Part I, Philosophy - this is going to be a long story
Posted by Bold Eagle on October 15, 2010 at 19:41:10
In the late 60's, I got introduced to
the KLH Model 20 Compact System. At $400 it was the best you could do for the
money at the time. To be sure, there were compromises made. Reading a review on
it, the reviewer opened my eyes with his comments on the design approach. The
amp, for instance, had non-flat response to help EQ the speakers. The speakers,
in turn were 4 ohm versions of the KLH 17 to get more power from the amp. What
this did was to make me realize that it was not about component quality; it was
really a Systems Engineering problem. I figured, "Hey, I'm an engineer, I
understand the principles of Systems Engineering. If I study up on the
technology and increase the $$$, I can do better and eliminate most of the
So I thought and read, and read, and thought. Finally, I realized that the ultimate limitation was the quality of the source material. It was pointless to build up a system that was significantly better than the source. In the late 60's we had reel to reel tapes, LPs and FM. A little research let me decide on the criteria for a system that would approximately equal those sources in the home.
I drew up four criteria.
1) frequency response of 40-15,000 Hz +/- 3 dB.
2) long term average sound level of 88 dB in a room of average absorption and 3000 cubic foot volume.
3) 13 dB dynamic headroom above the 88 dB average.
4) a four piece system (integrated amp, turntable and cartridge, and 2 speakers)or the equivalent.
I figured that $800, optimally spent(no money spent on frills, fancy cabinets, etc. Kits and used were fine) would allow me to meet these criteria. The specifications I set also meant I wasn't going after Bach's Tocatta and Fugue on a pipe organ as the lowest bass at high levels would add a LOT of cost.
And thus began a very long quest.
First off, I quickly realized I didn't know nearly enough to pull this off. Second, manufacturer's specs. were worse that useless.
Over time, I learned a little more, bit by bit; and eventually realized that my original criteria were incomplete and were too difficult to meet for $800 in 1970 dollars. The first change was to add criteria number 5, Stereo imaging. I needed a wide, deep, and stable soundstage. The second was to revise the room size down to 2400 cubic feet, and the third was to increase the dynamic headroom to 16 dB. And eventually, my four piece system (or equivalent) became a receiver, a CD player, and 2 speakers, and still later, an integrated amp and tuner replaced the receiver.
I "got there" quite a while ago, and today my Super Model 20 is a pair of much modified Advents, an NAD C350 integrated amp, a Rotel RCD-971 CD player, and an Onkyo T-4087 AM-FM tuner. In my 12' x 15' den it exceeds all the criteria, sounds really good, and is a very musical system. It will also play Bach's Tocatta and Fugue with the 32 Hz pedal notes that really annoy my condo neighbors.
I'm going to stop the Philosphy part here (it's long enough) and then I'll generate Part II, Technology in which I reveal my methodology for engineering my systems.
Part II, Technology
Posted by Bold Eagle on October 16, 2010 at 13:47:21
OK, this is where I try to explain how
I use Systems Engineering principles to put a system together.
This is a top down process - meaning that we start with an overall specification and then work our way down to components. In a system, the parts have to work together, so there needs to be consideration at each break point of the overall goal, and of the criteria needed to make the parts work together.
This is quite different from the usual hi-fi approach where a bunch of components are assembled, music comes out, and it's referred to as a "system". It isn't, it's a collection of parts that more or less work together. It's really a "bottom up" approach.
First consideration is your preference in how loud you want this to be. I picked 88 dB as my long term average because it is the preferred level for recording engineers and musicians, at least according to the data in the Radiotron Handbook which was last published in 1953, so hard rock was not included. That's pretty loud, and conversation is pretty limited; but I do like my music loud when the neighbors are gone. To put that in perspective, a rock concert or a disco has a long term average of over 100 dB. I measured 105 dB at a local disco about 30 years ago. At that level, you use flat weighting, and I used slow response. 88 dB is like being close to a jazz band when they be jammin'. A full symphony orchestra gets up into that range on the big, bombastic pieces like Wagner. In any case it's also true that music sounds best when played close to live levels as your hearing loses sensitivity in the bass and to a lesser extent in the treble at low levels. So you get the right balance of all the frequencies at the live level. You can restore the balance at lower levels by using EQ.
Next up is the room where the system is to be played. You need to measure the size and figure where you want the speakers and your preferred listening piosition to be. Rooms have walls which reflect energy back into the space, and they have openings (doors and windows) where the sound can escape. An open window is a perfect absorber. You also have drapes or curtains, carpeting, and furniture which act as absorbers; but primarily in the upper mids and treble range. Another factor is diffusion. This refers to the nature of reflections from the surfaces in the room. Basically, you want a lot of complex surfaces so the sound is reflected at many different angles. This is the key to the great concert halls of the world; most of which were built in the Baroque or Roccoco eras and have very complex and ornate interiors. All that gingerbread does a very effective job of creating very complex reflections. In a listening room, bookshelves do a good job. My den has book shelves on both long walls. Further, the shelves are not all the same depth. If your room is furnished in Danish Modern with two lightly padded chairs, a table, a small area rug, and a few pictures on nearly bare walls, you're in trouble. You can buy diffusers at ridiculous prices; but a few bookshelves on the walls behind you will work wonders.
