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: Concerning the maximal number of node when using THVD8000.

Part Number: THVD8000

Dear all!

First I want to say that THVD8000 is such a great gift for us!

As our team was developing a solution for polarity-free power over data bus, using standard RS485 transceiver with external circuitry doing exactly OOK modulation/demodulation.

Then comes the THVD8000!! Like a miracle! 

We'd like to deploy THVD8000 in many application, replacing traditional RS485 bus and SLC-like bus.

However, we have some question concerning the maximal number of nodes on the bus

I am aware of the following documents:

Datasheet : https://www.ti.com/lit/gpn/thvd8000 

Design Guide: https://www.ti.com/lit/pdf/slla496 

And EVM User Guide : https://www.ti.com/lit/pdf/sllu140 

===============================================

According to the DS or UG (also as standard RS485), suppose we are using 120Ohm termination, then the remaining bus loading have to be greater than 375Ohm.

Refer to Figure 4 in DS, or simply by calculation, if we choose the carrier frequency to be 1MHz, and the required number of nodes are 200, then the inductor would be 12mH! 

( or more precisely, 375Ohm * 200 / (2PI*1MHz) = 11.9mH )

This size of inductor seems hopeless, taking for example 

Bourns RLB-series inductor : https://www.bourns.com/docs/Product-Datasheets/rlb.pdf 

RLB1314-103KL has 10mH, but the SRF is only 0.53MHz, the its inductance would reduced dramatically when operating in 1MHz.

Same problem goes if we maximize the carrier frequency to 5MHz:

Even though the inductor size can be shrink to 2.4mH, but it seems even more difficult to get an inductor with that inductance and have sufficient SRF.

Our goal is to reach the maximal number of nodes up to 150!

I am thinking...

What if we mitigate bus loading requirement on inductor by eliminating the terminate resistors.

Say we leave the bus un-terminated, then in 1MHz carrier frequency, 200 units, the inductance would be 54Ohm * 200 / (2PI*1MHz) =1.7mH.

Similarly, 340uH for 5MHz carrier frequency with 200 units. This inductance looks promising when choosing inductors, right?

But here comes the problem of signal integrity, mainly the reflection issues. 

I've never designed a un-terminated RS485, it's kind of something unpredictable for me. 

However, as the carrier frequency is very very fast comparing to a single data bit, is it feasible as long as the reflection dies out within a small amount of a bit time. 

Do I understand this correctly?

===============================================

To recap my question:

1. If using termination, and number of units is up to 150. Is it simply not practical to choose a huge inductor to accomplish power over RS485?

2. If NOT using termination, is there anything we can do about the reflection issues? Simply reduce the data rate, say 1:100 or 1:1000?

===============================================

Thank you guys for reading this post, and thank you again for all the great products offering to the world!

Zt.

  • Zt,

    I'm glad to see you're interested in the device. What's the data rate, communication distance, and power (DC or AC) load in your application? Can you also confirm that the power delivery is needed among up to150 nodes? Please note that the inductor is only required for the nodes with the power delivery function.

    You calculation about the inductor value is correct. However please be aware that this initial evaluation is kind of pessimistic. The location (cable resistance) and other R/C components of the node all can impact the inductor's equivalent impedance. In reality, the inductor value could be much smaller. It's possible that you choose smaller inductor without the termination. You can tell this is a system design choice. If you select a higher carrier frequency, the inductor value is smaller but the reflection would be more severe without termination.   

     I suggest you start with a point to point setup then increase the node number. By doing this, you will check the basic functionality of the device with variable distance, data rate, etc before implementing the more complicated features.

    Regards,

    Hao  

  • Hi, thank you for the reply!

    In our application, the date rate could be low,

    1200 to 115200 can do the job.

    And yes, all nodes require power, and hence inductors are  required for all nodes.

    Zt.

  • Zt,

    What's the communication distance? Another thought is that you could lessen the 375Ohm requirement. This load value and 1.5V Vod are inherited from RS485 standard. The real goal depends on the application. Again, I suggest start with a point to point testing with several different carrier frequency. In each case, you will know what inductor value is proper. I hope this make sense to you.

    Regards,

    Hao

  • Hi, Hao!

    The maximal cable length is about 200m. 

    Thank you for reminding me the 1.5Vod spec, as this is way larger than 0.2V on RX side.

    We are going to buy a pair of EVM.

    Thank you again for the reply.

    Zt.

  • Zt,

    Another point I didn't elaborate in detail in the app note is that the impedance of parallel L and R needs to be calculated in the complex domain.

    https://www.allaboutcircuits.com/textbook/alternating-current/chpt-3/parallel-resistor-inductor-circuits/

    Please let me know if you have any further questions in the evaluation.

    Regards,

    Hao

  • Hi, Hao!

    This is a great point!

    In the DG or DS, the calculation makes us think that the impedance of equivalent inductor has be at least 375Ohm, in the carrier frequency.

    However, what we need is: the parallel impedance of 60Ohm and total equivalent inductor is at least 54Ohm.

    For example, if 1MHz carrier frequency is chosen, the minimal indutance is in fact, about 20uH, not 60uH.

    (60Ohm || 20uH @ 1MHz, Z_parallel = 1/sqrt(1/60^2 + 1/(2PI*1M*20u)^2) ~= 54Ohm.

    Do I understand this correctly?

    Also, I notice that once we take cable inductance into account,

    it seems that cable inductance is dominate if the cable length is long enough.

    Zt.

  • Zt,

    Yes, your calculation is correct (more accurate than the app note :)). Also your comment about the parasitic inductance of the cable is also correct. Unfortunately we haven't found an easy way to evaluate the inductance (or the whole impedance) of the system. The values sometime are chosen empirically.

    Regards,

    Hao