kenwstr Posted January 28, 2020 Posted January 28, 2020 OK, this is a controversial topic so, just bear with me a little. I do some PA and I do like my home stereo as well. It seems to me that the typical audio engineering view in PA is that it's pretty much just down to impedance but HiFi fans say there is a lot more to it. I have seen lab tests where expensive speaker cables are shown to cause artifacts in the signal that are not present in the recording, but this it seems is claimed to be interpreted as "detail". So anyway, I investigated the impedance thing and in the course of that, found a pro site that had done a small blind AB test on this. They found that most people reported improvements with reducing cable impedance down to 5% of nominal load but not beyond this. A different group of people used to critical listening all agreed there was improvement down to 2% of nominal load but opinions differed below this indicating personal preferences. As I had inherited clarity and tonality issues with the PA, I decided to try upgrading the cable gauge based on a 2% impedance chart. This resulted in a vast improvement and I was really happy with the results. Even got an audiophile's comment or "wow, true sound reinforcement!" That from an enthusiast of 40+ years in the biz who's main speakers $20k and he is not totally happy with them. So, now moving on, I still have my main stereo packed away and am using a small 2nd stereo in a spare room. This is driven by a pioneer A400 powering Techniques book shelf speakers that I got at the local op shop. So, on the basis of meeting 2% impedance, 1 m of 18 ga cable I had lying around should be eminently more than sufficient and that worked pretty fine. However, I decided to buy a pair of ProAc EBTs to replace the Techniques. This gave markedly better clarity but I found at higher volumes it became a bit harsh and after long sessions, left me feeling somewhat jangled and not relaxed. So, then I swapped in 1m of 14 ga cable and now the music is not harsh and is relaxing though it sounds much the same, except, doesn't go harsh when turned up. So I guess, my conclusion here is that it isn't a simply a matter of the linear relationship of impedance to gauge and run length. In terms of impedance, over 1 m, gauge would be pretty much irreverent for most cable gauges. While impedance matching is a major and well established audio principle and possibly the primary consideration, it seems to me something else important is going on as well. I am not sure what, inductance, capacitance, something? It also begs the question, Is the speaker's internal wiring adequate? From the impedance perspective, the answer should be yes, because the run length is so short as to make it's influence on impedance insignificant. However, that same argument doesn't appear to hold true when upgrading just 1m of cable from 18 ga to 14 ga is noticeably more pleasant. It's not a huge difference like the clarity and tonality that is obvious from the impedance perspective, it's a more subtle influence on how the music makes me feel. The next question is, how can this be defined as a relationship? Is there just a minimum gauge limit to be imposed upon the linear impedance relationship? Did the blind testing just concentrate on cognition and not emotional response? I am interested in opinions though I would ask comments be confined to copper cabling only, just to limit the number of factors to length and gauge and whether it's worth upgrading internal wiring in most cases? Please try to keep it sane, I am trying to identify possible relationships to help define a better predictive performance model here, well at least a starting point discussion anyway.
frednork Posted January 28, 2020 Posted January 28, 2020 Well done for having a listen and noticing the difference!! Should be some previous threads on speaker cables that you can peruse. Short answer is some here still say that any old cable as long as it is thicker than a certain gauge is fine and no further effort should be considered and internal wiring is more than adequate and any differences you hear are cognitive bias. Others spend $10k+ on cables as they see it as a crucial element of the system. Of course if you are looking at cables it is highly dependant on the system and the listener. My personal view is they are very important but they should match the rest of the system in terms of quality and synergy. As far as internal wiring and short lengths go, I found that some jumpers I was using (14ga copper 5cm long) made a huge negative difference to my setup when wired in line on the biamping terminals (to replace the even worse brass jumpers there previously) so for me short length will not prevent negative impact so if the internal wiring is not great..... Probably not that likely to find the "equation" that answers all your questions as I dont think things are that simple in reality (but would be great if someone did!!)
