Hey Jeff, just a small suggestion that I learned from Cathode Ray Dude's channel. When you put text on the screen to make a correction or add a note, it's useful to add a small sound effect so that people who are only listening with the video in the background know there's something to go see on screen.
Or...dub the correction. Many people do not/cannot watch the video. This was just a talking piece, MANY people are just going to listen to it while doing something else. Providing on-screen corrections only is something other channels have been called out for...
Yes! He did this on today's video and I nearly commented my thanks just for that because it worked exactly as intended to make me look over at the video. Edit: I do agree with the reply above me that a dubbed correction even better.
@@GoCreateTech please tell me how you are so sure that having a video on in the background doesn’t count as a view? That would screw podcasts to the point where I can’t imagine it would be worth it to post them in video form. Also “in the background” here does not necessarily mean the tab isn’t visible. I tend to have the video on my other monitor, clearly visible. Sometimes I’ll even have something playing on a completely different device (such as on my iPad while I’m on my laptop)
@@GoCreateTech I literally have no idea what you are talking about. If I open a video in a tab I can... 1. get up and do other things in my room while listening. 2. open another tab/window and do something on another monitor. 3. anything else that doesn't include my eyes pointing at the video. Jeff makes videos as his JOB. He has showed us his microphones. Jeff did literally nothing important in this video but talk. Which is TOTALLY fine. It gave you something to watch while he spoke. But it was essentially a spoken word FAQ/Press Release. Your "argument", and I struggle to even call it that, is nonsensical, or shows a STRONG sense of entitlement. "I don't have a problem" ...is not the same as... "There is no problem" I am ALL for discourse and improving everyone involved. But your immediate hostility, insults, and clear lack of understanding of what is even being discussed is not something I'm going to deal with right now. Good day.
While I am disappointed that the firmware is not being made open source at this time, I understand that most businesses don't see it feasible. From what I've seen so far you have designed an excellent product, have clear paths forward to new features/products, and you understand how to scope a project to be achievable. You've set yourself up for success, so good luck from here on out!
I greatly appreciate the focused and intentional nature behind the feature set at offer here. I encourage you to keep it as simple and as functional as possible. And maybe open source the firmware so people can hack stuff in as they want, but the device should focus on doing its one thing and doing it well.
@@CraftComputing You knew you were going to get _ie Mc_face. I'm just happy you haven't gotten some of the more explicit meme names like "granny _s" and "_ler did _thing wrong".
Why are you assembling them at all? Now I cant speak to everyone’s skill level who ordered them, but I would be fine getting the board, case and screws in a bag and put it together myself. I understand you wanting to ensure QC on the first 100 or so. From there, send a kit. You are going to have arthritis or at least really tired fingers after a few hundred. I ordered 2 of these and excited to get mine. From there I plan to roll them out at clients of mine once the full release is ready.
It is also clear you are creating a very bad device? Where is the subscriptions? Where is the horrendous upgrade cost? This lifetime support is madness? And where is the AI meaningless add-on or branding?
agreed- there is only one 'X' in the name. Needs at least 2 more to show everyone your super serious about high performance, enthusiast level, temperature monitoring.
if you are going to use a USB and a barrel jack be sure to put a diode on the usb side so when someone connected a usb and a barrel jack they do not have a bad day from shoving 56v into there computer's usb port
In the future, do batching - it's much more efficient. Batch 1: put every board in cases, batch 2: put the screws in but don't screw them down, batch 3: screw screw screw. Batching makes jobs like that go way faster.
For a future version of this, a decibel sensor would be nice. It's always noisy in data centers, but it's typically predicably noisy. Audio is just one more hint admins get that something isn't quite right.
In the future, you could also just sell the board/sensor assembled minus the case that way some buyers could print their own or even come up with custom housings
arbitrary dry contact inputs (1-4 of them) would allow leak ropes, limit switches, tilt sensors... basically every network climate monitoring device I used in the IT field had those, and usually at least one relay for an alarm output.
Jeff: I'm gonna make this awesome thing! I want to keep it simple. What would you like as options? Consumers: Can you just re-release a Pico with a temp sensor and maybe attach it to an entire computer so I can do everything with it. I need to check my email with my SNMP devices.
I like the direction you are going with this Jeff, love the keep it simple mentality. I think the POE version with 2 temp probes (One hot and one cold side) with intrusion dectetion (as you said with a momentary switch) is exactly what many of my collegues at work would kill for. Cheap and simple, those words are silk in the ears of any IT professional. For a home use I would love a air quality version. Living in australia bushfires are common so that would be awesome. I will still put the product i described above in my home server rack too.
Jeff, I'm a trademark attorney in the Eugene area and was the TM attorney for a brewery here. You've had beers I've TM'd. Let me know if you would like to discuss TM's. Thanks for making this video, have been wondering when my unit will get here.
Definitely going to be an extremely useful tool in the rack room. I do agree that you need to keep the cost down, even on the pro version. If it ever gets to $200+ you'll be competing with the likes of ControlByWeb and such which are about as fully featured as you can get. Also when I saw Axe Effect I actually thought of "Axe FX" which is a guitar DSP by Fractal Audio, lol
Names are hard! Sorry I cant help with a suggestion. I appreciate the update on the temp sensor, and for the ability to support the channel. I can understand why you would want to keep things simple for now. If you ever decide to go open source with the firmware, I'd love to contribute.
