Hello,
For the electronics-minded who want to experiment - I built some Madi to cat-cable converters a few years ago for a facility that needed them. Happy to share knowledge for someone to learn with.
I started with a passive version using ethernet transformers and some impedance matching resistance networks and while it worked, the range was too short due to a few things - the passive impedance matching networks caused additional loss, and most coax madi inputs don't use real cable EQ circuits to make up for lots of loss. Many use a variant of the Micrel SY100xxx differential receivers which work fine but can't compensate for the kind of cable loss that a cable EQ circuit can.
The successful device was an active version which worked well.
If you're going to experiment a few items to consider and optimize -
- Coax has a 75ohm characteristic impedance, cat-cable has a 100ohm characteristic impedance (mostly).
- cat-cable is balanced and functions best with ethernet PHY transformers. Not the PHY chip, just the transformers, which also gives galvanic isolation, important.
- transformers reflect impedance, and Madi is effectively RF, so you need to treat the cat-cable as transmission lines.
- there is a LOT of loss in a cat-cable so your receiver has to make up for that to recover a usable waveform.
- You must have an oscilloscope to see waveforms, the possibilities for making mediocre or bad waveforms are endless.
Simplified circuit flow -
Coax to cat-cable = coax input > LMH0344 cable eq > 2x OP858 opamps as a differential driver > 100ohm differential transmission line impedance > ethernet transformer > RJ-45.
Cat-cable to coax = RJ-45 > ethernet transformer > 100ohm differential termination impedance to 75ohm output impedance > LMH0344 cable EQ in differential mode > LMH0002 coax cable driver.
A few reasons for things -
A coax cable attenuates a signal with distance and the attenuation is frequency variable - high frequencies attenuate more. Cable EQs are variable gain+EQ input devices that expect a launch amplitude of 800mV (SDI video). Any level less than that, which is any signal that travels any distance, will cause the cable EQ chip to increase the gain and the high frequency EQ by the amount necessary to compensate for the loss in level and frequency response. Lower amplitude input will increase the high frequency gain geometrically more, approximating the inverse loss curve of a coax cable (mostly). All this is to say a cable EQ circuit can retrieve signals that would be lost with a simple input circuit.
Initially I fed the LMH0344 differential outputs of the transmitter directly to the ethernet transformer. While it worked, there were a few concerns -
A - the input termination network for the cat receiver has some loss and I wanted a bit more send level to compensate, the OP858 differential driver had a few dB of gain.
B - the LMH0344 differential outputs are not really designed to withstand the back-emf from cat-cable abuse over long runs, the OP858 as drivers can handle more of that with added output protection and filtering.
C - for the cat receiver I had better results with a passive impedance matching network feeding the receiver cable eq differentially than using an opamp buffer first. Either way would work.
The final product was 2 send 2 receive each end - there are 4 pairs in a cat-cable.
And an added cool item - I put a PoE PSE circuit in the end that requires power, and a PoE PD circuit in the other end, so it could be powered PoE-style over the cat-cable. Worked great.
So if anyone wants to give it a try have fun, the above gives you a start in the experimentation process. Enjoy.
Hugh
