I used to give seminars that included specs like Damping Factor as a Rep for various audio products way back in the '70's. (Marantz, Kenwood, Cerwin Vega, AR,...) A simple demo of what Damping Factor does is to take a speaker, cabinet and all, (higher efficient port loaded work better than sealed cabinets) by itself and tap on the woofer as you did. Then short out the speaker terminals, I'd often just pull out my car keys and short the back terminals, then tap again! The difference is extremely audible and effective to demo how well an actual DF helps control the drivers. This also proves that yes it is controlling the driver based on allowing the counter EMF directly back into the coil. Regardless of the coil's pure resistance. This removes any output stage circuitry, effects of active amplifier feedback loops, ... It is strictly a self contain system. If you were to drive your output stages to clipping you would find that there is in fact a voltage drop across those 10 ohm resistors. (or you could just measure it on the scope) Because the active feedback loop could not compensate by adding gain to the stage to increase the output to compensate. This is what is actually happening instead of a true low output stage resistance. This is why many extremely high DF amps sound so bad. An amp with no feedback with a DF of 10 is typically a better sounding amp than one with over 1,000 using massive feedback. A DF of about 100 is max for non-feedback amps. And yes the more resolving and demanding a speaker system is the more audible DF games become.
Another good one, John! For what it's worth, damping factor seemed to get more important when speaker manufacturers started applying Thiele/Small parameters to make woofers with small boxes and high-compliance drivers. That was roughly the time when large cone excursions became a status symbol of sorts. Traditional designs didn't need amplifier damping all that much. I don't know if damping factor affects the sound of speakers in linear operation, but in many cases it can help prevent the crack of a voice coil bottoming out, and the associated damage to the driver. Using a low cut filter at the right frequency can also help tame undamped cone excursions. While some think that a shaking speaker cone looks cool, those low frequencies can modulate higher frequencies and distort the sound.
Thank you John! That is perfectly clear. I never thought of negative feedback as something that improves damping factor, I thought that only the electrical component resistance was responsible for that.
You demonstrated that larger speakers need more dampening than smaller speaker when there was more ringing in the larger speaker with everything else equal. I have several amps from 150 to over 1200 dampening factor and the bass is by far the tightest and most controlled on my JC1 mono blocks (>1200 DF). In fact the bass control seems to be proportional to the dampening factor on my amps. However, the quality of the amps are also proportional to the DF, so that could be playing a bigger factor than the dampening factor. It makes sense that very large and powerful speakers are moving much more mass and would benefit more from higher dampening factor. For the kick drum to sound natural you can't have much ringing.
You should try shorter speaker wire. The shorter wire has a drastic improvement on the dampening factor. Also Thickest gauge possible! Maybe your amp that has 150DF can sound tight and controlled also.
Practical Guy I use short 12 AWG quality speaker wire (it is basically 0 ohms). My amp that has 150 DF sounds much more controlled (and better all around)than an mass market new receiver (even a $3000 Yamaha that claims 170W/ch which is a joke) I have heard. My amps are all high current and all mass market receiver are current limited. I am very susceptible to listening fatigue, so if an amp doesn't at least sound remotely natural, I can't listen for more than a couple songs before I have to turn it off. I know enough about electronics to fix most things and that is why I have way too many amps (pre-amps...) lol I bought a bunch of "fix or repair high(er) end amp for peanuts. I thought my amps sounded pretty good until I bought the Halo JC1 mono blocks (which have >1200 DF) and I was blown away when I heard the tight, natural and controlled bass (not to mention the much smoother mids and highs). For the first time I heard a natural sounding kick drum and I could distinctly hear every change in key, note, chord (I am not a musician so I don't know the exact term). Basically what I am saying is, that to notice the short comings of an amplifier, just listen to a better one. It will soon be apparent what you are missing. Bryston Amps also have very high DF and those amps are also renown to have excellent bass control. I certainly understand the gimmickry in marketing audio, but that only seems to be a problem in mass market amp (receivers...). Higher end companies tend to under rate their specs and leave considerable headroom (which I have verified in testing in certain things like clean output power before clipping). For instance I have several B&K Components and the receivers are rated at 150 watts/ch and the test at over 160 watts (no clipping) 2 channels driven into 8 ohms. Where as the Yammy receiver (170 watts/ch) at the store couldn't even properly drive Kef Q100s (86 db sensitivity) and especially the Kef LS 50s (85 db). All my amps sound good, but the JC1s sound amazing and they should because they cost a small fortune. Sorry for going off on a tangent (I tend to do that). My point is that I think high dampening factor is necessary with very powerful amps driving big speakers capable of high power (more mass and more momentum that needs to be stopped from ringing (oscillating). Even in John's test, the 1 DF stopped the small speaker, but wasn't enough to stop the larger speaker from ringing. I am sure there are other factors that contribute to tighter bass in higher end amps than Dampening factor, but there must be a good reason why they design more powerful amps with more dampening factor, because they don't seem to be fudging any numbers on the other specs. I could see the above mentioned Yammy claiming 1 million DF, because the power claims are ridiculously over rated at 170 watts/ch (lucky if it has 70 clean watts with 2 channels driven and it is still current limited and will have no dynamics, so the bass will suck). Not to mention the way they cheat with distortion spec as well, by taking them at 1 watt, rather than at rated power like higher end manufacturers do (or at least give both ratings).
Hey John, my observation with damping factor is that when you have a speaker with high mass and low mechanical and magnetic power then a higher damping factor is necessary, typically the speakers with a high frequency range and low efficiency i.e. the old Infinity, or Apogee, Dynaudio that type of speaker. But when you have a loud loudspeaker with low cone mass, relatively stiff suspension and lots of magnetic power then a lower damping factor will help. Speakers of that type are i.e. the JBL studio monitors like 4425, or Tannoy Westminster or Altec Voice of the Cinema these type of speakers. But as Michael Beeny already said, the higher the mass and the lower the frequency the more of an effect it has. A loudspeaker that deserves the name IMO always works best with an output transformer.
Thank u for that insight! Which amplifier would you recommend for the JBL XPL200 Stereo Speakers? My amp (Sansui AU-G90X) isnt able to control the big cones very well.. do u have a good recommendation for an amp or a damping factor that can handle those beasts more confidently? :-)
Very good presentation. Readers keep in mind, electrical is one part of damping, and some of the video points out, the box and mechanical parameter of woofer, determine final result.