You also need to think about where you sit. Sound waves, as they reflect from the walls, have a pressure maximum at the wall. You can try this for yourself. Play something with deep bass and start in the middle of the room and gradually back up to the back wall (the front wall is where the speakers are) and notice the bass increase as you near the wall. My room is 15' x 12' and the speakers are near the front wall. I want to sit near the other end of the room. However, at the back ends a 5' wide alcove about 3' deep where the double entrance doors are. Beyond the doors is a hallway that runs past the doors. With the doors open, I sit in front of the alcove, so there is no wall close behind me, and the bass smooths out. If the doors and alcove were not there, I'd move 3-4' from the back wall in a room that size. As it is, I sit about 13-14' from the speakers and the far wall of the hallway is about 5' behind me. Remember that 13' number, we'll use it later.
Oh, one more thing about rooms. Although my room looks geometrically symmetrical, it's not acoustically symmetric. The furniture arrangement isn't symmetric, and the solidness of the walls is different as the left side is a firewall and the right side has a sliding door which has no furniture in front of it. For best imaging, you want to find locations where your speakers sound the same (use a mono source and roll the balance control from one side to the other).
OK, so we have our list of system specifications, and we've assessed the room and where we want the speakers. Now we have to pick out the speakers and place them. If you already have the speakers, you'll want to know what the frequency response is, something about their lateral dispersion, their sensitivity, and their power handling. It's also good to know if the manufacturer has placement recommendations. In small rooms you'll want speakers that can be placed closer to the front wall and maybe that can be nearer to a side wall. You'll also want to understand if the speakers are suitable for near field listening, or if they need to be farther away. My Advents are a good example. The woofer and tweeter are quite far apart. If you get too close to them, the sound from the two drivers doesn't integrate well and you get an impression of two sources of the sound with a hole in the middle of the range. Usually speakers that are good for closer listening have the drivers clustered closely. Speakers where the drivers are spread out vertically need to be further away. For an example, we can use my Advents. They can be positioned fairly close to the wall behind them, they need to be on stands, and they don't want to be in a corner. They are a nominal 8 ohm speaker; but are closer to 6 ohms as an average. Sensitivity is 87 dB for 2.83 volts input, measured at 1 meter (3.3'). Frequency response is 30-15,000 Hz. +/- 3 dB and power handling is rated at 120 watts. Dispersion is good up to about 12 kHz and fair above that. Now, all that needs some explanation.
Sensitivity (often mistakenly called efficiency) can be either anechoic, or measured in a "typical" room. Some manufacturers use one kind of measurement, others use the other. The Advents are anechoic. My ADS L-710-3's were spec'd at 92 dB. No way were they 5 dB more sensitive than the Advents - +2 dB was a lot closer. Turns out they were spec'd in a real room. For my purposes, I want the anechoic sensitivity.
Frequency response in the case of the Advents is fairly close as the bass is -3 dB around 40 Hz, so -6 dB at 30 Hz isn't too far off. That's also without room gain and mutual coupling, both effects serving to boost the low bass. Measured at my listening position, the bass is abour +4 dB from 80 Hz to 25 Hz and still +1 at 20 Hz. A larger room will have less room gain. So, if I don't play too loudly, I can reproduce those organ pedal notes just fine.
Sensitivity is usually specified for the sound pressure level for 2.83 volts input. 2.83 volts is 1 watt into an 8 ohm load. But since speaker's impedance varies a lot with frequency (there are exceptions) it's much easier to talk about voltage sensitivity as it's independent of impedance.
Power ratings for consumer grade speakers like my Advents need a little translation. The 120 watt figure for the Advents is really saying that the speakers are safe for playing music with a 120 watt rated amplifier, provided it isn't driven into clipping. Since music has a very high ratio of peak levels to long term average level, the speaker will not be dissipating more than 10-12 watts of heat at most.
