Waking up a serial port

That sounds more like a USB-A connector pinout:

See if there's any data on the Rx pin. USB data runs at 12 to 480 Mbits/sec. What's the maker and model number of your oscilloscope? What's the scope bandwidth?

If the connector is USB, the data is differential between the + and - pins (which may be the Rx and Tx pins). RS-232 is specified at

The inverter manufacturers name and model number would also be helpful. Good luck.

Begin Mini_Rant()

1. What problem are you trying to solve? (Keep it simple)
2. What do you have to work with? (maker, model, versions, photos, URL's)
3. What have you done so far and what happened?
4. What is your ability or experience level?
5. Optional - All the background data and history. End Mini_Rant()

I'm trying to figure out if I can talk to the controller

Owon vds1022i usb scope, 100Ms/sec, likely 10 MHz bandwidth. Board images are at zefox.net/~bp/ampinvt/2nd_inverter/board_photos It sure doesn't look like a USB port to me. If it is, my scope is blind.

Connect the 'scope, power cycle the inverter, look for traffic on Tx.

Retired experimental physicist, from the pre-digital epoch.

The inverter behaved strangely (but usably) when acquired. Eventually I replaced it with another unit and would like to understand what ails this one if it isn't too difficult. It's still usable as a spare, so I'd rather not damage it.

My working guess is that the software is misconfigured. The vendor's behavior was completely obfuscatory (if that's a word) so I'm not asking them for anything. Their products are sold cheaply enough that buying three to get two that work was still the best deal available.

Thanks for writing,

bob prohaska

The wiring order and designations sure look like USB to me:

Is there +5VDC on the 5v pin? Try using the scope in differential mode. Channel A to + and Channel B to - with the scope looking at (Ch A minus Ch B).

Ok. I'll adjust my vocabulary to minimize digital technobabble.

I need some time to chase the vague references to the possible maker of the inverter (Ampinvt and Sigineer). Got any better info? Photo?

If the inverter was working (strangely), I suspect you might get some clues by looking at the output waveform with your oscilloscope (in differential mode). You can do some amazing things with software, but trashing the AC output waveform is not one of them. However, acting strangely (whatever that means) is possible.

Is your inverter one of these?

The URL is

The photos show markings on the board that might be related to the manufacturer, but I couldn't find any online references.

The inverter is sold by Ampinvt as model HT80112 on Amazon. The idea that it's a clone of a Sigineer design is something picked off the web and worth no more than I paid for it. The board photos might offer more clues, but I couldn't make anything of them.

The oddity was battery charging behavior, the output behavior seemed perfectly fine. One LED flashed to indicate charging and should become steady when charging completed, but instead it kept flashing. Eventually the LED simply went off permanently, at which point I replaced the unit. Far as I could tell the actual battery charging behavior was correct. It made me think a PID controller might be mistuned, when it went off that seemed more serious.

There's a jack for an RS485 remote control panel, I was hoping the apparent serial port might expose configuration parameters if it runs a Linux kernel with a serial console.

Thanks for writing!

bob prohaska

USB transmits and receives on the same two lines. The two lines are a differential signal, usually marked D+ and D-

If the pins are labeled Tx and Rx, I'd very much expect the port to be asynchronous serial, not USB.

I am not aware of any standards.

Some devices will respond to a few CR-LF, as in pressing the enter key a few times, but those are probably mostly intended to be accessible by the user.

A device like this may very well have the serial port disabled in firmware.

I did briefly try to connect a usb-serial adapter, hoping to see at least garbled boot messages during power up by varying baud settings. Nothing at all appeared on the terminal screen. That's when I tried the 'scope and found nothing.

There are some board photos at

Thanks for replying,

bob prohaska

I'm not seeing an MCU, although it could be underneath or on another board.

Sometimes a UART is for the instrument to communicate with some other device, which is something that would have to be configured somehow. Or maybe it's designed to be queried from another device, ie you need to speak before it will answer. It's quite likely not designed to work like a terminal with printable text. Maybe it's only for use in debugging or factory test versions of the firmware and you don't have one of those.

There's no way to know in general.

Theo

It turns out there is an smd package with at least 40 pins on the perpendicular daughter board at the lower of the photo at

Unfortunately it's on the side that can't be seen in the photo. It's mounted diagonally, as one sees on images of SBCs. I've tried to get a look at the package markings but can't make them out vi a dental mirror, too much stuff in the way. The packaging is rather tight, so improving the view won't be easy 8-(

Obviously it might help to identify the device. Then again, it might not.... I'll try to figure out how to get a good look. Something like a camera periscape the size of a pencil would be about right 8-) I'm hesitant to start taking things apart because the device is still usable, if not entirely perfect.

Thanks for writing!

bob prohaska

Try an endoscope camera:

If you leave it torn apart, you'll eventually lose some parts or hardware which potentially will make it useless. That's what recently happened to me with a Chinese 12VDC to 115VAC power supply. Best to make your decision fairly soon.

That looks qualitatively correct for the package shape. It's harder to judge exact size, but it's close.

I tried looking on Amazon (no Ebay account and no desire to get one) and would prefer something that does not rely on a phone, but rather a desktop computer running Linux. Apparently many such cameras support guvcview but haven't found one that explictly says so on Amazon. I've posted a question and am awaiting a response.

Agreed, it's back together for now.

Thanks for writing!

bob prohaska

Inverter/chargers are often provided with a basic communications method, if only to turn off and on remotely.

These can use any com hardware or protocol mixed and matched to do the job - usb sockets spouting rst232, network connectors doing the same.

Really popular and annoying are bluetooth dongles that will only talk through an 'app.

RL

This unit has a socket for connecting an RS-485 remote panel. Far as I can gather it merely duplates the function of the built-in panel, a five-button affair with one power button plus four: (up, down, enter, back). It's not clear how to tell if it's active, nor how to communicate if it is. Surely the manufacurer won't help... Any ideas are welcome!

Thanks for writing,

bob prohaska

A rough rule of thumb is if you can see a DRAM package (likely to be a BGA) and some storage (NOR/NAND/SPI flash) then there's a good chance it's running Linux and the serial port will give you messages from the bootloader if nothing else. If it just has an MCU then it's likely running bare metal firmware and I'd not expect to see anything out of the UART, although it wouldn't hurt to probe it and confirm.

The fact the power is 5V and not 3.3V does suggest it's more likely to be an MCU.

Theo

I've tried connecting a usb-serial adapter to the Tx pin and watched for output during power-on, at a variety of baud rates. I really expected to see something, even if it was garbage. Nothing at all. Then I tried to use my Owon VDS1022i and couldn't see anything. It's sounding like I'm barking up an empty tree. Next time there's an opportunity I'll try to get a better look at the daughterboard. But, not holding much hope....

Thanks to everybody who replied!

bob prohaska