The Development Log


 
Bryan,
I installed the latest version from 6 november. I don't know if the problem is in the firmware or my own laptop. but since this release the web GUI is not updating in the Chrome browser. so I have to keep pressing f5 to get a new temperature value and graph. the time on the top right side of the webinterface stays the same. it is not turning red. until I press the refresh button (f5) and then the time changes, the termperatures update and the graph updates.
So I don't know if the problem is in chrome or the HM....
 
Last edited:
Try running Chrome in incognito mode. If that works, you have a Chrome plugin getting in the way. If that doesn’t work, try with a different browser.
 
Steve_M said:
Try running Chrome in incognito mode. If that works, you have a Chrome plugin getting in the way. If that doesn’t work, try with a different browser.
Click to expand...

In incognito mode it is also not refreshing. in Internet Explorer there is no problem.
 
I can't reproduce this. I use Chrome as my main browser and it works on multiple devices (windows / Android) and HeaterMeters for me here. Is the time off between the two devices? If so that can make it do weird things or appear slow to load. You can also pop up the F12 developer console and see if there's any errors either in the console or network tabs, but it does not appear to be a global issue with Chrome and HeaterMeter.
 
Not a lot of updates recently, I've been in and out of town every week since the beginning of November and I'm starting to feel like I'm on crazy pills with how much time I am spending traveling. It should all calm down after the New Year though and I can start to get my head on straight again.

There's new AVR Firmware on the online repository. I had a user who had done a reset config on the device and ended up bricking the HeaterMeter. It wasn't broken but the firmware had gotten itself wedged to the point it wouldn't boot the AVR any more. I was able to reproduce something similar and put two and two together to find a solution. I've replaced the generic EEPROM writing code from Arduino with the suggested implementation from Atmel which is faster, takes less power, and reduces wear on the EEPROM cells vs the standard implementation.

I've also modified the RF12 wireless implementation so it doesn't hang indefinitely on startup waiting for a response from the RF module which will never happen if there isn't one present. Now it tries initialization at boot time once (at the 8ms mark, where the RF module may not be out of POR yet if power has just been applied) and now will just check later to see if it can initialize and keeps checking until it succeeds (usually less than 100ms after power is applied). This only is used if any probe type is set to "RF Wireless". Note that previously, if you mistakenly set your probe type to RF and did not have an RF module, you'd just be screwed because your HeaterMeter would never initialize. Now it is fine with the module not being there.

I've also added a "reset-eeprom.hex" file to the online repository which performs an erase on all the EEPROM area we use in case there are other ways to get the AVR stuck on initialization that I've missed.
 
Really (really) minor, but looks like the links to PushBullet and PushOver need a tweak

tuABJZc


https://imgur.com/tuABJZc

PushBullet link resolves to : http://192.168.1.220/luci/admin/lm/www.pushbullet.com
PushOver link resolves to : http://192.168.1.220/luci/admin/lm/www.pushover.net
 
  • Home page in the webui now shows the current control probe temperature in the browser's window title. I'd like some feedback on this on if it is annoying or a nice feature.
  • Added exFAT filesystem support for USB devices. Larger USB sticks are usually exFAT formatted and it would prevent the wifi autoconfig from actually taking effect when loading from USB if the device wasn't formatted FAT32/VFAT. Now, FAT16/FAT32/VFAT/exFAT filesystems are all supported.
 
Not strictly development related, but I was looking at Texas Instruments new line of DC-DC converter modules, the TPS82130. It would replace the OKI 5V converter for dropping the 12V down to 5V for the Pi and to feed into the 3.3V regulator... but it is 2.9mm x 3.1mm x 1.5mm. The Murata unit we use is literally 107 times the size of this thing, and this can put out twice the current (3A) peak, or 2A continuous. Clearly it must cost a fortune then, right? Nope, 10 cents less per unit at quantity 100 from Digikey. My mind is just BOGGLED how this is possible.

We're living in an amazing time, friends!
 
COOL. I just bought the entire Digikey stock. (just kidding).

i do see there are extra components to add (caps and resistors) to maintain the correct 5V output based on input.
 
Last edited:
Yeah they're too small to use as a component in a through-hole HeaterMeter. The current module has the advantage that it won't disappear if the exhale too hard around it. As for the extra components, we already have caps on our input and output, so you'd just need to add the soft-start capacitor, the two resistors to set the output voltage, and then I'm not sure the pullup is needed on the power-good line. Those would be less than 10 cents total in quantity. You just can't beat that much power at that teensy size though, it doesn't even compute how it is possible to build a 3A inductor into that package.

I use these little mini buck converters in all sorts of 3-5V projects that I'm powering from 12V (because I have a ton of those) although I just noticed that it is only rated for 23V and I just used one yesterday to get me 5V from 24V. It is only a matter of time before we start seeing even smaller ones for 36 cents a piece. They'll have to mount it on a big giant blank PCB just to give people room to solder onto it.
 
Bryan Mayland said:
Not strictly development related, but I was looking at Texas Instruments new line of DC-DC converter modules, the TPS82130. It would replace the OKI 5V converter for dropping the 12V down to 5V for the Pi and to feed into the 3.3V regulator... but it is 2.9mm x 3.1mm x 1.5mm. The Murata unit we use is literally 107 times the size of this thing, and this can put out twice the current (3A) peak, or 2A continuous. Clearly it must cost a fortune then, right? Nope, 10 cents less per unit at quantity 100 from Digikey. My mind is just BOGGLED how this is possible.

We're living in an amazing time, friends!
Click to expand...

Dang, that's quite the find! On the other hand, it's super small size will present a problem for those who might already be struggling with doing SMD work.
 
You must log in or register to reply here.

 

Back
Top