1st cook in 3 years


 
I just had to post, between kids and work, it's been over 3 years since I've cooked but staying at home has its advantages. I built this HM back in 2014/2015 time-frame iirc and it fired right up. I updated the code and working flawlessly. Ribs for dinner!

Thanks all!


IMG_20200409_121058.jpgIMG_20200409_121120.jpgIMG_20200409_142254.jpg
 
Heck yeah man, do it to it! I tell you, I have never eaten as much pork as I have in the past few weeks. How can I avoid going outside to the grocery store? Hrm, 7lbs of pulled pork should do it, but let's make a few racks of spareribs as well. Freezer = full.

I love seein' that rare 4 line display!
 
Lol, I know you did that just for me. Everyone I ever built had 4 lines. With ll the graphics horsepower of the RasPis, you should hang an LCD off one.
 
Lol, I know you did that just for me. Everyone I ever built had 4 lines. With ll the graphics horsepower of the RasPis, you should hang an LCD off one.
I did! Over a year ago I built a HeaterMeter with a 3.5" TFT LCD, built a little graphics library, touchscreen input and calibration transformation system, then set to work on the UI... aaaaand it was the worst. After getting so accustomed to the gorgeous displays cellphones have, a 320x240 display with extremely limited viewing angles, low contrast, and barely enough brightness to be used outdoor just was a real letdown. Most of the UI would get lost in the lack of contrast.
DSC06986.JPG

One problem is that most small TFTs are 240x320 and designed for 6 o'clock viewing so the UI needs to be portrait instead of landscape lit is is here. If I look from the left side of the screen it doesn't look bad, but starting at center it looks not good, then 5cm to the right it looks bad, and at 10cm everything except the contrastiest bits start to fade away. I don't know what a HeaterMeter PCB would look like for a portrait orientation since the Pi connectors shouldn't be up (to catch rain) or down (unusable). Then of course there's writing an entire new UI paradigm while also creating a visual style that's beyond my photoshop skills. Switching to this display type is more work than I can wrap my head around, it is just an immense project and would most likely require hiring a graphic designer to do a hundred of hours of work which is sort of outside the budget.

We also can't use any of the graphical horsepower of the Pi because this isn't connected to HDMI. I spent a lot of time searching for CSI interface LCDs (the connector the pi has for an external LCD) and never found one under $50 for the smallest one and it went up from there. An HDMI-capable board would be similarly expensive in the $40 range. HeaterMeter BOM cost=$30, Pi and display and sd card=$100? Seems a bit out of balance!
 
What about e-ink?
They're pretty expensive too. That one looks neat until you see that it takes 15 seconds to update the display which is unsuitable for something updating every second or more. I think all eInk displays have a very very low refresh rate, like a second or more for the larger panels. The TFT LCD even had animations (just some simple movements during touch calibration) so it would be nice to have a display that can keep up. I'd like an LCD to actually improve the the UI experience rather than just change one crippled display type to another crippled display type and taking hundreds of hours of work to make it happen.

I bought a whole bunch of different TFTs in different sizes and color depths and touch systems from a Chinese display manufacturer to try them all out and all of them were pretty disappointing. A lot of that has to do with the price point I'm trying to target to keep a HeaterMeter around $100 or less. The display is rarely used compared to the webui so I don't think it should cost more than all the other parts combined just so it looks neat when you power it up, once the display actually starts displaying. Another benefit of the character LCDs we're using is that they come on in milliseconds from power up and a TFT has to wait for the whole Pi to boot, all the services to load and the renderer to take over the screen. Something on the order of 30-40 seconds best case, which stinks if you just want to plug in the HeaterMeter and set a temperature.

I also looked at the smarter displays that have their own microcontrollers and scripting languages which were pretty nice but are definitely targeted toward major manufacturers, with toolchains costing thousands of dollars or not even listing prices on their website "Call us to set up an appointment with a design service engineer!" which is 😣. Their devices weren't cheap either. Itead has some neat cheaper panels ($25-50) with that sort of technology too, but it comes back down to not getting any more out of the UI for a much greater cost. I could see putting a nice 5"-7" touch panel on a device designed for indoor use but I fear that leaving these things outdoors overnight all the time is going to seriously compromise their lifespan as well.

EDIT: But yes the SHINY!! I do think they have an appeal that makes me want a HeaterMeter with a beautiful TFT display.
 

 

Back
Top