New Combustion Inc. Wireless Thermometer: first impressions

[glitchy]

Only have one cook down so far, but I’ll share my initial thoughts. The hardware does seem very well constructed. I love how much skinnier this is than the Meater+. Setup was easy, updates even easier. I didn’t test calibration, but it seems to be dead on. It was pretty cold out, so It was nice to put the probe in my Prime Rib inside and not stand outside fishing wires through the smoker and setting up a base unit. I have no reason to question their multi sensor algorithm, it at least initially appears very accurate.

When the Combustion Predictive thermometer was first announced, I was a little concerned about it being Bluetooth vs WiFi. I still have concerns around it, but not because of technical issues, but what that means for remote connectivity and being able to run an errand and check in on the cook. That being said, this being Bluetooth definitely made it the easiest 'smart' thermometer I've every used for setting, updating and getting cooking quickly. The reality is I rarely leave the house when a grill is running and still usually make sure someone else is home.

What’s needs improvement: With all of the waiting for this thermometer, I had built up huge expectations. I understand the supply chain issues the entire world is having and was not upset by the hardware delays. I just sort of expected that with the delays the accompanying mobile apps would be pretty robust. I feel let down by the over simplicity of the initial iOS app. Combustion claims on the website they left it simple on purpose to start. I don't necessarily think that makes sense with the type of people that pre-order a super expensive new thermometer. If they want to offer a simplified interface for dummies, make it an option in the interface to choose. I want graphs, histories, etc. I personally really want it day one on a premium product delivered half a year late. There also seems to be no notifications in the app. It tells you to enable them, but they don’t seem to be hooked in to be enabled nor functional. Many of the features, I was hoping for day one are on the short list they've already posted on their website as upcoming, so I probably just need more patience (but I always struggle with that).

Pics attached are pretty much the entire app, minus the screen that shows you the firmware versus of each probe and the base. You can see in my captures, I did not get any predictive information, just a percentage done on my first cook. Chris explains a few posts down how they are working on that.

[Donw]

Personally I’m glad they are taking it slow on all their features. Too many companies over promise and under deliver loading up their apps with features not well thought out and full of bugs. Let them make sure the basic functions are rock solid before adding other things. Just my 2¢.

[RonB]

Thanks - I may be able to squeeze a cook in tomorrow, but not really sure.

[BBQPhil]

The poor app is a showstopper for me. The Fireboard app is worth more than the hardware from my experience. Would be hard to go backwards.

[STEbbq]

Planned updates on the way:

https://combustion.inc/pages/product-release-notes

Planned features

Among other things, the team at Combustion is currently working on the following for early 2023.

• Faster/more stable instant-reads.
• More robust Bluetooth connectivity and better handling of temporary disconnects
• Carryover cooking (resting) predictions
• Improved battery life when in instant read mode

Planned features

• Graphing: this is very cool and kind of hypnotic, trust me

• Data exports: track your cooking results

• High/low temp alarms

• Multiple temperature alerts: nice for multistep cooks like reverse sear

•Signal strength + management

• What's cookin' where ID (working title): label a cook-in-progress with a photo

[STEbbq]

On the one hand glitchy, I totally see your point.

On the other hand, it seems like the supply chain challenges were pretty extensive. I have to imagine it is a pretty small team at this stage so maybe this was the best they could do with the time that they had. They have struck me as really transparent and honest thus far so my impression is the situation was far more challenging than they originally thought. They are doing their best and have huge ambitions.

Also, I have never preordered anything before. If the plan is to essentially take the app to amazing within a few months, that sounds pretty compelling to me. That does seem like the plan and I guess I haven’t see anything that would hinder app development at this point now they are shipping.

I guess if the app is bare bones in six months, I’d be more upset.

Of course, I am coming from a bare bones Inkbird so this initial app is about all I could do anyway!

[glitchy]

Why would supply chain issues affect app development? They’ve had working prototypes of the thermometers for probably years. In the year since we ordered, they’ve been testing tolerances and designing the hardware production process. Usually, the people writing apps are not the same team that is making the hardware.