1. The TDA2050 amp appears to be missing an output inductor (usually an R||L combo), which tends to contribute the bulk of commercial amplifiers' output impedance but is also important for maintaining stability under adverse conditions like capacitive loading due to long cables in addition to the customary Zobel network (RC to ground). Rolling your own actually is not too hard - 10-20 turns of AWG20-ish enamelled copper wire around a pencil or something (plus a 3.3-10 ohm, 2-3 W resistor) should do. 2. You can actually get fancy cable length vs. resistance calculators online, and tables of wire gauges tend to list resistances per length as well. So let's say you've got a twin lead of AWG16 wire at 13.6 Ohm/km (= mOhm/m), then that's 272 mOhms for a 10m run, or DF = 29.4 re: 8 ohms abs max. Pretty marginal. Here in 4 ohm land, longer runs like that tend to be done in 4 mm² (~AWG11), though often the ends will have to be thinned out to fit the speaker connectors - and that's fine. Nobody is going to care if 0.2% of the cable length is only half the diameter, that's an extra 0.6% of resistance. It's the bulk that matters. 3. Single drivers actually are not very picky about DF, loads of them will be perfectly fine at around 10. Crossovers can turn a speaker's impedance graph into a real rollercoaster though, necessitating values in excess of 20. In 8 ohm land that still is rather easily achieved with an amplifier DF of 60 (8 ohms) plus AWG12 wiring, in 4 ohm land that same amplifier would require some ridiculous wire gauge for long runs (like AWG6!). 4. Your headphone amplifier, as good as it may already sound at this point, still has some rough edges to iron out. a) If open-loop output impedance is 11 ohms, that's = 0.5 (r_e + R_e), with R_e = 10 ohms, making emitter follower output impedance r_e = 12 ohms each. Now that equals V_t / I_c, so your idle current comes out as about 2.16 mA. (Simulation with those same values and Bob Cordell's transistor models gives ~2.14 mA, so almost bang-on.) Your average BD139/140 pair (non-Philips) tends to appreciate a bit more than that as they're a tad slow. I'd say a minimum of 10 mA, 30 mA if you can afford it. Your average Class AB headphone booster tends not to use these transistors but rather some slightly exotic TO-92 jobs, with nice and constant beta outweighing their thermal disadvantage. (With nothing else at hand, one or two pairs of BC337/327 shouldn't be too bad.) b) There is significant asymmetry in transistor driving since the opamp cannot push current up through the upper bias diode, restricting current sourcing capaciting and generating even-order distortion. Place a capacitor of at least several 10 µF between the transistor bases to counteract this effect at mid-high frequencies. Alternatively, consider asymmetrical driving with both diodes on the lower side of the opamp output, in which case the upper bias resistor is also dispensable (yay, free opamp class A biasing!). This will cost you a diode drop's worth of output swing. It makes it easy to upgrade to hot class A operation with a 3rd diode (or a red low-leakage LED) though. c) For the sake of stability, employ local feedback from opamp output to inverting input (22-47 pF NP0, typ) and an output resistor for the opamp (10-100 ohms, typ) to counteract the inductive nature of opamp output impedance. You probably know how much emitter followers "love" inductive source impedance. You're probably getting away without since your gain is relatively high - at a more typical 8-14 dB and/or given a few nF worth of capacitive loading it may look different. d) Likewise, having at least a fraction of an ohm worth of output series resistance may be worth it, possible even something more advanced (full Zobel + output indoctor, though inductors tend to be tricky for headphone use as you generally have to resort to ferrite cores, and those can be a bit crummy). 5. Bonus fact: While high impedance (current) driving tends to mess up frequency response around resonance badly, it can drop distortion levels drastically (down into headphone territory), especially with high-end drivers employing a kapton or other nonconductive voice coil former (apparently the eddy currents in Al formers are not beneficial to voice coil linearity). You can design an amplifier that will transition from voltage drive at low frequencies to current drive at higher ones and so exploit the best of both worlds. I think there even is a small handful of high-end active speakers employing this technique..
There are quite a few errors or misstatements in your comments. Regarding 4 b), no, the design is not asymmetric with respect to driving the output transistors. The op amp can't sink current through the lower bias diode, either. *All* of the base drive current comes from the 4.7K resistors and both polarities of drive function precisely as complements. And placing a single capacitor between the bases will not help the op amp to drive more current to the transistors, unless the asymmetrical configuration you suggested is adopted first (i.e., those mods are mutually requisite, not alternatives). Placing a capacitor across *each* diode would permit the symmetrical configuration to be retained. In 4 a), quiescent Ic will be highly dependent on temperature difference between the diodes and the output transistors, which of course varies. Pegging the figure at 2.16mA +/- 0.02mA is pretentious. I have several other points as well regarding your comments, but then I also have better things to do. I will just caution the readers to be wary and not presume your statements are accurate.
@@marianneoelund2940 _Challenge accepted._ :) Loaded up a circuit like this in LTspice, plotted diode currents and voltages and... Well _I'll. Be. Dipped,_ as Derek @ Vice Grip Garage likes to say. (Funny guy, btw.) _Symmetry._ Distortion is dominant 3rd. _What the...?_ It's not like I haven't played with these circuits for a bit. **headscratch** (This one random circuit I pulled out combined a 1N914 model with a 1N4148 à la Bob Cordell, so that wouldn't help. I obviously corrected that first.) The good news is, bypass capacitors across the bias diodes still reduce distortion where they are effective, *and* you still get by with just one cap across both as one of the diodes is always conducting and low dynamic impedance. _Phew._ *Thank you* for this negative feedback - as Bruno Putzeys says, the more the merrier. ;) Usually this is _my_ specialty. Spouting nonsense is in fact not one of my favorite hobbies, and it still happens more often than I like. One time I derived a Friis formula equivalent for calculating the equivalent input noise of a string of voltage amplifiers and had it online for _almost two years_ before I noticed that it was missing a bunch of squares - I sure wished someone had told me earlier. (I guess nobody really reads Web 1.0 style homepages with loads of text any more.) Or the time I reckoned I needed a _current input_ to get a non-warped frequency response out of a headphone used as a microphone (actually, nope). At the end of the day I'm just some guy who knows a thing or two about a thing or two. I guess if I had the chops of Douglas Self, Bob Cordell or Bruno Putzeys I'd be _designing_ amplifiers for a living, not just writing about them on the interwebs as a hobby... (Honestly I'm glad I can do this next to my non-engineering day job. Doing things for a living can really take the fun out of them.)
@@PileOfEmptyTapes Glad to hear you're having fun with this hobby - and I have to commend you for the diversity of concepts you've embraced and are able to work with. I don't expect perfection from any individual. This is why engineering is always done by teams, and group-level design reviews followed by exhaustive testing are critical parts of the process.
Hi John, thanks for the interesting demonstration. You show 2 experiments with the large loudspeaker not connected to any amplifier. One with open contacts to show the effect of no electrical damping at all. Then you hook up an 8 ohm resistor and show the results. In this particular case the total impedance in the circuit (coil series resistance and load) is 8 + 8 = 16 ohms (damping factor 1) and the damping has improved substantially. Now, to my point: if you would short cut the loudspeaker (damping factor "infinitiy", 8/0) the total resistance in the circuit would be just 8 ohms (coil series resistance). The current in this case would be 2x as much as in the situation with the 8 ohms load. So the damping in this case would be 2x as high (double the current). This means that going from a damping factor of 1 to a damping factor of "infinity" will only improve the actual electrical damping by a factor of 2. This again shows that the concept of damping factor is pretty much nonsense. In an actual loudspeaker connected to an amp the total resistance in the circuit, and therefore the damping, is determined by the series coil resistance AND the impedance of the crossover filter which can be tenths of ohms in the region of the transition frequency. Speaker cable impedance and amplifier output resistance can be neglected as long as they stay below a few ohms. So, we don't disagree here at all, I just wanted to add some additional argumentation. Thanks, Robert
John - I think this is a great demonstration of damping factor and output Z determined by NFB. . Looks like a damping factor on that very stiff coned speaker turned out to be good critical damping at 1:1. Very interesting. I think the display on your first tapping of the boxed speaker is harmonic related and then covering the port gets you back to virtually no damping. I am guessing the second speaker is a "high compliance" speaker with a floppy cone from very soft suspension. I think the fact that SS amplifiers have so much NFB, we vacuum tube lovers, find them sounding compressed. I know I am digressing from your video. Anyway, I think your demonstration was very nice and made two great points.
I personally found that the damping factor generally affected small-diameter, low-efficiency loudspeakers. For large loudspeakers, with 38cm bass drivers for example, with high outputs, a low damping factor is not a fault, quite the contrary !
Of course the damping factor is only really important at low frequency's due to the large movement of the cone. The damping factor makes almost no effect at over probably 1khz. One reason why valve amplifiers tent to give the damping factor at 1Khz because at low frequency it will be very much lower. It's not only the cable resistance in series with the loudspeaker, in most high quality loudspeakers the first thing the signal sees is a choke as part of the crossover. Cheaper ones use thinner wire which gives higher resistance. Probably why an active loudspeaker tends to sound better. The amplifier goes straight to the drive units with very short cables and NO inductors or capacitors in series.
Howdy. Nice work. And yes. A loudspeaker element will display reactive power. The reactive current will flow backwards some part of the cycle and tends to raise the voltage at the amp. output. However. You are correct. The global negative feedback will sense this and restore the waveform fidelity. Some advokate the loudspeaker should be current driven (not voltage as is usually). That should compensate for the coil resistance allowing full benefit of a low output impedance. I don't really know for myself. I would imagine loudspeakers are designed for maximum sound pressure flatness to flat voltage. The current is let alone to vary at will. In my mind current driving would flip this setup resulting in an "anti" -flatness sound pressure response. Regards.
I think you got it right first time when you said that the negative feedback compensates for the voltage drop you would get from the potential divider network between the speaker impedance and the 10 ohm resistors. Maintaining the voltage between the 10 ohm resistors to compensate for the divider loss does not sound like reducing the output impedance,more like correcting the voltage drop.