Remember my 88 dB specification, and the 13' dimension? We're going to use those - so hang on. In a closed space, as in a room with not too much open area, the sound reflects off of surfaces and adds to the sound coming straight from the speakers. This is called the reverberant field. There is also the direct field. The direct field is what you measure in an anechoic (means "without echo") chamber. In free space, the direct sound for conventional speakers (not sound columns, or very tall planar speakers) the sound spreads out into the space like an expanding bubble, the surface of which is spherical. So the sound pressure level drops off by 6 dB for each doubling of distance. So, if my Advent produces 87 dB for 2.83 volts at 3.3' (1 meter), it will produce 81 dB at 2 meters. Now, 13' is pretty close to 4 meters; so at my listening position the direct field will have fallen to 75 dB. That's 75 dB in an anechoic environment. In a real room, the reverberant field from all the reflections from the room surfaces, plus the direct field is approximately equal to the direct sound halfway down the length of the room, and nearly constant after that. So, as an approximation, we'll take the 81 dB figure we calculated at 2 meters which pretty close to halfway down the length of the room from the front of the speakers.
Since we have two speakers, each producing the same sound more or less, we add 3 dB to our 81 and get 84 dB when both are playing. We want 88, so that's another 4 dB. If 2.83 volts (1 watt) gives 84 dB, then we'll need 2.5 watts continuous per speaker to get 88. Notice I've shifted from volts to watts.
My specification of 16 dB dynamic headroom is based on actual measurements published in Audio magazine in 1973. It says that the peaks were 40 times higher than the average. 40 times 2.5 watts is 100 watts. So in my little den I need the amp to be able to generate 100 watts on peaks to avoid clipping on some really dynamic music. If you only listen to string quartets you can ignore this.
The amp I use with the Advents is an NAD C350. It's rated at 60 watts per channel, continuous, has a Damping factor of 150 (source Impedance = 0.053 ohms), and is rated to be OK with 2 ohm loads. The Advents are rated for amps to 120 watts, and their minimum impedance is 5 ohms, so it's a good match. But what about our 100 watt figure? Not to worry, the C350 has 3.5 dB of dynamic headroom (DHR). That means that on short term peak demands, the amp can put out a voltage equivalent to 135 watts. So the amp's short term peak output comfortably exceeds the speaker's demand and even played very loudly we won't get into clipping.
We're cool on sound levels, and power handling, but what about the sound quality? That, dear friends, is up to you. I have four amps that will meet these criteria for power and dynamic headroom. The C350 sounds the best to me, so that's what I use.
Our speakers will work in my room, the amp will work with the speakers and meet our sound level and frequency response specification. But we have to connect the speakers to the amp. Since my electronics are located near my listening chair, the runs of cable to the speakers is 32' for the longest one. I'm very careful to make both cables the same length, as I swear it makes a difference in the imaging - even though calculations say it won't. There is an interaction between cable and amplifier impedances and the speakers. The rule of thumb is that the
sum of the amplifier source impedance and the cable resistance should not exceed 5% of the minimum impedance of the speaker if you are to avoid audible effects. Since the Advents have a minimum impedance of 5 ohms, 5% of that is 0.25 ohms. The source impedance of the amp is 0.053 ohms, so I have 0.2 ohms left for the cable. 32' feet of speaker cable is really double that since there is a total path length, out and back, of 64'. 12 gauge cable has a resistance of 1.6 ohms per 1000'. 1.6 x 64, divided by 1000 is right at 0.1 ohms. 15 gauge is 3.2 ohms per 1000', so we can use 14 gauge cable; but 16 gauge is a little too small. Besides, that's just the resistive portion, and cables also have inductance which adds impedance at higher frequencies, so I went with 12 gauge to have some wiggle room. I've tried running two parallel runs of 12 gauge, which cuts resistance and inductance in half, and heard no difference. So I think the method works.
I've become a fanatic about terminations on speaker cables. You really, really need good terminations at both ends. My Advents have been converted to take dual banana plugs, using gold plated 5 way binding posts on 3/4" centers. All internal connections are soldered. The cables are terminated on both ends in gold plated giant spade lugs which are crimped onto the 12 gauge, and them soldered. The spades can be used directly; but for convenience I very tightly clamped the spades under the head of the screws on gold plated dual banana plugs. Spades and plugs came from Radio Shack or Parts Express. The NAD has gold plated binding posts; so i have gold to gold connections on both ends. Why gold? Well, it's pretty; but it also doesn't oxidize, and if left alone for a while, forms a welded joint with no contact resistance. Copper is subject to creep, and if pains are not taken, a connection clamped onto copper will loosen over time. Hence the soldered spade lugs.
At this point, we have the system completed all the way back to the power amp input jacks. That's the most demanding part, so I'll stop here and finish up in a third installment.
Part III, More technology
Posted by Bold Eagle on October 16, 2010 at 14:55:38
When I left off, we had (finally)
reached the input jacks on the power amp. From here, we only have to get some
controls and sources and we're done - almost.
I chose an NAD C350 integrated amp for my system. Since it includes a preamp, I don't have to worry about selecting one or interfacing it to the power amp. If well engineered, it's a good argument for integrated amps and receivers as opposed to separates. However, if you do go with separates we need to worry about the interface between the preamp and power amp.
