Interesting, the Alternative B of IEEE802.3af standard suggesting to use method B in my drawing. Anyone know any reason why?Hello,
In Power Over Ethernet there are two versions.
One is using a set for power and one for combined power and signal using transformers for adding the power:
http://www.bb-elec.com/Learning-Center/All-White-Papers/Ethernet/Power-over-Ethernet-PoE.aspx
https://www.maximintegrated.com/en/app-notes/index.mvp/id/4091
https://en.wikipedia.org/wiki/Power_over_Ethernet
Bertus
That's a good question, I was paying too much attention until now. I don't know yet. I am thinking 24V or 48V. From what I can understand, there is no point using higher voltage if I am using leaner regulator at my end devices. As it won't reduce my current draw on the power lines.What supply voltage will be available to anode at the extreme? You must computer the IR drop of 8000 feet of cable to get a handle on this number. Depending on the current you require it might be a great deal less than you are expecting.
I am planing to use RS-485 on a cat6 at 200kbps. I don't need a lot of speed, even 100kbps is a lot for my application.Cat6 is only guarantee'd 1000mb up to 100 meters. Probably lose signal integrity at 1200 meters..
What chipset will you be using? For example, the MAX490CPA is capable of up to 2.5Mbps, but the MAX3082CPA is capable of only 115Kbps, and I wouldn't be using either at more than 50% their stated speed at the length you've just described.I am planing to use RS-485 on a cat6 at 200kbps. I don't need a lot of speed, even 100kbps is a lot for my application.
I am just researching at the moment, haven't look at any chip set yet. We only need about 2.4kbps max (in theory), so half of 115Kbps is still a lot. I am happy with anything that is more than 2.4Kbps.What chipset will you be using? For example, the MAX490CPA is capable of up to 2.5Mbps, but the MAX3082CPA is capable of only 115Kbps, and I wouldn't be using either at more than 50% their stated speed at the length you've just described.
Max knows a lot more than me about that question. But I'd probably use one of the fastest chips available, and then use it at the very slow speed you've just described. That ought to be reliable enough for your application. As for protocol, I'd include a CRC byte (preferably of the type X8 + X5 + X4 + 1) after each set of characters have been sent, plus timeout detection and error recovery routines.What chip would you use if you were me? Reliable is very important.
plus timeout detection and error recovery routines.
What are you going to use the line for? For control purposes, or only for monitoring?Thanks
- Timeout detection, you mean after a packet is sent, expect an ACK within some time, resend if ACK is not received?
- I am planing to follow the flow char here if I can't find a suitable protocal: https://en.wikipedia.org/wiki/Carrier_sense_multiple_access_with_collision_detection
- I am not fimilar error recovery routines, what would you suggest on this?
Cat6 isn't designed to be used for such long runs. Twists will loosen and cause signal distortion..I am planing to use RS-485 on a cat6 at 200kbps. I don't need a lot of speed, even 100kbps is a lot for my application.
I think it's a bit of both.What are you going to use the line for? For control purposes, or only for monitoring?
All my devices will be daisy chain together, will that work? The bus is not a single lone wire of 1000m. But the total length of the bus will be 1000m max.Cat6 isn't designed to be used for such long runs. Twists will loosen and cause signal distortion..
daisy chain should work, just make sure you use the right topology for RS485 when wiring, with proper termination resistors and all. Now I see that developing your own protocol will be a challenge. What I suggest, is have a single master continuously monitor the status of each device sequentially, and then have it transmit a single line of instructions that would be read and interpreted by the devices.All my devices will be daisy chain together, will that work? The bus is not a single lone wire of 1000m. But the total length of the bus will be 1000m max.
Cat5 Cable designed for long runs is solid conductor and requires plugs or jacks designed for same. It can be used reliably up to 100 meters.All my devices will be daisy chain together, will that work? The bus is not a single lone wire of 1000m. But the total length of the bus will be 1000m max.
Using a master to poll each device sequentially is not a good idea for me. As my slave will be added and remove. I believe I still need a CSMA/CD protocol. I will properly don't want to develop my own protocol, but instead I will adapt to something we already have. I am sure someone have done this long before me, I just need to find it.daisy chain should work, just make sure you use the right topology for RS485 when wiring, with proper termination resistors and all. Now I see that developing your own protocol will be a challenge. What I suggest, is have a single master continuously monitor the status of each device sequentially, and then have it transmit a single line of instructions that would be read and interpreted by the devices.
Considering that, CRC is a must for your application.
RS-485 at lower data rates can go much farther.Cat5 Cable designed for long runs is solid conductor and requires plugs or jacks designed for same. It can be used reliably up to 100 meters.
by Jake Hertz
by Aaron Carman
by Duane Benson