Put your money where your mouth is.....
OK, enough of this crap. I knew someone would fall Canare's line of BS. (What is that Swedish word again......?)
Here's what anyone who believes that can do:
Send me $500. Give me a date, time, and location to be at. (Only applies to USA for obvious reasons.) And a phone number in case the directions are too confusing.
I'll show up with my TDR, some 50, ohm BNCs, 75 ohm BNCs, a handful of coax cables and some adaptors. You bring your "75 ohm RCAs". If you are lucky, I will bring one of my old D/A boxes.
After I prove you wrong, and if there is any gas money left, I'll take you out for pizza.
Any takers............?
NO, didn't think so......
Jocko
OK, enough of this crap. I knew someone would fall Canare's line of BS. (What is that Swedish word again......?)
Here's what anyone who believes that can do:
Send me $500. Give me a date, time, and location to be at. (Only applies to USA for obvious reasons.) And a phone number in case the directions are too confusing.
I'll show up with my TDR, some 50, ohm BNCs, 75 ohm BNCs, a handful of coax cables and some adaptors. You bring your "75 ohm RCAs". If you are lucky, I will bring one of my old D/A boxes.
After I prove you wrong, and if there is any gas money left, I'll take you out for pizza.
Any takers............?
NO, didn't think so......
Jocko
sikkek:
You are excused. But the laws of physics do not lie, unlike most marketing weenies.
If they work fine for you, and if fine is an acceptable level of performance, then that is fine. Problem is that RCAs have too much capacitance to be anywhere near 75 ohms, no matter who says otherwise. And if you are using them on a 75 ohms line, you will have significant reflections.
In any transmission line- optical, coax, twisted pair, reflections are produced at any and all interface/transistion points. Reflections amount to a form of distortion, that in the case of SPDIF, that will affect the decision point in the RX chip, and create jitter.
So, how much jitter is too much? In the case of SPDIF, almost any is. The problem arises because the clock has to be recovered (with a PLL!.....yuk, more jitter......) from a serial data stream. The pulse distortion that is caused by reflections.....bouncing back and forth......exacerbates this problem.
(The effect of reflections can be heard.......and your $500 contribution will also include a demonstration of this.)
Anyway, back to what started this thread.......it is very likely that going to a LONGER fiber optic cable can actually be BETTER. There are several reasons. In the case of fiber, there are some unique problems that arise in the transmit end that don't exist in the copper world. But the solution for it, as well as for all the other reflection problems, can be fixed using the same techniques.
And that brings us back to the jet fighter analogy........
Jocko
You are excused. But the laws of physics do not lie, unlike most marketing weenies.
If they work fine for you, and if fine is an acceptable level of performance, then that is fine. Problem is that RCAs have too much capacitance to be anywhere near 75 ohms, no matter who says otherwise. And if you are using them on a 75 ohms line, you will have significant reflections.
In any transmission line- optical, coax, twisted pair, reflections are produced at any and all interface/transistion points. Reflections amount to a form of distortion, that in the case of SPDIF, that will affect the decision point in the RX chip, and create jitter.
So, how much jitter is too much? In the case of SPDIF, almost any is. The problem arises because the clock has to be recovered (with a PLL!.....yuk, more jitter......) from a serial data stream. The pulse distortion that is caused by reflections.....bouncing back and forth......exacerbates this problem.
(The effect of reflections can be heard.......and your $500 contribution will also include a demonstration of this.)
Anyway, back to what started this thread.......it is very likely that going to a LONGER fiber optic cable can actually be BETTER. There are several reasons. In the case of fiber, there are some unique problems that arise in the transmit end that don't exist in the copper world. But the solution for it, as well as for all the other reflection problems, can be fixed using the same techniques.
And that brings us back to the jet fighter analogy........
Jocko
SPDIF
"Most SPDIF signals are generated with some kind of transmitter. I'm sure we all can agree about that. This generator is probely like CS8404/04/05 which has jitter in the ns area."
I have NEVER seen a Crystal CS8404/04/05 in a CD player or transport. Too expensive. The jitter from that device is much lower
than the ":nS area", as are most SPDIF drivers. Pehaps you could stick to posting coments about something with which you have some (even a little) knowledge.
Thank you,
H.H.