The "rule of thumb" for impedance matches between a source and a receptor is that the input impedance of the receiving component should be 10x the output impedance of the source. If both impedances are pure resistive there will be no shift in frequency response; but there will a loss of 1 dB in level. If the impedances are complex and not constant with frequency there will be losses and a shift in frequency response. I much prefer at least a 20:1 ratio. I also like to look at the schematics and see what the preamp's output stage looks like, and what the power amp's input stage looks like. Alternatively, I'll take the time to measure the impedance Vs frequency to see how flat the impedance curve is. Since the NAD preamp's output impedance is 100 ohms, and the power amp's input impedance is 20,000 ohms, a 200:1 ratio; I didn't go any further.
Tube power amps generally have very high input impedances. Solid state preamps usually have rather low output impedances. But SS power amp input impedances vary all over the place; and most tube preamps have rather high output impedances. So you need to beware of driving a SS power amp with a tube preamp. It might be OK; but do your homework. In some older gear I've even seen impedance mismatches between the preamp and power amp in a receiver and had to build a buffer amp to make them compatible.
Interconnects. In an integrated amp or receiver, you often see those little jumpers connection the preamp and power amp. I have never had a problem with those. Further, using a shielded cable at that point can cause ground loops and subvert a careful internal grounding scheme. Just besure they are clean and make good contact. For separates, I use the shortest interconnects I can. I mostly use Radio Shack "Gold" and cut them down to the length I need and then install new plugs. The Radio Shack interconnects are well made, have quite low capacitance, and are inexpensive. My components are stacked vertically, so I'm using a 12" and a 18" cable - the idea being to minimize the capacitance in the cable. Same as for speaker cables, I use only gold plated plugs.
I cannot tell you exactly what's happening; but in three cases now, I've found that longer interconnects with higher capacitance cause a harsh, bright, edgy sound. In all three cases, going to a 12" interconnect got rid of it. John Dunlavy of Dunlavy speaker fame was also a firm believer in very low capacitance interconnects; and so was Gizmo Rosenburg.
Might be a bit extreme; but I measure frequency response and channel match on all my preamps at the volume control setting I most commonly use. Then I get into the circuit and balance the two channels as closely as I can for frequency response and gain, as I believe it helps the imaging. It is not uncommon to find that the balance is off, or that frequency response doesn't match, or isn't flat - particularly on vintage gear; but even on late model gear. It's also not uncommon for tone controls to be out of calibration, so that 12 0'clock isn't the flattest setting; and/or the two channel's tone controls don't match.
By the way, most Digital Multi-Meters cannot be used to measure frequency response above 500-2k Hz. For that you need a broadband meter, like a good old-fashioned VTVM or a TVM.
So far, I've assumed line level inputs on the preamp. Phono is a whole 'nuther kettle of fish, and I won't go into that here. But if there's enough interest, I'll write a separate thing on phono.
OK, sources. Tuners, tape decks, CD players, DAC's, outboard phono preamps. All line level outputs. The same considerations apply here as between the preamp output and power amp input. Keep the impedance ratio as high as you can. Use short, low capacitance interconnects. Throw away those skinny ones that are packed with most gear. A 3' long el cheapo interconnect I measured was 300 pF. It drove one of my CD players nuts. My 12" interconnect is 27 pF. A Radio Shack Gold interconnect in 3' length measured 70 pF. A Tara Labs interconnect that cost 3 times what the R-S did was 100 pF. A DiscWasher interconnect from the 70's was 54 pF.
All right, we're done for now. I'll surely think of something I missed or forgot, so watch for a follow up. Also, let me know what you think, and especially if you find errors.
I really do follow these guidelines, and have been doing so for over 30 years.
In part II, I went through speakers and amps for my little den. I think you can appreciate that if the room were larger, you'd need more power, and maybe speakers with more power handling. That's also true if the room is unusually heavily damped or if it has high ceilings. In fact, if your room is different from a rectangle with the speakers on the short wall, talk to me about it.
Part IV, Phono Systems
Posted by Bold Eagle on October 16, 2010 at 19:21:16
I wrestled with these for many years,
tic, pop, tic!
I assume everybody knows about setting overhang, and azimuth on their tone arms, and either have a good tracking protractor, or the manufacturer's gauge. Let's include vertical tracking angle in that.
One thing that's not so obvious is to know when you have it right and the biggest unknown is the anti-skating adjustments.
In a single pivot arm with an angled head shell, about 95% of all turntables (don't hold me to 95%), the arm tends to move to the center because of the friction between the cartridge and the record. Since the stylus is in a groove it prevents the arm from moving inward, but the forces are still there and it tends to decenter the stylus in the cartridge body and to give unequal forces on both walls of the groove.
