theory: f(x) = x
practice: f(x) = -8 + 16/(1+e^-10x)
Maybe it’s just mine but can we agree that a lot of showers do that?
A lot of showers need a new mixer valve...
this post was submitted on 02 Mar 2025
289 points (98.0% liked)
Memes
47737 readers
1866 users here now
Rules:
- Be civil and nice.
- Try not to excessively repost, as a rule of thumb, wait at least 2 months to do it if you have to.
founded 5 years ago
MODERATORS
I have this issue in a home built just barely pre-lockdown. So the issue isn't wear and tear.
What should I be looking for to fix this?
Sometimes even new(er) valves fail. If you have hard water then that is plenty of time to mess one up. In a best case scenario, you can lookup your make/model of faucet to buy a 'cartridge' for it and find the instructions on how to replace that. In more extreme cases / poor designs, you could have to tear into the wall and replace the entire thing.
You might also see if you can find a plumbing community and post as many details as you can... I'm an experienced DIY person who has done a lot hands on and read a lot, but I'm not a plumber so I still have blind-spots.
Based on the sounds from my computer, that's exactly the same curve the fans on my GPU and CPU uses, except for the X axis being temperature starting from 0°C going to 100°C and Y axis being fan power in percentage.
I think the problem there is that GPU temps are spiky themselves.
True. But I also tested when my GPU fans would turn on and it seems like the cut-off point was 45%, below that and they'd just stop completely. And normal idle temperature is around 40°C, and with the curve on the left it makes sense that even a 5°C increase would rev the fans up from 0% to 45% making it sound like a jet fighter about to take off.
If you're on windows, look up the program/github fancontrol. Nice easy way to adjust the settings, set custom curves based off different temp sensors (and combinations of temp sensors) all without having to muck around in your BIOS.
I'm on linux, switched last year. But thanks for the recommendation!
Did they actually stop spinning below 35% or did the software say that 45% was the “minimum” for them to turn on? My GPU will happily start its fans at about 20%. The pwm readings aren’t right but they’re spinning just fine.
Also try separating your fans if you can. My GPU has 3 fans. 2 are fan 1 and 1 is fan 2. I set fan 1 to kick on pretty early, then fan 2, then fan 1 ramps up, and then fan 2.
I have 2 fans on my GPU, and I can control the curve through CoolerControl on Linux. I've also looked at LACT which has GPU fan control, though a bit simpler. I kept the services separated, so I could test each without them interfering with each other, but I have to say, when I woke up today, that's what I thought was the problem, but no, after some testing I can see that it's just that my fans, perhaps due to firmware, just doesn't spin up unless it's above 45%.
I looked around to see if others were having issues, and this github issue says that Nvidia API caps it at 30%. Maybe it's capped at 45% for me on linux for some reason? I'm not too fussed about it, I've just made a curve that kicks in around the time I need it to.
CoolerControl
LACT
EDIT: I did see the fans try and kick in around 40%, that's why there are those spikes on the histogram on the left, that's me slowly increasing from 40% to 41%, to 42% etc. Was only stable at 45%.
Is there no fan speed hysteresis setting? On msi afterburner in windows I have like a 3 or 5 second hysteresis set, but my car has a HUGE heat sink that’s way overkill. On a smaller card I might do an even longer one.
fan speed hysteresis
As a noob on these things, I had to rummage around to find those settings, and yes, they do indeed exist.
I've changed it to 5 seconds, hopefully that'll calm things down a bit.
I greatly appreciate your comment, made me aware of what I could improve. I wouldn't even have noticed those settings if I hadn't looked up hysteresis, it's not a word I think I've ever come across before heh.
Get a thermostatic mixer valve. You'll never go back...
Will think about it ^^
Remember when those boring old two-tap showers never had this problem?
Hey now I have a new two tap shower... And still don't have this problem. I think that's just the solution, 2 taps are better than one.
You can plot my shower in three dimensions with water pressure.
The hotter the water, the weaker the pressure.
I drop the pressure, the water gets hotter :)
Oh god, same.
I have nearly figured out how to increase the pressure without changing the temperature.
I still boil/freeze myself sometimes though.
You could try restricting the shower head to a smaller output area. This way you would have a higher water flow
Ah, I meant increasing the pressure when I notice it's a bit too low for my liking.
My shower lever has to be pulled in a specific diagonal direction to increase pressure without causing the temperature to change drastically.
Turn your hot water heater down or replace your shower valve
No its just that our ideal water temperature is very very narrow
f(x)=x? In theory the water in a shower can get infinitely cold? That would be some shower that can go past absolute zero. It would be interesting to shower in a Bose-Einstein condensate.
If i remember correctly thats also close to hyperbolic tangent
Thats what i used in simulating very fast semiconductor switches so it wouldn’t cause solver issues. Might be better ways tho
Your pain receptors fire at some fixed threshold so anything beyond that very suddenly gets uncomfortable.