P.S. I used to design SPDIF AES/EBU interface devices and digital cables.
"Most SPDIF signals are generated with some kind of transmitter. I'm sure we all can agree about that. This generator is probely like CS8404/04/05 which has jitter in the ns area."
I have NEVER seen a Crystal CS8404/04/05 in a CD player or transport. Too expensive. The jitter from that device is much lower
than the ":nS area", as are most SPDIF drivers. Pehaps you could stick to posting coments about something with which you have some (even a little) knowledge.
Thank you,
H.H.
P.S. I used to design SPDIF AES/EBU interface devices and digital cables.
Jocko:
Do you mean the following?
A longer transmission line means a lower output at the end, due to losses in the cable, thus less reflected waves, these have to travel back to the transmitter (again more loss) and be 'bounced' back again to the receiver (also with more loss). With longer lines the signal/reflection ratio will be better...
Do you mean the following?
A longer transmission line means a lower output at the end, due to losses in the cable, thus less reflected waves, these have to travel back to the transmitter (again more loss) and be 'bounced' back again to the receiver (also with more loss). With longer lines the signal/reflection ratio will be better...
Interface
Hi Jocko,
My double shielded coax cable is one meter long i.e. 3 feet in USA measures.
But the nice thing about my interface is that it sounds the same with 13 meters of cheap coax. Thanks to you and Harry H.
http://www.diyaudio.com/forums/show...ge=15&highlight=elso AND digital&pagenumber=7
Replace the 74HCU04 by (MC)74VHC04
and you won't believe your ears.
Hi Jocko,
My double shielded coax cable is one meter long i.e. 3 feet in USA measures.
But the nice thing about my interface is that it sounds the same with 13 meters of cheap coax. Thanks to you and Harry H.
http://www.diyaudio.com/forums/show...ge=15&highlight=elso AND digital&pagenumber=7
Replace the 74HCU04 by (MC)74VHC04
and you won't believe your ears.Actually, it has been my experience (real, not the stuff you read in magazines), that the longer your cable for co-ax, the worst the issue of transmission line reflections. It is an issue of edge speed and transmission distance. Remember that the level of the reflection is a function of the impedance mismatch.
I agree whole heartedly with Jocko, I have never seen a real 75 ohm RCA. But the good ones that claim to be are actually much closer. It is not so much an issue of capacitance, but a matter of the physical dimensions of RCA where never intended for a 75 ohm connector.
That all being said, optical or co-ax, I find the biggest difference is not in the cables, but in the equipment itself. A cable that may be great with one piece is terrible with another and vice versa. It almost makes the whole cable industry suspect, and I can't really blame them. I think you can make a bad cable. But I think making a really good cable is equipment dependant because you have no issue which design flaw you are trying to fix.
Personally, I have this terrible habit of getting out my soldering iron and putting high speed low overshoot drivers with precision 75 ohm terminations and good low jitter receivers with matching impedance on the other end. When I do that, I find that any "good" cable works perfectly.
Now if someone could just make me a high impedance analog audio cable that I could use to match a pre-amp.......
Alvaius
I agree whole heartedly with Jocko, I have never seen a real 75 ohm RCA. But the good ones that claim to be are actually much closer. It is not so much an issue of capacitance, but a matter of the physical dimensions of RCA where never intended for a 75 ohm connector.
That all being said, optical or co-ax, I find the biggest difference is not in the cables, but in the equipment itself. A cable that may be great with one piece is terrible with another and vice versa. It almost makes the whole cable industry suspect, and I can't really blame them. I think you can make a bad cable. But I think making a really good cable is equipment dependant because you have no issue which design flaw you are trying to fix.
Personally, I have this terrible habit of getting out my soldering iron and putting high speed low overshoot drivers with precision 75 ohm terminations and good low jitter receivers with matching impedance on the other end. When I do that, I find that any "good" cable works perfectly.
Now if someone could just make me a high impedance analog audio cable that I could use to match a pre-amp.......
Alvaius
Coax length
"that the longer your cable for co-ax, the worst the issue of transmission line reflections."
NO! longer cables sound better. The reflection is delayed so not to coincide with the leading edge on the data transistion. Pick a cable length with a propagation delay of about about a quarter of the logic rise time and you will find out how bad a short cable can sound. Any one ever listen to a half meter digital interconnect?yuck! Read the book, I have two copies and went to his seminar.
www.sigcon.com
H.H.
