The honest (and painful) answer is, you get what you pay for, and that is true for tools of any kind.
I was a wiring technician in one of my previous lives, so I kind of know some stuff about crimp terminals.
The big issue with crimping is that you need a good solid gas tight connection. That requires the proper crimping tool for the particular type of crimp terminal being used, which usually means getting the crimp tool from the original crimp terminal manufacturer or a qualified alternate supplier. The so-called "universal" crimp tools rarely work very well, or at all.
Mike
For QC terminals, I use a Molex 0037 ratcheting crimp tool. $60 at Digikey (P/N: WM9986-ND). The $5 cereal box crimp tools that you find at home improvement and auto parts stores are completely worthless. Not only will your hands absolutely hate you if you have to crimp more than a couple, usually the connection is so loose that the terminal falls off the wire if you look at it funny. Try a ratcheting crimp tool and you'll never use anything else.
Tom
Tom
Bingo! There's no substitute for the right tool for the job, it's the worst possible place to cut corners.
Even when those cheap pliers-style crimpers seem to make a solid joint that passes the "wire pull" test, it's still not a gas tight joint, and that's important for long term reliability.
Mike
Last edited:
Antek transformer wires
I was very surprised to find that Antek uses 18AWG for connection to the AC entry terminals.
The reason for my surprise is that many "hi-end audiophile" power cables are 12 or 10 AWG and some are 1/2" and more in diameter. To top it off, some are only 3 or 4 feet long yet sell for $$$.
Maybe there is a technical reason but I feel there's a big market for snake oil.
Now back to reality. If a 3 or 4 foot long power cord is needed, for a MOD86 amp, wouldn't a 16AWG computer power cable be sufficient?
I was very surprised to find that Antek uses 18AWG for connection to the AC entry terminals.
The reason for my surprise is that many "hi-end audiophile" power cables are 12 or 10 AWG and some are 1/2" and more in diameter. To top it off, some are only 3 or 4 feet long yet sell for $$$.
Maybe there is a technical reason but I feel there's a big market for snake oil.
Now back to reality. If a 3 or 4 foot long power cord is needed, for a MOD86 amp, wouldn't a 16AWG computer power cable be sufficient?
I suspect the snake oil market is quite a bit larger than the high-end audio market.
That'll be fine.
When it comes to power cords the only thing I can think of that would make a difference would be the cross-sectional area of the ground conductor. A big fat ground lead combined with a quality power strip would result in a lower ground impedance between pieces of equipment, thus, resulting in lower error voltage between the different equipment pieces.
With the differential input of the Modulus-86, this is a moot point. You can switch power cords on the MOD86 all you want. It won't make any difference.
Tom
Sorry for this off topic post. I didn't know where else to put it and my post about Antek using 18awg wire for AC hookup is what prompted it.
That post got me thinking about speaker cables and snake oil I was influenced, no, probably fooled by, and the money I've wasted over the last 30 years.
If 18AWG is OK to carry 120V why use speaker cable the size of garden hose? For a short run of 5 feet what would be reasonable? 18awg, 20awg, 26awg?
I think this is somewhat related. I just remembered something I read, a few years back, about some "renown" speaker expert, or maker or someone involved in audio, doing a blind speaker cable test with a large number of people. I think it was done in the 70s or 80s.
He hooked up a speaker with cheap 16AWG zip cord and Hi-End cable via a switching box. He would switch, back and forth, between, what he called, cable A and cable B and ask attendees, after each switching, which was better…maybe they wrote down their answers. I recall that the result was a draw. Cheap zip cord was selected best as often as the hi-end cable.
So what should we be using? Honestly…, I didn't want to post this in the speaker section, because I've read many posts there, and it seems more than a few think variations of garden hose is the way to go
That post got me thinking about speaker cables and snake oil I was influenced, no, probably fooled by, and the money I've wasted over the last 30 years.
If 18AWG is OK to carry 120V why use speaker cable the size of garden hose? For a short run of 5 feet what would be reasonable? 18awg, 20awg, 26awg?
I think this is somewhat related. I just remembered something I read, a few years back, about some "renown" speaker expert, or maker or someone involved in audio, doing a blind speaker cable test with a large number of people. I think it was done in the 70s or 80s.
He hooked up a speaker with cheap 16AWG zip cord and Hi-End cable via a switching box. He would switch, back and forth, between, what he called, cable A and cable B and ask attendees, after each switching, which was better…maybe they wrote down their answers. I recall that the result was a draw. Cheap zip cord was selected best as often as the hi-end cable.
So what should we be using? Honestly…, I didn't want to post this in the speaker section, because I've read many posts there, and it seems more than a few think variations of garden hose is the way to go
Yeah. You gotta make sure your speaker cables are properly oiled. Otherwise the micro diodes get all confused and the electrons don't know what to do. Also make sure to use speaker cable stands. Do make sure you get the ones machined from blocks of Oregonian Myrtlewood. Anything else is far inferior and you really don't want your cables to touch your micro fibre carpet (see comment about confused electrons). </sarcasm>
Now on a serious note:
Because it's the voltage drop across the wire that matters. If you run a stereo MOD86 at full bore, you may draw 1 A on the mains through that 1.5 meter mains cable. So assuming you have 120 V at the light socket, you'll see a voltage at the transformer primary of,
Vprimary = Vmains - I * R -> Vprimary = 120 - 1 * (0.021*3) = 119.94 V.
