Hi
We are looking for full-duplex RS-485 driver and receivers meeting following requirement.
Bit rate : 2.5Mbps
Length : 300m
Connection : 1 : 1
We are considering SN65HVD3086E and SN65HVD54.
I think SN65HVD54 seems to be tough for the requirement due to its propagation delay, but SN65HVD3086E would be good.
What do you think which one is proper? Could you please tell us your view?
BestRegards
Propagation delay typically isn't a factor you need to consider when operating over longer links unless your application requires a certain latency. Even then, the delay through hundreds of meters of cable will generally be much greater than the delay through a transceiver IC. The dominant contributor to performance over cable length tends to be the cable itself. At low frequencies and long distances you care about the DC resistance, since this will result in signal attenuation (especially when coupled with low-impedance end-of-line termination). At higher frequencies you also need to consider the frequency-dependent loss of the cabling; generally cables have a "low-pass" characteristic that means their loss increases with frequency. This uneven loss distorts the signal waveforms and introduces jitter.
Here is a general guideline for the distance-versus-data-rate characteristic of a standard RS-485 link over common cabling (this is shown in the TIA/EIA-485 standard):
You can see that your use case would result in relatively small (<5%) amounts of jitter. We have also tested transceivers very similar to SN65HVD3086E across standard cables like CAT 5E and Belden 3105A and seen good performance across 300 m / 1000 ft.
By the way, the other device you mentioned (HVD54) could also support this distance. It didn't sound like you needed the high output differential voltage that device features, though, and so I thought HVD308x would be more cost-effective.
Regards,
Max
Hello,
Here is an example of the differential input and single-ended output eye diagram of one of our newer RS-485 devices when operated over a 1000-ft. CAT-5E cable:
Differential signal at end of cable, 2 Mbps
Corresponding receiver output, 2 Mbps
Differential signal at end of cable, 3 Mbps
Corresponding receiver output, 3 Mbps
Differential signal at end of cable, 4 Mbps
Corresponding receiver output, 4 Mbps
Beyond ~4 Mbps I would think the amount of jitter would be too high to be able to achieve a link with low bit error rate. At that point, using equalization (as discussed in the other thread) would be a way to improve signal integrity.
Regards,
Max