gwurb Posted January 29, 2020 Posted January 29, 2020 (edited) All of of cable performance relates to impedance, capacitance and inductance. Those properties vary with conductor composition and overall cable composition. A quick rule of thumb: Impedance will result in a drop in power between the amplifier and the speaker will change the signature of audio output unless the cable has a perfectly flat frequency response over a wide frequency range (nyquist, etc) will affect how linear the output increases with increase in volume Capacitance will affect how much detail can be represented as it effects how quickly the cable can discharge electric charge will affect how punchy the notes are more capacitance will result in bolder bass less capacitance will result in more finesse and detail Inductance will affect how much detail can be represented as it effects how quickly charge can build up in the cable less inductance will result in more detail as the music presentation can occur as quickly as it is intended and there is no lag These are rules of thumb. Ideally the cable will be a pure resistor. If that happens then the job of representing the music falls purely on the amplifier and the speaker. Otherwise the cable will have some sort of an effect. To get a better idea of what the cable would do you would need to measure its properties and perform some testing. There will be a sweet spot that would be preferred by most people. That sweet spot may be different to what any particular individual prefers. What is interesting is that different people like particular sounds. They also like the music represented a certain way. The most interesting thing about that is that when live music is played it is often amplified, and that has its own colouration. If its not amplified then the instruments themselves have particular tuning. So recordings and playback never represent some sort of 'perfect' thing but rather they represent a particular expression. Here is an interesting graph (on page 2): http://www.nfcorp.co.jp/english/doc_portal/download/pdf/appsheet/ImpedanceMeasurement_of_CopperWire.pdf A couple of points to note about it: Wires measured were approximately 8, 11, and 13 AWG The measurements were on a single conductor Lower diameter cable has a more linear impedance over the measured frequency range Lower diameter cable has a smaller phase shift that occurs at a higher frequency There is a note to state that the shape of the wire may have an effect The graph tells us very little about the whole of cable construction. None the less, those measurements do show something interesting. A single copper wire would give a more 'accurate' representation when its of smaller diameter. So the most 'accurate cable' would be of small (but not too small) of a gauge. Then the speakers would need to be driven by high voltage and high current amplifier. I would want to see measurement results for 14 to 18 AWG conductors, and even lower. If the relationship is that higher gauge results in flatter impedance, and less frequency shift then it would make sense to have a small conductor. Yet with a small conductor the overall impedance is higher and more power is lost along the wire and more heat is generated. So then it would make sense to have a lot of small conductors bundled together, which is what a lot of modern audio cables are. This means the amplifier does not need to have as much voltage and current output, which is much cheaper and easier to design and manufacture. The inductance and capacitance part would then be addressed by the phase shift of the signal, and if with really small wire diameters that is moved outside of audible range then perfect. After all of that there are still considerations relating to the overall cable, not just the conductors. To address the original questions: Physically its pretty complicated Purity of material would make a difference (the more pure the better) Mix of materials would also make a difference (e.g. copper and silver mix) Without measurements try getting a cable that has small wire strands but lots of them Edited January 29, 2020 by gwurb added two clarifying sentences 2
kenwstr Posted February 13, 2020 Author Posted February 13, 2020 (edited) Yep, Familiar with Impedance, inductance and capacitance. I mainly concentrated on impedance in my original post simply because it probably considered the major concern in PA due to the issue of impedance matching which should be familiar to anyone wanting to plug a pickup directly into a mixer. It's exactly the same deal with speakers because the load is so very low. It is very rare to get specs on these values though unless you go to certain well regarded manufacturers. I have never seen any of them specified for zip chord for example. OK, the is the crappiest basement end of the market but most likely we are all familiar with it. The example I gave was 18 g vs 14 g zip chord. I had thought that at 1m length this should be insignificant but the 14 g makes a significant difference. While thicker cable is usually chose for lowering impedance, it also changes the other 2 properties as the geometry of the cable changes. According to a twisted pair calculator I just found: As the distance between conductor centres increase: Impedance increases Inductance increases capacitance decreases Delay remains the same As the conductor area increases: Impedance decreases Inductance decreases capacitance increases Delay remains the same As the Dielectric constant increases: Impedance decreases