HERMS vs. "Counterflow" HERMS

[LizardRuss]

I'm planning to build a brew stand and re-purpose my all-copper CFC as a HEX, similar to what is described here. I'll be recirculating both the HTL and MLT, both with tangential return ports. This will hopefully remove the need for stirring.

I will also monitor the temperatures through a dial thermometer installed on the HLT, and then a thermometer at both the input and output of the MLT. The MLT is a rubbermaid drinking cooler, so I'm curious to see how much of a temperature difference there will be between the two.

I have a couple questions for those with experienced with this type of set up.

1. Does using a tangential return in the MLT still require stirring in practice? I can see where the grain bed would prevent a lot of circulation and lead to stratification. I have the MLT set up so I can still remove the lid during the mash to stir if required.
   
2. Is it possible to control this by hand? Automation isn't in the budget for a while. My planned strategy for manual control is to have a constant recirculation rate for both the HLT and MLT, and control the temperature using the HLT burner. In theory this would make the system an (in)direct fired mash, albeit with higher temperature losses, and I'd effectively be heating double the volume of water.
   
3. From the graphs posted here it looks like the control method is maintaining the HLT temperature and cycling its pump on and off. Is this the only practical way to achieve a step mash with this system, or is it just the way its been implemented?

My initial goal is only to keep the temperature stable. With my current set up, I don't have a way to monitor mash temperature, so I've only been making guesses. For example.. start the mash at 154, and finish the mash at 148... the target temperature being 151. I'll consider the new system a success if I can keep the mash temperature either steady, or at least above 150.

I don't have much hope for being able to do a step mash with my planned control method, since I'm essentially heating all of the water used in the brewing process with greater heat losses. I imagine I will just have to try it and see, unless someone's already tried this and can save me some time

Another downside to my control method is that the HLT will not be up to sparge temperature at the end of the mash. I may try a mash out, but it could take too much time to be practical. Worst case I can stop recirculating the HLT and heat it up to sparge temps.

 

[orangehero]

You don't want the mash to be constantly churning. You want to gently return the wort just under the surface of the mash, not with a fixed tangential inlet. You would stir in the grain to hydrate and break up any doughballs, then set the grainbed with slow recirculation. Unless you batch sparge you shouldn't have to stir again.

 

[LizardRuss]

Thanks for the reply. I'm planning to use loc line for the mash recirculation, and I'm going to keep the flow rate low. My post did make it seem like I was going to try whirlpooling the mash.. but I was just lazy and didn't go into detail.

 

[Vesku]

I did 4 brews with just an insulated bucket and some hot water to mix in. I had 2 sensors in the mash to see how much hotter the upper levels of mash are compared to the lower ones. Red and blue are in the mash. Here's a plot of 90 minutes mash:

They stayed pretty much identical once I got the temp right. Then in the end you can see when I started to recirculate the wort and the temps drop.

I'm having a pint of beer made by the old fashion bucket and a some hot water way and it's excellent

But, if you want to repeat the mash exactly the same way every time and do step mashes accurately AND do some serious geeking with the whole process...

I've an idea for a compact add on cf-herms that could be added to any all grain system, anyone keen to beta test it out?

 

[Vesku]

1st test with the pwm-controlled heating pump:

Interesting to see what it does with the grains, above with just water...

 

[alien]

That's awesome, Vesku. Any overshoot looks pretty local in that system. Can you graph the pump output too? What's the pump you're using?

What's your idea for the add on? My first guess would be a copper Liebig?

 

[grainbill]

I use my CF Chiller as well. Never had a problem. However, I do use a small kettle for a secondary HLT due to the nature of the water side of my CFC. The CFC does have an outflow thermo and I can monitor temps. I use about 2 1/2 gal in the small HLT, heated to 170. I adjust the temp of the mash by controlling the flow of the mash and/or hot water. When it is time to mash out, raise the temp on the water to 180 and I'm about done. Yes it does require two pumps. I have found with the 2 1/2 HLT I do not use as much gas to heat the small volume. I do need a fourth burner, which I have anyway for decoction batches. NOTE: I would not use a plate chiller as I would assume that you would plug it up in no time and from what I have read they are a mother, if not impossible, to clear...

