@sundberg84

OK, looks like I had multiple problems.
I somehow missed the instructions about "don't solder both", and was also slightly confused about the directions because I am using a battery (LIPO), but following the "5V regulated" instructions. When the instructions mention battery, it always mean a 3.3V battery. I had the BAT jumpered, and then I tried with and without the REG jumpered.
At the same time, I believe I damaged the MCP1700 regulator - I will have to take it off and test it.
So in the interest of getting this resolved quickly, I started all over again with a new board, new regulator, and jumping only REG. It worked! Here is the picture:

You can see that the radio will get 3.3V, and the rest of the board will get 4+ volts. (The TP4056 will cut it off as it goes down to around 3V).

So I will continue the build, and if everything goes according to spec, the radio, arduino, and sensors should all be OK. (Assuming the sensor can handle 4+ volts).

As far as the capacitors, it looks like the spec says 1uF on input and output. So I will put those where there are labels for .1 and 10uF. For the radio, I am well aware of its capacitor needs (learning the hard way). At this point, I always put the 4.7uF cap directly on the radio.

If everything goes well, I would recommend putting some sort of reference to LIPO batteries in the instructions. I couldn't be the only one who wants to do this.

Great support from everyone!

@sundberg84

I took a chance and tried these boosters. After adjusting the pot to boost to 3.3V, and using wires to connect, they seem to work the same as the "recommended" ones. I am hoping they produce less noise for the radio - I seem to have trouble with that, even with adding the capacitors. (Maybe my radios are not good).
Another experiment - boosting them to just over the 2.8V fuse limit. I would think that would be more energy efficient.

I use an accelerometer ( search for MMA8452Q in AliExpress), along with code that checks the orientation:

byte pl = accel.readPL();
   switch (pl)
   {
     case PORTRAIT_U:
       Serial.println("Portrait Up");
       break;
     case PORTRAIT_D:
       Serial.println("Portrait Down");
       break;
     case LANDSCAPE_R:
       Serial.println("Landscape Right");
       break;
     case LANDSCAPE_L:
       Serial.println("Landscape Left");
       break;
     case LOCKOUT:
       Serial.println("Flat");
       break;
   }

   if ( pl == LOCKOUT ) { // flat
     garage_state = GARAGE_CLOSED;
   } else {
     garage_state = GARAGE_OPEN;
   }

@tekka

OK, I changed it to isTransportReady() and it compiles fine.

Thank you!

@sundberg84

According to the new documentation, my build is now officially approved by you! Of course, it would be interesting to hear what the exact effect of jumping both jumpers is. All I know is that it works. There seem to be a lot of sensors that work fine down to 1.9V, so I would think that this configuration would be useful. (Assuming that reprogramming the bootloader is not too hard - I didn't think so).

So did you consider making the LIPO configuration "official"? So far, it seems that it works fine for me. It is good for when you really want a small footprint and/or your sensor needs 3.3V. Unfortunately, not as long a life as 2 AA, but you don't have to deal with the booster's expense and possible radio interference.

@toyman
Unfortunately, yes, I have returned the jumpers to their original settings, and have even tried all 4 combinations.
I also changed computers, changed OS's, changed cables.

The board LOOKS like it was well manufactured, but I suppose a bad board is a possibility. Other than that, I can't figure out what could be wrong.

@mr_red

I got to the point where the stm32loader.py runs and outputs:

Bootloader version 22
Chip id: 0x410 (STM32 Medium-density)
Write 256 bytes at 0x8000000
Write 256 bytes at 0x8000100
[..snip..]
Write 256 bytes at 0x8001900
Write 256 bytes at 0x8001A00
Read 256 bytes at 0x8000000
Read 256 bytes at 0x8000100
[..snip..]
Read 256 bytes at 0x8001900
Read 256 bytes at 0x8001A00
Verification OK

After that, you're supposed to plug in to the USB directly. I get " USB device not recognized". Zadig doesn't help. It lists it as:

"Unknown USB Device (Device Descriptor Request Failed)"

Any ideas?

Thanks.

@tmaster
Actually, the KY-038 works exactly what I want (except for the power consumption). I also have ones that are connect to mains power, and the false alarms are rare. Also, I have the Arduino programmed to sleep some time after detecting a sound and sending a message. In the case of a real alarm, it will keep sending regularly, which would make it obvious. The "batteries just end in a few months" is what I'm asking about - avoiding the .4ma used by the LM393. I'm hoping I can replace it with the LMV393, but I'm not sure if it will work.
Regarding the "you think you are safe", Mysensors has heartbeats, and battery percentage reporting to deal with all that.

@tmaster
I appreciate the reply, but I really do not want to open up or modify the smoke alarm in any way - just detect a loud sound. As I said, I really just want a KY-038 that is suitable for battery operation.

That scheme is interesting, but I don't want to actually fiddle with the detector - for fear I will cause a malfunction. Hence my request of "without using power from the detector itself." Also, it would be nice to detect any kind of alarm (loud noise), not just a smoke detector.