TempER, Temp/Environment Readout covers all the bases, gives you room for expansion of features. also gives you possibilities for a catchy slogan 'TempER your server rooms conditions' or similar.
Since you're already considering branching into non-IT areas, to my mind this might also be quite useful for freezer and refrigerator monitoring. I would think (not an expert) that both temperature and humidity sensors are useful there, and it would certainly make sense to expand it to have a sensor to register if the door is open or not for those as well. If the temperature range is high enough, and the enclosure and batteries robust enough (or use wired power), maybe look into expanding into restaurant and bakery ovens?
Did you base the ordered board amount to preorders or a gut feeling, that they’ll sell? Asking as I have a shelf full of stuff ordered/built on the assumption that someone will buy it.
Don't know if any of this is on your radar but in commercial environments and regulated industries you need calibration paperwork (NIST traceable) and recalibration abilities and/or processes or else you have to throw it out and replace after X months/years).
Note that you need the product CE certified when importing it to the EU. This means measuring immunity and radiated emissions. An ESP32 is a bit cheaper then the RP2040 as you don't need a crystal/oscillator and external flash. It also means you can build on top of existing IOT projects like Tasmota and ESPHome. That said, the RP2040 is available up to the year 2040 and might come down in price when it's successor arrives. The SDK is also way easier to comprehend.
Why not just a Shelly H&T at less than half the price? Standard product, fully supported with new firmware every now and then, works with MQTT, wireless, cheap, does the job..
for scientific use and data center use, it might be interesting to add an o2 sensor to see if a fire suppression system has been triggered or if the o2 level for, say, an NMR is too low to enter the enclosure to work on the machine.
Have you looked at the cost on those sensors? Most "real" sensors for counting gasses cost $60-$150 EACH in quantities of 1000. (Not the hacky Volatile Organics ones that get slapped into a random project with false claims and the company then disappears...) I ran into this situation when trying to throw together some cheap (think $5 BOM) environmental + air quality + CO2 monitors for my house. I "NOPE"d right out when I saw the CO2 sensors were $75 each in qty 1-5.
Keep in mind this reference is from someone who knows folks who work in the EDU environments and one who once was employed by EDU... where budgets are small.
Going to keep a close eye on this, really interested in using these in my home assistant instance. My current temperature and humidity sensors are so unstable and keep loosing their connection.
At this point I want my IoT devices to stay at “old” 2.4 GHz WiFi. I pretty much don’t have anything normal there, that would have to compete for the airtime, it’s actually easier to manage because I can deploy my IoT WiFi as 2.4 GHz and it slowly becomes less congested with “other” networks, as more and more devices go for 5+ GHz spectrum.
The multi sensor one I would buy a bunch for my freezers. Being able to just run ethernet to my freezer locations and poe power something that I can monitor (and alert) off temperature in my freezers would be a big buy at the reasonable price point I think you are targetting.
Enable that Pressure sensor! It would be perfect for NASA, servers on Submarines (both military and private science vessels) or even quantum computing setups that may be dependant on environmental variables.
I am wondering if the pressure sensor could be used as a warning for both fans and intrusion detection. Server fans tend to create low pressure and intrusions would instantly change the ' ambient ' pressure.
If the pressure sensor is in the airflow path, then the pressure it reads will be impacted by the airspeed interacting with the local geometry, most likely presenting as a reduction in pressure due to the venturi effect or some other similar effect I don't know the name of that produces low pressure areas adjacent to flowing fluid. The specific value of this is that it lets the pressure sensor validate air flow. I doubt you could meaningfully convert this into any flow unit, but you would be able to detect if flow stops, and you might be able to detect if flow changes in a substantial way, such as from one or more fans stopping. Would this data be useful? Probably not, those fans should have tachs they should be monitored by their hosts, and they are also probably being regulated so the flow rate isn't expected to be consistent. Perhaps for less intelligent devices, or if you're expecting the server fans to produce/induce some auxiliary airflow elsewhere in the case and you want to validate that's actually happening. Maybe you have forced air from an external device like data room air handlers or ducted in AC and you can't monitor that source device, maybe it's not yours, or it's the less intelligent device. The pressure sensor could let you monitor that the forced air is actually present, and you would be able to detect a loss of flow much faster via pressure. If Any Of This does stand up to scrutiny, a venturi tube could be added to the case model to amplify the pressure variation due to airflow, and make it less situational; but that would probably impact the sensor's temperature monitoring and might induce a whistle
It's on the list of features I'd love to have. If I can make it happen and stay within my budget, absolutely. But that might be an item that's pushed off to a future release, along with LED/Buzzers/etc.
Would it make sense to create a modular sensor system? Sell the brain unit and a variety of sensors that can just be plugged into it. As far as sensor ideas: - Power (like a Kill A Watt) - Sound level (decibels) - Air flow (MPH) - Accelerometer for vibration - Light level (lumens?)
So, there is your first improvement ... snap casing... screw-less lol Still one of the downsides of the current design is high density deployments (ie data-centres) Ideally one we can magnet/cable-tie onto the front door. So what you can do is just make different casings for it. Pressure .... add a sensor in the front and in the back in a DC and you can work out airflow movement ....
@@TheRogueBro There are many ways to do it at that size without it breaking.... and you don't really open the unit often.... but maybe because v1, just in case that is the case (lol)
This case started life as a press fit. I wasn't happy with how thick the walls became to prevent breaking. And honestly if I were to make one more revision, I'd add a 5th screw above the temp probe. A couple of the case lids want to lift still.
