Reliable network monitoring?

[cupoftea]

Hi,
If i want to connect back to a houselhold internet router 24/7, from an electrical equipment, using a powerline data link or a wireless repeater....then....
Which brands of these are reputable/reliable?

We are seeing a lot of them drop out.

they are then difficult to re-connect.
They connect to the equipment via an ethernet port.

Its just for a fairly small amount of monitored data...just bits and bytes, not images.

 

[betwixt]

Over what kind of range?
If you want to use WiFi and the equipment is within range of an access point (or router) I would suggest something a simple as an ESP8266-01. The programming is simple, free and they only cost about £2UK excluding the 3.3V power supply (~250mA needed). The antenna is built-in and they are smaller than a postage stamp. I use lots here so I can monitor remote systems anywhere that has WiFi access.

Brian.

 

[cupoftea]

Thanks,...i think that sounds an excellent idea.....come to think of it, thats how ive seen it done in other products. Do you know how they do the comms from the Zappi to the Hub?

(the range is up to 25m)

I remember one place, they used a very secure comms from an equipment, it used an antenna which looked like a spring of diam 5mm and length about 2.5cm.

Do you know what woudl be the best frequency.......

I now remember that places deliberately dont use internet for these short range connectrions.....but some other frequency, and protocol, thats more reliable than internet.

The device you speak of sounds great, but are there any , even more reliable ways?

 

[betwixt]

I'm not sure what the 'Zappi' is but any protocol using a physical interface is usually easy to identify. I use a tiny logic analyzer from Salea but there are many alternatives.

The antenna size you mention suggests it may be a 433.92MHz link. That frequency is almost universally free to use but with power restrictions. No frequency is 'best', they are all a compromise between antenna size, allowed power and licensing restrictions. The security of the data is normally achieved by adding error protection to the data. This can be as simple as a CRC check to reject bad data or embedded error protection bits that allow a degree of error correction.

The ESP module uses TCP/IP protocols to talk to the WiFi. They are very secure and have built in 'resend on error' functions. I usually use MQTT as a message carrier which allows even more flexibility and error protection but it does require a 'broker' on the network to act as the postman.

Brian.

 

[cupoftea]

Thanks, these are good ideas, i think we will come to use them.
In the meantime, i wonder if there is any device which we can send to the customer, which be can plugged in to the etherbnet cable, to check if the cable is good?

At the moment, the cable is plugged into the equipment, but we dont know if the cable has sprung a little wire inside.......we certainly arent getting connection to the equipment.....but we need to find out why, preferably without sending an electrician out there.

It might just be the ethernet cable. But we dont know, its in the customers house.

We wonder if sending one of these ethernet cables to the customer, and getting him to take the origianl ethernet cable out, and replace with the below , and see if we then get internet connection?...i guess this is the easiest way to check for cable problem? (without going to the property)

https://uk.farnell.com/pro-signal/psg02598/cat5-lead-1m-black/dp/3498225?st=ethernet%20cable

 

[betwixt]

Substitution for a known good cable will certainly work in theory but I wonder if some installations might pass cables through walls so a direct socket to socket link might not be possible. You would also have to test any loaned cables on return in case they had been damaged by the customer.

Testing CAT-5 (or CAT-6) cables is fairly easy. There are commercial units out there but you need a tester at one end and some other gadget at the other to ensure continuity. You could make a simple cable checker with a resistor network at one end and a simple voltage source and ADC at the other. As you only need to verify the cable is good or not, a 'go/no go' LED is all you need to present to the customer. You could build them for little more than the cost of a substitute cable.

I'm visualizing a chain of seven resistors across a ground and supply with the junctions connected to the other six wires at one end and an MCU with 8 ADC inputs and an LED output at the other. It would detect faults although not necessarily locate exactly where it was. I would guess a faulty cable would just be replaced anyway.

Brian.

 

[cupoftea]

Hi , The Myenergi company sell home devices which are all networked together. These are an EV chargepoint (zappi), a water heater (eddi), and a "hub" which connects to the household router by an RJ45 ethernet cable. The zappi and eddi communicate with the hub , not by wifi, but by a proprietary 868MHz wireless protocol.
Why didnt they just use wifi throughout?

You can see the 868MHz antenna in the bottom left of the Harvi device here at 4:04...
(a "spring" antenna approx 4mm diam by 20mm long)

At 12:38 of this video, you can see the 868MHz antenna in the top left of the product (a "spring" antenna approx 4mm diam by 20mm long)

At 00:08 of this video you can see the 868MHz antenna on top of the eddi

At 3:01 of this video, it explains why the 868MHz protocol is better than wifi.....

..so it does appear that the lower frequencies are better for going through walls and buildings etc.
.....so why is wifi so widely used?

What is confusing , is that all of our prodcuts are set up with wifi, and the installer doesnt sign the job off until they see full wifi connectivity......but sometimes the wifi drops out after the installer has left.......so howcome there is so much problem with wifi?....i mean, if wifi drops out, then why doesnt it just reconnect itself?

Why are we having so many customers who get their product installed with wifi that works on the install day...but then some days or weeks later, the wifi drops out, and our customers simply cannot re-connect the wifi...what is going on?


...but then you have to ask yourself...if 868MHz is so good at getting through walls etc, then why did this guy have to install a really long antenna cable and antenna so that the 868MHz antenna was within "Line of sight" of the zappi?...(otherwise his comnms dropped out)

...and why did his comms drop out when it was cold?

 

[BradtheRad]

Many ethernet switches (routers, hubs, etc.) have led's for simple monitoring. I have one like this. It's informative when I wish to check my internet traffic, or to network computers (even if it's merely 2 computers).

