💬 Easy/Newbie PCB for MySensors

[sundberg84]

@dakipro - hi!

but they just wont work reliably on lover battery power then 2.8v . Is that the limit?

No, the limit should be (if you have a genuine radio) 1,9v. Im not sure what kind of voltages a clone will manage, but my nodes goes below 2.8v!

I cannot debug using serial adapter, as when I connect it, then pro works fine as it gets 3.3v from the pc.

This should not be a problem - you can power the node from the "normal" way and just connect RX/TX and GND from the ftdi to the arduino ftdi header. I have done this several times.

D2-gnd show 2.5v

The IRQ line from the radio is 2.5v since thats what the radio gets. Im not exactly sure what this does, but in SPI which is what the radio are using, its either HIGH or LOW so 2.5v sounds right since the radio has that voltage. This should not disturb the arduino! This is changed in REV9 so you can disconnect D2 from the radio if wanted. You should not use D2 for any sensor below Rev9.

The led on arduino goes red for 3-4s, then goes off for 2s then again red for 3s, then off...

Sounds like the radio is transmitting... if this goes on and on it can mean the radio isnt getting connected/ack or establish a way to the gateway.

Do you have some suggestions on how I can debug what is going on?

You need to check the debut output (see above)!
Could you describe the issue? Are the node lost from the controller?

What is the lowest voltage the batteries can go (with the 3.3v booster)?

1.9v is the minimum to the radio.
0,8v is the minimum for the booster.

[dakipro]

Thank you for your answer, the tip about debugging helped a lot.
I now checked on one board and it looks that some pro minis are having problems when battery voltage goes bellow certain level (f.eks. 2.7v).
Testing results:
"good pro mini" works up to 2.0v , on 1.9v it shows fallowing errors

send: 4-4-0-0 s=1,c=0,t=7,pt=0,l=0,sg=0,st=fail:
find parent
send: 4-4-255-255 s=255,c=3,t=7,pt=0,l=0,sg=0,st=bc:
send: 4-4-0-0 s=2,c=0,t=6,pt=0,l=0,sg=0,st=fail:
temp: nan
nan
send: 4-4-0-0 s=255,c=3,t=0,pt=1,l=1,sg=0,st=fail:62
send: 4-4-0-0 s=2,c=1,t=0,pt=7,l=5,sg=0,st=fail:0.00
send: 4-4-0-0 s=1,c=1,t=1,pt=7,l=5,sg=0,st=fail:0.0

"bad pro mini" works fine above 2.7, on 2.6 it logs

84973 MCO:SLP:MS=15000,SMS=1,I1=255,M1=255,I2=255,M2=255
85047 !MCO:SLP:TNR
85874 TSM:FAIL:RE-INIT
85897 TSM:INIT
85919 TSM:INIT:TSP OK
85942 TSM:INIT:STATID=3
85968 TSF:SID:OK,ID=3
85991 TSM:FPAR
86042 TSF:MSG:SEND,3-3-255-255,s=255,c=3,t=7,pt=0,l=0,sg=0,ft=0,st=OK:
88117 !TSM:FPAR:NO REPLY
88141 TSM:FPAR
88193 TSF:MSG:SEND,3-3-255-255,s=255,c=3,t=7,pt=0,l=0,sg=0,ft=0,st=OK:
90269 !TSM:FPAR:NO REPLY
90294 TSM:FPAR
90345 TSF:MSG:SEND,3-3-255-255,s=255,c=3,t=7,pt=0,l=0,sg=0,ft=0,st=OK:
92422 !TSM:FPAR:NO REPLY
92446 TSM:FPAR
92497 TSF:MSG:SEND,3-3-255-255,s=255,c=3,t=7,pt=0,l=0,sg=0,ft=0,st=OK:
94574 !TSM:FPAR:FAIL
94595 TSM:FAIL:CNT=4
94617 TSM:FAIL:PDT
95066 MCO:SLP:MS=4999
95088 !TSF:SND:TNR
95610 MCO:SLP:TPD

What I see on the log parser is that the "No potential parents replied to find parent request" pops out as the biggest issue, suggesting that radio communication is not doing well.