@@Prophes0r You can make it so it's not impossible to open even with clickable .... it's just most products are designed not to be opened. There is also the other side where the screws can strip the plastic putting it in the same boat ...
I saw the Micronics box, did they 3D print them or did Joel reuse his box? I'd recommend a big farm like Slant3D, they would beat injection molding at your scale for sure.
In regards to future support for pressure sensor support: speaking from a decidedly non-developer standpoint, only having the vaguest inclinations of what a basic dev-cycle is like and the most rudimentary fundamentals of coding & associated costs, have you and/or your dev given any thought to making a lightweight core framework+plugin kind of build, so as to allow freedom to focus on the parts you want thoroughly fleshed out & polished? Then you don't have to worry (hopefully) about full stack rewrites just to add new features, and if one or more in the userbase wants to, they can start hammering out the additional extensibility. Just a notion, and if that's completely beyond the pale, that's fine.
Say as a potential upgrade for the final production of this device or future models ... is it possible to include a mini LCD screen to display temperature and humidity locally? For a future name how about EyeSNMP? RpiEYE? RpiESP?
So I'm looking to make a home lab server with a lot of the same specs as your 8 gamer server. was hoping to use a Dell T7910 Workstation and Tesla P100 GPU. everything I have been able to find on these idea seems like it should work, but I would like any outside input.
I don't know anything about Prometheus, but surely SNMP data can be bridged or gateway'd into MQTT; the second result for 'MQTT SNMP bridge' is even a DIN rail device that claims to serve this purpose, the first being a very compact github project. I'm pretty sure if someone wants to get the data from these into an MQTT setup, they can make that happen without difficulty and without compromising the security of the sensor.
maybe it's worth considering using drop shipping distribution points like localised amazon warehouses should the internation sales sufficiently rise? certainly works for a lot of chinese based distributors at least in my country.
BME280 is about $12 in single qty, and of course cheaper in bulk. Keep that in mind. Edit: They may enable the pressure sensor if you pay us more money. The sensor has temp, humidity and pressure built in. Why are they not all exposed?
I noticed my BME280 reads low lower than my DHT22 in regards to hummidity, all of them seem to get temp the same (D18B20, DHT22, BME 280) 10:49 maybe you are Linus and your water cooling loop experienced a catastrophic failure and is leaking gallons of water and knowing your server is flooding would be a very important thing to know
You also have things like electromagnetic emission standards you should try to comply to. I think the EU might even require those. those get more complicated with using PoE I think.
Probably kind of a silly idea, but it could be the 'Crafty' line of sensors. So this would be the Crafty Temperature & Humidity Sensor. Then it could just be a product line of Craft Computing Crafty sensors. Only other silly amalgam word I noodled was the 'Craft Sensputer' but that is probably a bit too out there lol.
it would be very interesting to know in what your developer would develop for esp32 if not in c? literally any software or library for esp32 that i know of, is written in c...
@@CraftComputing this answer does not even make any sense. i try to understand what you get at with your statements, i really do. but i can not quite make sense of it. so your developer told you, "snmp is written in c and porting it to esp32 is like teaching a monkey to read". snmp is a protocol, not a software. there are specifications, in rfc 1157 to be exact. everyone can read those and write code that does exactly what is specified in a weekend, i guess. that code can be written in c. or in "whatever". but c would be what you want for the esp32 platform anyway. i guess, the generic, existing arduino_snmp library that everybody uses would also work. it is also written in c. the code your "developer" would write to program the behavior of the esp32 would also be in c, because i have no idea in what language else anyone would program for esp32. so that whole statement does not make sense in the slightest. esp32 code is written in c, "snmp is written in c", every library for esp32 is written in c, where is there a problem? you could in theory compile freaking linux for esp32, i really really do not understand the argument made here...
@@CraftComputing this answer does not even make any sense. snmp is not a software, it is a protocol specified by RFC 1157. anyone could implement that specification in any programming language they want, in about a weekends worth of work. i would prefer to do that in c, as it its the native programming language for the esp32 platform. one could also just use the widely appreciated arduino_snmp library, that is also written in c. i have no idea, how your "developer" comes to the conclusion, that "snmp is written in c and porting it to esp32 is like teaching a monkey to read", as c IS the de facto standard for esp32 code. one could in theory compile linux for esp32 and run it on there, so there should be no reason, your software could not be compiled for it. i would be really really interested in the reasoning behind this "not so accurate" statement. deleting my comment is not gonna help me understand, though (i do not even understand that, i was polite and only presenting facts). "i decided not use the esp32 platform, because snmp is written in c" is anyhow not a logical statement whatsoever.
@@CraftComputing this answer does not even make any sense. snmp is not a software, it is a protocol specified by RFC 1157. anyone could implement that specification in any programming language they want, in about a weekends worth of work. i would prefer to do that in c, as it its the native programming language for the esp32 platform. one could also just use the widely appreciated arduino_snmp library, that is also written in c. i have no idea, how your "developer" comes to the conclusion, that "snmp is written in c and porting it to esp32 is like teaching a monkey to read", as c IS the de facto standard for esp32 code. one could in theory compile linux for esp32 and run it on there, so there should be no reason, your software could not be compiled for it. i would be really really interested in the reasoning behind this "not so accurate" statement. deleting my comment is not gonna help me understand, though (i do not even understand that, i was polite and only presenting facts). "i decided not use the esp32 platform, because snmp is written in c" is anyhow not a logical statement whatsoever.
