Hi, Arthur, I don't think we disagree on much except for a particular detail, so I'll only respond to that.
There are indeed many circuits that are actually "differential" in the world. Professional audio transmission is often done with true "differential" circuits. And even when the transmission of the signal has one side as ground (as in a shielded or coaxial cable), there are times that the receiving side will be truly "differential". In order to work, the circuit at the receiving side must in essence subtract the negative side signal from the positive side signal to give a result with regard to the receiver's ground. When that is done, then "ground bounce" or ground signal differences from transmitter to receiver are effectively subtracted out of the signal. The circuitry to do the subtraction is expensive, especially at video frequencies. But I don't deny that the monitor might have such a circuit--since neither of us have the schematics we can't determine that for certain for a given monitor. However, whether the monitor uses a differential receiver is not particularly relevant to our problem here. The reason is that any "adapter" would be applied at the transmitting end (the computer), not at the receiving end. As long as the transmitter side (such as our "adapter") does not add noise to the signal return grounds, even if the receiver has a differential circuit it will work without newly created noise. We know for certain that the return "grounds" are in fact connected to the computer's video section's ground plane, because the typical video card output circuit is a single integrated circuit with a common ground. Then as you comment grounding is very important. The R,G, and B circuits should be run on a ground plane that is continuous with the D/A chip in the card, until it reaches the video connector at the back of the computer (thus the grounds are tied together up to that point). And those traces if done correctly would be of matched "impedance" by having a certain trace width calibrated for the circuit board properties, as it passes over the ground plane, and normally video signal impedance would be 75 ohm. (But I am not sure on monitor video standards about the impedance because the multiple-wire cables may have difficulty with that low an impedance.) In any case it should be constant from circuit board to video cable, and 75 ohm is standard video signal impedance elsewhere. As long as any grounds that are tied together at the video connector of the PC are connected to the common ground plane of the video D/A converter, there will be little new noise introduced. An adapter with a properly done ground plane would simply be extending the common ground plane that occurs about 1/2 inch away in the computer. BUT that is why it is important to only tie together the signal grounds, and not the case ground. Normally there is a designated "case ground" applied to the cable shield, which indeed should be kept separate. Now as I commented, if you do separate the grounds in a WIRED adapter or cable, that is also entirely appropriate. The currents do indeed flow "differentially" whether or not the receiver uses differential circuits, and improperly wired can introduce noise. The main point is that an adapter properly done at the transmitting end is only extending the ground plane another inch or so, not likely to pick up crosstalk unless done poorly. Those adapters have a lot of switches and the like, and may not use multiple ground planes within them and may also do poorly at routing signals and impedance matching. So the best signal quality would be achieved by rewiring the monitor video cable with properly matched impedance cable and correct pinout. (Either with existing monitor cable or new cable properly matched). Fewer connectors will make for fewer reflections, as the D style connectors are not well matched in impedance to video, and double connectors means double reflections. (They try to put the D/A as close as possible to the D connector at the computer end so all impedance mismatches are at the original signal transition edge and not delayed to produce "ghosts".) You're on the right track, and method will work--I only disagree on some of the details of your comments. Improper grounding can most certainly introduce unwanted signal distortion. (PS I have designed a lot of equipment that has to meet better than 100 dB analog signal/noise standards, which are levels that can only be dreamed about by video designers. And also equipment with both differential and single ended signals approaching gigahertz frequencies.) Thanks Gordon Elliott ----- Original Message ----- From: "Artur Yelchishchev" <[EMAIL PROTECTED]> To: "Quadlist" <[EMAIL PROTECTED]> Sent: Tuesday, September 23, 2003 6:11 AM Subject: Re: Apple Multiple Scan 1705 Display > "Gordon Elliott" <[EMAIL PROTECTED]> wrote: > > > > Tying the signal grounds together in an adapter is just fine, and should > > cause no signal degradation. Other problems, like capacitance and impedance > > change may be at fault, but grounds are not the problem. > > Hi Gordon, > > Thanks for your input, but I don't think it's correct statement. My > former occupation was electronic engineering, and I know for sure, how > important the proper grounding is. Additional interconnection of ground > lines at any point except receiver and transmitter, causes unwanted > current fluctuations and signal modulation. > > > They are tied together in the monitor, and in the computer. > > Yes, but not directly. Each line is, so-to-speak, differential one, and > separate ground returns are used for proper compensation of electrical > noises. .... -- Quadlist is sponsored by <http://lowendmac.com/> and... Small Dog Electronics http://www.smalldog.com | Enter To Win A | -- Canon PowerShot Digital Cameras start at $299 | Free iBook! | Support Low End Mac <http://lowendmac.com/lists/support.html> Quadlist info: <http://lowendmac.com/lists/quadlist.shtml> --> AOL users, remove "mailto:"; Send list messages to: <mailto:[EMAIL PROTECTED]> To unsubscribe, email: <mailto:[EMAIL PROTECTED]> For digest mode, email: <mailto:[EMAIL PROTECTED]> Subscription questions: <mailto:[EMAIL PROTECTED]> Archive: <http://www.mail-archive.com/quadlist%40mail.maclaunch.com/> Using a Mac? Free email & more at Applelinks! http://www.applelinks.com