I am testing using the small bench power supply (the Chinese ones with the oled screen) so the current should not be the problem. Why would pro mini be the problem, when it is getting 3.3v directly from same stepup?
The radio and all other components are exactly the same in all the tests i just swap the minis, I think radio would have failed with both minis if it is the radio... or perhaps the signal to the radio is not good/stable enough? No idea to be honest :)

If that matters: they are all getting 2.7v (or what I set) on pin 2 even though I have physically cut the connection to the radio irq (with the knife, on the board, they are not touching anything else on the board).

p.s. now that I write this, i realize that another difference is the sketch on them, "the good one" is "vanilla" mysensors 1.5 code and all the bad ones are latest mysensors using the node-manager library. I do not know enough about chip programming, but as a web developer I would think that the code should not determine minimum voltage needed on board? (or there are some routines in some strange combination doing some magic... hm...)

[dakipro]

Well I just tested with latest example from mysensors library, and we can exclude my latest p.s., nothing to do with using the node-manager (as expected).
Only difference is the pro minis (but again, they should all get 3.3v via popup)?
and on the "good" node I have mysensors 1.5 while on the others I use 2.1.

[sundberg84]

@dakipro - It looks like when the voltage drops, you are having issues with your radio (No reply and ST=Fail).
This should not be a problem, since this is powered directly from the batteries and you say there are 2.8v left.

The first thing that comes to my mind is that as the voltage drops, the booster needs to work harder and introduces more noice into the board. As the noice increase, the radio gets in trouble.
The other I can imagine is that you have a bad clone which cant handle lower voltages.

Both problems are a bit tricky to diagnose. First I would try some different radios, if possible from different batches. Second I would look more close to the booster. They vary greatly in quality and some are just really bad. Sometimes a cheramic capacitor might help (from output to gnd on the booster). You could also just disconnect the booster and see if it works (the pro mini shoudl be able to handle down to 2.8v).

[dakipro]

After several hours of debugging and changing all possible components, i built exactly same circuit on prototype board and with exactly same (active) components I was able to go down to 1.9, sometimes to 1.8, while on the easyboard only down to 2.6. I also used exactly the same booster on both prototype and on the easyboard.

But then a breakthrough, I have found one passive component that was different between the two mentioned boards! It was the radio freaking capacitor, on the "good" board it was 47uf, and on the problematic board had 4.7uf!
I changed radio cap to 47uf and I am now able to go on both boards down to 2.3v. Which is ok, not the 1.9v but it is good enough (for now :) ) So those 42.3uf were missing for radio to go ~0.5v lower in voltage I guess.

The one thing I am missing on the prototyping board is the voltage measurement circuit, so I guess that gives some 0.4v lowest threshold or something, i will test these days on the prototyping board, just to verify.

Thank you @sundberg84 for attention and help :)

Are you able to provide 1.9v on battery input on some easy board with the dht22? Like the one from the photos on the openhardware?

[sundberg84]

@dakipro - the radio capacitor is crusial which has been proven many times before so good you found out your problem. What can be a pain with the radios is that some clones/batches seems to work with on capacitor value and the next need some higher/lower value on the capacitor to work at its best.

The source is most likley what i explained above, the booster noice which interfere with the radio. On a breadboard you have much more space and possibility to put the booster further away from the radio with longer wires. This is most likley your case here...

Are you able to provide 1.9v on battery input on some easy board with the dht22? Like the one from the photos on the openhardware?

I have not made any a large quantity of measurments, because i almost never have to change my AA batteries. I have 5 Easy nodes (different revs though) with DHT22 and the only time I measured it was around 2 if I remember right. Cant remember if it was 2.0 or 2.2 though... I will keep that in mind for the next battery change but the lowest node is at 2.5V at the moment so it might take a while. If I remember right according to Domoticz I will not have to change these batteries for atleast a couple of months.

So the short answer... im not far from 1.9V, but It will depend on the radio and the booster quality.

[dbemowsk]

@dakipro Jut to test, now that you have your voltages closer with the change in capacitor, you may want to just try swapping the radio modules between the two (provided you have them in sockets) and see if the voltage drop follows the radio.
.

[dakipro]

both boards are now identical, they go down to 2.3, but I will put 2.4 in scripts as the minimum, just to be sure.
If a board boots at 2.4 it can go down til 2.2 and maybe 2.1, and it will work, but cannot reboot at that voltage.

[sundberg84]

@dakipro - 2.4 volts will last you a long time.

If you want to go further I would suggest you look at the advanced user section of the EasyPCB and remove the booster and lower BOD instead. This will not work with the DHT22 but you could change that to BME280 and the radio @ 1.9v will be your lowest point. Either that or go for a more expensive booster. Im accually trying this in another project (https://www.sparkfun.com/products/10967) so I can give some feedback in a near future.

