So the shower at my gym has a sluggish response. It's a single rotary knob, I usually I have to twist it a few times before it settles at the right temperature; it delays, then overshoots. Alternate scalding and freezing, what fun ....
How to model this system? A zero in the right half plane?
The controller is P... no I necessary, maybe I need to add some D...
Try adding a simple delay that corresponds to the time it takes the water to travel from the faucet to the shower head. You adjust the knob, then the result is not felt for a few seconds while the new mixture makes its way through the pipe.
Transport delay in the pipes? Heating and cooling the pipes makes it more complex.
Some mixer valves try to regulate the temperature, too.
You could datalog the hot and cold turnon curves.
Well I wonder how to model the closed loop in the complex plane (and how to optimize the controller). My old controls text doesn't include a chapter on showers -
It's hard to compensate a pure delay, which seems to be the case here. It requires a pure anticipation, like one of those old time greek oracles -
Are you the closed-loop controller?
I run the hot water full-blast until it's pretty hot, then add cold to get something tolerable. Even that needs some tweaking for a couple of minutes. Our cold water has dynamics too, since the water inside the house has warmed up but the cold water from the street is very cold. The pipes to our upstairs shower are long runs, too.
If I let Mo shower first, my warmup is much nicer.
I think there are electric heaters right at the shower head, with local temperature control. Sounds too high-tech. Water valves are reliable.
Maybe there is an iValve that you can control with your phone. Like one of those fake LED fireplaces.
Anticipation is easy - downstream flow and temperature sensors and dead reckoning.
It's easy with hot water systems, because there is no integration of the control input. You tweak the knob, wait the required delay, and evaluate whether the effect was of the size you desire, and interpolate/extrapolate using a linear approach.
Plenty of control systems have delays that involve integration, and whose are truly difficult.
Clifford Heath
Be a Real Man and take a cold shower.
In LTspice, they have a transmission line, no?
I would expect the best controller to be one with quick response, meaning high gain proportional. Do you really think this is going to oscillate? Oh, wait, that's exactly what it will do. Best to adjust the controls before you get in the shower.
Can't you remember where the control ended up last time? Mark the control with a bit of tape or paint and just set it there, until the heat shows up. Then tweak without getting scalded.
It is quite simple really. Remember how much you have to tweak it to get from too cold or too hot to where it needs to be (and make smaller adjustments or you will be forever overshooting).
When installing a shower, I ran the hot and cold pipes separately up to where the shower head attaches and joined them just before the arm to the shower head. A plumber would not do that - it uses more pipe, but I cut the delay to less than half.
So your hot/cold water controls are at the point where the shower head comes out of the wall?
Clifford Heath snipped-for-privacy@please.net Wrote in message:r
Is it considered to be a dead beat system?
Did that really change the thermal delays?
Mixing at the valve will increase the water velocity up to the shower head, which would be less delay. But the time involved is small either way.
It won't change the time of hot reaching the shower from the heater. It will greatly change the delay time for adjustments changing the water temperature. This time is not so small if you are not running the water at full blast. Once the hot water has arrived from the heater, I turn the water flow back, and step into the shower to get the temperature right. I thought this is what he was talking about. Standing in the cold water flow while waiting for the hot, seems to be a bit pointless.
LTspice offers a plumbing plug-in?
well yeah, that's the no brainer solution. But it doesn't utilize my advanced expensive math education, and I want to get my money's worth -
Which reminds me - you ever notice how often a problem comes up, and you end up patching together a Joe Blow common sense solution, and your technical training is worthless? You know, string and popsicle sticks -
I've used it for thermal design.
One issue is that water flow in a pipe is nonlinear, which is trickier to model.
I do like my cold showers. But this discussion isn't about that.
Yes exactly. It won't get you hot water any sooner at the start of your shower, but if you let it run until hot water comes out, any subsequent adjustment will be felt more quickly.
Smaller diameter pipes would also help, but would violate local regulations.
Perhaps you are not familiar with the relationship between water flow and electricity. I've used VHDL to model an analog filter before. It's all just numbers, no?
If that's what floats your boat.
My technical training applies well to electronics. Other areas it is only an analog. I've never loosened a nut on a bolt with a voltmeter. Maybe I need to get a Boltmeter?
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.