• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 32 subscribers
  • Views 79 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

THVD1429: Differential waveform looks worse than expected

Florian Kaffanke
Prodigy 170 points
Part Number: THVD1429
Other Parts Discussed in Thread: THVD1450, THVD1420, THVD1400, , SN65HVD23

Tool/software:

Hello,

we use two RS-485 transceivers to transmit data with  max. 12.288 MHz (differential manchester encoding) over Cat 5e cable. One transceiver is the transmitter, one is the receiver. The roles are randomly determined by connecting two devices for the first time. We initially aimed for 50 m cable length. During development we tried different transceivers in the following order:

  • THVD1450: Worked very well up to 50 m with margin. But we didn't pass the radiated EMC test. We assumed that the fast edges of this 50 Mbps transceiver played a role.
  • THVD1400: We changed than to THVD1420 with only 12 Mpbs, being aware that our datarate is a already out of spec. But we passed EMC tests and the communication link worked. Later it turned out, that  the link doesn't work reliably. With some devices it works over 50 m with others only over 10 m. Waveforms looked almost identical to THVD1450 so we assumed we're hitting the 12 Mpbs spec limit.
  • THVD1429: We changed to the THVD1429 (20 Mbps) which works reliably over 30 m which at this point was OK for us. But strangely enough the waveforms looked really bad, see below.

The schematics are the same for the transmitting and receiving end.

Oscillograms from the THVD1420 over 50 m. Both differential signals on the left side, eye diagramm of the differential signal on the right side

THVD1429 with exactly the same test setup. It looks considerably worse, especially for the short bit times. The waveform looks not like the typical RC charging waveform. It's more like a straight at the beginning which indicates a current source. But I couldn't find any clues in the datasheet like a lower drive strenght that would explain this.

Any help is appreciated. Right now I'm not so sure the THVD1429 is the right part for the job or if there's something wrong with the schematic.

Best regards,

Florian

  • 0
    TI__Guru 56270 points

    Hi Florian,

    This is kind of a tricky situation. To get further distance, you need a device that has a faster slew rate. But using a higher slew rate will generate more EMI. Using a slower device with a slower slew rate will end up with less distance it can support (it takes longer for the device to charge/discharge the bus capacitance). 

    From the schematic, it seems like the customer's using good design for lower EMI (split termination with cap in the middle and what appears to be a common mode choke to help against EMI). Other minor things would be like ensuring the PCB traces are exactly the same length, using good cables that the same length (twisted pairs can sometimes be slightly mismatched lengths) and making sure the termination resistors are low tolerance. 

    I'm not sure if this is an option but we do have 2 devices in our portfolio with an equalizer circuit on the receiver (it basically boosts what it sees on the input so it can support higher distances or in your case maybe pick up more of the signal at 50m). SN65HVD23 and 24 (seems like the 23 would be your choice based on the speed you want to support). If your set up has the receiver set up so it doesn't need to change to become a transmitter then you could try to use that device and try to pair it with the THVD1429 to see if that combination would work. 

    -Bobby

  • +1 in reply to BOBBY
    Prodigy 170 points

    Hi Bobby,

    thanks for your answer. I will definetly look into the two devices you mentioned. Right now we will stick to the THVD1429. We found out that the specific cable type has a huge impact on the signal shape. With a high quality cable we can reach 50 m.

    Below are two eye diagrams. The first is from a random 40 m CAT 7 S/FTP cable bought from Amazon which didn't work. The second is from a (pricey) 50 m CAT 6a S/FTP cable our company sells and which still works fine after a couple of days of testing. I did not expect to see such a big performace difference in similar specified cables.

    40 m, doesn't work

    50 m, works fine

    Best regards,

    Florian