Im also building a variable power supply which might help me to try different volt levels on my boards as well... all I need now is a oscilloscope - anyone wants to fund (and teach) me :) ?

[dakipro]

I will consider that for new nodes. Where is a "advanced user section of the EasyPCB", i tried searching the forum and openhardware, but not sure what exactly you are referring to?

[gohan]

@sundberg84 wow, that's a quite expensive booster :cold_sweat:

[sundberg84]

@gohan - yes almost 5 times more expensive but I'm building it myself and it's probably five times better...

@dakipro https://www.openhardware.io/view/4/EasyNewbie-PCB-for-MySensors check "Battery without step up booster (advanced users)"

[Nca78]

@gohan said in 💬 Easy/Newbie PCB for MySensors:

@sundberg84 wow, that's a quite expensive booster :cold_sweat:

The chip itself is less than 0.7$ on Arrow.com (so, legit one) so with additional components it should be 2$.

But I see in the datasheet that typical operating current for 3.3V version is 45uA, added to the 15% of losses (max 85% efficiency) I'm a bit skeptical about the battery life you can get using it ?

@sundberg84 if you find someone funding oscilloscopes, please send me the contact :D Else you have the captures in the datasheet for the most critical cases :P

[sundberg84]

@Nca78 45 uA will last me a long time. On my dht nodes I aim for <100uA so if you sleep between readings that's ok. If this circuit would be much better than these cheap eBay boosters I'm happy even if they run at a couple of more uA

Will let you know if I find oscilloscope funder 😉

[Nca78]

@sundberg84 said in 💬 Easy/Newbie PCB for MySensors:

@Nca78 45 uA will last me a long time. On my dht nodes I aim for <100uA so if you sleep between readings that's ok. If this circuit would be much better than these cheap eBay boosters I'm happy even if they run at a couple of more uA

Will let you know if I find oscilloscope funder 😉

But I still fail to see the interest of all the pain with the boosters and keeping the DHT22 which need "high" voltage of 3.3V when a sht21/si7021 is 3$ or less on aliexpress and can run way over 1 year with a simple cr2032 ? And if you use 2 AA it will run for 5 years. No problem with noise/stability, no booster to buy, just a 0.x$ big capacitor to add in parallel with the battery if you're using a CR2032 (and nothing if you're using 2 AA/AAA).

[dbemowsk]

@Nca78 Another option that I went with is an HDC1080. The only problem with this one is that according to the datasheet, the recommended operating minimum voltage is 2.7v, where the si7021's minimum is 1.9v. When I was looking on ebay, these said that they were a replacement for the si7021, and it was my mistake not checking the datasheet before buying. They do seem to be working good though and I have had one of them running for about 5 months with no trouble so far.

[sundberg84]

@Nca78 that's why I added the option without boosters. The only downside is that it's not that easy changing bod. When you get the booster function to work you can also run a motion dec in 3.3v for example. 3.3v will give you more sensor options.

[dakipro]

@sundberg84 said in 💬 Easy/Newbie PCB for MySensors:

Battery without step up booster (advanced users)

I could just (do it very very ugly) and cut the vcc line near the board and solder it directly to the battery input?
That way both arduino and radio will get battery power, and dht22 would still receive 3.3v? (then I can lower the BOD)
Or would that not work due to some other connection?

Also, would flashing pro mini to 1Mhz benefit battery consumption at all? (while still using 3.3 booster)

(p.s. I stole your signature :) )

[dbemowsk]

@sundberg84 That is one of the nice things about your board is it's flexibility.

[Nca78]

@sundberg84 said in 💬 Easy/Newbie PCB for MySensors:

@Nca78 that's why I added the option without boosters. The only downside is that it's not that easy changing bod. When you get the booster function to work you can also run a motion dec in 3.3v for example. 3.3v will give you more sensor options.

Well, it is pretty easy IMHO, you just have to follow on of the many tutorials, and you learn a lot on the way :)
For motion detection isn't a booster a bad idea ? You can use AM312 anyway, very small and very stable, no soldering necessary to run it down to 2.7V meaning you can use most of the capacity of a lithum cell like CR123.

@dbemowsk thanks for the information I didn't even know about this sensor, looks like a very precise one with a very low power consumption. Can you link to where you bought it from I see nothing below 4$ on aliexpress ?