jabney wrote on Wed, 10 August 2005 06:40 |
My first post here.
I have a SUMO Theorem DAC and a Drawmer M-Clock. The M-Clock has BNC clock outs and the SUMO Theorem has an RCA clock in. I had a BNC to BNC cable made with 75 ohm plenum-grade cable and co-axially crimped-on BNCs. I cut one end and soldered an RCA male to it. I tried my best to get the cable's copper shield to cover the RCA's ground in a 360 degree fashion, but even with a Hako iron and Kester 44, it ain't very coax
Everything's working (I think) but before I invest in a store-bought BNC to RCA cable (assuming such a thing exists) is it worth it?
1. Should my almost-coax RCA be OK, or is true coaxial coverage of both ends of a clock cable that critical?
2. Is the SUMO Theorem - which is supposed to have Burr-Brown D/A, a discrete output, but relatively high jitter using its own clock - worth bothering with?
I'm not a clock fanatic, but I have several digital devices, and an external clock seems the only way to get them all to live in peace. My other DACs include RME ADI-8 Pros & AEs, CreamWare Scope, Alesis HD24 (and a couple of brands I won't mention).
|
What is a sumo theorm DAC?
Here are some general comments about cable imperfection:
I assume you paid attention to having both the termination impedance and cable impedance be the same. In your case it is probably 75 Ohms.
Having said that, the issue of imperfection in the soldered region will effect the transmission path (cable and connectors) impedance in that location. So your question boils down to “how much damage to the signal” will take place when a certain length of the path is at the wrong impedance.
Impedance mismatch does cause some signal reflections. But a short length of mismatch is not as severe as a long length. Also, the slower the signal rise time, the less reflections.
So here is the very condensed answer:
1.If you are using Word Clock, do not worry about the reflections at all, because the repetition frequency is so slow that the reflections due to one clock edge will die out before the next cycle.
2. If you are using AES signals (or higher frequency signals in general), the cycle time is much faster and reflections will have more impact, but a “1 inch long violation” is not going to harm you at all. A 1 inch is about 150 ps (pico second) delay, and I bet your signal rise time is a few nano seconds, therefore you are relatively safe.
The makers of BNC connectors and hardware designed their hardware carefully to accommodate GH signal transmission. A GH (giga Hz) signal rise times are in the few hundred pico seconds or faster. You do not need to worry about an 1/2 inch of bare wire for digital audio transmission, or for audio word clock.
(By "bare wire" I do not mean a palce where you lose continuity, What I mean is "exposed wire". You do want the shield (your return path) to conduct...)
Regards
Dan Lavry
www.lavryengineering.com