Inductance remains the same capacitance increases Delay Increases So it is difficult to choose a different cable without changing all the properties. I assume these relationships would work in a similar way for zip chord, just worse values. In the case I cited, I do not know the brand of 16 AWG I used but the 14 AWG is Studio Acoustic which I believe uses a flexible silicon dielectric while the 16 AWG is a much harder stiffer dielectric. So, there are some unknown here and the comparison in this case is probably not valid for drawing conclusions. It seems that while twisted pairs are mainly used to decrease noise ingress and egress, it makes capacitance more consistent along the cable and reduces inductance by virtue of keeping the conductors closely spaced. However, in zip chord they are closely spaced anyway. This is somewhat curious because some high end cable has the conductors space apart deliberately, which would seem to only improve capacitance at the expense of impedance and inductance. There is a whole thing on braiding cables, but there are various ways in which that may be done and very little about the effects on the cable properties we are talking about here. It seems impossible to test these things quantitatively without very expensive equipment. I don't think I want to go buy a scope just for this. Probably someone somewhere has tested all this before? I just don't know where to look. Edited February 14, 2020 by kenwstr
stereo coffee Posted February 14, 2020 Posted February 14, 2020 Sadly 99 % of amplifiers know nothing of the loudspeaker that they attach to. We need to embrace current amplification, and much less so voltage amplification. http://www.customanalogue.com/various/Current-Amplification.pdf There is need of a second Zobel network at the loudspeaker end of the cable - but this is rarely implemented as outlined by Cyril Bateman on page 20 of the cables at AF article. http://web.archive.org/web/20091027010126/http://uk.geocities.com/cyrilb2@btinternet.com/downloads_5.html
gwurb Posted February 14, 2020 Posted February 14, 2020 31 minutes ago, stereo coffee said: There is need of a second Zobel network at the loudspeaker end of the cable - but this is rarely implemented as outlined by Cyril Bateman on page 20 of the cables at AF article. http://web.archive.org/web/20091027010126/http://uk.geocities.com/cyrilb2@btinternet.com/downloads_5.html Zobel network is implemented between the crossover and the driver, not prior to crossover. The goal is to make the driver impedance as linear as possible (if so desired), not the to make the crossover network linear. Good reading: https://leachlegacy.ece.gatech.edu/ece4445/downloads/zobel.pdf
stereo coffee Posted February 14, 2020 Posted February 14, 2020 (edited) 24 minutes ago, gwurb said: Zobel network is implemented between the crossover and the driver, not prior to crossover. The goal is to make the driver impedance as linear as possible (if so desired), not the to make the crossover network linear. Good reading: https://leachlegacy.ece.gatech.edu/ece4445/downloads/zobel.pdf In any audio system there are three Zobels needed, one at the amplifier to assist the amplifier with the speaker cable ( note NOT the speaker itself ) a second at the speaker lead end, and the third you refer to. Edited February 14, 2020 by stereo coffee
gwurb Posted February 14, 2020 Posted February 14, 2020 5 hours ago, kenwstr said: So it is difficult to choose a different cable without changing all the properties. I assume these relationships would work in a similar way for zip chord, just worse values. In the case I cited, I do not know the brand of 16 AWG I used but the 14 AWG is Audio Acoustic which I believe uses a flexible silicon dielectric while the 16 AWG is a much harder stiffer dielectric. So, there are some unknown here and the comparison in this case is probably not valid for drawing conclusions. Hey, Great post! I am going to extrapolate some of the data from the link that I shared before; the impedance of copper wire. To be precise in my message I would need to take measurements but I don't have the same equipment so I can't. I think the link shows enough of a trend to use extrapolation and theoretically their results make sense anyway. Lets build up a theoretical cable from basic parts. It may help to explain what you are seeing in the findings. The original link showed that smaller gauge, single strand, copper approaches ideal in terms of being as transparent as possible. In other words thinner strands have more uniform impedance over a frequency range, and introduce less phase shift in to our signal. First step, lets start off with thin strands but not too thin as they will break under any movement. Unfortunately thin strands have higher impedance than thicker strands, so we have a problem as there will be too many losses due to heat when we try and conduct electricity through them, and they they may melt. But, we still don't want thick strands. Also, unless shielded well enough single strands will be very susceptible to external noise. A proposed solution is to combine multiple strands together. Second step, combining strands in to a conductor: Here is where we need to think more about capacitance and inductance (we didn't before because we accepted copper as the only material that we'll use and thus we accepted its material properties related to capacitance and inductance). Placing two strands near each other and sending current through them will create interaction between the