@@CraftComputing this answer does not even make any sense. snmp is not a software, it is a protocol specified by RFC 1157. anyone could implement that specification in any programming language they want, in about a weekends worth of work. i would prefer to do that in c, as it its the native programming language for the esp32 platform. one could also just use the widely appreciated arduino_snmp library, that is also written in c. i have no idea, how your "developer" comes to the conclusion, that "snmp is written in c and porting it to esp32 is like teaching a monkey to read", as c IS the de facto standard for esp32 code. one could in theory compile linux for esp32 and run it on there, so there should be no reason, your software could not be compiled for it. i would be really really interested in the reasoning behind this "not so accurate" statement. deleting my comment is not gonna help me understand, though (i do not even understand that, i was polite and only presenting facts). "i decided not use the esp32 platform, because snmp is written in c" is anyhow not a logical statement whatsoever.
If its passive PoE might aswell have no PoE. If I cant plug it into a run of the mill PoE af/at switch It aint worth it. Also I would want the IPv6 Stack in software
@@CraftComputing But the crucial question is whether your customers appreciate a bright green gadget in their server rack. If there was a choice, I am convinced that most people would choose a more muted color, such as e.g. black or grey.
Yo. Ive got a spare delvo 7349-bme (electronics esd precision electric screwdriver) if you want to borrow it. And good luck with this, I'll be following it closely. Haven't got the nerve to order one of my UPS's purpose built temp monitors since the card cost more than I got the thing for lmao. Additionally, a buzzer to start yelling at me would be awesome. My servers would start screaming due to the temps before that, but I'd like to know its temp rather than just them working hard.
I wouldn’t mind buying something like that with 4 T/H probes that could be connected to an HVAC system in the house. POE, or wifi would work. Even 24v ac maybe.
AXETemp would be more descriptive, or some other synonym for temperature. I kinda like pyrexia, which basically is a synonym of fever. If you really really want to keep the axe part of course, it's neat IMO how it fits with the shape. AXEpyrex. Why does that sound familiar? I'm probably thinking of aspirate? Or some disinfectant, something like that. Oh well. In the end it's just a name, just pick whatever you like. :)
Hey Jeff, just a small suggestion that I learned from Cathode Ray Dude's channel. When you put text on the screen to make a correction or add a note, it's useful to add a small sound effect so that people who are only listening with the video in the background know there's something to go see on screen.
Or...dub the correction.
Many people do not/cannot watch the video.
This was just a talking piece, MANY people are just going to listen to it while doing something else.
Providing on-screen corrections only is something other channels have been called out for...
Yes! He did this on today's video and I nearly commented my thanks just for that because it worked exactly as intended to make me look over at the video.
Edit: I do agree with the reply above me that a dubbed correction even better.
@@GoCreateTech please tell me how you are so sure that having a video on in the background doesn’t count as a view? That would screw podcasts to the point where I can’t imagine it would be worth it to post them in video form. Also “in the background” here does not necessarily mean the tab isn’t visible. I tend to have the video on my other monitor, clearly visible. Sometimes I’ll even have something playing on a completely different device (such as on my iPad while I’m on my laptop)
@@GoCreateTech I literally have no idea what you are talking about. If I open a video in a tab I can...
1. get up and do other things in my room while listening.
2. open another tab/window and do something on another monitor.
3. anything else that doesn't include my eyes pointing at the video.
Jeff makes videos as his JOB. He has showed us his microphones.
Jeff did literally nothing important in this video but talk. Which is TOTALLY fine. It gave you something to watch while he spoke. But it was essentially a spoken word FAQ/Press Release.
Your "argument", and I struggle to even call it that, is nonsensical, or shows a STRONG sense of entitlement.
"I don't have a problem"
...is not the same as...
"There is no problem"
I am ALL for discourse and improving everyone involved. But your immediate hostility, insults, and clear lack of understanding of what is even being discussed is not something I'm going to deal with right now.
Good day.
@@GoCreateTech"supposedly support" is a backhanded insult.
While I am disappointed that the firmware is not being made open source at this time, I understand that most businesses don't see it feasible. From what I've seen so far you have designed an excellent product, have clear paths forward to new features/products, and you understand how to scope a project to be achievable. You've set yourself up for success, so good luck from here on out!
I greatly appreciate the focused and intentional nature behind the feature set at offer here. I encourage you to keep it as simple and as functional as possible. And maybe open source the firmware so people can hack stuff in as they want, but the device should focus on doing its one thing and doing it well.
GUITARHS - Generally Useful Internal Temperature and Relative Humidity Sensor
Actually, as a fellow guitar owner/enjoyer, this is a stellar idea!
Tempie McSensorface would be a good name
....God dammit 😂
I know people that would buy this based on that name alone 😂
@@CraftComputing mAXwell?
@@CraftComputing You knew you were going to get _ie Mc_face. I'm just happy you haven't gotten some of the more explicit meme names like "granny _s" and "_ler did _thing wrong".