@ileneken3
I want to detect when a smoke detector is sounding an alarm without using AC power, and without using power from the detector itself. I currently detect an alarm (really, any loud sound) using the standard KY-038 module, like this:

KY-038

But this can't be used for a battery application because the KY-038 has a LM393 on it, which draws .4 ma.
I ordered a LMV393, which is more suitable for battery operations, to replace it. But I'm not sure if that would work, or if I could manage the SMD soldering without breaking the board.
I played with a microphone like this:

https://www.adafruit.com/product/1063

but I can't get it to trigger the interrupt correctly while the Arduino is sleeping. (Maybe an extra capacitor?).

Any ideas?
Thanks!

Re: 'MySensoring' a Kidde Smoke Detector. (Completed)

@mr_red
I can add some of my own "documentation" on this issue. I only recently was able to program an Ebyte E73 module, and it was a long struggle. I used the NRF DK, and had help from a Nordic engineer.

• The recommended way to connect to the external device is:

Using the P20 and P1 headers, connect:
P20 VDD -> 3.3v on EBYTE 2G4M04S1B module
P20 SWDIO -> SWDIO on EBYTE 2G4M04S1B module
P20 SWDCLK -> SWDCLK on EBYTE 2G4M04S1B module
P20 VTG -> P1 VDD
P1 GND -> GND on EBYTE 2G4M04S1B module

To unlock the chip, execute:

nrfjprog -f nrf52 --recover

They also recommend using nrfjprog to program the device, because nRFgo Studio is deprecated. But I used the Studio and it worked fine. I gave Studio the .hex produced by the Arduino IDE using "Sketch" -> "Export compiled Binary".

As mentioned in the previous post, it is unlikely to unlock the board with anything but the NRF DK or a genuine J-Link. (Maybe a good fake J-Link).

Mistakes others may be able to learn from:

• Soldering the module onto a custom board can be hard. If you miss a little solder on SWDCLK or SWDIO, there is no hope. Check continuity on every pin you need!
• Make sure you don't order a NRF52810 module - only NRF52832 . You can program them, but Mysensors is not compatible. It's easy to hit the wrong link on AliExpress as they look identical.
• The NRF DK seems to deliver only 2.8V to the device. I first thought there was a problem, and tried to connect external power sources. But it seems to work powered from the DK.

@neverdie

Yes, I understood what you meant. I will only do the "Export compiled Binary" from the IDE, and leave the drivers as J-Link commander or nrfGo Studio wants them.

In: https://www.openhardware.io/view/376/MySensors-NRF5-Platform it says:
"Currently, the nRF52832 and nRF51822 are supported so we recommend sticking with those for now."

I noticed that my board: (E73-2G4M04S) is labeled "NRF52810". So is that why after programming it nothing works (nothing seen on gateway) ?

If so, I messed up on that order.

But I am also trying the E73 2G4M04S1B, which at least from the AliExpress order
https://www.aliexpress.com/item/CDEBYTE-E73-2G4M04S-BLE-4-2-5-0-long-distance-100m-2-4GHz-SMD-ARM-Core/32820692238.html?spm=a2g0s.9042311.0.0.27424c4dSKBLn7

says it is nRF52832. (It isn't labeled on the board though). For this module, I can't unlock it using "Recover" from nRfgoStudio. It always fails with a message "Recover failed: Unknown error".

From the command line, if I do

nrfjprog.exe -f NRF52 --recover --log

and it comes back immediately with:

Recovering device. This operation might take 30s.
ERROR: Recover failed. Please make sure that the correct device family is given
ERROR: and try again.

The key error from the log seems to be:

2018-Dec-27 12:25:22 . . . . . nRF52_power_debug_and_system_regions
2018-Dec-27 12:25:22 . . . . . . nRF52_write_debug_port_register
2018-Dec-27 12:25:22 . . . . . . nRF52_write_debug_port_register: JLink: JLINK_CORESIGHT_WriteAPDPReg(DP reg 0x02, 0x00000000)
2018-Dec-27 12:25:22 . . . . . . nRF52_write_debug_port_register: JLink: returns -1
2018-Dec-27 12:25:23 . . . . . . nRF52_write_debug_port_register: JLink: (0005ms, 0288ms total)
2018-Dec-27 12:25:23 . . . . . . nRF52_write_debug_port_register: JLinkARM.dll CORESIGHT_WriteAPDPReg returned error -1.

2018-Dec-27 12:25:23 . . . . . nRF52_power_debug_and_system_regions: JLinkARM.dll CORESIGHT_WriteAPDPReg returned error -102.

If I do: nrfjprog.exe -f NRF51 --recover --log

it actually makes LD5 on the DK board flash for 30 seconds, but still returns failure with the exact same message.

Any other tricks for "recover/unlock" of those boards?

Thanks!

@neverdie
Unfortunately, nothing is that simple - at least for me. I have gone in circles with the J-LINK drivers. In order to use the Arduino IDE for programming, you must use the WinUSB driver installed by Zadig. But if you do that, the nRFgo Studio no longer recognizes any J-LINK device. So I re-install RFgo Studio, and then I can't use the Arduino IDE anymore!

But I can work around those issues, because everything programs and works right when programming the device on the NRF DK board. The problem is I can't do anything with the two external boards I tried. With one, I was able to unlock it with a "Recover" in nRFgo Studio, but the programming doesn't seem to cause the chip to do anything (nothing logged by the gateway). With the other, I can't unlock the board - it always says: "Recover failed: Unknown error".
Both boards are E73 - but one is marked as E73-2G4M04S and the other is E73-2G4M04S1B.

Any ideas?