Its' interesting to bring in classic spring damper theory to this and to connect the dots. The differential equation for a spring damper system is mx'' + cx' + kx =0; where m is mass and x''is the second derivative of displacement in the x -direction, i.e. acceleration, giving F = mx''. c is the damping factor which damps the sinusoidal motion as a function of velocity x', and k is the stiffness spring constant from Hooke's law (the stiffness of the suspension in this case). The amplifier signal provides the force to accelerate the cone, the stiffness of the cone (k) resists the motion, and the damping factor of the system (c) damps the harmonic motion of the system. However, c, is not just the damping factor Rl/Rs, the suspension of the speaker will also be self damping to some extent since it's not an ideal spring. Theoretically then, the higher the damping factor of the amp/speaker combination, the more closely the the speaker will track the waveform; this will be important at HIGHER frequencies for a woofer/midrange driver because the acceleration of the speaker cone is greater as the frequency increases, and cone overshoot at increasing frequency=distortion. At lower frequencies the displacement will be greater but the acceleration much lower.
Very intresting matter. Clear explanation. I do recall the Philips MFB (Motional Feed Back) speaker system. They had a piezo glued in the woofer that measured the actual movement of the speakercone. As a result, the speakercone itself was part of the feedback loop. They had their poweramplifier built-in and were pretty expensive. These systems were built arond the early 70's (now start guessing my age... ) Try googling on Philips MFB to find lots of information (most in Dutch) I'm not sure if the Philips MFB was marketed outside the Netherlands. Now we are googling... also google on speaker resonance frequency and q-factor. Question of the week: Why is the damping factor an issue only for low frequencies? In John's video we see the higher frequencies are not effected.
Nice video! It would be interesting to connect an amp with no input signal (or with a resistor shorting the input) to a speaker and do the pencil bounce test while watching the speaker terminals on the scope.
It's funny in a Cerwin Vega Facebook group people claim that you need high damping factor for better cone control of the 12 or 15 inch models. Perhaps those amps with high damping factor are just better than most low damping factor amplifiers, but I don't think it matters
There are speaker drivers with low system Q, they are over-damped and amplifiers with low damping factor make them sound a little bit warmer (or "less tight"), with more bass response due to bass notes not being damped quick enough
An interesting experiment and a great video. In the ported enclosure experiment, the complex waveform at 7:57(please see the waveform) is most likely due to the interference of the air issuing out of the port with the front side wave and its effect on the cone movement. That is why we see the rising portion of the second half cycle (of the first cycle) dipping towards the X-axis. Without the interference, the “complete” exponentially damped sinusoid waveform should be seen, which is confirmed by your port closure experiment subsequently. In ported enclosures, for frequencies below the resonance, the phase of the port output will reduce the cone radiation. My two cents worth. Please continue your interesting experiments.
The damping factor decrease due to the speaker cable can be audible, at least in a direct comparison of two separate setups. In principle you can solve this problem with a separate wire connected between the signal input clip of the speaker and the feedback input which is not connected connected to the amplifiers output. This third wire doesn't need to be with a large diameter, because the input resistance of the feedback loop is high compared to the wire resistance. I am also very surprized that amplifier manufacturers don't supply this option of a 3-wire-configuration. They only need to connect the feedback loop internally across a 10 Ohm resistor to the amplifier output to operate it also in 2-wire-configuration. You can furthermore operate the speaker also in 4-wire configuration providing a virtual ground for the feedback loop, which is also connected to the real ground across a 10 Ohm resistor in order to allow all the other configurations from 2 wires up to 4 wires. In the 4-wire configuration you can entirely eliminate the effect of the speaker cable.
That's why I always made sure my speaker wires were the same length. I didn't want to introduce too much of a variance between the two speaker channels.
You could make them excessively thick too, i mean AWG10, 3mOhm/m or so. If you fold the cables, you form an inductor. Not that either of both terribly matter, but still.
My amps DF tends to range from 200 to 500. The one with 500 sounds the best and has the tightest chest thumping bass. That could be just a coincidence. What I think is really going on is that the speaker cables are lowering the DF even further. This would depend on length and thickness of the cable. So if you want tight bass, keep your speaker cables short and thick. Nothing less than 12 g will do ...
Wow, a very interesting and well made video with clear explanation. You're good! :) I remember seeing some of your first videos and they alone convinced me to subscribe to your YT channel. It's a very uplifting thing to see you're getting better and better in what you do (and probably what you love to do). :) Keep up the good work and fun. :) Cheers!
Hey John, this video is some months old, but it's only been recently I discovered your videos in general. It's been a pleasure to watch them and your informal yet uncompromised speech. On this video, however, it became sort of unclear what is the actual way how damping factor is / should be measured, i.e. with global NFB applied around the output stage or not. The common sense tells it should be measured with the amplifier unmodified, as it's actually going to be used, but wouldn't that mean the numbers would be much higher than, say, the example you shown of the Kenwood? After all, any amplifier from that vintage already employed global negative feedback and therefore, would be expected to achieve much higher DF, to the brand's benefit so, why it was spec'd so "low"? I would love to see your comments. P..S. Snickers is lovely and a great touch in your videos. Regards from Brazil, Fabio
22:11 I'll chip in and take Benchmarks view on Damping Factor, which is that it should be as high as possible to negate the variable impedance of a speaker/crossover system. Benchmark's argument is that the output impedance and speaker impedance act as a voltage divider. Taking Z1 as amplifier impedance and Z2 as speaker impedance and assuming Vin is the voltage that the amplifier would produce without a load; Vout is the voltage at the terminals on the back of the power amplifier when the speaker (Z2) is connected Vout/Vin = Z2/(Z1 + Z2). By calculating 20log(Vout/Vin) at the max and min impedance of the speaker (which will have a nominal impedance of, say, 8Ω) the attenuation at these frequencies can be seen. The delta of these attenuations may be somewhere around 2dB for an amp with an output impedance of 0.8Ω and a speaker with a min impedance of 3Ω and a max of 18Ω; this will be audible. Factoring in speaker lead resistance, Benchmark's view is that a damping factor of ~ 170 will give a Δfrequency response variation of ~ 0.13dB which would be near inaudible.
I haven't seen an ss amp with an output impedance of .8Ω it is usually in the range of mΩ where amp damping factor is somewhat irrelevant bc as the video suggests, achieving a >20 df is totally simple. Now the catch is in speaker cables - the longer it is the more resistive it becomes BUT can the measurable difference of
One thing I've always wondered about is how does it factor into the linearity of the speaker? The back-EMF will make it behave like an electric motor in that the current draw will be substantially different when the voice coil and audio signal are out of phase with each other. I've been wanting to do distortion measurements with varying damping factors for a while now to see how much effect it really has on the speaker's performance.
I have a BSS EPC 780 with a damping factor over ,>1800.. into a sub 2 ohm load! This amp delivers 1500w@2ohms and 1000w@4ohms.. 8ohms is not listed as it is a.touring amp but it has been measured into a resistive load at around 650w. ...So a damping factor of 3600 at 4ohms.. and speculatively 7200 at 8ohms. Let me tell you now.. I have owned so much and upgraded my hifi equipment so many times since I bought this amp back in 2009. I have had MA silvers, B&W 604s, and crap Sony speakers that came with a midi system from the 90s. I also have an Echo interface as a DAC (Marantz PM6007 also) and a Formula Sound broadcast VCA driven mixer as a preamp. This amp has an impact and immediacy that is IMMEDIATELY apparent... even on the crap Sony speakers. There is just NO congestion in the sound at all and the sheer grip on the bass is just jawdropping. Even non audiophile friends and people who have come over to my house have commented on how good it sounds. No other amplifier I have ever heard, home, car or massive club sounds anything like this. So yes,.I believe the resistance to the inductance created my the moving electromagnet allows the amp to drive the capacitive load (one way) with ultimate precision. Seriously, buy one of these... Go on eBay. It will be the last amp you ever buy until the day you leave this planet. Dan.
@@pipesnnozzles7999 I actually did testing with various speakers and impedances some time ago. The only thing that the amplifier impedance will impact (and this is important with subs with 1-2 ohm impedances) is the frequency response. The impedance of the sub can pull the amps output around if its too high relative to the sub. Usually a DF of 100 or more will make this completely disappear. Distortion and nonlinearity were more or less unchanged as the motor nonlinearities become a factor once the DF of the amp is high enough to have an impact on the back-EMF. Yes it helps damp it, but the motor itself then causes additional distortions that offset any gain.