two strands. The interaction effects will be multiplied when there are more than two strands. The space between the strands creates a capacitor (conductor, dielectric (air), conductor is a simple capacitor). The closer the strands are to each other the more they will interact and there will be less impedance between them, and the closer the strands are to each other the greater the capacitance. We want to pass as much current as possible at a decent voltage level. When we place multiple strands together we create a common conductor. This decreases the overall conductor impedance, which we want. When we induce current flow in the conductor we actually induce current flow in the strands so we are still keeping the benefits of having thinner strands. We also observe interactions between the strands, which creates its own undesired current flows. To try and gain benefits from having multiple strands together we twist them together. This increases the capacitance, reduces inductance and forms a conductor that is less susceptible to external noise. In other words our twisted together strands become a conductor that is able to offer less impedance, is carrying a single that experiences less interference, and is able to hold charge well but also allows it to dissipate quickly (less inductance). We now have a conductor that is made of multiple strands but we need more than a single conductor. In unbalanced dual conductor cable we need two. Third step, combining conductors to make a cable. We take our conductors and place them in a dielectric like a polymer. The better the dielectric the less external magnetic field is created outside of the polymer insulator that is insulating the conductor. We also want to increase the distance between the two conductors. We want less capacitance between the two conductors (increase dielectric and distance) and increase inductance as we want them to have less influence on each other. With magnetic field having less influence on the partner conductor there is less noise introduced in to the conductor from its partner. Single insulation is good but we can push the boundary if we want to and start double insulation. Fourth step, double insulation. The two layers of dielectric are different. They have different mechanical and electrical properties. The less conductive but less pliable dielectric would be surrounding the conductor but its thin. The thicker and more pliable dielectric would then add an extra shield. We can go even further with insulation and shielding. Shielding from external interference and using conducting shields. Optional fifth step, more shielding. Good where it counts, not needed where its not needed as it can end up doing nothing significant to the signal path but adds extra cost. Around each of the two conductors we place insulation (dielectric). We could then place another insulator but also go further. We could then place further conductors (they are far enough apart now in terms of electric propagation) that form an external shield to our primary conductor pair. The shield will have electric flow induced in it from external magnetic and electric interference and shield the inner pair. But that shielding is best protected by another layer of dielectric so it itself is somewhat protected from external interference. So now we have a cable. The cable is made of thin strands that have been combined to form a large gauge conductor. They have been insulated and possibly shielded. Ideally now the conductors do not 'colour' the signal but are also offering little to no impedance to the signal path. The signal is not being interfered with by the secondary conductor, and is not being effected by external noise. I'm out of time for now but I'll make more comments later.
gwurb Posted February 14, 2020 Posted February 14, 2020 6 hours ago, kenwstr said: It is very rare to get specs on these values though unless you go to certain well regarded manufacturers. I have never seen any of them specified for zip chord for example. OK, the is the crappiest basement end of the market but most likely we are all familiar with it. How come coca cola never release their recipe? It means that if people like it they will buy it but it makes it difficult for anyone else to copy. If all manufacturers published all of their specs then there would be less subjective bias, but alas perceptions with ambiguity in solid data make money. When increasing the conductor the insulation needs to increase. The more of both the bigger the cable. At that stage there is a lot of playing around with sizing to maintain performance but keep physical dimensions, and costs, down. The calculator that you found seems to be correct (but I am guessing what is being kept constant and what is changing in terms of all of the physical properties). Why increase the gap between strands or between conductors? Increasing it between conductors is beneficial. Increasing it between strands, I am not sure; I would need to see data. My guess would be 'punchier' sound as there is more charge built up between strands and when its discharged its results in more charge reaching the speaker. That would also result in worse off transients but there may be a balance where transient response can be low, and the sound may be pleasing. I wouldn't think that would be a better cable per se, I would think it would be a cable to is preferred by some people because of the way it colours the sound. Ideal is perfect signal transit, everything else plays in to preferences.
