Transmitting linear 4-20mA d.c. signal over fibre optic cable

I posted this question on the sci.electronics.design forum and it was suggested that I come here.
I have a signal conditioner which outputs a 4-20mA d.c. signal
linearly proportional to a measured input.
About 4 miles away is panel meter which displays the 4-20mA d.c. signal. The amplifier is connected to the meter by a screened twisted copper pair.
I have now been requested to decommission the copper wire and replace it with a fibre optic cable.
The following product has been recommended to me.
http://www.addvid.co.za/html/ll120.htm
It's exactly what I need except that it can only transmit up to two miles. Can anyone suggest an alternative, please?
(There is power available at both the transmitter end and the panel meter end.)
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Start by asking the supplier if the 3km length limit in their data sheet is for their multimode optics unit or if it applies to their singlemode offering too.
The reason I suggest this is that they quote an optical link budget of 12dB. That might be consumed by the product of bandwidth and loss on multimode due to speckle and plain multipath attentuation/pulse spread. If the transmission path is attentuation or spread limited then 12dB on singlemode is going to be a whole lot further, provided the launch is good into the fibre. The downside might be that in the digital coding they have assumed a maximum turn around time that length limits at 3km - I would guess unlikely - they shouldn't be cloccking the link that fast for the quoted performance and this is a single ended application so there are no collisions to worry about unlike a true network.
If your link is 3 or 4 miles through duct and the like, you want to be in single-mode anyway. Yes the investment in splicing equipment for the fibre is greater, but you are then firmly in the telecoms pool of expertise rather than the jumped up "long-LAN" pool of expertise. Frost and water damage to multi-mode is a greater problem than with singlemode in my bitter experience. Single mode splicing professionals tend to do the job once and right.
As an aside, how often does your panel meter need to update? Have you considered low power radio or is that out of the question for some reason? That would be the utility/scada/rtu way to go...
Top hit on a Google for "4 20mA low power radio":
http://www.omniinstruments.co.uk/radiotel/tx7120rx7120.pdf
That would seem to do the job, if the pointer on the metercan tolerate moving once per second - you also have a back channel for control purposes and some digitals for heartbeat and the like.
Hope this helps.
Dave Slee Riding Mill Energy Contracts Ltd.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@boltblue.com says...

I looked at the datasheet for that device, and it isn't clear what the resolution of the converted signal is. The specs list "960 samples/sec", which implies an A/D converter, but the only accuracy spec is "100 uA deviation"...
A bit of googling returned Weed Instruments, who offer analog to fiber multiplexers capable of 40 km: http://weedinstrument.com/product_groups/fiber/multi.html
--Gene
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.