The solution for this was an outward force applied to the arm itself to counteract the skating force. There are basically two ways to see if the anti-skating forces are counteracted accurately. One is to use an ungroooved record, set the arm near the middle of the area where there would normally be grooves, and see if the arm moves in or out. The problem with this is that an ungrooved record has less friction than a grooved record.
The second method is to get a small flashlight and look at the arm from the front end on a grooved record, and see if the stylus cantelever is aligned with the center of the cartridge or offset to one side. You want it centered.
Confusing this is the brushes on Pickering, Stanton and some Shure cartridges. You never know how the lateral forces are being shared between the brush and the stylus. Plus, they make setting vertical tracking force more difficult. My advice? Remove the Pickering/Stanton brushes, and clip the Shure brush in the up position.
One of the artifacts of getting the anti-skate wrong is to have the side forces on the stylus cause the stylus cantilever to twist in its mounting causing severe miss-alignment of the stylus in the groove, and severe record wear. This is a LOT more common than you think. Check for it by dropping the arm on a mirror and looking at the stylus from head on. The mirrored image of the stylus MUST line up perfectly with the actual stylus.
We also have two basic types of cartridge: moving magnet, and moving coil. To some extent that's a misnomer, as all magnetic cartridges have magnets, coils, and iron to carry the flux. It's really more of a matter of output level, impedance, and which of the three elements moves.
The high output cartridges are relatively high impedance and need a fairly high load around 47,000 ohms to work properly. They also are sensitive to the capacitive load they see which is comprised of the capacitance of the cables in the headshell and arm, the interconnects from the TT to the amp, and the input capacitance of the preamp input. If the load capacitance is too low, the cartridge develops a high frequency peak and a sag below the peak. if it's too high, the high end rolls off prematurely. Get it right, and you have flat, extended response. The manufacturer should tell you what the proper loading for his cartridge is. The trick is figuring out what the capacitance of your system is. If you have a meter that measures capacitance, you can measure the capacitance of the tone arm wiring and interconnect. The hard part is measuring the capacitance of the phono preamp's input. Hopefully the preamp's manufacturer will tell you; but if you have older gear, you may be out of luck. Older Shure and Pickering cartridges wanted 450 pF, later ones are around 250.
Back when I was "into" phono, I tried to use cartridges that needed a lot of capacitance, and then used low capacitance cables. Then I could install a box that let me switch in different amounts of capacitance and either listen or make frequency response measurements with a test record, and then add in cables that brought the capacitance up to the right value. Some cartridges specify a capacitance of 100 pF. That's hopeless, as most any tone arm wiring is higher than that and you still need interconnects and the preamp input to account for.
I also found that most early phono stages had significant errors (like 2-3 dB) in the phono EQ. Trust me, it takes good instrumentation, a box full of parts, and hours of time to correct the RIAA phono equalization. My Pioneer SX-990 was way off, and I spent the best part of 2 days getting it to within +/- 0.25 dB. Then I got the next generation SX-626 and it was dead on right out of the box. You can take it to the bank that any 60's phono section will be off by quite a bit. It might sound wonderful; but it's not to spec. If you like it, leave it - I suspect a lot of the errors are deliberate to make it sound "better".
Low output cartridges are another matter. The recommended loading is all over the map, and because they are lower output, they need a step up device. Some work best with transformers, some with a pre-preamp. This is where forums, dealers and friends are important, to help you find a good match. But all of this only brings the output voltage level up to the level of a high output cartridge. Then you still need to feed it into a regular phono input with RIAA equalization. Some outboard phono preamps include the regular phono preamp and RIAA EQ, and those connect to a line level input on the preamp.
OK, so now we have the cartridge properly mounted and aligned, anti-skating set, and the cartridge properly loaded and with the right capacitance. Ready? No, not quite.
Way back at the beginning we needed to select a cartridge whose mass and stylus compliance (the ease or difficulty of deflecting the stylus)match the mass of the tone arm. Put simply, high mass arms need a low compliance cartridge, and vice versa. The idea is to get the resonant frequency of the arm cartridge combination to be between 5 and 10 Hz. That keeps the arm from leaping over record warps, and keeps it out of the audible range. it also keeps it above the isolation frequency of the suspensions on Duals and AR's and a lot of others.
If the turntable has no suspension like a Rega Planar and others, it needs to have some kind of isolation. You can buy expensive isolators; but you can make one out of a concrete paving block big enough to hold the TT, and a small, fat innertube, such as used in wheelbarrows. Put just enough air in the innertube to raise the block and TT. If you bounce the TT it should have a resonance of 1-2 Hz. Cones and spikes are NOT isolators.
Putting the TT on a wall shelf, so it's not connected to the floor is also a good idea.