McSnmpFace
Why are you assembling them at all? Now I cant speak to everyone’s skill level who ordered them, but I would be fine getting the board, case and screws in a bag and put it together myself. I understand you wanting to ensure QC on the first 100 or so. From there, send a kit. You are going to have arthritis or at least really tired fingers after a few hundred. I ordered 2 of these and excited to get mine. From there I plan to roll them out at clients of mine once the full release is ready.
Arthritis after a few hundred? OK sure, dude who has never worked in manufacturing
😆 all the way through the video I'm thinking.. "I wonder why he's not using a power screwdriver"
It is also clear you are creating a very bad device? Where is the subscriptions? Where is the horrendous upgrade cost? This lifetime support is madness? And where is the AI meaningless add-on or branding?
agreed- there is only one 'X' in the name. Needs at least 2 more to show everyone your super serious about high performance, enthusiast level, temperature monitoring.
I can feel the joy you are feeling. Congrats for your product man, you must be very proud. Enjoy it!
if you are going to use a USB and a barrel jack be sure to put a diode on the usb side so when someone connected a usb and a barrel jack they do not have a bad day from shoving 56v into there computer's usb port
In the future, do batching - it's much more efficient. Batch 1: put every board in cases, batch 2: put the screws in but don't screw them down, batch 3: screw screw screw. Batching makes jobs like that go way faster.
For a future version of this, a decibel sensor would be nice. It's always noisy in data centers, but it's typically predicably noisy. Audio is just one more hint admins get that something isn't quite right.
Name idea: Networked Environment Recording Device, abbreviated as N.E.R.D.
In the future, you could also just sell the board/sensor assembled minus the case that way some buyers could print their own or even come up with custom housings
"Server Sniffer" Axe effect as you said was a marketing campaign of Axe they would not like your name one bit.
arbitrary dry contact inputs (1-4 of them) would allow leak ropes, limit switches, tilt sensors... basically every network climate monitoring device I used in the IT field had those, and usually at least one relay for an alarm output.
This project of yours pleases my inner-geek so much! I can't wait until I get my device!
Can’t blame you Torx is the best.
My validation models used hex heads, and I hated them. So uncivilized.
Jeff: I'm gonna make this awesome thing! I want to keep it simple. What would you like as options?
Consumers: Can you just re-release a Pico with a temp sensor and maybe attach it to an entire computer so I can do everything with it. I need to check my email with my SNMP devices.
Beautiful work Jeff.
Pressure is useful because it allows you to easily calc dew point to make sure nothing gets wet.
I like the direction you are going with this Jeff, love the keep it simple mentality. I think the POE version with 2 temp probes (One hot and one cold side) with intrusion dectetion (as you said with a momentary switch) is exactly what many of my collegues at work would kill for. Cheap and simple, those words are silk in the ears of any IT professional.
For a home use I would love a air quality version. Living in australia bushfires are common so that would be awesome. I will still put the product i described above in my home server rack too.
"S#!t, I'm five percent of the way through..." I'm so sorry Jeff, but I almost laughed my butt off on that one. 😁 Been there!
Jeff, I'm a trademark attorney in the Eugene area and was the TM attorney for a brewery here. You've had beers I've TM'd. Let me know if you would like to discuss TM's. Thanks for making this video, have been wondering when my unit will get here.
Wow that sponsorship so early... just joking. Keep making amazing content Jeff! I love your videos.
You make me want to drink my Treehouse on Wednesday.
The Tempinator
Read that in the Dr. Doofenshmirts voice
Definitely going to be an extremely useful tool in the rack room.
I do agree that you need to keep the cost down, even on the pro version. If it ever gets to $200+ you'll be competing with the likes of ControlByWeb and such which are about as fully featured as you can get.
Also when I saw Axe Effect I actually thought of "Axe FX" which is a guitar DSP by Fractal Audio, lol
Names are hard! Sorry I cant help with a suggestion. I appreciate the update on the temp sensor, and for the ability to support the channel. I can understand why you would want to keep things simple for now. If you ever decide to go open source with the firmware, I'd love to contribute.
12:10 2,4 Ghz is not less secure, old WIFI is. For example the "MikroTik hAP ax lite" only has 2.4 GHz which however is WIFI 6 and supports WPA3.
Nice, Good Luck on the Venture 🙂
TempER, Temp/Environment Readout covers all the bases, gives you room for expansion of features. also gives you possibilities for a catchy slogan 'TempER your server rooms conditions' or similar.
Since you're already considering branching into non-IT areas, to my mind this might also be quite useful for freezer and refrigerator monitoring. I would think (not an expert) that both temperature and humidity sensors are useful there, and it would certainly make sense to expand it to have a sensor to register if the door is open or not for those as well. If the temperature range is high enough, and the enclosure and batteries robust enough (or use wired power), maybe look into expanding into restaurant and bakery ovens?
Did you base the ordered board amount to preorders or a gut feeling, that they’ll sell? Asking as I have a shelf full of stuff ordered/built on the assumption that someone will buy it.
Honestly, it was gut feeling. I chose a number I felt was attainable. As of this morning, I've sold 394/400 🙂
Don't know if any of this is on your radar but in commercial environments and regulated industries you need calibration paperwork (NIST traceable) and recalibration abilities and/or processes or else you have to throw it out and replace after X months/years).
If you're doing a hub already. You could always keep it simple. Then sell like a hubplus with extras like trip sensors or circuit cutting or air etc.