[STEbbq]

Yeah, I have no idea how the software side works and I think you work in tech (?) so it is more of an area of expertise. If so, this is a valid criticism for Chris and team.

[STEbbq]

So I do agree it is a fair critique. Hopefully everything will be upgraded in a few months and we will get our money’s worth.

[chefchrisyoung]

Hi Folks, Chris Young from Combustion here.

I understand the disappointment with the state of the mobile apps. They will get better, but it will take some time. Combustion is a small, bootstrapped team and I have to make some hard choices on what we prioritize. The delays from the supply chain problems consumed most of our time and financial resources this year.

But there is more to it than just that. From my experience building the Joule sous vide circulator, I am deeply aware that maintaining a lavish, feature-rich app requires constant effort of a team of software engineers. ChefSteps spent close to $2M per year maintaining the Joule app and cloud connectivity services. Once Breville took over and reduced the cost by reducing engineering resources, the app quickly became very buggy. It really annoys me how glitchy the app has become, and it's a major goal of mine to avoid this for Combustion. Doing this means being very thoughtful about what we build and what we don't build.

I've choose to prioritize a few things:

(1) High-reliability Bluetooth connectivity that isn't laggy and (as much as possible) just works.

I'm sure this seems trivial, after all lots of folks do Bluetooth! But achieving high-reliability across different phones, with different Bluetooth chips, and running different versions of their operating system can quickly become a PIA. I think we've made some very good decisions on how we architected this, and it generally it takes less than 1 second from removing the thermometer from its charging sleeve to seeing temperature data on the display or with the app. And this all happens without the need to "pair" your thermometer, which is a major source of failed IoT connectivity.

And the way our Bluetooth architecture works, additional devices (some not yet announced) will add to a mesh network (we call it Meat Net) that expands the reach of your thermometer so that you have fewer dead spots and longer range.

(2) Useful cooking time predictions. This was by far the most ambitious thing we set out to do, and there is still a lot to improve. For example, we've seen that the Prediction Engine gets confused by low-and-slow cooks where the heating rate is very slow. Lots of folks ran into this on Christmas when doing very low-temp reverse-seared prime ribs.

We biased the algorithm against making stupid predictions, so in this case it just display a progress indication when it gets confused. That's not the experience I'd like folks to have had. And it's unacceptable for barbecue! We're already working on a fix and hope to get that out in a couple of weeks. There is quite a bit more to do here, but this is also pure research so it's very hard to know how long it will take to get things where I think they should be.

(3) High accuracy and generally fast temperature instant read measurements. The core temperature sensors (first 4 sensors) achieve an accuracy of no less than 0.2°C as measured against some NIST traceable calibration equipment, and they have swift response times. But we've now seen that the instant read algorithm undershoots or overshoots on some thermometers. This is down to small variations in tolerances during manufacturing. I spent a lot of the last month gathering a bunch more data from a larger number of production thermometers so that the thermometer can self-tune its instant reading algorithm with the goal of sub-3 seconds instant read responses.

(4) A well-documented, easy to implement open source communication framework. This is really important so that other folks (individuals or companies) can integrate support for our thermometer. I'm pleased to see that someone has ready developed an Apple Watch app, which is cool. And there will be more to share in 2023 about appliances supporting the thermometer.

Doing these things has given us a rock solid foundation to build on. But this meant that the apps couldn't be a high priority. I understand that after nearly a year of waiting everyone would like the product to be mature and feature complete. I'm pretty impatient, and I'd love that too, but given this context I hope some of you will agree that having a well-built product, with a solid technical foundation that lets the product work reliably over the long haul was the better thing to start with.

Either way, please keep the feedback coming: it's incredibly useful in helping me make good decisions about what to invest in.

[chefchrisyoung]

Will try to keep an eye on things here and answer questions as they come up. I'm a bit faster on Twitter and I read (nearly) all emails that go to [email protected].