We're now up and running. So what's left? Well, turntable mats for one thing. This gets very personal, so I'll tell you what I think I know. A vinyl record presents a large flexible surface area to the sound in the room. It's a great source of feedback with the record acting as a microphone diaphragm. I think you really want what's on the record without the feedback; which adds a kind of artificial reverberation. The reverb does make the sound more lively and with more "ambience"; but if you want that, add an outboard reverb unit. Just don't tell me the mat is allowing the stylus to "extract mode information from the groove". Yet, that is a claim made by several TT mat makers. My favorite mat is a Sorbothane mat once made by AudioQuest. It supports the record over the whole surface and has a very quiet background and no artificial reverb effects. It also does a good job of damping the platter itself.
Any TT mat has to make allowance for the fact that the label area and the rim of an LP are thicker than the groove area. If they don't (as in the Marcoff Glass Mat) the LP sits up and is not supported or damped. If you listened carefully with that mat, it sounded very "lively"; but certain frequencies were amplified and had a ringing sound. Right - undamped resonances in the LP. So in the end, I come down on the side of the "dead" mats that damp resonances in the LP.
One more thing. Stylus type and tracking forces. OK, that's two more.
There are three basic stylus types: spherical, elliptical, and line contact. If you picture a V grove with a ball sitting in it, you have the spherical. It's pretty much insensitive to the orientation of the sphere, so the cartridge doesn't need to be perfectly square to the groove. It's the most forgiving of set up and also has the lowest contact pressure. The drawback is reduced frequency response on the inner grooves and higher harmonic distortion. Sphericals typically have a very smooth and musical sound and are not analytical sounding.
Elliptical styli were an attempt to more accurately track the inner grooves. The stylus has an oval cross-section and the narrow end of the oval contacts the groove. Sizes range from 0.4 x 0.7 mils, 0.3 x 0.9 mils, and 0.2 x 0.7 mills which are the side and front radii of the oval. Ellipticals are quite sensitive to alignment errors and you need to be sure the stylus is square to the groove in both planes. Narrower ovals can theoretically track finer grooves; but have much higher contact pressures. Years ago, I calculated the Hertzian stresses of several contact radii. A 0.2 mill ellipse at 1 gram tracking force will do significantly more damage to the groove than a 0.7 mil spherical at 3 grams. I also conducted a series of tests on identical pressings that verified this.
The third type is the line contact. Names and specifics vary (Shibata, Hyper-elliptical, Van den Hul, etc); but they all contact the groove along a line up the sides of the groove (a V fitting within a V) and spread the load over a larger area, and often an area not damaged by an elliptical. They have a very clean sound, open and clear; but are by far the most fussy about alignment and particularly anything that causes the stylus to be off of vertical when viewed head on. Remember the mirror trick?
If I had my way, I'd put the whole turntable into a closet with the door closed while playing. It's a lot more sensitive to acoustical feedback than you'd believe. Placing it between the speakers is a very bad idea. It needs to be in a low bass location, or well isolated from air borne excitation. Thin plastic covers keep the dust off; but are only moderately effective in keeping sound waves away.
Before we moved to the condo, I had almost 800 LP's; but I had started to convert to CD's. I'm now up to about 500 CD's. What I found in my basement rec room, which was sound treated, I found that a really well set up TT system with a Sorbothane TT mat and very good isolation both acoustically and from vibration, sounded very much like my CD's of the same material. In an A-B comparison, my 15 year old son was openly disappointed that the CD's sounded the same as the LP. In contrast, I was delighted as it meant I had got the phono system well set up.
Having said that, I have had a number of CD's which were genuinely awful. However, I blame the engineering of the recordings, rather than the medium, as I have a number of CD's that are really good.
Thus endeth my sermon on phono.
It's up to you guys to chime in now, and I'll collect the comments an publish a follow up.
Technology, Part V
Posted by Bold Eagle on October 19, 2010 at 19:28:23
This section is going to be a dog's breakfast
of odds and ends. I want to talk some about line conditioners and line noise,
FM and antennas (antennae?), planar speakers, cables, and tube amps. I may add
other stuff as I think of it.
In one of the early posts, I excluded line arrays, and tall electrostatic and planar magnetic speakers. The obvious question is why. The reason is simple: if the speaker is tall and radiates the same frequencies over its entire height, then the sound waves don't obey spherical spreading laws. Rather the sound waves spread as an expanding half cylinder ijn the forward direction. The result is that the sound drops off 3 DB (instead of 6) for each doubling of distance. This means that a woofer added to boost the bass spreads it's sound into the room with a different set of rules. Makes integration of a tall speaker of this type with a subwoofer much more difficult. Years ago the Infinity Reference Standard which was a line array, got around this by making the bass section a separate cabinet with 6 12" woofers in a vertical array.