@@mysock351C yes, I know what you mean by this.. an example I'll give is driving a set a 250ohm Sennheiser 650s from a cheap Android phone. It's not that they don't go loud, they just sound crap. Really insipid bass and no punch or attack... Even at very low listening levels so as to be within the reserves of the power supply/capacitor bank reserved for transients. So yes I agree this is a real phenomenon in the case of a low output to load impedance ratio or even worse a negative one... Which was likely the case with that phone. All I will say is that I wish you would hear this amplifier as many many people, mostly non audiophiles who don't usually care about sound but importantly a friend who works in the live sound industry.. who got me the amp in the first place.. all at first demo were just mesmerized by the sheer grip and control and just effortless immediacy it has on seemingly any spealer system. You are aware that hifi speakers with passive crossovers pretty much nullify the majority of any damping factor that is present right? Hence why good hifi amps typically spec it at 100 upto 250 max for premium stuff. Things change when you have am unheard of at the time number of extremely high slew rate MOSFET output devices connected series-parallel like in a 10/90 odd ratio. 0ohm output impedances into resistive loads ie heating elements has marginal relevance when dealing with a complex reactive load like an amp/speaker combo... Particularly when passive filters are involved. I have connected this amp to several car audio sub drivers also... No passive crossovers.. a direct connection. Turn the amp on... No signal input.. and the cone goes noticeably stiffer... This is class A topology in action. I have never owned a tube amp but I imagine the same thing would happen. I could be wrong here, however if so, there is another phenomenon at play if it isn't damping factor per se. I don't doubt myself at my level of knowledge however I am smart enough to know that there is always more to things and not to reduce things down to the simplest "smart" explaination. What else could this be?
Okay. That made perfect sense and helped to further my understanding of feedback. Good Job! Now why do my speakers sound better, (to me at least), if I use 12 gauge zip cord (appliance wire) instead of the 18 gauge speaker wire provided with some of the cheaper speakers?
Night Stranger It may be that larger gauge speaker wire (thicker conduter) that you used has better impedance for your application. I usually use 10 feet of 16 gauge wire and 14 gauge if any longer than 15 feet. I always assumed that the less resistive wire at greater length would sound better, in my experience it has.
I guess the relatively low damping factor in guitar tube amplifiers (i.e., a high output impedance, with little control over the speaker) contributes a lot to the sound of them, especially when compared to solid state guitar amps that have very high damping factors and want to control the speaker tightly. Would I be correct in that assumption?
I appreciate your videos so much! I'm getting into audio electronics just for the love of both music and circuits. When you go to UA-cam to develop or relearn and interest, you want accurate information via quality instruction, and most important, prefer not to watch an audition for amateur comedy open Mic or take a tour of anyone's house or meet a cat named mosfet. It's a cool name but I could watch a video of a cat doing the hokey pokey if I want, and mosfet did nothing cool. He peed on your keyboard. So thanks for your videos.
Great video, though you packed it pretty densely in post, I had to do a double and even triple take in some sections, though I was multitasking...could have been a two-parter. Still disagree with one of your last points, but will have to whip out the math or maybe a way to demonstrate.
I'm guessing the only way to make the output impedance measurable would be to 'overload' the output with a 1-ohm load. Set the output to about a volt - don't want to draw a lot of current!
With it's ported and in the Box, you tap it, then the odd wavelength is... It vibrates twice before the air exits the Box, then it vibrates twice as the speaker slowly sucks air back in Box. Kind of like a piston..
I wonder though, common Class D amplifiers can't likely have a damping factor this good? I mean they seem to lack global feedback, because it was patented for Class D by someone in Holland.
Siana Gearz Hi, well the XTZ Edge a2-300 is class D with a damping factor >1500 @ 8 Ohm 20hz-500hz for just 495 Euro. Triple loop feedback and 150 watt @ 8 Ohm. It sounds amazing to be honest and is High-End for a silly price. Cheers.
I was just given an old Rotel amp and in the specs it listed the damping factor at 1000. They tripled up the power transistors on each rail, and they extended the negative feedback sense lead right to the speaker terminals. No doubt it is capable of some mean current and has low output impedance. Does it sound any better for it? I dunno.
hi john when you talk about 1.29 amperes in your calculations i suppose that its the current flow from fonction generator but id like to know how much current you inject in your amp from generator with no load and so there is also current lost by insertion in the amp but no matter what in the calculation.
Does this mean current model solid-state McIntosh amplifiers such as the MC452 and the MC462, both with output transformers and a damping factor of “>40 Wideband”, might have difficulties driving and controlling modern speakers?
Robert Langford, sir, this is not the place to be taught or to teach pronunciation of the English language, my friend. Actually, if you study History, different dialects of this language you will find many differences from the language you learned (or did not learn) in school. This lesson also stands for other subjects as well! Don't be so quick to correct people who may be more educated than yourself. HUMILITY!
A few quick questions... Does a higher dampening factor result in a reduced T.H.D.? In general terms? Also, longer speaker wire lengths, can they reduce a speakers dampening factor?
I've had this Yamaha CP2000 since 2004... Damping = 200 ...450W RMS X 450W RMS@8Ohm > 650WX650W RMS@4Ohm >2000W RMS @ 4ohm BRIDGED< 1%THD 20Hz-50KHz +|- .5 DB...#) what, WHAT...? Did you say something ? Tinnitus y' know... Hope talking about it doesn't let the magic smoke out ........ THANX AGAIN >John A_T !
High damping factor will do what to the audio signal? be a clean sine waveform. A low damping factor will do what to the audio signal be a distorted waveform?
My take on 'Damping' is more about keeping the speaker's acceleration and deceleration seriously clamped to the changing waveform. Fuzzy logic being amplifier power gets the speaker moving. Damping stops over shoot.
strange that someone first demonstrated how speaker is non linear active load and than test amplifier with pure resistant linear load. test in this video is totally wrong for a subject that intend to demonstrate . Conclusions are wrong completely.
dude DUUDE! How about getting a dual voice coil (Coil HA, like my name) Uhum... Dual Voice Coil speaker and revisit "SPEAKER DAMPING FACTOR" again. One driver coil to move the speaker and the other to monitor it's movements or (Better) ATTACH VARYING RESISTANCE ACROSS IT TO DIAL IN DAMPING !?! >> Fakk .... A Year Ago ? No one will ever see this. SO , getting a good scope soon ...Maybe I'll make my first video... THANX JOHN A_T
I've had this Yamaha CP2000 since 2004... Damping = 200 ...450W RMS X 450W RMS@8Ohm > 650WX650W RMS@4Ohm >2000W RMS @ 4ohm BRIDGED< 1%THD 20Hz-50KHz +|- .5 DB...#) what, WHAT...? Did you say something ? Tinnitus y' know... Hope talking about it doesn't let out the magic smoke........ THANX AGAIN, John A_T
This make sense if all your going to do is tap your speaker with a pencil all day, but hooked up to a subwoofer with high back EMF is real world and those back EMF WILL enter the feedback circuit more or less.
A difficult area very well illustrated and presented!
I used to give seminars that included specs like Damping Factor as a Rep for various audio products way back in the '70's. (Marantz, Kenwood, Cerwin Vega, AR,...) A simple demo of what Damping Factor does is to take a speaker, cabinet and all, (higher efficient port loaded work better than sealed cabinets) by itself and tap on the woofer as you did. Then short out the speaker terminals, I'd often just pull out my car keys and short the back terminals, then tap again! The difference is extremely audible and effective to demo how well an actual DF helps control the drivers. This also proves that yes it is controlling the driver based on allowing the counter EMF directly back into the coil. Regardless of the coil's pure resistance. This removes any output stage circuitry, effects of active amplifier feedback loops, ... It is strictly a self contain system.
If you were to drive your output stages to clipping you would find that there is in fact a voltage drop across those 10 ohm resistors. (or you could just measure it on the scope) Because the active feedback loop could not compensate by adding gain to the stage to increase the output to compensate. This is what is actually happening instead of a true low output stage resistance. This is why many extremely high DF amps sound so bad.
An amp with no feedback with a DF of 10 is typically a better sounding amp than one with over 1,000 using massive feedback. A DF of about 100 is max for non-feedback amps.
And yes the more resolving and demanding a speaker system is the more audible DF games become.
