I’m the exact opposite - I’d much rather solder a big SMD resistor than a bridge I will add the extra jumper in and populate the burden resistors by default, though - much cheaper to have it onboard rather than having to order an extra bag. For future reference, eBay is a good source for small component orders, rather than trying to deal with the big distributors.
I’m not sure it will be a regular use case, but I’ll make a note on how to do it - essentially it’s a remapping of the pins and External Interrupt Controller (if you want to look in the datasheet). In the firmware, the pin mappings are a set of defines and arrays in board_def.* - you need to remove the analog input from the ADC array and add it to the pinGPIO_In array. (Just looking at it now, the ADC pins haven’t been named - I’ll add the logical mappings…).
No need to apologise, I’ll make sure it’s clear in the documentation
Wow this thread has done a sharp U-Turn from my negative start to a very bright future!
Overall I think backwards compatibility is important if it is not cost prohibitive, I agree the one wire temp sensors are less critical these days, but I have some on my hot water as part of diversion monitoring and if I upgraded would want to keep them on a EmonPi3
replacing an EmonPi1
Temperature measurements, I never started with EmonTH but I would share as a general recommendation what I am using in case it helps others - which is EocWitt who make weather stations.
You don’t need an entire weather station, they have a USB gateway GW1100/GW1200 that receives the 868Mhz (Europe) sensors and uploads to their cloud but also will integrate to Home Assistant. https://www.ecowitt.com/shop/goodsDetail/299 its about £20-25 on Amazon. The system supports up to 8 WH31 sensors that are around £13-15 on Amazon and report temperature and humidity, AA batteries last about 1 year before charging (Eneloops) outside and longer inside where its warmer.
You can of course start with a full weather station as well, but even if you just want temperature its not expensive option, the 868Mhz radio is reliable and the gateway presents to HA for integration in automation or dashboards.
There are various other sensors, usual wind rain etc, but also temperature of water (on a flying lead), soil moisture for the gardeners, leak detection lightening etc.
The sensors can be had cheaper from China, but the official Amazon store is price comparable when sale prices are available.
@awjlogan Thanks again for the reply, and for adding the jumpers and loading the burden resistors. Also for working out how to measure two pulse inputs and one DS18B20 input.
And thanks for the tip about component sources.
I hope all goes smoothly on the emonPi3 development.
Added the jumper option, you can see it on commit 62dad58. I took the opportunity to tidy up the schematic a little as well, breaking the repetitive CT front end into a hierarchy.
Yes, definitely considered but there are two main barriers: safety and cost. Generally, they are connected directly to mains voltages (see the application information). This means the safety and regulatory work is more intensive (and expensive) - I wouldn’t mind doing it for myself, but as an OEM product, it would not be feasible. And cost: they are 2 or 3 channel devices, whereas the SAMD used on the emonPi3 is capable of 3 phase, 12 CT input on its own. Of course, the ASIC solution is better specification wise, but I don’t think it works in this setting.
I was thinking more along the lins of what forum user John Deglavina did.
His device uses CTs, and an ESP32. Although it could have used other MCUs as well.
It’s capable of being expanded to 42 channels.
Ref:
and
(He sells this version for 55 USD, the one linked above for 75 USD)
If nothing else, some good info in the threads linked above.
Thanks Bill - nice reading material, and John has done a good job with it. In principle, it could even be extended further, but I suspect he’s run out of pins on the ESP I will certainly do some more reading - I think for this version, at least, it will remain with the microcontroller front end and processing, but there’s certainly advantages to the dedicated ASIC.
@awjlogan Sorry - one more (rather late!) comment -
Looking again at the picture of the emonPi3, I see that it has screw terminal blocks. I’m guessing these will need the case to be removed to allow access to the clamping screws, which is not very convenient. Also if you need to move an installed emonPi3 for any reason, you have to work out again which wire goes into which terminal.
I much prefer the plug-in terminal blocks used on the emonTx4, as they don’t have these problems, although they add a little to the cost. And they may make testing the emonPi3 easier/quicker - no need for loose wires for the terminal blocks?
@rupert I think those screw terminals are just in the render for convenience, the actual part used will be the plug in type used on the emonTx4 and emonPi2
Yes, as @TrystanLea said, just for rendering purposes - I’ll see if I can find the correct STEP file, it would be better.
@Gwil - managed to find space for a single mounting hole for the OLED display. Needed to rearrange a few tracks and fills, but nothing too drastic! It’s in commit 9e6175c. Unfortunately the other 3 holes completely overlap something solid, but it should provide a bit of extra stability.
If an ESP is used (on the serial pins) then I suspect one of the buttons will be inaccessible. Might have missed it, but there seem to be 2 ‘reset’ buttons?
Hi Brian - the bottom (as viewed) button (SW2) is an optional “user” button for anything that might come up in future, or you want to add yourself - the reset button is the top one (you can just see RESET on the silkscreen). There’s an additional couple of LEDs there (LED3+4) for general user defined purposes. I suspect they will not be useful for the emonPi3 in the current case anyway as it will be covered by the case and I doubt there’ll be a respin of the case as it’s expensive as is.
For the ESP module, I’ve got a little design that slots underneath (same as the Pi) and extends out at the same point as the Pi’s USB/ethernet. I think before getting too far down this route, a suitable case should be found - that’s always the start as PCBA is much cheaper than small volume case production. Ideally, it would be plastic so the WiFi will propagate. A metal case means extending the antenna outside the case, realistically. I’ll have a look for cases as well.
A little AVR-DB hardware update: We are gearing up to having the emonTx5 unit and 12 CT option (which can be applied to the emonPi2 or the emonTx5) available in the shop. The ETA should be by the end of the month / early May.
We are still on the first batch of emonpi2 boards, the emontx5 will be using current stock. I need to do the repeated test over multiple boards that you recommended.
If there is time, do you think there might be some space to punch out a hole for an LED indicator on the non-CT panel? There’s obviously not a huge amount of space, but it might be possible to squeeze something next to the SMA connector and pulse/temperature sensors. Useful for a powered on sanity check, if nothing else.
I will add the extra jumper in and populate the burden resistors by default, though - much cheaper to have it onboard rather than having to order an extra bag. For future reference, eBay is a good source for small component orders, rather than trying to deal with the big distributors.
Thanks for the update.
)