gwurb Posted February 14, 2020 Posted February 14, 2020 1 hour ago, stereo coffee said: In any audio system there are three Zobels needed, one at the amplifier to assist the amplifier with the speaker cable ( note NOT the speaker itself ) a second at the speaker lead end, and the third you refer to. Are you saying that all cables have a zobel network on the end? I haven't seen all cables but nothing that is sold by the meter has them unless you introduce them yourself. I am yet to see one in the cables that I have seen. Maybe I have not looked hard enough or simply overlooked them. But for example I don't see any reference to it in something like this: https://www.chord.co.uk/product/c-screen-speaker-cable/ Is it introduced above a certain price point?
stereo coffee Posted February 14, 2020 Posted February 14, 2020 23 minutes ago, gwurb said: Are you saying that all cables have a zobel network on the end? I haven't seen all cables but nothing that is sold by the meter has them unless you introduce them yourself. I am yet to see one in the cables that I have seen. Maybe I have not looked hard enough or simply overlooked them. But for example I don't see any reference to it in something like this: https://www.chord.co.uk/product/c-screen-speaker-cable/ Is it introduced above a certain price point? No, they should - but sadly DO NOT have a zobel on the speaker end of the lead per the attached diagram, for each channel. parts are 18nf capacitor non polarized 100v, a 68R 1 watt, and a 100R 5 watt non inductive type - so NOT wire wound. No, you will have to do this yourself - DIY
kenwstr Posted February 14, 2020 Author Posted February 14, 2020 3 hours ago, stereo coffee said: Sadly 99 % of amplifiers know nothing of the loudspeaker that they attach to. We need to embrace current amplification, and much less so voltage amplification. http://www.customanalogue.com/various/Current-Amplification.pdf There is need of a second Zobel network at the loudspeaker end of the cable - but this is rarely implemented as outlined by Cyril Bateman on page 20 of the cables at AF article. http://web.archive.org/web/20091027010126/http://uk.geocities.com/cyrilb2@btinternet.com/downloads_5.html Yes, This is a major argument in favor of active speakers as the whole design could be integrated synergistically. The problem is that manufacturers know most customers will not pay for everything they can't see in the box (it still just looks like a speaker). This of course leads to too many corners being cut for the potential to become reality. 3 hours ago, stereo coffee said:
gwurb Posted February 14, 2020 Posted February 14, 2020 (edited) 2 hours ago, kenwstr said: Yes, This is a major argument in favor of active speakers as the whole design could be integrated synergistically. There is still a line level signal that needs to get to the active speaker. That needs to be carried by a well built cable. It would be better to get small amount of noise on an amplified signal going from an amp to a speaker as the SNR is higher. Getting noise on to a line level signal results in a low SNR, which then ends up amplified. To fix that poor SNR input filtering needs to happens, which changes the original audio. Alternatively digital signal needs to be fed to active speakers, filtering that is easier. Power amps have voltage gain first and current gain second. Output impedance should be as low as possible, cable should be as transparent as possible. All of that leads to a well controlled voltage at the speaker input, and good amount of current driving the speaker. Edited February 14, 2020 by gwurb Rephrase
stereo coffee Posted February 14, 2020 Posted February 14, 2020 48 minutes ago, gwurb said: There is still a line level signal that needs to get to the active speaker. That needs to be carried by a well built cable. It would be better to get small amount of noise on an amplified signal going from an amp to a speaker as the SNR is higher. Getting noise on to a line level signal results in a low SNR, which then ends up amplified. To fix that poor SNR input filtering needs to happens, which changes the original audio. Alternatively digital signal needs to be fed to active speakers, filtering that is easier. Power amps have voltage gain first and current gain second. Output impedance should be as low as possible, cable should be as transparent as possible. All of that leads to a well controlled voltage at the speaker input, and good amount of current driving the speaker. In a unbalanced audio system, perfect silence is achievable, so SNR does not contribute. You just have to consider attenuation methods you might not have known or considered. Reinforcing the voltage amplifiers long history, does little to embrace the next page , which will see the amplifier fully capable to sense the loudspeaker, current driven and long overdue, then having capability to correct for speaker non linearity - based on comparison to the input signal.