That's not enough droop to care, which is why 18 AWG is fine on the transformer primary.
Now, before I get schooled by some of the more detail oriented readers here, let me point out that in reality the math isn't quite that simple as the voltage and current are not in phase. Above math is accurate for a purely resistive load (current and voltage in phase). The MOD86 is not a resistive load on the mains, so the power factor should be taken into account. I'll leave that as an exercise for the reader.
Now do the same math at the speaker using a cable run of 1.5 m (two conductors -> total length in series with load = 3 m). Assume a 4 Ω*resistive load and you get 7 A peak when the MOD86 output swings towards the 28-30 V voltage rail. That's about 5 A RMS.
Vspeaker = Vout - I * R -> Vspeaker = 28 - 5 * (3 * 0.021) = 27.69 V.
That's over 1 % of droop (-39 dB). We might care about that. Now switch to 12 AWG (0.005211 ohm/meter):
Vspeaker = Vout - I * R -> Vspeaker = 28 - 5 * (3 * 0.0052) = 27.92 V.
0.3 % droop (-51 dB).
Again: Back-of-envelope math here.
Another way to look at it is to compare the speaker cable resistance with the output impedance of the amplifier. In the Modulus-86, the output impedance is dominated by the impedance of the inductor used in the Thiele network. The DCR measures a few 10s of mΩ. Connectors have a finite and non-zero impedance as well.
If I was to give a rule of thumb, I'd probably aim to keep the cable series impedance below the sum of the connector impedances and amplifier output impedance. In reality, I usually end up going with the largest cable that will fit in the connectors I'm using or that I'm willing to afford, whichever is lower. I'm running a 4x14 AWG cable for my LXmini speakers...
The next thing to consider as that most people don't run their amps with sine waves or at full power. Once you factor in the crest factor, you can probably dial the cable droop back a good 10-15 dB. Then compensate for the fact that critical listening tends to take place at reasonable volume levels and you get another 10-20 dB. You quickly realize that the cross-sectional area of the cable doesn't matter as much as the cable manufacturers want you to believe, as long as you keep it reasonably large.
This leads me to believe that my rule of thumb (Zcable < Zout(amp) + Zconnector) is probably the way to go.
Note that I'm talking the series impedance of the cable. Not the characteristic impedance of it. The characteristic impedance is relevant once the length of the cable approaches the wavelength of the signal being transmitted through the cable. It's not applicable for home audio.
You can see the resistance per length of the various wire gauges here: https://en.wikipedia.org/wiki/American_wire_gauge
I'm not surprised. A way to improve that experiment would be to make the experimenter blind as well. Have a computer select A or B, randomly and keep a record of whether A or B was selected. Then compare with the participants' record of whether they thought they were listening to A or B.
There is some science in cable geometries. Note that some geometries (like the woven or braided cables) either don't help at all or actually make the performance (cable insertion loss) worse. Others, such as the Goertz MI-series actually do make sense. The inter-conductor capacitance ends up cancelling the inductance of the cable, resulting in lower cable loss at 20 kHz. I don't think the difference is likely to be meaningful, but is is there.
Tom
Last edited:
I've got mucho feet of XLO reference cable I bought over 20 years ago. It is made of many 28 or so awg solid strands each coated in teflon, then all wound in a complex, good???, geometry and sheathed in teflon. It seems to function real well. I am just looking for something smaller and less obtrusive. Belden Brilliance OFHC comes to mind. Good wire and coating, no fancy winding geometry, and a low price per foot. Thinking of getting 16AWG. Anyone using that stuff?
Another speaker cable option, not as cheap as zip cord, but not as expensive as audiophile cables, is to use Canare's 'star quad' speaker cable. You can purchase from Markertek or Blue Jeans Cable.
Another point, which is acknowledged by cable 'reviewers' is the shorter the speaker cable the less effect they perceive. I.E., if it is short even the folks who are into cable swapping don't hear any differences.
Another point, which is acknowledged by cable 'reviewers' is the shorter the speaker cable the less effect they perceive. I.E., if it is short even the folks who are into cable swapping don't hear any differences.
For mounting LM3886 I use these clips MAX08NG Aavid Thermalloy | Mouser Also available from Fisher at Farnell/RS. Those put pressure on the right place of the chip and allow for "thermal play".
Yes, the 'only' way to go! Makes things so much easier, especially if you're into changing components on the board.