Note that you need the product CE certified when importing it to the EU. This means measuring immunity and radiated emissions. An ESP32 is a bit cheaper then the RP2040 as you don't need a crystal/oscillator and external flash. It also means you can build on top of existing IOT projects like Tasmota and ESPHome.
That said, the RP2040 is available up to the year 2040 and might come down in price when it's successor arrives. The SDK is also way easier to comprehend.
TempSense or T-Sense (TSWB, TSPoE, TSPEH) like (WB-xxx) for Wireless Beta one. (PE-xxx) Power over Ethernet. (PEH-xxx) Humidity one. Etc. Good luck.
In the future you should definitely consider adding DI/DO functionality. Able to read a door sensor or motion sensor, actuate relays and so on.
Why not just a Shelly H&T at less than half the price? Standard product, fully supported with new firmware every now and then, works with MQTT, wireless, cheap, does the job..
for scientific use and data center use, it might be interesting to add an o2 sensor to see if a fire suppression system has been triggered or if the o2 level for, say, an NMR is too low to enter the enclosure to work on the machine.
Have you looked at the cost on those sensors?
Most "real" sensors for counting gasses cost $60-$150 EACH in quantities of 1000.
(Not the hacky Volatile Organics ones that get slapped into a random project with false claims and the company then disappears...)
I ran into this situation when trying to throw together some cheap (think $5 BOM) environmental + air quality + CO2 monitors for my house.
I "NOPE"d right out when I saw the CO2 sensors were $75 each in qty 1-5.
@@Prophes0r well worth the price of investment, and there are good sources.
Keep in mind this reference is from someone who knows folks who work in the EDU environments and one who once was employed by EDU... where budgets are small.
I already want at least two of the POE versions! Can't wait!
Going to keep a close eye on this, really interested in using these in my home assistant instance. My current temperature and humidity sensors are so unstable and keep loosing their connection.
“The Temperature Sensor” is an awesome name.
Once you move to the RP2040 unit, would be handy to have a low power mode, so a remote unit can be powered by battery/solar.
A power led would be nice to see if the device is on and maybe be able to turn it off.
A bridge detector would be nice for in a garage to see if the garage door is open and the humidity and tenpature too.
Maybe a dumb idea, but what about a raspberry pi HAT or module to be used with home assistant?
can't wait for the PoE version
At this point I want my IoT devices to stay at “old” 2.4 GHz WiFi. I pretty much don’t have anything normal there, that would have to compete for the airtime, it’s actually easier to manage because I can deploy my IoT WiFi as 2.4 GHz and it slowly becomes less congested with “other” networks, as more and more devices go for 5+ GHz spectrum.
The more 5GHz devices we have, the more IOT 2.4GHz devices get that spectrum all themselves.
The poe unit sounds amazing, that one is going to be big.
JAX Effect
Jeff's Analytical eXperiment. Effecting your knowledge of your rack.
The multi sensor one I would buy a bunch for my freezers. Being able to just run ethernet to my freezer locations and poe power something that I can monitor (and alert) off temperature in my freezers would be a big buy at the reasonable price point I think you are targetting.
Do you plan to have p/n and s/n stickers?
You explain thing idk about so well that I want to learn about it more
Enable that Pressure sensor! It would be perfect for NASA, servers on Submarines (both military and private science vessels) or even quantum computing setups that may be dependant on environmental variables.
I am wondering if the pressure sensor could be used as a warning for both fans and intrusion detection. Server fans tend to create low pressure and intrusions would instantly change the ' ambient ' pressure.
If the pressure sensor is in the airflow path, then the pressure it reads will be impacted by the airspeed interacting with the local geometry, most likely presenting as a reduction in pressure due to the venturi effect or some other similar effect I don't know the name of that produces low pressure areas adjacent to flowing fluid. The specific value of this is that it lets the pressure sensor validate air flow. I doubt you could meaningfully convert this into any flow unit, but you would be able to detect if flow stops, and you might be able to detect if flow changes in a substantial way, such as from one or more fans stopping.
Would this data be useful? Probably not, those fans should have tachs they should be monitored by their hosts, and they are also probably being regulated so the flow rate isn't expected to be consistent. Perhaps for less intelligent devices, or if you're expecting the server fans to produce/induce some auxiliary airflow elsewhere in the case and you want to validate that's actually happening. Maybe you have forced air from an external device like data room air handlers or ducted in AC and you can't monitor that source device, maybe it's not yours, or it's the less intelligent device. The pressure sensor could let you monitor that the forced air is actually present, and you would be able to detect a loss of flow much faster via pressure.
If Any Of This does stand up to scrutiny, a venturi tube could be added to the case model to amplify the pressure variation due to airflow, and make it less situational; but that would probably impact the sensor's temperature monitoring and might induce a whistle
CLIMATH - Compact Low-power Instrument for Monitoring Ambient Temperature and Humidity
have you considered adding a small oled screen to display live temp/humidity?
It's on the list of features I'd love to have. If I can make it happen and stay within my budget, absolutely. But that might be an item that's pushed off to a future release, along with LED/Buzzers/etc.
probably better e-ink
@@marcogenovesi8570ooh, yeah, e-ink would be dope.
Would it make sense to create a modular sensor system? Sell the brain unit and a variety of sensors that can just be plugged into it.
As far as sensor ideas:
- Power (like a Kill A Watt)
- Sound level (decibels)
- Air flow (MPH)
- Accelerometer for vibration
- Light level (lumens?)