gwurb Posted February 14, 2020 Posted February 14, 2020 1 hour ago, stereo coffee said: In a unbalanced audio system, perfect silence is achievable, so SNR does not contribute. You just have to consider attenuation methods you might not have known or considered. Reinforcing the voltage amplifiers long history, does little to embrace the next page , which will see the amplifier fully capable to sense the loudspeaker, current driven and long overdue, then having capability to correct for speaker non linearity - based on comparison to the input signal. What is an alternative way to drive speakers if not using both voltage and current gains? I think affordable real time audio feedback systems are a long way away. Measure and then preset components are around but they are not for the mass market yet due to their price. The step to make it real time is far away.
stereo coffee Posted February 14, 2020 Posted February 14, 2020 6 hours ago, gwurb said: What is an alternative way to drive speakers if not using both voltage and current gains? I think affordable real time audio feedback systems are a long way away. Measure and then preset components are around but they are not for the mass market yet due to their price. The step to make it real time is far away. To use adequate source impedance to arrange the complex load of the loudspeaker as part of the amplifier circuit. A good implementation would enable the amplifier to sense how much current and how much voltage, and at what time, it should amplify, based on that load, with comparison to the input signal. https://www.current-drive.info/
kenwstr Posted February 15, 2020 Author Posted February 15, 2020 18 hours ago, gwurb said: Hey, Great post! I am going to extrapolate some of the data from the link that I shared before; the impedance of copper wire. To be precise in my message I would need to take measurements but I don't have the same equipment so I can't. I think the link shows enough of a trend to use extrapolation and theoretically their results make sense anyway. Lets build up a theoretical cable from basic parts. It may help to explain what you are seeing in the findings. (SNIP) In the post you responded too, I made a mistake, the 14 AWG I referred to is Studio Acoustic, not Audio Acustic. You put a great deal of effort into you response which I appreciate. I just want to say, I have read that a shielded speaker cable construction will significantly raise, either capacitance or inductance (I don't recall which) to a seriously low pass threshold. For this reason, a twisted or braided construction will perform better. Twisted pair construction was originally developed for long signal level (>2V) cable where RF ingress is proportionately significant. However, passive speaker cable carries a much higher voltages, sometimes even lethal levels. For this reason, induced noise ingress is proportionally much smaller and twisted pairs attenuate it further. However, there is no harm trying anything you like to satisfy curiosity. 18 hours ago, gwurb said:
Southerly Posted February 15, 2020 Posted February 15, 2020 This is for me a very timely thread as I am removing the internal x/overs in my Heybrook Sextet Mk4 s for external ones. I should say that it is my intention to one day replace passive x/overs for active ones having been influenced by Rob Elliots excellent articles on the same. My opinions are from purely empirical experiments rather than qualifications and theoretical arguments and they are based on 4 years spent building analogue interconnects. The difference between i/connects and speaker cables is surely one of current and voltage otherwise other factors are equal? gwurb raises many good questions and a few answers - heat is not a problem with i/connects but definitely is with sp/cables. Multi strand cables are a complete no-no unless separated by 'effective' dialectric. High purity is important, within a short time of starting my experiments with i/cs this became audibly apparent. This comment relates only to i/cs - as I abandoned 'accepted wisdom' so the sound improved - shielding distorted everything, very quickly abandoned. Number of conductors was a revelation - I ended up using 4 x for signal and return - 4 x 26AWG solid core = 1mm in total. Dialectric - hugely important. I simply cannot understand why Teflon is de rigeur, from the same fluoropolymer family, FEP is transparent in a way that Teflon simply cannot compete. If air is the best dialectric and coupled with Oliver Heaviside's unchallenged work on electric conduction that - electricity flows not only through a conductor but on the surface and around the conductor - oversized FEP tubing as dialectric was emphatically superior on first trial. Not mentioned on this thread are binding posts as antiquated and interfering with signal flow as are RCA,XLR and h/phone jacks - in need of serious updating and real change. Chris - zobel networks - surely interfering with signal flow. The mantra - KISS has for me always been a winner. Do you still find the i/cs I sent you worthwhile? gwurb raises a very good point - what about internal wiring. Continuity is totally abandoned why? there is a complete discontinuity between internal/external wiring - shouldn't external and internal be the same for continuity of signal flow. I have used the UBYTE 11 (TNT audio) for many years trouncing money wasted on commercial cables. Experimental construction is - 2 x 1mm solid core copper for signal and return inside oversized FEP (standard wall) tubing, straight runs no twisting, no connectors at either end, binding posts on amps but soldered to same gauge solid core wire on external passive x/over board suspended level with Tonigen tweeter level. With 4-5ft length so no contact with floor. Internal wiring - no problem with woofers or mid range, uncertain about the Tonigen tweeter. If the result is good I will then invest in 6N solid core copper. Anyone curious about FEP tubing - Adtech Engineering.
gwurb Posted February 15, 2020 Posted February 15, 2020 32 minutes ago, Southerly said: This is for me a very timely thread as I am removing the internal x/overs in my Heybrook Sextet Mk4 s for external ones. I should say that it is my intention to one day replace passive x/overs for active ones having been influenced by Rob Elliots excellent articles on the same. Are you going to move to external passive for now? Next step will be active analogue or active digital? I was having a read of TI example regarding analogue active crossover: http://www.ti.com/tool/TIPD134 It had me pondering! I was thinking I may want to build my own speaker to test our analogue active design.
Ittaku Posted February 15, 2020 Posted February 15, 2020 (edited) The lowest characteristic impedance speaker wires I'm aware of are from Alpha Core Goertz. They have a bit of a cult following and have characteristic impedance of 2-4 ohm to match most speaker loads. However it comes at a cost, and that cost is quite high capacitance (almost 1000x as much as other cables). When used at modest to long length with high bandwidth amplifiers it is absolutely vital that there is a zobel network at the speaker end or they can put them into oscillation and destroy them. https://www.thecableco.com/cables/speaker-cables/mi-3-divinity-speaker-cable-pair.html Edited February 15, 2020 by Ittaku
HypnoToad Posted February 15, 2020 Posted February 15, 2020 20 hours ago, Southerly said: gwurb raises a very good point - what about internal wiring. Continuity is totally abandoned why? there is a complete discontinuity between internal/external wiring - shouldn't external and internal be the same for continuity of signal flow. Another thing to consider is not just the wiring in the speaker box but the voice coil in the speaker itself, for example it would take around 6 metres of 34 awg copper wire to make a 5.6 ohm speaker coil.