They also have the smaller MAX07NG for the TO-220 devices like diodes, etc - can't see where they've hidden the 'bar clamps' that're really useful for '2 in 1' devices (diode bridges, for example)
I've wasted a silly amount of time in those double blind tests - you just confuse yourself with short exposure to music from different components even when you're used to critical listening -
When you take your time (at home), with your own system, most of the more extreme cable 'sound effects' become obvious and you can then settle for the choice that more suits your budget, or whatever - I used to hate selling speaker wire as many choices are just fashion statements and many speakers just don't show any differences in the sound with extreme differences in speaker wire at normal listening levels.
For some people, CAT 5, (or 6, not sure) seems to work quite well in some systems, surprisingly - I even tried solid silver in cotton/Teflon for my high efficiency drivers for a short time, but not for long as they were far too bright in the system - for bass, I'm still using #16 solid copper (UPOCC) in the plastic with a rubber sleeve (well, I did splurge all those dollars ...!)
Now, headphones are a bit different ....
Yes, the 'only' way to go! Makes things so much easier, especially if you're into changing components on the board.
They also have the smaller MAX07NG for the TO-220 devices like diodes, etc - can't see where they've hidden the 'bar clamps' that're really useful for '2 in 1' devices (diode bridges, for example)
I've wasted a silly amount of time in those double blind tests - you just confuse yourself with short exposure to music from different components even when you're used to critical listening -
When you take your time (at home), with your own system, most of the more extreme cable 'sound effects' become obvious and you can then settle for the choice that more suits your budget, or whatever - I used to hate selling speaker wire as many choices are just fashion statements and many speakers just don't show any differences in the sound with extreme differences in speaker wire at normal listening levels.
For some people, CAT 5, (or 6, not sure) seems to work quite well in some systems, surprisingly - I even tried solid silver in cotton/Teflon for my high efficiency drivers for a short time, but not for long as they were far too bright in the system - for bass, I'm still using #16 solid copper (UPOCC) in the plastic with a rubber sleeve (well, I did splurge all those dollars ...!)
Now, headphones are a bit different ....
Add a SilPad sleeve and you'll be set.
I hate to burst your bubble, but you're probably dealing with confirmation bias. The phonological loop (auditory short-term memory) is only about 30 seconds long, so if you want to be able to compare two sound bites, each bite should be less than 30 seconds long.
I'm not surprised you found the scientific tests to be a waste of time. In most cases there'd be no significant difference, hence no significant results. That's how science works.
CAT5, 5e, and 6 all use 24 AWG wire so it wouldn't matter which one is used. Yeah, I've seen the mention of network cable being used a speaker wire on DIY Audio a few times, usually accompanied by the notion that "it sounds better". I've yet to see any of these "tests" being done in a scientific way (blind, controlled, random assignment, etc.)
Tom
I hate to burst your bubble, but you're probably dealing with confirmation bias.
Yeah Tom, could be something to that - I don't get too wound up about all that as my own hearing preferences are the ones that mainly matter to me but must accept that after the accumulated years, it's pretty much guaranteed that my hearing isn't in the golden brigade by any means, but it's the only one I have - pretty sure some opinions are in the same questionable state!
I find the dominance of the speakers and the room response (and it's control) are more significant to me these days and leave most discussions about interconnects and signal accuracy to the world of design and specification where it's quite significant - I still unfortunately open my 'yap' and air my ignorance far too often
Yeah Tom, could be something to that - I don't get too wound up about all that as my own hearing preferences are the ones that mainly matter to me but must accept that after the accumulated years, it's pretty much guaranteed that my hearing isn't in the golden brigade by any means, but it's the only one I have - pretty sure some opinions are in the same questionable state!
I find the dominance of the speakers and the room response (and it's control) are more significant to me these days and leave most discussions about interconnects and signal accuracy to the world of design and specification where it's quite significant - I still unfortunately open my 'yap' and air my ignorance far too often
I can find no valid logic in this. My audiotory system seems to need a "settling time" of at least 10...20 seconds to become conciously aware of a change in a change of sound during a gapless playback (with soft transitions, eg. crossfade), ever after a lot of training. Also I find repeating the same snippet over and over is a bad idea, doesn't work for me.
When I evaluated the effect of phase response of speakers I found I hardly can tell a difference when looping a 30 second snippet and switching in a double-blind test. When the length is extended to tens of minutes and no looping is involved, my hit-or-miss rate was much better. Practically I listen to a whole album with only a few switching points (still in double-blind mode) and note my impressions. Then I compare this to what the arbitrator (a program) had choosen (including fake switching) to select next. Typically 90% hit rate ... issue settled (phase/polarity matters). No way I would have come to this result using the looping/short-term A/B-switching method.
So did Bruno, which is why it was mentioned. However you forgot to note that twisted pair still suffers from problems, which is why starquad was invented. If you are minimising any possible differential mode pickup its only pennies more per metre in non-audiophool outlets
. Audible difference zero, knowledge you have done the best you can for your signal worth at least 2 veils
The settling time is your brain, not your ear. Of course your sample of one proves all the research into this over the years is wrong
But you don't understand...Some folk's ears are super-human, they can hear stuff that none of us mere mortals can.
The DIY membership here seems to embody a statistically anomalous status in having so many golden ears in such a small place. Mike
Last edited: