HM expansion, can it do more than 4 temp sensors?

Hi, I've been researching many temp controller possibilities and have looked extensively at what you guys have done with the HeaterMeter (Kudos!). I've read through many of the posts as well. One thing I haven't figured out and is likely a stupid question is why the decision was made (Bryan) to use both an Arduino board and a Raspberry Pi instead of just the Pi itself? Since the Pi has a number of GPIO pins couldn't the same thing be done with just the Pi? What were the advantages of the Arduino? I've seen just a Pi used albeit with only 2 sensors. I would think this would give you the advantage of using the HDMI out also to display your temps on a secondary screen if you wanted.

I'm asking because I'm looking for an all in one temp controller solution for a large smoking trailer. It's got 2 cook chambers, one of which is 6' wide. I'd like to have a pit sensor in 3 places and several meat sensors (the more the better) for all of the large cuts being smoked at the same time. I've seen where someone here built an expansion board to the HM but it didn't seem to add any additional temp sensors. What would have to be done to wire up more temp sensors (3 pit sensors, 4 meat sensors and the blower at a minimum?) Right now the Stoker is the only solution I've found that can do it all but at $600+ with all the expensive proprietary sensors and expansion adapters. The HM is the way I'd prefer to go but would need to figure out how I can get more probes on it. Thanks for any input!

PS. Yes, I know the blowers work better for charcoal than stick burners but I'm working on a solution for that as well. The blower would be controlled by only one of the pit sensor temps.

The HM pre-dates the RasPi, so that's one reason. The RasPi GPIO is all digital, not analog. The temp probes are analog and so you need an ADC to interface with the Pi. The arduino is doing all the heavy lifting and the RasPi provides the IP connectivity, web interface and email alerting.

The HM can only have 1 active pit probe, so if you're looking to average 3 probes to control the pit, that's also outside of scope of the HM's functionality.

You could always use 2 HMs, giving you 1 pit control probe and 7 meat probes.

The Pi may have lots of GPIO but still lacks an ADC so you'd have to add one. The cost of an SPI ADC is roughly the cost of the ATmega so you're not going to save any money there. What you lose is real-time data processing. There would be no way to do A/C line noise removal reliably on the Pi, because the sampling of the ADC needs to be at exact fixed interval, something you can't get on off-the-shelf Linux. You'd also have to rewrite a lot of the HeaterMeter source code which is non-trivial. I'm also not sure what the PWM capabilities of the Pi are, if different pins can have different frequencies, which we use.

HDMI output, I would find it to be a highly irregular configuration if there is a TV within cable-reach of the grill. And what, dedicate an entire television/monitor to see the current temperature? Seems like a waste. If that is desired, just have something plugged into the TV that displays the website. If you mean put a TFT on the device itself, maybe, but man wouldn't it be great if you didn't have to actually go to the device to see what the temperatures are? Holy smokes! You can have a TFT LCD display anywhere via wifi and see the temperatures without having to be tethered to the device itself. I feel like a dedicated tablet is A) cheaper than a color LCD B) portable C) doesn't require maintaining a "thick client" which is Linux GUI app and also running an fbdev or X11 to display it.

As far as adding more probes and multi-control, the internals of HeaterMeter are pretty tied to using 4 probes (and all the ADC pins), as well as the database all the data goes into, and the webui is not flexible with regards to how many probes it displays. One could theoretically design a web interface that supports an arbitrary number of probes, but I feel like you'd end up with just a table of numbers which is just sort of generic looking.

Thanks guys. So maybe I should have prefaced by saying I'm not looking to recreate my own solution. I'm looking for the best solution to accomplish several goals without breaking the bank (and I know there are others like me.) @Steve, 2 HMs has already been considered but with all the probes and the bigger cfm fan thats already required its getting pricey again. Plus as Bryan says, "the webui is not flexible with regards to how many probes it displays" and I'm not sure how efficiently I could flip between 2 HMs on one iPad. I've soldered before but lack the time and confidence to put my own board together instead of buying it assembled. I figured the A-D conversion was the issue and what I'm getting from Bryan's first paragraph is you can't get the accurate sampling needed from the inputs using the Pi alone and the programming in Linux is trickier. There is an 8 port, Pi based WLAN controller on Hackaday that Armin did but the how-to is all in German. It doesn't solve some of my other problems but was giving me hope (and inspiring my original question) about getting more ports out of the HM. Do you think his temp accuracy is way off then?

