jneutron said:
Going back to the AlphaCore web site, I noticed that they indicate their cables have 2~4 ohms characteristic impedance and around 2.2 mOhm/ft, and others mostly have over 50ohm characteeristic impedance. What would your thinking be considering that these differ from the 8 ohm impedance we normally try to match in speaker crossover design?
For those who use CAT 5 cable, it would be interesting to know what characteristic impedance values are met and what the series resistance is.
I checked CAT 5 is 100Ohm characteristic impedance?
Konnichiwa,
If you parallel multiple Cat 5 Pairs, connected in what I term "interleaved" connection (same BTW for computer ribbon cables!) each doubling of pairs halves the CI. The DCR can be easily calculated from the fact that the wire is 24 Gauge (84mOhm/m) and again scales with the number of pairs.
My "standard" recommendation has 2 Cat 5 Cables and 8 Pairs, so 12.5 Ohm CI and 21mOhm/m. The slightly upgraded "FFRC" has 3 Cat 5 Cables and thus 12 Piars and hence 8.2 Ohm CI and 14mOhm/m.
The TNT "Tripple T" which has 6 Cat 5 Cables braided (and is a bit extreme IMHO) has 24 Pairs and thus 4.1 Ohm CI and 7mOhm/m.
I hope this helps.
BTW, my computer ribbon cable based (3MP) flat cable offers 2.94 Ohm CI and 20mOhm/m.
Sayonara
soongsc said:For those who use CAT 5 cable, it would be interesting to know what characteristic impedance values are met and what the series resistance is.
I checked CAT 5 is 100Ohm characteristic impedance?
If you parallel multiple Cat 5 Pairs, connected in what I term "interleaved" connection (same BTW for computer ribbon cables!) each doubling of pairs halves the CI. The DCR can be easily calculated from the fact that the wire is 24 Gauge (84mOhm/m) and again scales with the number of pairs.
My "standard" recommendation has 2 Cat 5 Cables and 8 Pairs, so 12.5 Ohm CI and 21mOhm/m. The slightly upgraded "FFRC" has 3 Cat 5 Cables and thus 12 Piars and hence 8.2 Ohm CI and 14mOhm/m.
The TNT "Tripple T" which has 6 Cat 5 Cables braided (and is a bit extreme IMHO) has 24 Pairs and thus 4.1 Ohm CI and 7mOhm/m.
I hope this helps.
BTW, my computer ribbon cable based (3MP) flat cable offers 2.94 Ohm CI and 20mOhm/m.
Sayonara
To John:
John,
The eye opener I picked up from this thread is the importance of the charateristic impedance for energy storage. Knowing that characteristic impedance *for reflection phenomena* doesn't come intp play unless tha cable is very, very long, I had dismissed the issue. But your graph in post # 27 is indeed clear.
So, this is the engineering issue. What is not clear to me is the practical effect on the energy transfer. Do you have an indication of the amount of energy loss, so to say, if the charateristic impedance is, say, off by an order of magnitude versus when it is spot on? What are we talking about?
Jan Didden
Edit: just looked again at that graph, unit of energy storage is microjoules. I guess, since the total assymptotically follows the inductive storage with rising char. imp., we can exptrapolate along that line. Now, I only need to work out what a microjoule is in comparison to, say 1 watt in 8 ohms. I guess that a realistic comparison?
John,
The eye opener I picked up from this thread is the importance of the charateristic impedance for energy storage. Knowing that characteristic impedance *for reflection phenomena* doesn't come intp play unless tha cable is very, very long, I had dismissed the issue. But your graph in post # 27 is indeed clear.
So, this is the engineering issue. What is not clear to me is the practical effect on the energy transfer. Do you have an indication of the amount of energy loss, so to say, if the charateristic impedance is, say, off by an order of magnitude versus when it is spot on? What are we talking about?
Jan Didden
Edit: just looked again at that graph, unit of energy storage is microjoules. I guess, since the total assymptotically follows the inductive storage with rising char. imp., we can exptrapolate along that line. Now, I only need to work out what a microjoule is in comparison to, say 1 watt in 8 ohms. I guess that a realistic comparison?
Konnichiwa,
Yes. This is IMHO the preferred solution.
I also find with such cables of comparably low CI that it is worthwhile terminating both ends with the CI, as in the regions where the CI becomes relevant most amplifiers and speakers have high impedances.