 

[Vesku]

The pump is a cheap 24VDC, the ones you can get from several Ebay sellers. I did some more testing and I think that my CFHX- is a bit too powerful, it would be much easier to control the pumping power near the setpoint, if it would take a bit more pumping of the hot water than what it takes now. Now it overshoots at lower temperatures even when I've the pump's pwm output set to 1 of 255. Ideally, the heating pump should be running all the time, then all the sine waves should go away or at least be really small and much lower frequency, because the delay before the wort sensor starts to warm up would be a much smaller.

I can graph out pretty much anything that is going on with the MegunoLinkPro (and it's easy too and very handy when testing out stuff), but did you mean the PWM-curve or what? I've that on "a table" in the MegunoLP.

The wort temperature probe is straight after the HX. I moved the mash-sensor on the output of the mashtun. I now only use the wort-probe for setting the mash temperature.

I'm using this for CFHX, it's about 2 meters on length:

I'll make a shorter one, about 1 meter and see how that goes. I'll use the old for a whirlpool cooling ... now that I've one extra pump available

On the downside, the heating is of course slower now ... reversing the flow would make a lot faster heating times possible, I might try that in some stage.

It will be interesting to taste, if it makes any difference in the beer, now that the wort doesn't "overheat".

More test curves:

 

[alien]

Yeah, I was wondering what the graph of the PWM output looks like.

The biggest cause of the lag seems to be mixing in the mash, not heat transfer. So maybe the way to go is a shorter, fatter Liebig for faster recirculation.

 

[Vesku]

I'll take a graph of the PWM next time that I have time to test the system.

The problem with a faster recirculating, if it's "gravity fed", is that it's really easy to end up with a stuck mash. That's why I've been thinking about the reverse flow and maybe switch to gravity fed after the mash off temperature is achieved to get a clear as possible wort or just slow down the down to up flow. I've DC-pump for recirculation too, so that would be easy to do (the slowing). I've to test this...

The HX that I have has 15mm inner copper tube, it's not limiting the flow at the moment, the grain bed is.

 

[alien]

Sounds to me like you might have hit the limits, but I'll be interested to see what you can come up with.

I wasn't able to PWM my 12V brushless pump without making a horrible noise, what frequency are you using?

 

[Vesku]

I haven't checked the frequency myself, but internet tells that it's 490Hz by default with Arduino Mega that I'm using. I've sketch that changes the pwm range when setpoint comes closer and it sounds quite nice, like a Lo-Hi tune from Kuusumu :fro: It also sounds like the motor is not liking it, but at least it's survived over 10 hours without any problems. With the lowest setting it sounds like the motor is not revolving at all, it's that bad...

 

[alien]

Maybe it is just an aesthetic problem. Who said pumps were supposed to sound nice, anyway. Or Finnish art bands.

 

[Vesku]

Here's a graph with the PWM numbers (green), it's the 65 -> 70degC step from a real mash, 5 gallon patch.

Here's the mash temp only from the whole mash, measured from mash tun's output:

Wort + mash:

Pretty happy with it, but I think that I can make it a lot smoother with a smaller HX + tweaking the code (maybe using a PID library).

 

[alien]

Yes, it looks good to me.

 

[Vesku]

Productive day, I made a new shorter HX:

And I added a PID-library to my Arduino program, here's a test run with parameters from this example, I only changed the "gap"-parameter to 2 degC
http://playground.arduino.cc//Code/PIDLibraryAdaptiveTuningsExample

Now I just have to get into the PID parameters to fine tune it. I should have gone the PID-way right from the start... Now it's time for a beer

 

[alien]

I'm definitely thinking of giving this a whirl.

I was originally planning a 10mm copper coil inside a 1.5 litre electric kettle for the HEX, but a short Liebig like you have made might give better flow rates as well has having a higher capacity for heat transfer.

I have a couple of 12V pumps and logic level FETs and have ordered some gate drivers to help do the PWM.

 

[Vesku]

Did some testing and reading last night and now a brew. The new 1 meter HX is still way too much. What I can guesstimate from my graphs is that about 10 centimeters is enough, if the length/efficiency-ratio is linear on HX's. I think I do a 30cm one and see how it goes.

Now it's almost all the time out of the "range" and just oscillating.

