LinkMeter v2 Homebrew BBQ Controller - Part 1


 
Status
Not open for further replies.
Originally posted by D Peart:
It says I should install openwrt first, then install linkmeter, but I thought I also read that you had your own custom ROM that we should install instead.
Click to expand...
Now LinkMeter is included in the custom ROM but release 2 has a bit of a problem in that the build process messed up the order in which the packages are installed so you have to manually edit the conf file. If you install release 2
http://capnbry.net/linkmeter/release/2/

Then edit /etc/config/lucid and add to the bottom:
<pre class="ip-ubbcode-code-pre">
config 'daemon' 'lmserver'
option 'enabled' '1'
option 'slave' 'linkmeter'

config 'linkmeterd' 'linkmeter'
option 'serial_device' '/dev/ttyS1'
option 'serial_baud' '115200'
option 'rrd_file' '/tmp/hm.rrd'
option 'stashpath' '/root'
</pre>
And restart lucid by rebooting or `/etc/init.d/lucid restart`

I know it is complicated but it isn't supposed to be
icon_smile.gif
I didn't realize that OpenWrt when building a ROM installs things in willy-nilly order rather than dependency order which makes the lucid install erase the LinkMeter config.
 
Originally posted by Bryan Mayland:
Now LinkMeter is included in the custom ROM but release 2 has a bit of a problem in that the build process messed up the order in which the packages are installed so you have to manually edit the conf file. If you install release 2
http://capnbry.net/linkmeter/release/2/

Then edit /etc/config/lucid and add to the bottom:
<pre class="ip-ubbcode-code-pre">
config 'daemon' 'lmserver'
option 'enabled' '1'
option 'slave' 'linkmeter'

config 'linkmeterd' 'linkmeter'
option 'serial_device' '/dev/ttyS1'
option 'serial_baud' '115200'
option 'rrd_file' '/tmp/hm.rrd'
option 'stashpath' '/root'
</pre>
And restart lucid by rebooting or `/etc/init.d/lucid restart`

I know it is complicated but it isn't supposed to be
icon_smile.gif
I didn't realize that OpenWrt when building a ROM installs things in willy-nilly order rather than dependency order which makes the lucid install erase the LinkMeter config.
Click to expand...

thanks got it all up and running, well running is relative as I don't have a heatermeter.

dave
 
Originally posted by Bryan Mayland:
I've never been a fan of the blower driver circuit at all. I'd prefer using some sort of adjustable switching supply that puts out a (relatively) smooth voltage rather than the PWM junk we do now.
Click to expand...
I have no problem with the PWM approach to the blower speed. Varying the full torque duty cycle is a proven way of getting good speed control. I'm using Hruska's spec'd blower and the PWM is very smooth and controllable though the PWM is running at a 30 Hz rate, not 25 kHz or even 500 Hz. I don't know what effect that may have on things. I can detect a pulsing at low duty cycles, say 5-10%.
 
Originally posted by Bryan Mayland:
Yeah the design is supposed to be modular in that way. Some parts are optional depending on which config you're doing to be doing. I think the parts list is final. I'll be making a page in the wiki with the parts needed for each configuration if you want to wait a day or two at the most. I have a new MOSFET I want to try for a day before I recommend it... which I just noticed that since I've purchased it 2 weeks ago Mouser went from having close to 1,000 in stock now only has 24 and lists production delays. It was a 60A/60W part with a 0.012 ohm Rds(on) at 4.5V gate and it was $1.07. It was like my dream part! Whelp.
Click to expand...

Sounds good - I am probably going to try and add my switching circuit to the same board/schematic and put it on order later this week. From what I can find from the often used cheap board houses, there is no price difference between a 10x10 and a 10x5, so I have a lot of real estate to play with. Not sure if I am going to truly integrate it and reduce the number of connectors or just go with a minimalist approach and change as little as possible with your design.
 