If you have amplifiers having marginal stability into capacitive loads it is also worthwhile to add some inductance (usually has to be seperate magnet wire - around 6 - 10 Turns around a Biro in each leg will do) and a further 0.1u/10R zobel on the amplifier side.
The resulting cable is a predictable lowpass but of course no longer low inductance, which kinda defeats the objective.
Sayonara
soongsc said:So it seems like you take one conductor out of each pair as the + side and the other conductor of each pair as the - side?
Yes. This is IMHO the preferred solution.
I also find with such cables of comparably low CI that it is worthwhile terminating both ends with the CI, as in the regions where the CI becomes relevant most amplifiers and speakers have high impedances.
If you have amplifiers having marginal stability into capacitive loads it is also worthwhile to add some inductance (usually has to be seperate magnet wire - around 6 - 10 Turns around a Biro in each leg will do) and a further 0.1u/10R zobel on the amplifier side.
The resulting cable is a predictable lowpass but of course no longer low inductance, which kinda defeats the objective.
Sayonara
Konnichiwa,
Make sure the cable you use is the "trunk" type, that is with solid core wires, plus make sure the wire is actually copper!
By all means. Over at www.tnt-audio.com you can find a number of DIY Speaker Cable ideas using Cat 5 Cable.
Sayonara
soongsc said:I think I have use for that roll of CAT 5 cable. I knew I would use it somewhere some day.
Make sure the cable you use is the "trunk" type, that is with solid core wires, plus make sure the wire is actually copper!
soongsc said:
By all means. Over at www.tnt-audio.com you can find a number of DIY Speaker Cable ideas using Cat 5 Cable.
Sayonara
Kuei Yang Wang said:
Now! We get to the part where we have the same CI but different material.
I tend to think this is related with durrent redistribution. Different material have different current redistribution capabilities due to conductivity. However, in order for current to redistribute, the surface area or cross section needs to be sufficient. So this brings us to another question. Taken the CI are equivalent, comparing solid core ribbon structure versus CAT 5 type structure. Is there a technical advantage in either structure type?
1. Same structure type, same CI, different material.
2. Same material, same CI, different structure type.
Kuei Yang Wang said:
Foil ribbons are available from 3M, I beleive also there are some other articles that mentioned foil type interconnects. John had also shown coaxial braids. It seems we have a little more control with CAT 5 and foil ribbon.
The double braided coax shown by John also looks interesting if there is a way to have some control over the distance between the coax braids. (wonder why he did'nt use copper braids?)
Konnichiwa,
But they are not insulated.
I have used dual crossconnected coaxial cables and the principle again scales quite well, using smaller diameter high quality coax.
An assembly of 2 pcs of RG-214/MIL coax will have a very large crossectional area (>> 12 Gauge) and around 22 Ohm CI, which is not that bad. Adding two more RG-214/MIL and we have around 11 Ohm CI.
As RG214 is truely massive, RG-223 may be a better choice for use in multiples, using 6 length RG223 is a "starquad flower" will give 7.33 Ohm CI and prodigous crossectional area.
There are many ways to skin that particular cat, all with practical Pro's and con's.
Something that seems clear (to me) is that we want a symmetrical cable structure ideally with a minimum of surface are per structural part, from that view multiple Cat 5 or computer Ribbon Cable (IDE133/SCSI3) seems well placed, better than foil ribbons.
Sayonara
soongsc said:
But they are not insulated.
soongsc said:
I have used dual crossconnected coaxial cables and the principle again scales quite well, using smaller diameter high quality coax.
An assembly of 2 pcs of RG-214/MIL coax will have a very large crossectional area (>> 12 Gauge) and around 22 Ohm CI, which is not that bad. Adding two more RG-214/MIL and we have around 11 Ohm CI.
As RG214 is truely massive, RG-223 may be a better choice for use in multiples, using 6 length RG223 is a "starquad flower" will give 7.33 Ohm CI and prodigous crossectional area.
There are many ways to skin that particular cat, all with practical Pro's and con's.
Something that seems clear (to me) is that we want a symmetrical cable structure ideally with a minimum of surface are per structural part, from that view multiple Cat 5 or computer Ribbon Cable (IDE133/SCSI3) seems well placed, better than foil ribbons.
Sayonara
janneman said:
Sorry to quote my own post...