"that the longer your cable for co-ax, the worst the issue of transmission line reflections."
NO! longer cables sound better. The reflection is delayed so not to coincide with the leading edge on the data transistion. Pick a cable length with a propagation delay of about about a quarter of the logic rise time and you will find out how bad a short cable can sound. Any one ever listen to a half meter digital interconnect?yuck! Read the book, I have two copies and went to his seminar.
www.sigcon.com
H.H.
Attachments
Re: Only for Europeans...?
Jocko,
there are 3 1/2 countries left in the world still using the inch system ... the half one is obvious, it is GB, (they are fervently trying every other day and succeeding better and better) then there is Birma, and Burundi still using it, but, (help me, folks) cannot remember which the 3rd country was ..... which was it, can you help me?
Harry,
As far as i can judge (not very far), what you say makes sense. But you could tell just that and leave the invitation to feel bad in your junk box. How about that?
Jocko,
Jocko Homo said:
there are 3 1/2 countries left in the world still using the inch system ... the half one is obvious, it is GB, (they are fervently trying every other day and succeeding better and better) then there is Birma, and Burundi still using it, but, (help me, folks) cannot remember which the 3rd country was ..... which was it, can you help me?
Harry,
bitchy as usual?
As far as i can judge (not very far), what you say makes sense. But you could tell just that and leave the invitation to feel bad in your junk box. How about that?
How about that?
How about I just sit back and let people, passing themselves off as technically inclined, post all sorts of misinformation. I spent several years working on digital signal transmission and the SPDIF interface in particular. But I guess anyone else's opinion is as valid as mine since we are all just here to get along and useful content is not important. All I did was disagree strongly and give a reference to those who might like to learn something. I don't remember calling anyone any names or insulting anyone...... Remember hate mail goes to fdieck@hotmail.com
H.H.
How about I just sit back and let people, passing themselves off as technically inclined, post all sorts of misinformation. I spent several years working on digital signal transmission and the SPDIF interface in particular. But I guess anyone else's opinion is as valid as mine since we are all just here to get along and useful content is not important. All I did was disagree strongly and give a reference to those who might like to learn something. I don't remember calling anyone any names or insulting anyone...... Remember hate mail goes to fdieck@hotmail.com
H.H.
I will trade your years of SPDIF for years of SPDIF, SMPTE-259 and 292 as well as gigabit interconnect.....
I have the book as well. It is good, and I recommend most people to read it to pick up the basics.
I have found that the rise times of many SPDIF outputs to be quite slow, on the order of 10's of nsec.
For those who don't know what we are talking about. Where you have a mismatched termination (generally from the receive end), you will get a reflection approximately 10nsec/meter of cable later. The size of the reflection is proportional to the mismatch in the impedances. What you don't want to have happen is to have that reflection happen at approximately the logic transition level.
To Harry's argument, if you have rise times of say 20 nsec, then a cable on the order of 1 meter would generally be poor, i.e. the reflection would occur right in the middle of the rise at 10nsec. So, I will take back my previous statement. For much of the audio world, 1 meter may not be a good thing.
I just wish that the consumer electronics people could stick to this. Harry, help me on this, I think the IEC958 spec if 10nsec rise times? That would generally make 1/2 meter cables really bad.
So yes, really short cables are not good. However, conversely, really long ones can be bad too. When the edge speeds start to drop, you can get jitter with some receivers.
Alvaius
I have the book as well. It is good, and I recommend most people to read it to pick up the basics.
I have found that the rise times of many SPDIF outputs to be quite slow, on the order of 10's of nsec.
For those who don't know what we are talking about. Where you have a mismatched termination (generally from the receive end), you will get a reflection approximately 10nsec/meter of cable later. The size of the reflection is proportional to the mismatch in the impedances. What you don't want to have happen is to have that reflection happen at approximately the logic transition level.
To Harry's argument, if you have rise times of say 20 nsec, then a cable on the order of 1 meter would generally be poor, i.e. the reflection would occur right in the middle of the rise at 10nsec. So, I will take back my previous statement. For much of the audio world, 1 meter may not be a good thing.