One thing I think folks should try is the surface temperature feature. This is the true cooking temperature your meat is feeling in the smoker (or oven), no matter what the pit temperature is. Lately, I've been adjusting my smoker's temperature to control the surface temperature of the meat. Really happy with the results.

[BourBonQ]

Thanks for joining Chris, and appreciate the feedback/updates, as well as the support along the way.

[Huskee]

Welcome, Chris. Happy to have you aboard.

[RonB]

Donw

STEbbq
Baltimorelger
BourBonQ

OK folks - I think each of us, (and anyone I missed), should give our first impressions in this post so they will all be together - one stop shopping.

Here's my review of a rotisserie chicken I cooked on my Kettle tonight:

First, I have an Android phone, so no app yet. All I did was use it to monitor the bird's IT. That was until my wife asked me how much longer. And it hit me like a brick - I might be able to tell her. I didn't remember how to set up time left, so I had to look it up and it was very simple to do. By the time I remembered, the IT of the bird was above 150° and the base unit said less than 5 min left to reach the set temp of 156°, (Combustion recommends setting the alarm several degrees below the pull temp). It took 10 sec longer than predicted to reach the set temp. I pulled it at 158° . That alone was enough to make this thing a winner in my book. I know it made SWIMBO happy.

So, without the app it worked well for me and I was pleased. Time will tell about the app.

In case anyone missed it, Combustion sez they are waiting on Google to approve the app.

[chefchrisyoung]

Ron, thanks for the feedback. The Android app is running behind—had to pull the engineers off the app to fix a nasty firmware bug. We did just release an Android app that can update the thermometer and display to the latest firmware (which I highly recommend doing), but we're probably two weeks out from the first version of the Android app.

[RonB]

Yes - I did the updates today before the first use. Through my computer illiterate eyes, it seemed to go well and was easy to do.

[Baltimorelger]

Thanks for the prompt RonB. I used mine for a prime rib. The app wasn’t available during my cook but I set it up afterwards with no issue.

Overall I was happy with it. Temp seemed accurate (I thermopenned a few times) and had an accurate countdown to finished time. I’d like to experiment with it more.

I found initially cycling through the options and two different thermometers difficult as I couldn’t remember the buttons, but got the hang of it quickly.

One big plus was zero issue with connectivity. Because we use a mesh system most Wi-Fi based thermometers don’t work in my house (signals is a full no-go, Meater was spotty at best). This worked perfectly. I still need to test the distance but overall it’s got huge potential.

[grantgallagher]

Weird, i run a mesh system and no problems with signals or meater+

[Baltimorelger]

I wish that was the case for me. It just was spotty as heck and kept losing the signal.

[BourBonQ]

1st impression... I like it. My first use was to finish pork tenderloin that had been SV'd the day before, so not a full cook, per se. Android phone, so no app yet, but stand alone operation worked fine.

Pork tenderloin had been SV'd to 140 the day before and refrigerated overnight. To finish for serving, I inserted the probe into the largest of 4 tenderloins and set the target temp to 120 with the intent of serving at 125-130.

Placed them in the gasser on a med-hi heat setting with a few wood chips over one of the burners.

The display initially read the cook as % complete, then after a few minutes changed to indicate time left to hit target, and it was pretty accurate.

Wireless connected quickly, but range was very limited with the grill lid down and the display indoors, maybe 12-15 feet. Looking forward to the Android app for sure.

One tiny complaint was cycling through the different temp settings (internal/surface/cooker). The temp button was a little stubborn and takes quite a bit of force to actually press, and doesnt really 'click', just kind of a mushy press that takes some effort, at least on my unit.

But the functionality is great, and looking forward to running it through its paces on some longer cooks.

[glitchy]

I just wanted to note that I did some fairly heavy wordsmithing of my original post. I don't think I changed my overall feedback, but definitely the tone. Since the forums are member's only, I regularly post how I would say things to my buddies. However, this thread seems to be getting hits on search engines even behind the paywall. I wanted to clean it up a bit as I definitely don't want to discourage anyone at all from further researching or buying this product. It's very impressive out of the box, but I just feel the apps need some (already planned apparently) work to be a really great product.