I found out that one Joule = one wattsecond. So, selecting a cable to minimize the CI (see post 27 figure) saves us say, 10 microjoules energy storage. That's 10^-5 wattseconds. Referred to an energy delivered by the amp to the speaker at 1 watt level for a second (1 wattsecond) a loss of 0.001% of energy by using the wrong CI. (We had already agreed that the other CI issue, reflections, was a non-issue).
Now, suppose the cable has a DC resistance of 0.08 ohms. That represents, for an 8 ohms load, a loss of 1%; a factor of 1000 higher than the mismatched CI. Why do we still need to take care of that CI? it seems a total non-issue if there ever was one. Can anybody enlighten me? I mean, this IS an engineering discussion.
Jan Didden
soongsc said:There are lots of things that relate with localization:
IID and ITD as you have mentioned, but also the difference between a transient and reflections. Man is a continously learning machine that coorelates what is heard and what is seen, learned, then used to localize based on what one has learned. These studies would best be used to determine placing of multiple mics in the recording process.
If someone could come up with a software that could scan in a room interior, allow the user to point to where the performers will be, and then allow computerized calculation of optimum mic locationing, then this might be a big revolution in the recording industry.
This subject would belong in a different thread though.
That is certainly a subject for another thread..the complexity of secondary reflections is beyond the scope of my model, as is the modelling of human localization transfer function..
Cheers, John
janneman said:
You beat me to it Jan, I was getting to that part too. It does seem that for small detail signals, resistance is the more critical factor. It seems that we can only determine what differences can be detected only if we are able to make a bunch of different cables and test them in a systematic manner. I refrain from saying which is best, becasue we must first determine the difference is detectable first. Then we can find out what causes differences.
soongsc said:
Going back to the AlphaCore web site, I noticed that they indicate their cables have 2~4 ohms characteristic impedance and around 2.2 mOhm/ft, and others mostly have over 50ohm characteeristic impedance. What would your thinking be considering that these differ from the 8 ohm impedance we normally try to match in speaker crossover design?
For those who use CAT 5 cable, it would be interesting to know what characteristic impedance values are met and what the series resistance is.
I checked CAT 5 is 100Ohm characteristic impedance?
Going below load Z also increases the storage, but with capacitance...from what I recall, many amps have issues with this regime of operation more so than inductive storage..
Line Z for cat 5 is per pair..and the newer versions of cat 5 has all 4 pairs with different pitches to drop pair coupling..paralleling four within a cable will bring Z to about 25 ohms. Paralleling more than that will suffer a bit in terms of inductive coupling, but because they are small wires, the dipole fields drop off in a reasonable fashion, so a triad of cat 5's would be closest to load Z..
Cheers, John
janneman said:
Yes...welcome, I had feared you bolted.
I had done the calcs a while ago, so bear with me, I'm kinda winging it..(the numbers are arbitrary for convenience)..
A 10 Khz sine wave at 100 watts rms into an 8 ohm load..28 volts rms, 40 volts peak...Ipeak is 5 amperes.
100 watts is 100 joules per second. 1 lobe of the wave is 1/20,000 of that number, or 100/20K, =.005, or 5 millijoules per lobe.
A ten foot run of zip sports .2 uH per foot, or 2uH total.
Energy within the inductance of the wire, peak:
1/2 L I2
.5 * 2 e-6 * 25 = 25 * 10-6 joules, 25 microjoules..
Fraction of power being delivered to the load:
25 * 10-6 /5 * 10-3
=5 * 10-3....5 thousanths of the signal..
Excruciatingly small, one would think
In a monophonic setting, not of any interest..I don't think anyone can hear that..
For localization....if one channel varies .5% from the other, what happens to the image??
No clue...I do not yet have all the math done for modelling the ideal receiver localization parameters, nevermind humans..
Cheers, John
soongsc said:
I purchased 1 kilofoot of microphone cable out of parts express, as I had need for about 250 feet in a current application..and the price break made it very attractive to go with 1 kilofeet..
I just stripped the braid off a portion to make the double braid..it is copper braid, just tinned..
John
jneutron said:
About the capacitance storage, you are right that soem amps have problems with them, so it was necessary to jump an R across the terminals I beleive. I have not ran into that problem though.
With the CAT 5's is it the CAT 5E STD you are talking about that have different pitches? How about CAT 6?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.