I should have also been more clear on the HDMI and how I got there. One of the things I'm looking for is a big, bright local display so when I'm standing at the smoker I can quickly see temps without having to depend on the analog gauges or use a sun-washed out (and possibly greasy) iPad. It would also be nice to see it from a short distance away if I'm sitting nearby the trailer. Some controllers like the QMaster accomplish this with 7 segment LCDs and I looked into larger 1" - 2" versions of those panels. But ~$15 multiplied times 7 or 8 temp inputs I realized it would probably just be cheaper to display it in a web page or something on an old LCD I've got laying around. So as funny as it sounds, yes having a large display of the temps that I could quickly glance at on the smoker (like you'd see in any restaurant) is appealing to me without having to pull out my $500 iPad or phone and get it anywhere around a dirty trailer. The washout of the LCD in the sun is a concern however so it's not a solution yet. Still like the idea of large segment LCD panels. I don't like the idea of scrolling through 8 different temps. (Already I'd be looking for a 4 line LCD on the HM at the least.). BTW, I've used an older iGrill for a couple years so these are all issues I've experienced.

Admittedly my use case for this is different than what most people in this forum are looking for. You built your device around a small Weber and I'm looking for something on a large smoking rig to use for overnights on a larger scale with several meats. You've done an amazing job which is why I'm hopeful it could be used for my needs as well in some form. So my primary goal is trying to find out if anyone has done or is working on a way to expand the HM beyond 4 temp ports. I've already read you aren't doing any 4-line LCDs yourself Bryan but there are some out there. Is it possible to go larger than that? 2 x 4-line LCDs? Larger panels? Thanks again for all the input!

Steve_M said:

The HM can only have 1 active pit probe, so if you're looking to average 3 probes to control the pit, that's also outside of scope of the HM's functionality.

Click to expand...

I wouldn't want to average, just use the temp from one probe to control the blower. Likely the one in the middle or wherever I wanted the pit to be 225.

Android tablets are cheap and plentiful these days. I picked one up for $70 Canadian a couple of months ago. You can easily display multiple HMs on a single webpage through the use of HTML iframes.

ie:


or

Interesting. So this would work on a WLAN but not in ad hoc mode right because you'd have to be able to connect to 2 different networks. If I were cooking away from home I'd have to set up my own little wireless (non-internet connected network.

John C in Houston said:

Interesting. So this would work on a WLAN but not in ad hoc mode right because you'd have to be able to connect to 2 different networks. If I were cooking away from home I'd have to set up my own little wireless (non-internet connected network.

Click to expand...

In the away from home scenario you can just bring a wireless router with you and connect all devices to it.

John C in Houston said:

I figured the A-D conversion was the issue and what I'm getting from Bryan's first paragraph is you can't get the accurate sampling needed from the inputs using the Pi alone and the programming in Linux is trickier. There is an 8 port, Pi based WLAN controller on Hackaday that Armin did but the how-to is all in German. It doesn't solve some of my other problems but was giving me hope (and inspiring my original question) about getting more ports out of the HM. Do you think his temp accuracy is way off then?

Click to expand...

The quality really varies by the source power. You can get a tremendous amount of feedback in the circuit from a ground loop + mains hum which we can cancel out, but some outlets work perfectly so you never really run into this problem.

The washout of the LCD in the sun is a concern however so it's not a solution yet. Still like the idea of large segment LCD panels. I don't like the idea of scrolling through 8 different temps. (Already I'd be looking for a 4 line LCD on the HM at the least.).

Click to expand...

I can totally see that. What would be badass is a bank of 7-segment LCDs driven by an ESP8266, the ESP can just connect to the HeaterMeter and stream the data and update the bank of 7-segments in real time. That would be portable too so you could take it inside with you and stay connected. I think that's a really cool showy idea and now I want to build one of these too.

The easiest way to get this all connected would probably be a portable wifi ap, like Ralph suggested. It wouldn't be too difficult though to get one HeaterMeter running in AP mode (not ad-hoc) and then have a second HeaterMeter connect as a client to it, then everything would be on the same wireless network with the AP device being at 192.168.201.1 and the client being 192.168.201.x.

@Steve_M: Love that side by side light page iframe setup!

EDIT: Some quick shopping around found me for a 4 probe 7-segment LED display you'd want:
1x HT16K33 16x8 LED matrix driver $2
1x ESP8266 $3-$5 depending on form factor
4x 0.52" 4 digit 7-segment display $1 each

So under $15 and you can power it from a 5V USB battery pack or a 1A USB charger. Somebody make this!

Bigger! Glad you're seeing the benefits of what I'm looking for now. But I'd want to see it from 15' away. Your pricing looks better than my quick search, what would the cost diff be for 1" or 2" tall elements? But having a bank of them was exactly what I had in mind. I pictured 8 of them with 1 main pit temp in one color (i.e. red) that was used for the blower, 2 other pit temps (orange) and 5 meat temps (green, etc), all in 3 different colors to distinguish. (But I guess that would all be up to whoever was using it.) Wireless is even better which I hadn't considered as long as it didn't interfere with the iPad app.