Dipole speakers (most planar electrostatics and planar magnetics) radiate both backward and forward. This adds a sense of spaciousness; but you do have to worry about what's happening to the back wave. Usually the speakers are 3-5' out into the room to get them away from the front wall so the reflections from that wall are delayed in getting to you. They work best in big rooms, and need a reflective wall behind them They also have fewer problems with reflections from the side walls as they tend to be beamy with restricted dispersion, and also because they are dipoles, and the back wave is out of phase with the front wave, so there is a null at the sides.
One of the arguments for planar speakers is that they have lower mass and therefore, superior transient response. This is not true, regardless of what your ears tell you. Acoustic Research and Phillips both demonstrated that a well designed moving coil speaker has superior transient response, and the Phillips data was in an article in the Journal of the Acoustical Society of America. Measurements aside, the speakers sound "faster". If the room is big enough, tall planar speakers that are dipoles have a sense of ambience lacking in direct radiator types.
One more electrostatic thing and I'll move on. The Quad electrostatic speakers are shorter and the later ones, thanks to some electronic magic, don't act like line arrays. Rather, they emulate a point source located about 18" behind the plane of the diaphragm that has spread spherically through the diaphragm. They are still dipoles, ands still have an out of phase back wave and nulls at the sides, but they have very different directional properties and the forward sound does drop off 6 dB for each doubling of distance.
This is getting long, and since I got up at 0:dark:30 this AM to get the grandkids off to school; I'm running out of gas.
Trust me, I will be back with the rest of the topics and I might even think of some more.
Technology, Part VI
Posted by Bold Eagle on October 21, 2010 at 10:02:50
Unless someone comes up with a really
good topic, this will be the last in the series. So today, I'm going to address
tube amps and power conditioners and line noise.
Tube amps have regained a lot of popularity. In the early 70's you could buy tube amps for a song. If you step back and look at an amplifier as a black box, you put a signal in one end, and get a bigger replica of that signal out of the other end. Forgetting about what goes on inside the box; how do tube black boxes differ from solid state black boxes? At the input, the signal level required from the preamp is about the same for both, something around a volt for full output. However, the tube box usually has a very high input impedance, 100,000 ohms or so at the low end and going up to around 500,000 ohms. Makes it a really easy load to drive. The SS black box has input impedances ranging from 100,000 ohms and going down to as little as 7,000; requiring more current from the preamp and sometimes becoming a difficult load to drive. Further, if one looks at the input impedance versus frequency, the tube amps vary very little, while some SS amps have significant changes in impedance with frequency. The input impedance for the "laboratory" input for the Phase Linear 700 was 10,000 ohms +/-30%. With the wrong preamp, that will (and actually did) cause a significant shift in the sound.
At the output of our black boxes, they both have the same kind of power output and the same sort of voltage swings; but what differs is the impedance in series with the output (AKA, Source Impedance). Some manufacturers actually tell you what it is; but most give it to you in disguised form. They call it Damping Factor. Somebody in the distant past thought this up as a figure of merit. I really, really, really wish they never did. Calling it Damping Factor directs one's mind to damping of cone motions - I suppose with the connotation of better transient response. but actually, it's not that. In reality, it's a number calculated from the amplifier's series impedance at the output; and the series impedance has implications for shifts in frequency response over the whole frequency range.
Tube amps have series impedances starting around 0.4 ohms and going on up as high as 5 ohms. So if your nice new high tech tube amp has a series impedance of 2 ohms (like the Manley Stingray) you don't need to worry yourself about another tenth or two in the speaker cables. Remember the discussion in the first technical post about the 5% rule for the sum of the impedances of the cables and the amp? Well, that doesn't work at all well for tube amps, and for a lot of modern tube amps like the SETs and low negative feedback (NFB) designs, it doesn't work at all.
Okay, so what about SS amps? The McIntosh line of SS amps with the autoformer coupled outputs are about as high as SS amps get with source impedances of around 0.4 ohms (similar to their tube line)and ranging down to impedances of 0.01 for some early power amps like the Phase Linear 700. More typical numbers are 0.1 to 0.03 - a heck of a lot lower than the tube amps and far less likely to affect the frequency response of the speakers. Here the 5% rule works, and you should pay attention to the cable impedance since it's often larger than the amplifier impedance.
What else can we say about our two black boxes? For one thing, tube amps clip less sharply than SS amps; but that's only a factor if you drive the amp to overload. Tube amps typically have higher harmonic distortion, and in some of the SET and low NFB designs harmonic distortion reaches audible levels, adding a little warmth and "character"; but not sounding distorted.
Something you need to be aware of and watch for is the effect on source impedance caused by changes in the circuitry. Amps that have switchable tetrode/triode modes also will show a change in source impedance. As I mentioned, SETs have rather high source impedances. Amps that feature "low negative feedback" have high source impedances. What all this means is that the sound differences you hear from one of these amps may not be due to the changes in the circuit design; but, rather, due to changes in the source impedance. If you want to see the effects of this try putting a 4 ohm resistor in series with your speakers.
John Atkinson in Stereophile has been testing amplifiers for years and years with a simulated speaker load, and publishing the graphs of the response of the amplifier under such a load. Sadly, nobody seems to pay much attention. but in the test of one such amplifier, a SET, the response curve showed a 4 dB deviation from flat!!! That's about like turning the tone controls up about halfway.
Having said all that, speaker designers in the 60's knew all about the effects of tube amps, and so they adjusted the frequency response to get what they wanted with those amps. Great, so now I come along with my SS amp and those speakers sound terrible. Transistor sound, must be what those transistors are doing to the sound. Well, not exactly! What you are hearing are the compensations that the manufacturer of the speakers put into the design to compensate for the effects of the high source impedance of the tube amps. Compensations that you don't need for your SS amp. The solution? There are three possibilities: 1) get a tube amp of the same vintage, 2) put a 1 ohm resistor in series with the speaker to simulate the source impedance of a tube amp, and 3) use a graphic equalizer. This happens to be one of my "causes", so please excuse the time spent on it.
Time for conditioners and line noise. I have put an oscilloscope on my power lines (risky, and not for the amateur as you risk your scope)and saw no line noise. So, I don't use filters or conditioners; but I do use a good quality surge protector on all our electronics. However, if you live in an area with a lot of heavy industry, or a major metropolitan area, than there is a lot of noise on the power lines and it can get into your electronics. At which point a good filter and/or line conditioner makes sense. A good conditioner will also maintain constant voltage and many areas of the country will have fluctuating line voltages during periods of peak demand.
Another potential source of line noise could lie in your own gear. My Rotel CD player has a very elaborate line filter for the power cord. It's bi-directional so it prevents digital noise going in or out. My older players didn't have a line filter at all, relying on the power supplies to filter out the noise - so the CD player itself was a possible source of line noise.
I'm not a believer in exotic power cords. Seems pointless to have 6' of line shielded and high quality while the rest of the house wiring feeding the component isn't. I also don't believe in converting components to IEC connectors. It just adds one more set of mechanical connections in the line that you didn't have before. If your power cord is soldered to the transformer leads inside the amp, why would you want to replace that with slip fit connections?
A quick check on power cords is to play your system for a while and then feel the power cord going to the power amp. If it's warm to the touch, the wire gauge is marginal and you migh want to replace it with a heavier gauge. If it's hot, shut it off and then replace it with a much heavier cord. That's pretty rare; but it did happen years ago to me and I think it was due to corrosion of the wire inside the jacket.
I do believe in good quality wall outlets. You don't need to spend a fortune on some exotic audiophile grade; but a tight fit on the blades of the plug is important, and a hospital grade outlet is a good way to plus up the connection.
This is the end unless someone has a hot topic I know something about.
In closing let me leave you with this:
All the technology in the world won't help if it doesn't sound good. You are the final judge of that. It doesn't hurt to get some friends to help you listen and maybe point out something you missed (like E-Stat noticed the directional issue on my JBL's at the top of the midrange). A reviewer for a magazine may have golden ears; but he's in a different room with different tastes and different "associated equipment" and can't know your tastes. Pretty much it's up to you alone.
I started off with my philosophy for setting up a system. I do use instruments and calculations; but a lot of listening goes into the process, and if the measurements disagree I go with my ears.
While on the subject of ears; you do need to train yourself to listen to a system. It's not quite the same as listening to music for enjoyment - so don't get too caught up in the audiophile disease of listening for minutia while playing music for enjoyment.
If you're going to do a lot of listening while setting up a system you need material that suitable, right? The best sources are acoustic instruments, human voices, and complex sounds like audience applause and rain falling. I use about a dozen or so CDs for set up. Diana Krall's "Live in Paris" is good for piano and voice on band 11. I use several string trio and quartet CD's. I use the Swedish jazz recording "Jazz at the Pawnshop" for several kinds of instruments and applause sounds. I use a Telarc recording of "Peter and the Wolf" for its introduction where the solo instruments are introduced. Nice drums!
For male vocals I use a CD of the Blind Boys of Alabama, "Spirit of the Century" which has a range of voices with rich textures, a lot of instruments, and some good string bass lines. Recently I picked up a copy of a New Age CD with rain and a thunderstorm. Rain is hard to get sounding right.
If I can get all that to sound right; I know I've gotten pretty darn close. CD's and LPs and FM all vary a lot in quality from sample to sample, so you need a set of material that you know and trust.
Finally, remember that it's really all about matching of the pieces. There is no good correlation between sound quality and price. Want proof? Just go look at Stereophile's list of recommended components and look at the range of prices in any category.
So happy hunting, and I hope all these articles have been helpful to some of you. I built my first kit in 1953 and have been trying to figure out how this stuff really works ever since.