stereo coffee Posted February 16, 2020 Posted February 16, 2020 (edited) On 15/02/2020 at 1:36 PM, Southerly said: This is for me a very timely thread as I am removing the internal x/overs in my Heybrook Sextet Mk4 s for external ones. I should say that it is my intention to one day replace passive x/overs for active ones having been influenced by Rob Elliots excellent articles on the same. My opinions are from purely empirical experiments rather than qualifications and theoretical arguments and they are based on 4 years spent building analogue interconnects. The difference between i/connects and speaker cables is surely one of current and voltage otherwise other factors are equal? gwurb raises many good questions and a few answers - heat is not a problem with i/connects but definitely is with sp/cables. Multi strand cables are a complete no-no unless separated by 'effective' dialectric. High purity is important, within a short time of starting my experiments with i/cs this became audibly apparent. This comment relates only to i/cs - as I abandoned 'accepted wisdom' so the sound improved - shielding distorted everything, very quickly abandoned. Number of conductors was a revelation - I ended up using 4 x for signal and return - 4 x 26AWG solid core = 1mm in total. Dialectric - hugely important. I simply cannot understand why Teflon is de rigeur, from the same fluoropolymer family, FEP is transparent in a way that Teflon simply cannot compete. If air is the best dialectric and coupled with Oliver Heaviside's unchallenged work on electric conduction that - electricity flows not only through a conductor but on the surface and around the conductor - oversized FEP tubing as dialectric was emphatically superior on first trial. Not mentioned on this thread are binding posts as antiquated and interfering with signal flow as are RCA,XLR and h/phone jacks - in need of serious updating and real change. Chris - zobel networks - surely interfering with signal flow. The mantra - KISS has for me always been a winner. Do you still find the i/cs I sent you worthwhile? gwurb raises a very good point - what about internal wiring. Continuity is totally abandoned why? there is a complete discontinuity between internal/external wiring - shouldn't external and internal be the same for continuity of signal flow. I have used the UBYTE 11 (TNT audio) for many years trouncing money wasted on commercial cables. Experimental construction is - 2 x 1mm solid core copper for signal and return inside oversized FEP (standard wall) tubing, straight runs no twisting, no connectors at either end, binding posts on amps but soldered to same gauge solid core wire on external passive x/over board suspended level with Tonigen tweeter level. With 4-5ft length so no contact with floor. Internal wiring - no problem with woofers or mid range, uncertain about the Tonigen tweeter. If the result is good I will then invest in 6N solid core copper. Anyone curious about FEP tubing - Adtech Engineering. Hi Stuart Yes the cables are great, I use them constantly thank you, they provide insight into music other cables cannot reach. Re Zobels their use is an admission of a voltage power amplification system where they are absolutely required. The amplifier end Zobel presents as a High frequency load to the power amp and is required to curb parasitic oscillations that otherwise occur- particularly where a speaker load is not present. The speaker lead would otherwise act as an antenna of sorts. Think of what precedes the speaker binding posts and we see circuit board tracks connecting to various rates of transistor or mosfet switching. All solid state amps are unstable to varying degrees when a loudspeaker lead is connected. A zobel at the amplifier end of the cable assists the amplifier to exist and survive when a speaker cable is attached. The reader might ask what happens with Zobels and valve amps ? A valve amp in 99% of cases uses a transformer as its speaker interface and is to the best of my knowledge stable with or without speaker leads. The zobel rarely used at the speaker lead end, we can see acts as a load for the earlier zobel and further properly assists the proper transfer of full frequency audio information, to arrive intact at the speaker binding posts. We stop here, for a moment to pause and reflect that we have a pretty good system of voltage delivery.... but everything changes ... as what follows is a highly reactive speaker. A third zobel exists to try to account for the speaker impedance located inside the speaker cabinet. The passive crossover is with 99% of such designs begun with a inductor - its purpose is to control the back electromagnetic force EMF of particularly the bass driver(s) so pretty alarming compromises occur because we choose to use voltage amplification - the best a voltage amplification system can do is just stiffen that voltage delivery - but the amplifier knows nothing of the speaker. Hence much needs to be done to change the landscape of power amp design, to begin to know the speaker they attach to. Edited February 16, 2020 by stereo coffee
muon* Posted February 16, 2020 Posted February 16, 2020 20 minutes ago, stereo coffee said: A valve amp in 99% of cases uses a transformer as its speaker interface and is to the best of my knowledge stable with or without speaker leads. Just don't run them without a load and an active signal :)
stereo coffee Posted February 16, 2020 Posted February 16, 2020 32 minutes ago, muon* said: Just don't run them without a load and an active signal Interesting ! thanks
Recommended Posts