Originally posted by Brad Baker:
I have no problem with the PWM approach to the blower speed. Varying the full torque duty cycle is a proven way of getting good speed control. I'm using Hruska's spec'd blower and the PWM is very smooth and controllable though the PWM is running at a 30 Hz rate, not 25 kHz or even 500 Hz. I don't know what effect that may have on things. I can detect a pulsing at low duty cycles, say 5-10%.
Click to expand...
It would be a good way to achieve good speed control on a brushed motor, but the brushless motor already has a sort of pulse controller built in it so turning that controller on and off is PWMing the PWM. We still get acceptable results but I'd say it is well short of ideal.
 
Can anyone find female right angle pin headers on mouser? Like these from Sparkfun except 5 pin. I used one of the Sparkfun 6 pin versions and just cut one off but it would be nice if I could get them as part of a Mouser order.

I'm trying to make a Mouser "project" that has all the parts in it but the female header situation is infuriating. I thought I had found some but they turned out to be SRRA when I got them. I'm not sure what that means, but it has pins coming out both sides. Freak show.
 
Originally posted by Bryan Mayland:
Can anyone find female right angle pin headers on mouser? Like these from Sparkfun except 5 pin. I used one of the Sparkfun 6 pin versions and just cut one off but it would be nice if I could get them as part of a Mouser order.

I'm trying to make a Mouser "project" that has all the parts in it but the female header situation is infuriating. I thought I had found some but they turned out to be SRRA when I got them. I'm not sure what that means, but it has pins coming out both sides. Freak show.
Click to expand...


OK I found this at Digi-Key
http://search.digikey.com/scri...Detail&name=S5441-ND

It is what we want, but I called Mouser and they don't have it. The best they can do it this:

http://www.mouser.com/ProductD...aYLpIVKuSOLWKQ%3d%3d

Which may or may not work depending on where it is on the board, due to the board clamp things.

dave
 
Originally posted by Bryan Mayland:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by D Peart:
OK I found this at Digi-Key
http://search.digikey.com/scri...Detail&name=S5441-ND
Click to expand...
Yeah! That's the exact part I looked for first. It seems Mouser doesn't carry anything from that manufacturer. They have 8,000 varieties of 0.1" pin header male with all sorts of lead and tail lengths and bends but not any female ends. </div></BLOCKQUOTE>

Can you get everything else from Digi-Key?
 
I believe I'm using the latest build. I have not had the lock problem, but since I changed the mosfet. The only problem I see is it will not turn on right away, and once in a while I still get the gibberish on the lcd, a good hit, lol and fine after that.
 
Originally posted by D Peart:
Can you get everything else from Digi-Key?
Click to expand...
It is possible, most components are pretty much "something like this part" in that they're pretty generic. I much prefer Mouser though, due to their non-idiotic BOM manager, their much lower prices, and a search that I am more familiar with.

The digikey site doesn't let you add things from a product page to a BOM project. You've got to copy and paste the part number, and sometimes it doesn't find the part it is really annoying.

Also as an example, an ATmega328P-PU is $5 at Digi and $3.50 at Mouser. I suppose 20% more for every part isn't a big deal when you're talking about such a small project though.
 
Originally posted by Ed Pinnell:
Ok, that killed it.
Click to expand...
Did you upgrade just the package or the whole firmware?

I'm sorry that the project has been stuck in its tracks recently. Seems I'm running into problems on every front and I just can't make much headway on them:
-- LuCI, while being an awesome extensible system that is easy to work with, is resource intensive and slow. Causes stability and performance problems.
-- OpenWrt installs packages to the firmware image in [unknown] order instead of in order of dependencies. Breaks the config file.
-- My slow HTML design skills and desire to make things look good makes every minor change into an afternoon-long affair.
-- I'm trying to fill out the wiki as I go but the different build options make talking about it plainly difficult.
-- I don't know hardware so when something doesn't work I'm always at a loss as to how to fix it or why it is even broken.

