Not sure if this is in the correct place?
Anyway I want the brew pi spark to control two 3500W heating elements in a still, does anyone see any problems using the brew pi spark for this?
The boiler will be a kegmentor 58L with two 3500W heating elements.
I’d be interested too. I did a small experiement on my euro-30. What I’ve done so far is to put a 1-wire thermometer into the fitting for a regular themometer in top of of the still and measure what is going on. Reading from a regular thermometer was always a bit difficult. I was going to move onto other matters once I had done some measuring.
In my case, it was very interesting to see the temperature slowly moving up for a period and then suddenly shoot up to a very level 79 before remaining very steady. This was with constant water flow. No fiddling needed with the water flow at all and I was getting around 93% alchohol. Towards the end, the temperature started to rise slighty, and my still is designed to essentially stop prooducing when getting to the end point of the run. The element simply doesn’t have enough power, by design, to push any more energy in for the temperature to continue rising.
So I decided my old, simple still balanced very easily and I didn’t really need to do any more.
I did try a test run with just water, setting up a brewpi fridge and the temperature to 77 and that worked fine with the element comming on as needed. The system reacted quite quickly to changing the fridge temperature, faster than I expected.
I probably don’t need to, but for the next run, I was going to try controlling the heating and keeping the temperature to 77 rather than the natural 79 using the brewpi. In my case I cannot increase the water flow any more from the design of the still.
I don’t see why you could do much the same with your rather larger setup. I’m not sure if it’s possible to drive each of the heaters seperately, but it would be neat to drive both up to the point of getting a good initially flow and then cutting back for the main run.
My cooling flow is very simply, but if you have a mains pump that would accept the controlling influence of the brewpi, that would make sense.
I’d be interested in seeing anything you come up with!
I think it would be very interesting to see how we can control a still with the Spark.
I think it would be interesting to use a 2 PIDs, where one measures the temperature in the top of the still and adjusts the setpoint for the PID at the bottom of the still.
I would not use the fridge settings for this, the HERMS wizard is much closed to what you want. The HLT would be the bottom of the still and the MT the indirectly heated top of the still. I don’t have experience with distilling, so can’t comment much.
But if you describe the ideal control scenario with all the bells and whistles that you would like, if maybe possible, then I can think of a block chain that would achieve it.
I’m brand new to distilling & haven’t got the still set up yet, It sounds promising, I will link this thread with another forum & see if some of the more seasoned distillers can offer some insight into a ideal control scenario, thanks for the input so far.
Ok how does the Brewpi control the heating elements, is it either On or Off or does it manipulate the voltage to reduce the heat output?
Seems that this is the way it needs to work?
Everyone is pointing me towards a voltage controller, which sees crude to me?
We use PWM signals to get % output from an on/off actuator.
https://brewblox.netlify.com/user/all_blocks.html#analog-actuators
Had this reply on the other forum
You are much better off using that beer pi logic to hit a wattage and not a temperature. Power voltage cycles up and down a fair bit for me during a run and having a dead stable amount of watts applied would help column stability no end
You will find you tube vids of people running stills by controlling temp, this stuff needs to be put into the urban myth category. What you can’t tell from those vids is the taste of the product or as bluc said, the smearing evident
Ethanol and the other 30 odd Alcohols we deal with do have different boiling points, but they wont just boil off perfectly in order by setting the temperature. Some are bonded to each other, some form as other components dissapear of get broken down. This is what a reflux column is for, the constant phase changes up and down the column let the different components separate for us to draw off slowly and manually seperate by cuts
I would like to get a little more science into the discussion.
Of course we can run a constant amount of watts, but I would think a control system is better. So what do we want to control if it is not temperature? What do we measure?
Or is temp control failing because it is not stable enough in those Youtube vids? Is the temperature
maybe not equally distributed?
If we can measure it, we can control it. I don’t believe that not measuring is a better approach.
Hi Elco, I wish I knew enough about the process to answer, I sort of get the idea that the guys who distill do it by feel, but they don’t seem to like change, i will link to the topic on the other forum, maybe you could ask the questions on there? or at least have a read up on the actual processes involved? I did link to this topic but it seems they don’t wish to come & give any input.
Typically experienced distillers want new people to focus on the taste and smell of the products comming out to let them know when to change the water flow etc, so to them its more of an art than a science. Other advice I’ve seen is to throw the thermometer away and reply on smelling/tasting the end result to judge where to make cuts. This is particularly true when using a pot still to fashion whiskey or rum. The contents of a fermented wash with these can vary a great deal in terms of content. Consider the difference between fermenting pure sugar vs a mash of your own fruits from the garden.
But with a reflux still where one is more focused on high proof alcohol and removing components of the origonal wash, I bet a brewpi would work well.
I suspect a number of people have tried to engineer their way into distilling and don’t get too much encouragement to apply science. I’d love to have a go at improving my final product using my reflux still and I do plan to try a hermes based wizard approach next time I need to do a run.
Have you made any progress on setting up the spark for controlling a reflux still - I have just got hold of a column and would like to set it up on Brewblox, but being new to distilling would like to see if some others have tried it first.
Cheers
I would think you’re pressure sensors would be great for pressure, level sensing, and density to run mash/wash programs.
Alcohol detection, pH and to a lesser extent DO w/ relay control would certainly be cool additions.
Reply to my own comment here. I’ve now tried using a simple pot still with brewblox for a stripping run instead of the reflux still I used before.
This simple still has just one thermometer, power and a cooling condenser. For the condenser I have it running with quite a decent amount of cold water at a constant speed to condense anything the still generates.
The thermometer is at the top of the still, just next to the entry to the condenser and I’d like this to stay at 79 degrees C. The thermometer is a brewblox one-wire with a very good fit.
The power is hooked up to the brewblox and what I’ve found is that I get quite a lot of cycling around the 79 degrees rather than it staying close to the 79 degrees. As the temperature rises up to around 77 or above, I get product from the condenser. The power is then cut by brewblox and the temperature drops down to around 73. This is too low and nothing comes out of the condenser until I get another burst of power some 30 secs of so later and we cycle round again.
I tried to show this on the enclosed and I’ve disabled some of the readings by clicking on the graph legend, but as soon as the screen auto refreshes, the other, slightly confusing lines get drawn back (as an aside, is there a way to stop this from happening as it didn’t used to do this in previous versions). I’ll fiddle around some more before uploading the screenshot.
Ideally I’d like the temperature maintained to around the 79 and not going over 80 and not dropping below 76.
Any thoughts on tweaks to any settings appreciated!
That sounds like a bug. I’ll look into it. If you select a subset of the graph by clicking and dragging horizontally, it also will not re-render when it receives new data.
For feedback on PID settings, we’d indeed have to see the graphs, but if a constant baseline is required, you could experiment with the boil mode settings. Set the boil mode offset to -21*C, and it will kick in for setpoints >= 79*C.
Thanks Bob Here is a screen shot of the graph. Air Sensor is the still temperature I’m happy to dabble with any settings you might suggest
Definitely something that @Elco may want to take a look at - comparatively I’m a dilettante when it comes to tuning.
Could you also take a screenshots of the PID graph?
What I can tell is that your system is much, much more responsive than what the PID is expecting. It looks to be mostly P-driven with a bit of overshoot. You’ll want a <1m Ti to dampen the oscillation in the PWM setting / temperature.
What is the current PWM period for your heater?
Unfortunately I’m not very experienced with brewblox Bob. I used a fridge wizard some time back and have used it for pretty much everything I’ve needed for fermenting and bread making with any tampering that I’m aware of 
So I’m afraid I wouldn’t know how to answer the questions about PWM and so on, but I would be happy to simple run a wizard to set things up from scratch - does that sound sensible, say a fridge wizard?
I’m going to assume you’re using the default PWM period of 4s then =).
You can find the PID graphs by going to the Spark service page, selecting the heat PID block, and opening the graph from the actions menu (3 dots in toolbar).
Edit: try setting a Ti of 10s, and a Td of 0s in the PID. You should then see a more constant value of PID I, and less oscillation.
I guessed at that moments before your reply, so the GUI must be very well designed Graph uploaded