Another good one, John! For what it's worth, damping factor seemed to get more important when speaker manufacturers started applying Thiele/Small parameters to make woofers with small boxes and high-compliance drivers. That was roughly the time when large cone excursions became a status symbol of sorts. Traditional designs didn't need amplifier damping all that much. I don't know if damping factor affects the sound of speakers in linear operation, but in many cases it can help prevent the crack of a voice coil bottoming out, and the associated damage to the driver. Using a low cut filter at the right frequency can also help tame undamped cone excursions. While some think that a shaking speaker cone looks cool, those low frequencies can modulate higher frequencies and distort the sound.
Thank you John! That is perfectly clear. I never thought of negative feedback as something that improves damping factor, I thought that only the electrical component resistance was responsible for that.
You demonstrated that larger speakers need more dampening than smaller speaker when there was more ringing in the larger speaker with everything else equal. I have several amps from 150 to over 1200 dampening factor and the bass is by far the tightest and most controlled on my JC1 mono blocks (>1200 DF). In fact the bass control seems to be proportional to the dampening factor on my amps. However, the quality of the amps are also proportional to the DF, so that could be playing a bigger factor than the dampening factor. It makes sense that very large and powerful speakers are moving much more mass and would benefit more from higher dampening factor. For the kick drum to sound natural you can't have much ringing.
You should try shorter speaker wire. The shorter wire has a drastic improvement on the dampening factor. Also Thickest gauge possible! Maybe your amp that has 150DF can sound tight and controlled also.
Practical Guy I use short 12 AWG quality speaker wire (it is basically 0 ohms). My amp that has 150 DF sounds much more controlled (and better all around)than an mass market new receiver (even a $3000 Yamaha that claims 170W/ch which is a joke) I have heard. My amps are all high current and all mass market receiver are current limited. I am very susceptible to listening fatigue, so if an amp doesn't at least sound remotely natural, I can't listen for more than a couple songs before I have to turn it off. I know enough about electronics to fix most things and that is why I have way too many amps (pre-amps...) lol I bought a bunch of "fix or repair high(er) end amp for peanuts. I thought my amps sounded pretty good until I bought the Halo JC1 mono blocks (which have >1200 DF) and I was blown away when I heard the tight, natural and controlled bass (not to mention the much smoother mids and highs). For the first time I heard a natural sounding kick drum and I could distinctly hear every change in key, note, chord (I am not a musician so I don't know the exact term).
Basically what I am saying is, that to notice the short comings of an amplifier, just listen to a better one. It will soon be apparent what you are missing. Bryston Amps also have very high DF and those amps are also renown to have excellent bass control. I certainly understand the gimmickry in marketing audio, but that only seems to be a problem in mass market amp (receivers...). Higher end companies tend to under rate their specs and leave considerable headroom (which I have verified in testing in certain things like clean output power before clipping). For instance I have several B&K Components and the receivers are rated at 150 watts/ch and the test at over 160 watts (no clipping) 2 channels driven into 8 ohms. Where as the Yammy receiver (170 watts/ch) at the store couldn't even properly drive Kef Q100s (86 db sensitivity) and especially the Kef LS 50s (85 db).
All my amps sound good, but the JC1s sound amazing and they should because they cost a small fortune. Sorry for going off on a tangent (I tend to do that). My point is that I think high dampening factor is necessary with very powerful amps driving big speakers capable of high power (more mass and more momentum that needs to be stopped from ringing (oscillating). Even in John's test, the 1 DF stopped the small speaker, but wasn't enough to stop the larger speaker from ringing. I am sure there are other factors that contribute to tighter bass in higher end amps than Dampening factor, but there must be a good reason why they design more powerful amps with more dampening factor, because they don't seem to be fudging any numbers on the other specs. I could see the above mentioned Yammy claiming 1 million DF, because the power claims are ridiculously over rated at 170 watts/ch (lucky if it has 70 clean watts with 2 channels driven and it is still current limited and will have no dynamics, so the bass will suck). Not to mention the way they cheat with distortion spec as well, by taking them at 1 watt, rather than at rated power like higher end manufacturers do (or at least give both ratings).
@@gordthor5351 thanks for sharing your knowledge👍
Thank you very much indeed! Your video is a brilliant example of how to explain technical stuff in a way which is simple and straightforward.
Hey John, my observation with damping factor is that when you have a speaker with high mass and low mechanical and magnetic power then a higher damping factor is necessary, typically the speakers with a high frequency range and low efficiency i.e. the old Infinity, or Apogee, Dynaudio that type of speaker. But when you have a loud loudspeaker with low cone mass, relatively stiff suspension and lots of magnetic power then a lower damping factor will help. Speakers of that type are i.e. the JBL studio monitors like 4425, or Tannoy Westminster or Altec Voice of the Cinema these type of speakers. But as Michael Beeny already said, the higher the mass and the lower the frequency the more of an effect it has. A loudspeaker that deserves the name IMO always works best with an output transformer.
Thank u for that insight! Which amplifier would you recommend for the JBL XPL200 Stereo Speakers? My amp (Sansui AU-G90X) isnt able to control the big cones very well.. do u have a good recommendation for an amp or a damping factor that can handle those beasts more confidently? :-)
Very good presentation. Readers keep in mind, electrical is one part of damping, and some of the video points out, the box and mechanical parameter of woofer, determine final result.
1. The TDA2050 amp appears to be missing an output inductor (usually an R||L combo), which tends to contribute the bulk of commercial amplifiers' output impedance but is also important for maintaining stability under adverse conditions like capacitive loading due to long cables in addition to the customary Zobel network (RC to ground). Rolling your own actually is not too hard - 10-20 turns of AWG20-ish enamelled copper wire around a pencil or something (plus a 3.3-10 ohm, 2-3 W resistor) should do.
2. You can actually get fancy cable length vs. resistance calculators online, and tables of wire gauges tend to list resistances per length as well. So let's say you've got a twin lead of AWG16 wire at 13.6 Ohm/km (= mOhm/m), then that's 272 mOhms for a 10m run, or DF = 29.4 re: 8 ohms abs max. Pretty marginal. Here in 4 ohm land, longer runs like that tend to be done in 4 mm² (~AWG11), though often the ends will have to be thinned out to fit the speaker connectors - and that's fine. Nobody is going to care if 0.2% of the cable length is only half the diameter, that's an extra 0.6% of resistance. It's the bulk that matters.
3. Single drivers actually are not very picky about DF, loads of them will be perfectly fine at around 10. Crossovers can turn a speaker's impedance graph into a real rollercoaster though, necessitating values in excess of 20. In 8 ohm land that still is rather easily achieved with an amplifier DF of 60 (8 ohms) plus AWG12 wiring, in 4 ohm land that same amplifier would require some ridiculous wire gauge for long runs (like AWG6!).
4. Your headphone amplifier, as good as it may already sound at this point, still has some rough edges to iron out.
a) If open-loop output impedance is 11 ohms, that's = 0.5 (r_e + R_e), with R_e = 10 ohms, making emitter follower output impedance r_e = 12 ohms each. Now that equals V_t / I_c, so your idle current comes out as about 2.16 mA. (Simulation with those same values and Bob Cordell's transistor models gives ~2.14 mA, so almost bang-on.) Your average BD139/140 pair (non-Philips) tends to appreciate a bit more than that as they're a tad slow. I'd say a minimum of 10 mA, 30 mA if you can afford it. Your average Class AB headphone booster tends not to use these transistors but rather some slightly exotic TO-92 jobs, with nice and constant beta outweighing their thermal disadvantage. (With nothing else at hand, one or two pairs of BC337/327 shouldn't be too bad.)
b) There is significant asymmetry in transistor driving since the opamp cannot push current up through the upper bias diode, restricting current sourcing capaciting and generating even-order distortion. Place a capacitor of at least several 10 µF between the transistor bases to counteract this effect at mid-high frequencies. Alternatively, consider asymmetrical driving with both diodes on the lower side of the opamp output, in which case the upper bias resistor is also dispensable (yay, free opamp class A biasing!). This will cost you a diode drop's worth of output swing. It makes it easy to upgrade to hot class A operation with a 3rd diode (or a red low-leakage LED) though.
c) For the sake of stability, employ local feedback from opamp output to inverting input (22-47 pF NP0, typ) and an output resistor for the opamp (10-100 ohms, typ) to counteract the inductive nature of opamp output impedance. You probably know how much emitter followers "love" inductive source impedance. You're probably getting away without since your gain is relatively high - at a more typical 8-14 dB and/or given a few nF worth of capacitive loading it may look different.
d) Likewise, having at least a fraction of an ohm worth of output series resistance may be worth it, possible even something more advanced (full Zobel + output indoctor, though inductors tend to be tricky for headphone use as you generally have to resort to ferrite cores, and those can be a bit crummy).
5. Bonus fact: While high impedance (current) driving tends to mess up frequency response around resonance badly, it can drop distortion levels drastically (down into headphone territory), especially with high-end drivers employing a kapton or other nonconductive voice coil former (apparently the eddy currents in Al formers are not beneficial to voice coil linearity). You can design an amplifier that will transition from voltage drive at low frequencies to current drive at higher ones and so exploit the best of both worlds. I think there even is a small handful of high-end active speakers employing this technique..
There are quite a few errors or misstatements in your comments.
Regarding 4 b), no, the design is not asymmetric with respect to driving the output transistors. The op amp can't sink current through the lower bias diode, either. *All* of the base drive current comes from the 4.7K resistors and both polarities of drive function precisely as complements. And placing a single capacitor between the bases will not help the op amp to drive more current to the transistors, unless the asymmetrical configuration you suggested is adopted first (i.e., those mods are mutually requisite, not alternatives). Placing a capacitor across *each* diode would permit the symmetrical configuration to be retained.
In 4 a), quiescent Ic will be highly dependent on temperature difference between the diodes and the output transistors, which of course varies. Pegging the figure at 2.16mA +/- 0.02mA is pretentious.
I have several other points as well regarding your comments, but then I also have better things to do. I will just caution the readers to be wary and not presume your statements are accurate.
@@marianneoelund2940 _Challenge accepted._ :) Loaded up a circuit like this in LTspice, plotted diode currents and voltages and...
Well _I'll. Be. Dipped,_ as Derek @ Vice Grip Garage likes to say. (Funny guy, btw.)
_Symmetry._ Distortion is dominant 3rd. _What the...?_ It's not like I haven't played with these circuits for a bit. **headscratch** (This one random circuit I pulled out combined a 1N914 model with a 1N4148 à la Bob Cordell, so that wouldn't help. I obviously corrected that first.)
The good news is, bypass capacitors across the bias diodes still reduce distortion where they are effective, *and* you still get by with just one cap across both as one of the diodes is always conducting and low dynamic impedance. _Phew._
*Thank you* for this negative feedback - as Bruno Putzeys says, the more the merrier. ;) Usually this is _my_ specialty.
Spouting nonsense is in fact not one of my favorite hobbies, and it still happens more often than I like. One time I derived a Friis formula equivalent for calculating the equivalent input noise of a string of voltage amplifiers and had it online for _almost two years_ before I noticed that it was missing a bunch of squares - I sure wished someone had told me earlier. (I guess nobody really reads Web 1.0 style homepages with loads of text any more.) Or the time I reckoned I needed a _current input_ to get a non-warped frequency response out of a headphone used as a microphone (actually, nope).
At the end of the day I'm just some guy who knows a thing or two about a thing or two. I guess if I had the chops of Douglas Self, Bob Cordell or Bruno Putzeys I'd be _designing_ amplifiers for a living, not just writing about them on the interwebs as a hobby... (Honestly I'm glad I can do this next to my non-engineering day job. Doing things for a living can really take the fun out of them.)
@@PileOfEmptyTapes
Glad to hear you're having fun with this hobby - and I have to commend you for the diversity of concepts you've embraced and are able to work with.
I don't expect perfection from any individual. This is why engineering is always done by teams, and group-level design reviews followed by exhaustive testing are critical parts of the process.
Excellent video! I was surprised by the effect of the port! One thing I would like to have seen was an 80Ω load.
Hi John, thanks for the interesting demonstration. You show 2 experiments with the large loudspeaker not connected to any amplifier. One with open contacts to show the effect of no electrical damping at all. Then you hook up an 8 ohm resistor and show the results. In this particular case the total impedance in the circuit (coil series resistance and load) is 8 + 8 = 16 ohms (damping factor 1) and the damping has improved substantially.
Now, to my point: if you would short cut the loudspeaker (damping factor "infinitiy", 8/0) the total resistance in the circuit would be just 8 ohms (coil series resistance). The current in this case would be 2x as much as in the situation with the 8 ohms load. So the damping in this case would be 2x as high (double the current).
This means that going from a damping factor of 1 to a damping factor of "infinity" will only improve the actual electrical damping by a factor of 2.
This again shows that the concept of damping factor is pretty much nonsense. In an actual loudspeaker connected to an amp the total resistance in the circuit, and therefore the damping, is determined by the series coil resistance AND the impedance of the crossover filter which can be tenths of ohms in the region of the transition frequency. Speaker cable impedance and amplifier output resistance can be neglected as long as they stay below a few ohms.
So, we don't disagree here at all, I just wanted to add some additional argumentation.
Thanks, Robert
Not in the case of there being no passive crossover.
What a great explanation. Best video I every saw describing what damping factor is.? Thx
Another great video John. Thanks and keep up the great work! We nubs need them!!
Great and simple explanation, as usual
John - I think this is a great demonstration of damping factor and output Z determined by NFB. . Looks like a damping factor on that very stiff coned speaker turned out to be good critical damping at 1:1. Very interesting. I think the display on your first tapping of the boxed speaker is harmonic related and then covering the port gets you back to virtually no damping. I am guessing the second speaker is a "high compliance" speaker with a floppy cone from very soft suspension. I think the fact that SS amplifiers have so much NFB, we vacuum tube lovers, find them sounding compressed. I know I am digressing from your video. Anyway, I think your demonstration was very nice and made two great points.
It is nice to have other channel creators watching and commenting. Thanks!
I personally found that the damping factor generally affected small-diameter, low-efficiency loudspeakers. For large loudspeakers, with 38cm bass drivers for example, with high outputs, a low damping factor is not a fault, quite the contrary !
Of course the damping factor is only really important at low frequency's due to the large movement of the cone. The damping factor makes almost no effect at over probably 1khz. One reason why valve amplifiers tent to give the damping factor at 1Khz because at low frequency it will be very much lower.
It's not only the cable resistance in series with the loudspeaker, in most high quality loudspeakers the first thing the signal sees is a choke as part of the crossover. Cheaper ones use thinner wire which gives higher resistance. Probably why an active loudspeaker tends to sound better. The amplifier goes straight to the drive units with very short cables and NO inductors or capacitors in series.
Michael Beeny , True, true. And they DO sound better! Not just to your ears, either... but the measurement equipment, also.
Howdy. Nice work.
And yes. A loudspeaker element will display reactive power. The reactive current will flow backwards some part of the cycle and tends to raise the voltage at the amp. output.
However. You are correct. The global negative feedback will sense this and restore the waveform fidelity.
Some advokate the loudspeaker should be current driven (not voltage as is usually). That should compensate for the coil resistance allowing full benefit of a low output impedance.
I don't really know for myself. I would imagine loudspeakers are designed for maximum sound pressure flatness to flat voltage. The current is let alone to vary at will.
In my mind current driving would flip this setup resulting in an "anti" -flatness sound pressure response.
Regards.
I think you got it right first time when you said that the negative feedback compensates for the voltage drop you would get from the potential divider network between the speaker impedance and the 10 ohm resistors.
Maintaining the voltage between the 10 ohm resistors to compensate for the divider loss does not sound like reducing the output impedance,more like correcting the voltage drop.
Its' interesting to bring in classic spring damper theory to this and to connect the dots. The differential equation for a spring damper system is mx'' + cx' + kx =0; where m is mass and x''is the second derivative of displacement in the x -direction, i.e. acceleration, giving F = mx''. c is the damping factor which damps the sinusoidal motion as a function of velocity x', and k is the stiffness spring constant from Hooke's law (the stiffness of the suspension in this case). The amplifier signal provides the force to accelerate the cone, the stiffness of the cone (k) resists the motion, and the damping factor of the system (c) damps the harmonic motion of the system. However, c, is not just the damping factor Rl/Rs, the suspension of the speaker will also be self damping to some extent since it's not an ideal spring. Theoretically then, the higher the damping factor of the amp/speaker combination, the more closely the the speaker will track the waveform; this will be important at HIGHER frequencies for a woofer/midrange driver because the acceleration of the speaker cone is greater as the frequency increases, and cone overshoot at increasing frequency=distortion. At lower frequencies the displacement will be greater but the acceleration much lower.
Very intresting matter. Clear explanation.
I do recall the Philips MFB (Motional Feed Back) speaker system. They had a piezo glued in the woofer that measured the actual movement of the speakercone. As a result, the speakercone itself was part of the feedback loop. They had their poweramplifier built-in and were pretty expensive. These systems were built arond the early 70's (now start guessing my age... )
Try googling on Philips MFB to find lots of information (most in Dutch)
I'm not sure if the Philips MFB was marketed outside the Netherlands.
Now we are googling... also google on speaker resonance frequency and q-factor.
Question of the week: Why is the damping factor an issue only for low frequencies? In John's video we see the higher frequencies are not effected.
I remember powered Advent speakers around 1975.
Nice video! It would be interesting to connect an amp with no input signal (or with a resistor shorting the input) to a speaker and do the pencil bounce test while watching the speaker terminals on the scope.
Thanks John, really enjoy your videos.
Would be interesting to measure the damping factor at the resonant frequency of the speaker - this is where it should matter the most.
The crossover frequencies are more important, this is where the they tend to veer away from their nominal impedance se my reply above.
Thank you for all the hard work you put into this video, " great job "
Very informative, you should take up teaching, you have a gift.
It's funny in a Cerwin Vega Facebook group people claim that you need high damping factor for better cone control of the 12 or 15 inch models. Perhaps those amps with high damping factor are just better than most low damping factor amplifiers, but I don't think it matters
There are speaker drivers with low system Q, they are over-damped and amplifiers with low damping factor make them sound a little bit warmer (or "less tight"), with more bass response due to bass notes not being damped quick enough
An interesting experiment and a great video. In the ported enclosure experiment, the complex waveform at 7:57(please see the waveform) is most likely due to the interference of the air issuing out of the port with the front side wave and its effect on the cone movement. That is why we see the rising portion of the second half cycle (of the first cycle) dipping towards the X-axis. Without the interference, the “complete” exponentially damped sinusoid waveform should be seen, which is confirmed by your port closure experiment subsequently. In ported enclosures, for frequencies below the resonance, the phase of the port output will reduce the cone radiation. My two cents worth. Please continue your interesting experiments.
The damping factor decrease due to the speaker cable can be audible, at least in a direct comparison of two separate setups. In principle you can solve this problem with a separate wire connected between the signal input clip of the speaker and the feedback input which is not connected connected to the amplifiers output. This third wire doesn't need to be with a large diameter, because the input resistance of the feedback loop is high compared to the wire resistance. I am also very surprized that amplifier manufacturers don't supply this option of a 3-wire-configuration. They only need to connect the feedback loop internally across a 10 Ohm resistor to the amplifier output to operate it also in 2-wire-configuration.
You can furthermore operate the speaker also in 4-wire configuration providing a virtual ground for the feedback loop, which is also connected to the real ground across a 10 Ohm resistor in order to allow all the other configurations from 2 wires up to 4 wires. In the 4-wire configuration you can entirely eliminate the effect of the speaker cable.
That's why I always made sure my speaker wires were the same length. I didn't want to introduce too much of a variance between the two speaker channels.
You could make them excessively thick too, i mean AWG10, 3mOhm/m or so. If you fold the cables, you form an inductor. Not that either of both terribly matter, but still.
Ty for this video-there are noobs here that appreciate your explainings :)
Thanks for the tutorial Sir...really learnt a Lot!!!
Interesting, I’ve been enjoying 1970’s amps with damping factors of 20-50.
My amps DF tends to range from 200 to 500. The one with 500 sounds the best and has the tightest chest thumping bass. That could be just a coincidence. What I think is really going on is that the speaker cables are lowering the DF even further. This would depend on length and thickness of the cable.
So if you want tight bass, keep your speaker cables short and thick. Nothing less than 12 g will do ...
Wow, a very interesting and well made video with clear explanation. You're good! :) I remember seeing some of your first videos and they alone convinced me to subscribe to your YT channel. It's a very uplifting thing to see you're getting better and better in what you do (and probably what you love to do). :) Keep up the good work and fun. :) Cheers!
CJWarlock , and showing that junk level test gear works to do meaningful tests DOES help! LOL
Hey John, this video is some months old, but it's only been recently I discovered your videos in general.
It's been a pleasure to watch them and your informal yet uncompromised speech.
On this video, however, it became sort of unclear what is the actual way how damping factor is / should be measured, i.e. with global NFB applied around the output stage or not.
The common sense tells it should be measured with the amplifier unmodified, as it's actually going to be used, but wouldn't that mean the numbers would be much higher than, say, the example you shown of the Kenwood?
After all, any amplifier from that vintage already employed global negative feedback and therefore, would be expected to achieve much higher DF, to the brand's benefit so, why it was spec'd so "low"?
I would love to see your comments.
P..S. Snickers is lovely and a great touch in your videos.
Regards from Brazil,
Fabio
Enjoyable and educational, thanks!
22:11 I'll chip in and take Benchmarks view on Damping Factor, which is that it should be as high as possible to negate the variable impedance of a speaker/crossover system. Benchmark's argument is that the output impedance and speaker impedance act as a voltage divider. Taking Z1 as amplifier impedance and Z2 as speaker impedance and assuming Vin is the voltage that the amplifier would produce without a load; Vout is the voltage at the terminals on the back of the power amplifier when the speaker (Z2) is connected Vout/Vin = Z2/(Z1 + Z2). By calculating 20log(Vout/Vin) at the max and min impedance of the speaker (which will have a nominal impedance of, say, 8Ω) the attenuation at these frequencies can be seen. The delta of these attenuations may be somewhere around 2dB for an amp with an output impedance of 0.8Ω and a speaker with a min impedance of 3Ω and a max of 18Ω; this will be audible. Factoring in speaker lead resistance, Benchmark's view is that a damping factor of ~ 170 will give a Δfrequency response variation of ~ 0.13dB which would be near inaudible.
I haven't seen an ss amp with an output impedance of .8Ω it is usually in the range of mΩ where amp damping factor is somewhat irrelevant bc as the video suggests, achieving a >20 df is totally simple. Now the catch is in speaker cables - the longer it is the more resistive it becomes BUT can the measurable difference of
At 10:30 why didn't you add the resistor and again show how a damping factor of one stops that resonance? But great video.
So is damping like the shock/struts on a car analogy, to prevent the car from bouncing up/down more than once after you've hit a speed bump/pothole?
One thing I've always wondered about is how does it factor into the linearity of the speaker? The back-EMF will make it behave like an electric motor in that the current draw will be substantially different when the voice coil and audio signal are out of phase with each other. I've been wanting to do distortion measurements with varying damping factors for a while now to see how much effect it really has on the speaker's performance.
I have a BSS EPC 780 with a damping factor over ,>1800.. into a sub 2 ohm load! This amp delivers 1500w@2ohms and 1000w@4ohms.. 8ohms is not listed as it is a.touring amp but it has been measured into a resistive load at around 650w.
...So a damping factor of 3600 at 4ohms.. and speculatively 7200 at 8ohms.
Let me tell you now.. I have owned so much and upgraded my hifi equipment so many times since I bought this amp back in 2009.
I have had MA silvers, B&W 604s, and crap Sony speakers that came with a midi system from the 90s. I also have an Echo interface as a DAC (Marantz PM6007 also) and a Formula Sound broadcast VCA driven mixer as a preamp.
This amp has an impact and immediacy that is IMMEDIATELY apparent... even on the crap Sony speakers. There is just NO congestion in the sound at all and the sheer grip on the bass is just jawdropping.
Even non audiophile friends and people who have come over to my house have commented on how good it sounds.
No other amplifier I have ever heard, home, car or massive club sounds anything like this.
So yes,.I believe the resistance to the inductance created my the moving electromagnet allows the amp to drive the capacitive load (one way) with ultimate precision.
Seriously, buy one of these... Go on eBay. It will be the last amp you ever buy until the day you leave this planet.