So yes it has been going nowhere recently but not due to any lack of effort on my part. The original HeaterMeter seemed to move much faster because it was already done before I started talking about it. This project has me struggling even more due to the complication caused by working in HTML, JavaScript, LUA, shell scripting, Make, and C++, Wiki Markdown, and hardware. I was really hoping that the project would at some point attract a professional web designer but I guess those guys don't eat BBQ
icon_smile.gif
 
Oh Ed, you've been a fantastic contributor to this project and even if you can't fix things yourself you are certainly skilled enough to track down the problem so I can figure out how to fix it.

I have a really hard time managing a project without having people in the same building as myself. It is hard to know what skillsets people have, what sort of time they have to work on things, or even whether I'll ever see them again. Things like the part list I end up getting my own set of requirements as to what parts should be used, considering cost, dimensions, tolerances and other things. I'd like to go with parts I myself have exactly used to remove some of the variability of things not working properly I can at least rule out some things. I feel that ultimately if something doesn't work the responsibility falls on my shoulders, mainly because I know it is a large time commitment to build one of these things and I know how frustrating it can be when something doesn't work.

Anyway, I'll just keep plugging away at it!
 
... all that AND Bryan/HM has helped me more than once to save countless hours of frustration on a totally different project - which I had to start form scratch because I have different hardware and could not comprehend the HM code yet. As I learn from the forum and the hands-on discovery with my project, it will help me take the wifi step.

Actually, the forum has been a great resource, because I find that other interest groups about BBQ do not have the HW and SW techie skills as are here. This thread is a great help to see what it takes to go end-to-end with something that is repeatable, modular and robust, a tall order for a DIY project that is electronic using multiple hardware and software components.

Not to add to the pile, but how is the starting time of day established in HM - just counting using millis ? Ever consider using a response from a time server using NTP ?

I have a time chip on my project, but would rather have wifi capability that would contact the NTP server, then have the Arduino count milliseconds after that.
 
Originally posted by Gary Graham:
Not to add to the pile, but how is the starting time of day established in HM - just counting using millis ? Ever consider using a response from a time server using NTP ?
Click to expand...
The Arduino doesn't track time of day at all. The timestamps which are in the JSON updates are added by the router, which gets its time from an NTP server. HeaterMeter doesn't have any concept of time beyond changing the fan speed every ~2 seconds.
 
Originally posted by D Peart:
Do you expect your boards back this week?
Click to expand...
I thought they'd be in last week actually, based on the previous turn-around time. It isn't guaranteed though so hopefully soon. It's just as well, as I had no time this weekend to even think about HeaterMeter aside from using one to make 11lbs of pulled pork for Sunday's football game.
 
Speak of the devil, look what was waiting in my mailbox when I got home from work today. The CMOS shift register in the latest schematic didn't work for the LCD though, so I've switched back to the original recommended TTL shift register. Everything works (LinkMeter, no RF variant) from probes to blower to buttons.


You can see I didn't populate the 3.3V power section nor the RFM12B pad. The barrel jack probe is also absent as it is only suggested if you're building a standalone HeaterMeter. You can see the bull**** color bands used by KOA Speer's 1% tolerance resistors. Red looks identical to Brown, Orange is what Red should look like. I'm not happy with:
-- The placement of the FTDI connector, the cable doesn't seat all the way because it hits the edge of the reset button and the 25V cap is really close.
-- The placement of the 25V cap makes it tough to pull the ATmega if needed.
-- BC337 kinda gets in the way of just testing an LCD by placing it on the 16 pin connector.
-- My ceramic cap under the 7805 should be a MLC because an old school ceramic is so big it makes the LM7805 not able to lay down
-- I'd prefer 6.3mmx7mm caps instead of the 11mm tall ones I used
-- The silk on the back was mirrored. The positions were right but the lettering is backward. Fixed in git.

That said, it works and the latest EAGLE files are in git. If I have time tomorrow I'll list the materials I used.
 
Status
Not open for further replies.

 

Back
Top