Yeah the 1" digits make a huge difference. The 1.2" size are over $1.50 per digit and are hard to come by on eBay in all the colors. Finding them in Mouser means going through 15 pages of options because they don't have the size properly indexed, but they're probably $5+ given that they have 0.56" digits for over $3 (and the paper catalog only shows up to 0.56" units). Looks like 1.8" are a little better on eBay at just over a dollar a digit, but again, not many sellers.

hehe if you wanted to make a PCB that held all the 1.8" digits for 4 probes, that would be $126 just for the set of prototype PCB at OSH Park. Can someone say wirewrap?

Now you see how HDMI out to an old LCD monitor was conceived. But $1.50 x 3 digits x 8 probes rolleyes is only $36. What other parts are needed, driver boards, PCB, etc? Even better if the 1.8" LEDs are closer to $1. It still seems doable but I guess I'm missing some things.

I really like the idea, I've thought about it during many cooks. I even made a mock up of what I was envisioning, albeit I was imagining 2" LEDs. Geeking out, I know.

In the picture (and for my personal needs anyway) these would be:

Meat probe 1 Pit probe 1 (main chamber closest to firebox)
Meat probe 2 Pit probe 2 (main chamber other side, controls blower)
Meat probe 3 Pit probe 3 (secondary "ribs" chamber)
Meat probe 4 Meat probe 5



Hmm, image didn't post, will try direct link.
https://drive.google.com/open?id=0B0tPX4ce1Gp2YmVWb015NW5FWXc

Apart from the parts I listed, you just need something to wire it all into. If you buy individual digits, remember you have to run 9 wires to each digit so it becomes a bit of a wiring nightmare. You need one HT16K33 per every 16 digits, which is why I'd go with 4 digits per probe. And because HeaterMeter only supports 4 probes, I'd say make a unit that is 4 probes, 4 digits each, with one HT16K33 and one ESP8266 to drive them. Then run two units, and two HeaterMeters for 8 total probes.

A tablet and a bit of html/css tinkering to the lmlight page to have the numbers display in whatever colour you want is still probably the most direct approach, and if your $50 tablet breaks, it's way easier to replace than the franken-LED display!

You may be right about keeping it at 4 probes and using 2 of each. Starting to see that benefit for simplicity if the HM is locked to 4 probes. I guess I could still make a housing that combines 2 displays into one unit. Or make 2 housings that link together and share power.

I know I saw larger LEDs already assembled on boards to avoid the wiring nightmare. I'll look again. The 4th digit seems extraneous since you'll never get a smoker that hot. I guess you could space it away from the other 3 a little and use it as a label, maybe in a different color as the rest.

1 225
2 215
3 182
4 199

So all I see for parts for one 4 probe unit are:
1 x HT16K33
1 x ESP8266
1 x PCB
4 x 7 segment LEDs
OR
1 x 4 digit, 7 segment LED board
1 x power supply
1 x case

What am I missing?

@Steve, where's the fun in that?

Besides, I think it can be done for cheaper but it'll take some parts hunting. I can barely see my iPad in the bright sun much less a $50 tablet.

Oh I was thinking you'd want 4 digits to include a decimal. Seeing something at 159 for 10 minutes is far less exciting than seeing 159.3 for just 1 minute before it changes to 159.4! But yeah the the two housings could be joined together so it looks like one unit, nobody would know that there's actually two separate systems inside there. I didn't see any of the larger 7-segments that came as 3 or 4 digits without them starting to be like $15 a set which turns it from a $20 project into a $60 project (or $150 for 8 displays).

You still need to write all the software to make it work, but it isn't complicated. The easiest way is to poll the HeaterMeter's status, but I'm sure there's a way to make a quick streaming client as well. Then you need to make a little webui on the ESP to be able to configure what it connects to and you're all done, right?! I would say this is a non-trivial project in terms of time but everything is pretty much off-the-shelf in terms of software libraries and hardware. It just needs to be all linked together.

Ha, I don't need to watch a smoke that closely to see the decimals change, the idea is to have to watch it less. I'll look back where I found the large digits but yeah, they weren't super cheap. I just realized if I used 2 HMs I'd need 2 RPi's too (and 2 wifi, 2 power, etc) wouldn't I? Now the cost is getting up there again. Ugh!

use rPi zero-w, really cheap and complete with wifi.
You could also probably get away with running the second unit headless (no LCD board) which would cut the cost a bit more.

That said, it would seem that a HM "slave" unit configuration might be possible, one HM transmitting to another via digital I/O, though Bryan probably weighed that option already? Seems like that would take a LOT of coding to create, the communication protocol, the gui redesign etc...