Two led's per port.

* One led rapidly flashes green when a port transmits normally.

* The other led lights orange whenever a data 'collision' occurs in the port. This happens every second or two when communicating.

* If a port detects a new device plugged in, one led flashes rapidly as the connection is tested.

* If transmission is impossible, both led's remain dark. One led may light periodically to indicate ping attempts.

 

[betwixt]

There is no 'best' frequency. All frequencies, at least up to millimeter wave (10s of GHz and higher) work equally well under ideal conditions. It's the surroundings that make the biggest difference to range but beyond a few GHz manufacturing costs also increase sharply. As for why 433MHz and 868MHz (and 316MHz in some countries) are used, that is purely a political reason, those frequencies are generally license exempt for low power equipment but of course being so, they can be quite crowded with transmissions in urban areas. Usually only brief transmissions are allowed to minimize collisions and its up to the designer to add error protection to ensure data is interpreted correctly.

As Brad points out, most RJ45 ethernet ports have built in LEDs to show status but these are driven by the devices interface IC or from ports on the controller, they can have different functionality. On my hub (a Netgear GS316) the LEDs are both green and for 'link established' and 'gigabit/100M' link speed with one or both flashing to indicate activity. On a bad ethernet connection the LEDs should not light at all, regardless of mode.

I would still consider WiFi as a viable alternative. TCP/IP has very robust error protection and is very easy to use.
I have a link here running over about 50 metres from inside a brick building, across open space then to inside my house which is timber built but with aluminum backed insulation slabs in all the walls. It works flawlessly 24/7 using only the module internal antennas.

Brian.

 

[cupoftea]

There is no 'best' frequency.

Thanks, i see your point, though would you agree that generally speaking, because lower frequencies suffer less attenuation in brick etc buildings, then lower frequencies tend to be better?....this is discussed at 3:01 in the 4th video down (above) [post #7]

So, as long as you have an antenna big enough, it does appear that lower frequencies are better.....providing that no one else nearby is transmitting on the same band.

Though as you know, there are special coded digital modulations which allow multiple people to transmit on the same band in the nearby vicinty of each other (without interfering with each other)

Do you think that Myenergi chose 868MHz because their Zappi product goes outdoors, and wifi devices tend not to be very waterproof, and not generally made for outdoors?....eg, putting a wifi access point outdoors is unlikely to be a good idea?.....so therefore myenergi put the comms transsceiver inside the zappi chargepoint where it is waterproof...and then made it 868MHz because low frequency is better at going through walls etc?...what do you think their reasoning was?
But yes, i imagine standard wifi (2.4GHz) is the cheapest as its more ubiquitous.?

 

[betwixt]

I think it more likely 868MHz was chosen because ready made transmitter and receiver modules are mass produced and therefore cheap. They probably could have done it a little cheaper at 433.92MHz but that frequency is very crowded so perhaps they wanted a quieter band to work in. To be honest, the difference in range between 868MHz and WiFi's 2.45GHz is minimal.

WiFi transceivers are very small and inexpensive so encasing them in a watertight container is no problem. If anything, they have issues with current consumption, less than 10mA at 3.3V when idle but about 250mA when transmitting at full power. The image should give you an idea of their size, this gadget not only has WiFi, it has a general purpose MCU with enough processing power and RAM to run quite complex programs.

All wireless links use pulsed transmission, including WiFi. That's why error protection is so important, data could be lost in a collision so detection, correction and resend should be built into them all.

Brian.

 

[cupoftea]

To be honest, the difference in range between 868MHz and WiFi's 2.45GHz is minimal.

Thanks, thats interesting to know. What about ability to go through walls etc?...would you say the 868MHz would be significantly better than wifi (2.4GHz) in that regard?

Also, which is the easiest one for your general software engineer to manage?.....an 868MHz link, or a wifi link?.....rememvering that ultimately, the data has to go into the internet at some point.
(i am talking about the route to getting to the household wifi router....ie using either 868MHz , or wifi)

 

[betwixt]

One of those little modules will talk directly to the internet through a normal WiFi router!

As difference between 868 and 2450MHz, it isn't that great. There is no 'brick wall' effect that suddenly stops a frequency penetrating a substance, it is a gradual increase until you get to molecular resonance effects. I find those ESP modules are good up to about 100m from the router in open space but still perfectly functional from room to room inside and nearby outside the house.

If 868MHz is used, some error protection has to be built in to the data stream. I strongly advise using FSK modulation rather than OOK and adding at the very least a CRC check on the data packet. CRC won't fix errors but it does allow them to be detected so a mechanism for re-sending the data has to be implemented.

If using WiFi I suggest you read up on the MQTT protocol. It is very simple to implement and as well as the low level protection inherent in TCP/IP, it guarantees delivery. For example, if the power goes out at the router, it will store messages until it is restored. You can even share or set up your own company cloud broker so units anywhere can use it as a 'post office' for handling data. There are several, personally I use Mosquitto which is free and really easy to use. A Raspberry Pi handles all my device communications as well as running other services.

Brian.

 

[FvM]

Major advantage of sub-GHz radio is (or has been at the time of development) lower operation power, specifically for battery operated devices. Thus it is used for most home automation systems. Slightly higher range inside buildings has been mentioned, also lower software complexity compared to WiFi and Bluetooth plays a role. The fact that product XY uses sub-GHz doesn't mean that it can't work better these days at 2.4 GHz. I'd subsume the verbose complains about failing WiFi communication under configuration problems rather than inherent behaviour.