I just wish that the consumer electronics people could stick to this. Harry, help me on this, I think the IEC958 spec if 10nsec rise times? That would generally make 1/2 meter cables really bad.
So yes, really short cables are not good. However, conversely, really long ones can be bad too. When the edge speeds start to drop, you can get jitter with some receivers.
Alvaius
SPDIF
I wasn't taking seminar for SPDIF. I confess to being a telecom engineer as well as an audio geek. Actually since you are quite up to speed on cable TDR theory, I will through a few more variables in the equation. HCMOS has a rise time of in the 2 to 5ns range and is pretty typical of the SPDIF interface as implemented in much of the equipment out there. Add a pulse transformer and you can see figures of 10nS to 30nS. Optimum cable length is also a function of the propagation velocity of the dielectric used in the cable. Two cables can be the same physical length and have different propagtion times. Electrically lossy cables often work well since they attenuate reflections better. You can have too much bandwith in a cable also. Try ACMOS and a really low loss cable and see how long you can listen to it. One of the most fatiguing cable I ever heard was some Gore laboratory grade coax that was $180 a foot. Detailed yes, but ten minutes and you would run from the room. Very short cables also work well. Try about a 5 or 6 cable some time. Oh ya..... mechanical damping is critical a a very good digital cable design.
H.H.
I wasn't taking seminar for SPDIF. I confess to being a telecom engineer as well as an audio geek. Actually since you are quite up to speed on cable TDR theory, I will through a few more variables in the equation. HCMOS has a rise time of in the 2 to 5ns range and is pretty typical of the SPDIF interface as implemented in much of the equipment out there. Add a pulse transformer and you can see figures of 10nS to 30nS. Optimum cable length is also a function of the propagation velocity of the dielectric used in the cable. Two cables can be the same physical length and have different propagtion times. Electrically lossy cables often work well since they attenuate reflections better. You can have too much bandwith in a cable also. Try ACMOS and a really low loss cable and see how long you can listen to it. One of the most fatiguing cable I ever heard was some Gore laboratory grade coax that was $180 a foot. Detailed yes, but ten minutes and you would run from the room. Very short cables also work well. Try about a 5 or 6 cable some time. Oh ya..... mechanical damping is critical a a very good digital cable design.
H.H.
We can't forget the very poor power decoupling on that logic in host equipment that makes for terrible ringing. That is one thing you need to watch with long cables. I have seen a reflection reinforce a ring causing a logic toggle.. very annoying.
I am thinking of ripping out all my SPDIF and replacing it all with LVDS.
Alvaius
I am thinking of ripping out all my SPDIF and replacing it all with LVDS.
Alvaius
If guys like us designed SPDIF in the first place.............
The only thing worse is the AES/EBU version. You would think that they would have asked guys like us, but no..........
Anyway, I stick to my position of longer cables. The transit delay for reflections more than makes up for any slowing of rise times. Plus it eats up more of those stinking reflections.
Jocko
Missed one....
Right. Since Z = (L/C)^1/2, it is the capacitance that makes RCAs in the 30 ohm or so range.
Look at a 75 ohm and 50 ohm BNC side-by-side. You wil see that the only difference is that the 75 ohm has some of the Teflon center insulator removed. (75 ohm TNCs actually use a smaller center pin as well.) Now, it should be farily obvious that the physical dimensions of an RCA are nowhere near correct.
But if you guys want to continue using them..........
It is about this time, some wiseguy like HairyHoller will tout that they can make a network to compensate for it. Yes, he can. But that is not the same as the connector actually being 75 ohm.
Still waiting for $500 contributions...........[joke]
Jocko
Right. Since Z = (L/C)^1/2, it is the capacitance that makes RCAs in the 30 ohm or so range.
Look at a 75 ohm and 50 ohm BNC side-by-side. You wil see that the only difference is that the 75 ohm has some of the Teflon center insulator removed. (75 ohm TNCs actually use a smaller center pin as well.) Now, it should be farily obvious that the physical dimensions of an RCA are nowhere near correct.
But if you guys want to continue using them..........
It is about this time, some wiseguy like HairyHoller will tout that they can make a network to compensate for it. Yes, he can. But that is not the same as the connector actually being 75 ohm.
Still waiting for $500 contributions...........[joke]
Jocko
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.