• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 33 subscribers
  • Views 101 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

THVD8000: Noise immunity for THVD 8000 installations

sajul babu
Prodigy 85 points
Part Number: THVD8000
Other Parts Discussed in Thread: THVD8010

Tool/software:

Hi,

  We are using THVD8000 as power over RS485 for our application and this works just fine in Engineering testing scenarios where the differential A & B lines handle all the noise for up to 400+ft of cable(this is the max length we use in field) . Our product is a Thermostat which powers from the Roof top unit over the 2 wires and recently in some site conditions we are facing issue with high noise on the RX lines which interfere with our communicated signals and cause disconnections. 

The implementation of Power over RS 485 was done based on the app notes available from TI and circuit at Thermostat end and Roof top unit controller are attached.

In both devices 48V TVS diodes are provided near terminal block for A & B lines

In the case of severe noise, we identified that the noise is coupled from the cable ( mostly 18/2 thermostat wires which are non twisted and non shielded) and when we use a shielded cable instead with the shield grounded at the controller end which source power, the noise is gone.

Below are the communication specifications

USART Baud rate to THVD800 is 9600

Carrier frequency is 250Khz

DC voltage coupled is 28V DC

The type of noise see at receiver end looks like 4-5 uS dips as shown duration pulse low signals

  • 0
    TI__Guru 52172 points

    Hi Sajul,

    So there are two ways to go about fixing this issue that are most likely to bring success. 

    First - I'd maybe look into the THVD8010; it is pin to pin with THVD8000 and you wouldn't have to change anything else about the schematic. The THVD8010 is much more immune to noise in general than the THVD8000. The trade off is that you can't get a modulation frequency above 300KHz and you wouldn't be able to max out the distance because the data input thresholds require a stronger received signal - but when I say max out bus distance I mean > 1km /3280ft. Since you are using 250KHz modulation  and have a max of 400ft the THVD8010 sounds like it would be a good fit and you wouldn't have to change much else. Many applications using the THVD80x0 family of devices does use 18/2 thermostat cable - but in noisier environments it can couple in noise and the THVD8000 has a very sensitive receiver - the THVD8010 doesn't and that may be better in this specific application. 

    The second option is more involved - it basically breaks down to noise mitigation:

    1. Shielded cabling is a good way to avoid modifying circuit - but from experience I know that is generally more expensive and there are generally cheaper ways to modify the board - basically the sheilded cable works but I don't know many applications that use that due to the expense so I wouldn't be surprised if that was the case here. 

    2. Next is adding additional filtering to the bus - there are 3 major things I'd focus on here:

    2a) Adding filter caps - so close to the A/B pins of the device having a 50pF to 100pF from A to GND and B to GND can help reduce magnitude of higher frequency noise coming into the bus.  

    2b) Having a split termination - you currently have a 120 ohm and 0 ohm resistor in series - if you change those to both be 60 ohms and have a capacitor to ground in-between you can create a split termination - which acts as a low pass filter where fc = 1/(60 * 2 * pi * C_split) - where you pick C_split as the corner frequency -I'd pick it above 250KHz - ideally that would be 250KHz + 25% (modulator tolerance is +/-25%) would be what I target for which would give a C_Split value of ~8.5nF.

    2c) I have also seen success in adding a common mode choke to the communication nodes.

    I think that would help reduce the total noise - but honestly I think the easiest and cheapest way forward is to just switch to the THVD8010 because it is way less sensitive to noise. 

    Please let me know if you have any other questions!

    Best,

    Parker Dodson