I would love a sump pump monitor mode!! I just had a sump pump failure.
So, there is your first improvement ... snap casing... screw-less lol
Still one of the downsides of the current design is high density deployments (ie data-centres)
Ideally one we can magnet/cable-tie onto the front door.
So what you can do is just make different casings for it.
Pressure .... add a sensor in the front and in the back in a DC and you can work out airflow movement ....
Not sure why he didn't do press-fit. Maybe because he went the 3d printed route? They tend to break pretty easily when that small.
@@TheRogueBro There are many ways to do it at that size without it breaking.... and you don't really open the unit often.... but maybe because v1, just in case that is the case (lol)
This case started life as a press fit. I wasn't happy with how thick the walls became to prevent breaking. And honestly if I were to make one more revision, I'd add a 5th screw above the temp probe. A couple of the case lids want to lift still.
Absolutely not.
Saving some assembly time and
@@Prophes0r You can make it so it's not impossible to open even with clickable .... it's just most products are designed not to be opened.
There is also the other side where the screws can strip the plastic putting it in the same boat ...
Name idea: Blade Effect? Just an evolution of the current name but invokes the image of server racks.
I saw the Micronics box, did they 3D print them or did Joel reuse his box? I'd recommend a big farm like Slant3D, they would beat injection molding at your scale for sure.
LOL, that's exactly what happened. Joel shipped my prints in his Micronics box. It's all we have to remember them by now :-D
The chicken coop usecase is real, my DHT-11 based sensors need replacing :)
Yay!!! I can't wait to get mine!
the names for project: Regulate Hub, Heat Shield but i do like Axe Effect alot
In regards to future support for pressure sensor support: speaking from a decidedly non-developer standpoint, only having the vaguest inclinations of what a basic dev-cycle is like and the most rudimentary fundamentals of coding & associated costs, have you and/or your dev given any thought to making a lightweight core framework+plugin kind of build, so as to allow freedom to focus on the parts you want thoroughly fleshed out & polished? Then you don't have to worry (hopefully) about full stack rewrites just to add new features, and if one or more in the userbase wants to, they can start hammering out the additional extensibility. Just a notion, and if that's completely beyond the pale, that's fine.
nice logistics company you have here
Say as a potential upgrade for the final production of this device or future models ... is it possible to include a mini LCD screen to display temperature and humidity locally? For a future name how about EyeSNMP? RpiEYE? RpiESP?
I'd love to add a small status screen in the future. Definitely something we've got on the list.
It would be great to have DC voltage reporting. Great for comm rooms and wisp remote equipment enclosure with dc power systems.
Multiple temp sensors would be great for HVAC systems and radio/TV broadcast sites who need to monitor air temp going IN and OUT of transmitters.
Raised floor environments definitely need water sensors under the floor!
So I'm looking to make a home lab server with a lot of the same specs as your 8 gamer server. was hoping to use a Dell T7910 Workstation and Tesla P100 GPU. everything I have been able to find on these idea seems like it should work, but I would like any outside input.
What about smoke detection if the spirit of devices is trying to escape?
Stoked for the POE version… really hoping for active….
I don't know anything about Prometheus, but surely SNMP data can be bridged or gateway'd into MQTT; the second result for 'MQTT SNMP bridge' is even a DIN rail device that claims to serve this purpose, the first being a very compact github project. I'm pretty sure if someone wants to get the data from these into an MQTT setup, they can make that happen without difficulty and without compromising the security of the sensor.
I think Jeff wants to make the best SNMP temperature and humidity sensor.
I would like to see an smart outlet that can do SNMP. I have a portable AC unit and I would like to monitor the power and etc.
maybe it's worth considering using drop shipping distribution points like localised amazon warehouses should the internation sales sufficiently rise? certainly works for a lot of chinese based distributors at least in my country.
Definitely something I'm looking into.
BME280 is about $12 in single qty, and of course cheaper in bulk. Keep that in mind. Edit: They may enable the pressure sensor if you pay us more money. The sensor has temp, humidity and pressure built in. Why are they not all exposed?
Axe effect pro max ultra is probably a good name. At least if you ask big tech.
You forgot AI, but other than that nailed it.
I noticed my BME280 reads low lower than my DHT22 in regards to hummidity, all of them seem to get temp the same (D18B20, DHT22, BME 280)
10:49 maybe you are Linus and your water cooling loop experienced a catastrophic failure and is leaking gallons of water and knowing your server is flooding would be a very important thing to know
You also have things like electromagnetic emission standards you should try to comply to. I think the EU might even require those. those get more complicated with using PoE I think.
I honestly love the name. I say keep it.
Probably kind of a silly idea, but it could be the 'Crafty' line of sensors. So this would be the Crafty Temperature & Humidity Sensor. Then it could just be a product line of Craft Computing Crafty sensors. Only other silly amalgam word I noodled was the 'Craft Sensputer' but that is probably a bit too out there lol.
it would be very interesting to know in what your developer would develop for esp32 if not in c?
literally any software or library for esp32 that i know of, is written in c...
