMan]

So I'm part way through mashing as we speak. And I'm reading your tuning information. I'm an industrial electrician myself and deal with heating every other day. And I got to thinking how .3 +/- isn't that great. So start messing with PID settings. Only to find that it's still set to on/off ( looks away sheepishly) I set it to PID mode and cut both I and D out to 0. Set P to 1 and voila now controls at exact set point. This equals me being more than happy. Thanks for all your info and models once again. If you hadn't written all that info for me to be reading I probably would have had the damn thing set to on off for ever in a day.

 

[Adr_0]

That's awesome... No worries and very glad it performs as it should.

Sounds like you have a good system. You can probably increase the gain to 2-5 if you want to get a bit more aggressive closer to the set point. Might save you a tiny bit of time... Up to you,won't really make a massive difference I don't think.

Hope you age enjoying your brew day!

 

[Rosscoe]

Finally found this thread again...

I have to apologise to TheWiggman and CanMan for the bit about short coils possibly being better, where I thought the time to change the temperature would offset the temperature gained. Wrong.

Interestingly in all of my modelling (using feedback from about 5-6 guys and also my own RIMS testing) there is a definite 'trouble' spot with HERMS if your HEX/HLT (wherever the coil is) starts to get too big and your coil is too short. This 'trouble' is that the system will inherently overshoot. Although you can tune this out with PD control, you sacrifice a lot of ramp time in the process. A bigger element helps the ramp time but likely makes the overshoot worse, which needs more D and a sacrifice in ramp time to prevent the overshoot.

If your system is inherently stable/no overshoot, then you can change elements, tuning, pump circulation and the system should respond well regardless. You can probably almost do it just on P, but probably need a bit of D control as well.

Seems like if your HEX is more than 1/3 of the volume of your mash and you have a short coil (e.g. 2m of copper) you will run into some overshoot problems. Your ramp rate will also be pretty crappy.

And having said all that, since you have a 30L HLT/HEX and 15m of stainless with 2 x 2200W elements, you will be fine. Theoretically 5-8m of stainless might have some problems but 15m is great.

Happy brewing...

Hi guys, have been working on redesigning my gear to go herms with pid control, upgrading from my single infusion setup. Have found this thread pretty interesting.

I was looking at a very similar setup to the author of this thread, but as I do double batches, was going to use a 40ltr HLT and use an immersion coil. The one I have is 7.6mtr of 1/2" S/S. I've got 2 x 2.4kw elements in it.

Do you think this will be too short for the size of the vessel?

Cheers
Roscoe