[chefchrisyoung]

BourBonQ thanks for the feedback. Agree the buttons on the top of the display are not quite where we wanted them in terms of actuation force. A click would be nice, but is a bit at odds with the elastomeric shell we used for weather proofing and shock protection from falls. Definitely something I'd like to make a lot better in a future version.

As for BLE range, once the Android app is released you should get more range by leveraging the display as a repeater. Android app will probably have an initial release mid-January, and become a lot more feature complete over a couple of months.

glitchy appreciate the update to your post.

A few comments for context and optimism: Yup, we set mile-high expectations and we haven't hit them all—yet.

Getting all the "core" infrastructure done for connectivity so that things just work was a huge lift—about two engineering years so far. It's obviously that our competitors skip this work and use what ever they get "out of the box" from their Bluetooth or WiFi chip. Hence the often-broken connectivity experience that many people report in reviews for these products.

A lot of hard and expensive lessons at ChefSteps/Joule convinced me this is a massive mistake. Connectivity needs to be incredibly reliable and that means doing a lot of foundational and custom work. For example, our goal is sub-1 second reconnection times to the display or app when you pull the thermometer out of the charging sleeve. We wanted this to feel instant, like just hitting an on/off button. And we want all of our devices (including those we haven't announced yet) to effortlessly expand the reach of this MeatNet network, like a Tile Tracker or AirTag. Sounds simple. Not simple.

This, by the way, is why we abandoned WiFi. We totally get the benefit—when it works! From experience at ChefSteps, WiFi was the number one customer support problem and nearly always because of how the customer had their wifi configured. I think we had two software engineers working close to full time on WiFi support. At Combustion this would drown our team and guarantee we can't work on features we feel are higher value—like improved predictions, carry over cooking, data visualization.

WiFi may happen at some point, but initially it's likely to sneak in the side door. For example, something I'd like to get done sooner rather than later is enabling the apps to route the data up to the cloud and then down to another device running an app and connected to the internet. This offloads the WiFi configuration problem to Apple and Android who are far more equipped to support reliable WiFi.

Looking a bit further out, a lot of technology companies are adding the ability for third-party companies like Combustion to get our Bluetooth signal to something like a Ring Camera and then have the Ring Camera use its internet connection to route the data up to the cloud. Supporting these kinds of capabilities is the direction we're most likely to support over the next year, since it makes it much easier for us to expand the coverage for our devices, while not abandoning the simplicity of having everything just work, while letting Combustion focus on solving problems that only we will solve like improved predictions or better data visualization features.

The goods news here is that behind the very simple looking app is a very robust foundation, so adding a lot of the missing features will happen relatively quickly over the next couple of weeks.

[STEbbq]

Thanks Chris. This is really helpful insight behind the curtain if it were. Very appreciated!

[BourBonQ]

Thanks for the follow up Chris. As a design engineer who also sells things (albeit in a much different industry, but with some similarities), I do appreciate what it must've taken you to get to this point.

Looking forward to the continued evolution of this impressive tool!

[BBQPhil]

As a design engineer of close to 100 products (I'm old) and author about them (From Concept to Consumer), I've learned hardware is hard and software is never done. I think it's great we're getting feedback from the company/CEO, and avoiding speculation. One never knows all the problems until a product gets in the hands of hundreds or thousands. And in a small startup company you can't do everything and need to pick your battles. I did a WiFi thermometer for a big cooking store chain and the variations amoung routers and setups caused huge connectivity issues. I wish Mr Young the best of luck because he's taken on a big challenge with many innovative elements.

[chefchrisyoung]

BBQPhil thanks for the kind comments. And 100 products is... impressive!

I have a guess about which WiFi thermometer you would have worked on. If it's the one I'm thinking of, I think that was a cautionary tale that retailers think hardware is easy, while hardware companies think retail is easy. Grass always looks greener on the other side.