2
\$\begingroup\$

I recently started to explore energy meters with the ability to transmit data over RS485.

The specific meter I'm using is an Eastron SDM72D V2. I made a simple program to request data from the meter by requesting specific registers and it works without any problem. To perform continuous measurements I hook up a raspberry pi to an "TTL to RS485" board like the one attached and started to receive data. After a few days the "TTL to RS485" becomes inactive and to my surprise it had literally blown up. I have used 3 boards by now always with the same result after a few days (blowup MAX485 chip). I'm not sure the chip is a clone as it has no markings.

The connections are very simple just directly connected ground and vcc to the raspberry pi and TX and RX to the pins for the serial out on the raspberry. Then connected the A and B directly to the eastron energy meter.

The raspberry pi is powered by a din rail power supply and the energy meter is powered directly from the powerlines it is measuring.

This worked well at the beginning without any blowout but it seems to have start happening after a second EV charger install where we sometimes have 2 7.4kW chargers working. I suspect probably some static discharge but I really don't know the cause or how to prevent this situation.

Please let me know any ideas to protect the "TTL to RS485" against this blowouts. Thanks!

Image of a blown out "TTL to RS485" (visible hole in the "MAX485" chip, sometimes it only creates a small bump): enter image description here

EDIT: I have tested both with or without ground connection to the meter with the same results. The boards have been purchased from alibaba so probably counterfeit as they were also very cheap. I have a different model purchased from a local electronics store which is slightly different and has the ic markings (please check attached photo). I might test this one next to see if the same outcome arises.

enter image description here

Update: Sorry for the late update but had higher priority tasks, now I'm back at tackling this issue. Here is what I have tried since last time:

  • Connecting the ground between meter and the "TTL to RS485"
  • Connecting the ground between meter, "TTL to RS485" and Earth
  • Used the "TTL to RS485" with the IC markings from a local reputable electronics store (no grounds)

All cases resulted in a blown up "TTL to RS485" module..

I'm now thinking that it may be the meter that is out of spec in a way. Nevertheless I have yet a final test which is to use an USB to RS485 adapter, a configuration I'm using in a working installation for more than an year with the same model of meters (3 of then instead of just 1 on a single RS485 bus) but with a rpi4B.

I will post the results once I managed to complete both experiments (using the same meter and the USB adapter and changing the meter altogether).

Also is there any way to validade if the meter's RS485 is busted in any way that is making the "TTL to RS485" to blow up?

\$\endgroup\$
11
  • 2
    \$\begingroup\$ How come there was no0V connection to the meter? The protection will probably not work in that situation. You must connect the 0V from the board to the meter. \$\endgroup\$
    – Kartman
    Commented Nov 18, 2022 at 13:21
  • 1
    \$\begingroup\$ I understand that 接大地 means "connect to ground" \$\endgroup\$
    – jonathanjo
    Commented Nov 18, 2022 at 13:23
  • 1
    \$\begingroup\$ Seems there is no 0V connection on the meter. I’d suggest using an isolated RS485 interface. \$\endgroup\$
    – Kartman
    Commented Nov 18, 2022 at 13:25
  • 1
    \$\begingroup\$ I have tested with the 0V connection to the meter and without it has well. The only visual damaged caused by the fault is really the hole on the ic the other markings are from soldering paste. I have purchased this boards from alibaba but at the time really didn't know much about RS485 so looked for the cheapest \$\endgroup\$
    – pedroply
    Commented Nov 18, 2022 at 13:34
  • 2
    \$\begingroup\$ Any form of data communication assumes that all nodes have the same voltage and ground potentials. Otherwise nothing will work and things might blow up. In case of RS485 that does indeed mean either ground everything together or use galvanic isolation with a secondary grounded towards the target. \$\endgroup\$
    – Lundin
    Commented Nov 18, 2022 at 14:46

2 Answers 2

5
\$\begingroup\$

It would be nice if you would provide a pseudo-schematic or block diagram of your setup, with details of how the power and grounds are hooked up.

But even without seeing that, I'm going to guess that for some reason you're exceeding the common mode voltage range (Vcm) of your '485 receiver. Most '485 receivers have a Vcm of -7 V to +12 V, though some may be better than this. If you exceed this range, you could damage the receiver.

The hole in the IC package would also be consistent with this kind of failure, in that you need enough energy to blow a hole in the package. Localized heating of the IC die could do this. So could a blown/fused bond wire. I do not think (though I cannot say for sure) that a typical ESD like event would do this (lightning strikes aside). But a load dump (suddenly removing a high current load from the system) could cause a voltage shift between the various grounds with enough energy available that could cause this kind of damage.

You said the problem started or has gotten worse after the second EV charger was installed. This also points to a ground (return) shift, or ground difference between the meter and your '485 board.

\$\endgroup\$
3
\$\begingroup\$

If the basic functionality works, but several boards have been destroyed, either the boards are below specification or your environment is more hazardous than typical.

Consider using

  • high-quality branded convertors
  • better ESD protection
  • isolation

I've had good results from Digilen isolated RS-485 boards, but many manufacturers are available.

Also, I understand that 接大地 means "connect to ground", but it appears that it hasn't been soldered to. You can get difficulties with this.

I strongly recommend reading Bob Perrrin's indispensible Art and Science of RS-485, especially the portions about grounding.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Thank you for the suggestions. I will for sure look into the Art and Science of RS-485. I'm getting convinced that boards I purchased are the culprit. \$\endgroup\$
    – pedroply
    Commented Nov 18, 2022 at 13:46
  • 1
    \$\begingroup\$ The most common property of cheap board and parts is that they work under good conditions and fail more readily under poorer conditions. The property of well-made devices is they really do perform like the datasheets say, and their extra cost goes into testing, QA, documentation etc. \$\endgroup\$
    – jonathanjo
    Commented Nov 18, 2022 at 13:56
  • 1
    \$\begingroup\$ The use of this particular phrase "大地" specifically refers to the Earth ground, otherwise it would just be a ground (地). So yeah, it's possibly an ESD issue due to insufficient protection. Connecting it to PE might reduce the possibility of damage, and this might be why they've put a notice here... \$\endgroup\$ Commented Nov 18, 2022 at 22:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.