Yes, ESP libraries are written in C. Those libraries don't include network monitoring tools developed in the late 90s ;-)
@@CraftComputing this answer does not even make any sense.
i try to understand what you get at with your statements, i really do. but i can not quite make sense of it.
so your developer told you, "snmp is written in c and porting it to esp32 is like teaching a monkey to read".
snmp is a protocol, not a software. there are specifications, in rfc 1157 to be exact. everyone can read those and write code that does exactly what is specified in a weekend, i guess.
that code can be written in c. or in "whatever". but c would be what you want for the esp32 platform anyway.
i guess, the generic, existing arduino_snmp library that everybody uses would also work.
it is also written in c. the code your "developer" would write to program the behavior of the esp32 would also be in c, because i have no idea in what language else anyone would program for esp32. so that whole statement does not make sense in the slightest.
esp32 code is written in c, "snmp is written in c", every library for esp32 is written in c, where is there a problem?
you could in theory compile freaking linux for esp32, i really really do not understand the argument made here...
@@CraftComputing this answer does not even make any sense.
snmp is not a software, it is a protocol specified by RFC 1157. anyone could implement that specification in any programming language they want, in about a weekends worth of work.
i would prefer to do that in c, as it its the native programming language for the esp32 platform. one could also just use the widely appreciated arduino_snmp library, that is also written in c.
i have no idea, how your "developer" comes to the conclusion, that "snmp is written in c and porting it to esp32 is like teaching a monkey to read", as c IS the de facto standard for esp32 code.
one could in theory compile linux for esp32 and run it on there, so there should be no reason, your software could not be compiled for it.
i would be really really interested in the reasoning behind this "not so accurate" statement. deleting my comment is not gonna help me understand, though (i do not even understand that, i was polite and only presenting facts).
"i decided not use the esp32 platform, because snmp is written in c" is anyhow not a logical statement whatsoever.
@@CraftComputing this answer does not even make any sense.
snmp is not a software, it is a protocol specified by RFC 1157. anyone could implement that specification in any programming language they want, in about a weekends worth of work.
i would prefer to do that in c, as it its the native programming language for the esp32 platform. one could also just use the widely appreciated arduino_snmp library, that is also written in c.
i have no idea, how your "developer" comes to the conclusion, that "snmp is written in c and porting it to esp32 is like teaching a monkey to read", as c IS the de facto standard for esp32 code.
one could in theory compile linux for esp32 and run it on there, so there should be no reason, your software could not be compiled for it.
i would be really really interested in the reasoning behind this "not so accurate" statement. deleting my comment is not gonna help me understand, though (i do not even understand that, i was polite and only presenting facts).
"i decided not use the esp32 platform, because snmp is written in c" is anyhow not a logical statement whatsoever.
@@CraftComputing this answer does not even make any sense.
snmp is not a software, it is a protocol specified by RFC 1157. anyone could implement that specification in any programming language they want, in about a weekends worth of work.
i would prefer to do that in c, as it its the native programming language for the esp32 platform. one could also just use the widely appreciated arduino_snmp library, that is also written in c.
i have no idea, how your "developer" comes to the conclusion, that "snmp is written in c and porting it to esp32 is like teaching a monkey to read", as c IS the de facto standard for esp32 code.
one could in theory compile linux for esp32 and run it on there, so there should be no reason, your software could not be compiled for it.
i would be really really interested in the reasoning behind this "not so accurate" statement. deleting my comment is not gonna help me understand, though (i do not even understand that, i was polite and only presenting facts).
"i decided not use the esp32 platform, because snmp is written in c" is anyhow not a logical statement whatsoever.
axe effect for the beta model. " the craft probe " maybe. brewmaster's beacon. lager liaison.
I am still unable to access your site keeps ion saying I'm doing something dodgy, possible for an email link.
If its passive PoE might aswell have no PoE. If I cant plug it into a run of the mill PoE af/at switch It aint worth it.
Also I would want the IPv6 Stack in software
This sensor + zabbix + notifications on discord = win.
I have only one question - the color of the casing!?!
Is Jeff possibly colorblind?
Literally the same color as the outline on my logo.
@@CraftComputing But the crucial question is whether your customers appreciate a bright green gadget in their server rack. If there was a choice, I am convinced that most people would choose a more muted color, such as e.g. black or grey.
the presser censer could be used if you us this to try to tune cars at drag events but idk if anyone knows that lol
June 12th is my birthday!
Yo. Ive got a spare delvo 7349-bme (electronics esd precision electric screwdriver) if you want to borrow it. And good luck with this, I'll be following it closely. Haven't got the nerve to order one of my UPS's purpose built temp monitors since the card cost more than I got the thing for lmao. Additionally, a buzzer to start yelling at me would be awesome. My servers would start screaming due to the temps before that, but I'd like to know its temp rather than just them working hard.
Craft sensors: Temperature Axe
I wouldn’t mind buying something like that with 4 T/H probes that could be connected to an HVAC system in the house. POE, or wifi would work. Even 24v ac maybe.
IPv6 support? Our monitoring environment is single stack IPv6.
Would love to have a unit that could do power detection or read from the USB/Serial of a UPS unit.
😉 Stay tuned.... That's on the list for a future device.
AXETemp would be more descriptive, or some other synonym for temperature. I kinda like pyrexia, which basically is a synonym of fever. If you really really want to keep the axe part of course, it's neat IMO how it fits with the shape. AXEpyrex. Why does that sound familiar? I'm probably thinking of aspirate? Or some disinfectant, something like that. Oh well. In the end it's just a name, just pick whatever you like. :)
Pressure sensor would be useful for remotely finding out if your server lifted off the ground😂