 

[Vesku]

I couldn't make the PID work as I wanted, so I ditched it and I'm using my own code again. I also tried with even shorter HX, but step times suffered too much. Here's the latest test, I think I leave it like this for now (I'll just make it a tiny bit more aggressive on steps + different profiles depending whats the temperature difference between the setpoint and HLT temperaure):

And after some tweaking:

 

[alien]

It still looks over damped to me.

I believe the brewtroller code implements PID for this somehow.

 

[Vesku]

Yeah, it does, oscillating like crazy.

I just put in a 12V pump instead of the 24V. It looks like I hit the sweet spot now. I've PID back on again and the pump is running all the time, no oscillation (on/off). I need to wait to my HLT to hit 80 degC and I can see, if it works in all of the heating steps. And then really try my luck with setting the PID-parameters

1st test with the "stock" PID-parameters :rockin::

 

[Vesku]

Yep, it seems to work now. I didn't have the patience to wait long enough to the HLT to heat up. The pump was just a tiny bit underpowered when heating from 50C to 65C and it didn't reach 65C straight away, but that will be cured by just g(l)etting the HLT to heat up to the 80C or just above. The PID-parameters overshoot the wort temp by 1C when heating to the next step, but I'm sure that this is relatively easy to dial out now that the system is stable

 

[alien]

That looks good. I reckon that is pretty much dialed in.

I'm encouraged by the 12V pump too because that's what I've got.

 

[Vesku]

Yet another test run 50-65 deg C, maybe a bit too aggressive (some oscillation), but it might be just right once the grains are in:

Once settled, it keeps the wort with in +- 0.1degC and the mash under +- 0.05degC :rockin: I guess this starts to be it or I have to get more accurate sensors

Now I can start to convert the system to the reverse flow CF-HERMS. I might try it once with the grains before that though

 

[PDX_T]

it's pretty common in industrial steam PIDs to have steam pressure under slave control of the main temperature PID. this would be like controlling the HLT temp as an output of mash temp instead of controlling pump flow or HLT heating element with the mash temperature controller. much more stable, no feedback loops.

 

[alien]

it's pretty common in industrial steam PIDs to have steam pressure under slave control of the main temperature PID. this would be like controlling the HLT temp as an output of mash temp instead of controlling pump flow or HLT heating element with the mash temperature controller. much more stable, no feedback loops.

I don't follow. What is the relevance to this?

 

[PDX_T]

I don't follow. What is the relevance to this?

I could have misread your setup, but it might help reduce oscillation and overshoot if you keep the product flow constant at a rate optimized to HX length and bed pressure, and have the mash temp PID control heated water temperature in the HX kettle, not flow rate of hot water. maybe I'm confused about how you have this hooked up but by varying pump speed you have essentially no thermal ballast, so your system is over-responding to the mash temp output by cycling the pump on and off. to slow down oscillations, vary HX water temperature, not flow. this is like an industrial application where the controlled variable is steam pressure, not valve output. just that in a cascade system steam valves are PID controlled as a second level, by the steam pressure setting. If you have enough thermal ballast in your HX source kettle you don't have to do that here unless you want to set up your system to allow you to control HX heating element as a function of HX temperature as a function of mash temperature. it will be more controlled because the HX heating water temperature will not oscillate. it just needs to have sufficient delta from the mash and sufficient power available for the heating task you require.

 

[Vesku]

The main reason behind all of this (there's a lot of info about this earlier in this thread, my previous setups with a constant pump speed and before that I had several HLT controlled systems) is to get an instant respond to the setpoint changes when doing step mashes. This can be done with a HLT-controlled system too, but it would require insane amounts of power to get a fast respond.

Here's a graph from "HLT-controlled" setup (2kW induction stove, 1 cup of water in a small kettle with a HX):

It takes ~ 12 minutes for the wort temperature to reach the setpoint (50C to 66C). With the CF-HERMS it takes a few seconds.

 

[Vesku]

I had (another) go with a reverse flow HERMS, pumping the wort under the grain bed and collecting from the top to get faster heating from step to step.

It failed miserable (again), but after the pump run out of steam I just tried to do it the normal way (up -> down) while using the full power of the pump on step-temperature changes. It worked very well! No signs of a stuck mash or any other ill effects. The fastest mash temperature change that I've ever achieved:

I just need to re-tune the PID parameters...

The mash when the pump was on full power: