This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

THVD1452: Short-circuit protection

Part Number: THVD1452
Other Parts Discussed in Thread: TMUX7462F, THVD2442, THVD2452V, THVD2429

Hi all,

 

Maybe a "stupid" question, but I would like to be sure.


We are using the THVD1452 in combination with TMUX7462F.

The THVD1452 should provide protection for shorts between A<->B, A<->Y, A<->Z. What happens with a short with GND?

The TMUX7462F, will it react fast enough to protect the THVD1452 if, for example, 24V is applied to A,B,Y or Z.

 

Greetings,

Tim

 

 

  • Hi Tim,

    That is not a stupid question at all - RS-485 isn't really something someone ever learns until they need to so no worries - and we definitely would prefer to help iron out all the design questions before you hit problems later on so please don't hesitate ever if you need to reach out to us. 

    THVD1452 is RS-485 compliant which means the IC is protected from short circuits to any voltage in-between -7V and 12V where the max current through the driver would be capped at 250mA per standard. 

    The abs max on the bus facing pins (which I am assuming is where the mux would be placed)  is +/-18V so 24V is not allowed on the A/B/Y/Z pins (except very short bursts that would be covered under EFT or ESD ratings). Here is the issue on that one - the mux you are looking at has a fault protection and the fault response time is typical 150ns but can be above 400ns - EFT is the "longest" lasting transient rating that we have on the 1452 and the duration tested is 100ns bursts - so due to the duration if the mux took too long to respond to the fault you could see potential damage on the transceiver and if you came back to us we'd probably point out that the fault protection is too slow and it creates an area of potential operation that could damage transceiver we generally would consider that an apps issue. 

    That being said - I imagine you are looking for fault protection - and that is why you have that switch there since it is a 4 channel 1:1 - in general multiplexers and RS-485 transceivers don't mix well because you really need to run the mux at +/-15V to ensure you can support RS-485 signaling - common mode ground shifting is extremely common (even moreso if you are looking at a pretty heavy duty fault protection). 

    If you are really looking for fault protection we have better stand alone solutions - the only real reason you would be adding a mux to the line for fault protection is in case you need the line to be broken by the switch due to down-stream devices - but since it is an RS-485 bus they should be other RS-485 device (however I am not sure if you control the entire bus - a lot of RS-485 designers don't - and if that is the case there could be a need for a switch):

    Option 1: THVD2442 - 20Mbps Full duplex RS-485 device with +/-70V stand off protection. If you don't need the full 50Mbps of the THVD1452 the THVD2442 is basically very similar except: it has +/-70V stand off protection on bus pins, its open/idle/short fail-safe actually gives you a higher hysteresis  without any passive components needed on the bus. The 1k cost on TI.com is actually cheaper than just the mux you are looking at - so you would get the high voltage protection because essentially these are just more robust versions of other RS-485 devices. Mux is $2.737 for 1k units on TI.com and THVD2442 is $1.749 for 1k units

    Option 2:  THVD2452V - if you need the full 50Mbps along with the protection of the 2442 this is the right device. There are some extra features (integrated level translator allows for split supply option and slew rate limiting feature if you aren't using the full 50Mbps to help with EMC concerns) which you wouldn't need - so that will make this part a bit more expensive a $2.17 for 1k units on TI.com - which still puts it under the mux price. 

    Option 3: THVD2429 - this device probably won't fit if you need the full duplex as it half duplex (you can make a full-duplex system but it requires multiple IC's per node so most likely not a great option) - however it integrates surge protection and has a bus voltage support of up to +/-42V but it only goes up to 20Mbps. This will be our highest protection because it also has surge protection - it also has the cheapest option with the DRC package at $1.683 on TI.com for 1k units (however if you need full duplex you would 2x the parts so not really cheaper then) 

    So I think option 1 or 2 would be best - it can help simplify the solution to a lower BOM cost. I am telling you this because I have never seen anyone use a mux/switch for fault protection on an RS-485 bus - there are a lot of downsides to mixing multiplexers and RS-485 because of how the power tree ends up looking (you need 3.3V to 5V for RS-485 but you would need a mux that does +/-15V to properly support signaling) and on resistance is a little high (but that isn't the real issue). The RS-485 device will never output too large of a signal (it won't output more than 5V with 5V supply) so I am struggling to see why you would need the mux (which seems to be used as a resettable fuse essentially) to cut off the bus if the local node can handle the higher voltage - however you are the expert in your system and not me - but I will say from our side the current suggested solution has risk of transceiver damage because the duration of potential electrical overstress (EOS) exceeds the max duration of our EFT ratings for a single pulse so technically your application would fall out of "proper implementation"  if you move forward as is. 

    Please let me know if you have any other questions and I will see what I can do!

    Best,

    Parker Dodson

  • Hi Parker Dodson,

    Thank you very much for your detailed answer. Both option 1 and option 2 sound very intersting (full-duplex is indeed needed). The design choise behind the mux (indeed acting as an eFuse) was to use the exact same transceiver used in other parts of our system. This was to minimize risk in design errors. However, as you pointed out, the "eFuse" is not acting fast enough AND is not designed to be used with the transceiver (ground shifting was not really a big concern with only +- 50 cm of cabling).

    However: it is unfeasable to change the transceiver (as this is not on the PCB we want to change). Is there any other solution to protect the THVD1452 to signals up to 24V.

    Greetings,

    Tim

  • Hi Tim,

    Completely understand the limitations on the layout changes. 

    The reference design you found is a good compromise if you can't change the transceiver - the SIDAC and resettable fuse combination help alleviate a lot of the major concerns that just a fuse alone (which essentially is what the TMUX would acting like) don't address. One other thing that can help improve robustness is to have pulse proof 10 ohm resistors between the IC and SIDAC (usually wire wound or thick film - thick film is more common in modern applications) that help reduce stress from transients during the initial clamping period- these aren't explicitly required and if you find it doesn't make practical sense to add these in your application it should be fine as most applications with some sort of clamping don't include them - but they do add robustness overall to the system at the cost of some slight reduction in the output voltage for a terminated system and unless you are really maxing out length of the application it usually isn't a concern. 

    But please let me know if you have any other questions and I will see what I can do!

    Best,

    Parker Dodson