Dan.
@@pipesnnozzles7999 I actually did testing with various speakers and impedances some time ago. The only thing that the amplifier impedance will impact (and this is important with subs with 1-2 ohm impedances) is the frequency response. The impedance of the sub can pull the amps output around if its too high relative to the sub. Usually a DF of 100 or more will make this completely disappear. Distortion and nonlinearity were more or less unchanged as the motor nonlinearities become a factor once the DF of the amp is high enough to have an impact on the back-EMF. Yes it helps damp it, but the motor itself then causes additional distortions that offset any gain.
@@mysock351C yes, I know what you mean by this.. an example I'll give is driving a set a 250ohm Sennheiser 650s from a cheap Android phone. It's not that they don't go loud, they just sound crap. Really insipid bass and no punch or attack... Even at very low listening levels so as to be within the reserves of the power supply/capacitor bank reserved for transients.
So yes I agree this is a real phenomenon in the case of a low output to load impedance ratio or even worse a negative one... Which was likely the case with that phone.
All I will say is that I wish you would hear this amplifier as many many people, mostly non audiophiles who don't usually care about sound but importantly a friend who works in the live sound industry.. who got me the amp in the first place.. all at first demo were just mesmerized by the sheer grip and control and just effortless immediacy it has on seemingly any spealer system.
You are aware that hifi speakers with passive crossovers pretty much nullify the majority of any damping factor that is present right? Hence why good hifi amps typically spec it at 100 upto 250 max for premium stuff.
Things change when you have am unheard of at the time number of extremely high slew rate MOSFET output devices connected series-parallel like in a 10/90 odd ratio. 0ohm output impedances into resistive loads ie heating elements has marginal relevance when dealing with a complex reactive load like an amp/speaker combo... Particularly when passive filters are involved.
I have connected this amp to several car audio sub drivers also... No passive crossovers.. a direct connection. Turn the amp on... No signal input.. and the cone goes noticeably stiffer... This is class A topology in action. I have never owned a tube amp but I imagine the same thing would happen.
I could be wrong here, however if so, there is another phenomenon at play if it isn't damping factor per se. I don't doubt myself at my level of knowledge however I am smart enough to know that there is always more to things and not to reduce things down to the simplest "smart" explaination.
What else could this be?
Okay. That made perfect sense and helped to further my understanding of feedback. Good Job! Now why do my speakers sound better, (to me at least), if I use 12 gauge zip cord (appliance wire) instead of the 18 gauge speaker wire provided with some of the cheaper speakers?
Night Stranger It may be that larger gauge speaker wire (thicker conduter) that you used has better impedance for your application. I usually use 10 feet of 16 gauge wire and 14 gauge if any longer than 15 feet. I always assumed that the less resistive wire at greater length would sound better, in my experience it has.
Thanks for demonstration.
Thank you very much for explaining damping sir.
I got 4000 damping factor in my hegel H190, very good controll
Good video, however you should draw with a red pen the changing in connections, not just tell them by speaking
Thanks a lot
Very informative
I guess the relatively low damping factor in guitar tube amplifiers (i.e., a high output impedance, with little control over the speaker) contributes a lot to the sound of them, especially when compared to solid state guitar amps that have very high damping factors and want to control the speaker tightly.
Would I be correct in that assumption?
very well explained
I appreciate your videos so much! I'm getting into audio electronics just for the love of both music and circuits. When you go to UA-cam to develop or relearn and interest, you want accurate information via quality instruction, and most important, prefer not to watch an audition for amateur comedy open Mic or take a tour of anyone's house or meet a cat named mosfet. It's a cool name but I could watch a video of a cat doing the hokey pokey if I want, and mosfet did nothing cool. He peed on your keyboard. So thanks for your videos.
Great video, though you packed it pretty densely in post, I had to do a double and even triple take in some sections, though I was multitasking...could have been a two-parter. Still disagree with one of your last points, but will have to whip out the math or maybe a way to demonstrate.
I'm guessing the only way to make the output impedance measurable would be to 'overload' the output with a 1-ohm load. Set the output to about a volt - don't want to draw a lot of current!
Great topic and explanation. Interconnect wire can do the same for the preamp as the speaker wire does for the speaker - right? Always educational!
Kiss Analog , Not really, as the preamp has a HUGE input impedance compared to the speaker.
With it's ported and in the Box, you tap it, then the odd wavelength is... It vibrates twice before the air exits the Box, then it vibrates twice as the speaker slowly sucks air back in Box. Kind of like a piston..
excelente
I wonder though, common Class D amplifiers can't likely have a damping factor this good? I mean they seem to lack global feedback, because it was patented for Class D by someone in Holland.
Siana Gearz Hi, well the XTZ Edge a2-300 is class D with a damping factor >1500 @ 8 Ohm 20hz-500hz for just 495 Euro. Triple loop feedback and 150 watt @ 8 Ohm. It sounds amazing to be honest and is High-End for a silly price. Cheers.
I was just given an old Rotel amp and in the specs it listed the damping factor at 1000. They tripled up the power transistors on each rail, and they extended the negative feedback sense lead right to the speaker terminals. No doubt it is capable of some mean current and has low output impedance. Does it sound any better for it? I dunno.
What's your opinion on wiring subwoofers to the amplifier at a lower impedance so as to mitigate against impedance rise power reduction?
Great video! Is there a way to manipulate or change the damping factor by changing parts inside the amp. (Specifically Crown Com-Tech 810 )
Thanks
I don't know the design of the Crown amp, but DF is largely a function of feedback and you probably shouldn't change that.
Yeah, it would probably be easier and better to just buy a different amp with a DF that I like more better 🤣
Thanks John
hi john when you talk about 1.29 amperes in your calculations i suppose that its the current flow from fonction generator but id like to know how much current you inject in your amp from generator with no load and so there is also current lost by insertion in the amp but no matter what in the calculation.
But how does this oscillation/back EMF work for electrostatic or magnetostatic speakers?
Just the same as with a voicecoil?
Could you have checked the voltage drop across an emitter resistor to figure impedance?
Does this mean current model solid-state McIntosh amplifiers such as the MC452 and the MC462, both with output transformers and a damping factor of “>40 Wideband”, might have difficulties driving and controlling modern speakers?
They should have no problem with modern speakers.
Can you try the pencil tap with the low impedance amplifier?
Robert Langford, sir, this is not the place to be taught or to teach pronunciation of the English language, my friend. Actually, if you study History, different dialects of this language you will find many differences from the language you learned (or did not learn) in school. This lesson also stands for other subjects as well! Don't be so quick to correct people who may be more educated than yourself. HUMILITY!
didn't you just do what you are recommending not to do :) ha ha good comment though
A few quick questions... Does a higher dampening factor result in a reduced T.H.D.? In general terms? Also, longer speaker wire lengths, can they reduce a speakers dampening factor?
Esoteric Audiophile , Use the words "meaningfully reduced" here instead of reduced. That is "the rub."
thanks
Sorry, John.
What is causing the damping factor in amplifiers, is it the power amplifiers output impedance or what?
It is the ratio of load impedance to the amplifier's output impedance. Feedback is what can make it very high.
I've had this Yamaha CP2000 since 2004... Damping = 200 ...450W RMS X 450W RMS@8Ohm > 650WX650W RMS@4Ohm >2000W RMS @ 4ohm BRIDGED< 1%THD 20Hz-50KHz +|- .5 DB...#) what, WHAT...? Did you say something ? Tinnitus y' know... Hope talking about it doesn't let the magic smoke out ........ THANX AGAIN >John A_T !
High damping factor will do what to the audio signal? be a clean sine waveform. A low damping factor will do what to the audio signal be a distorted waveform?
My take on 'Damping' is more about keeping the speaker's acceleration and deceleration seriously clamped to the changing waveform. Fuzzy logic being amplifier power gets the speaker moving. Damping stops over shoot.
And, of course, this is absolutely pointless if the room you're listening in sounds horrible.
How about not absolutely? There's still direct sound, and speaker design and placement can maximise it and minimise the room effect.
Bogdan Serban , Putting some onion in your soup WILL change the taste... but there is no test like a taste test!
some receiver Technics have 50 damping factor
strange that someone first demonstrated how speaker is non linear active load and than test amplifier with pure resistant linear load. test in this video is totally wrong for a subject that intend to demonstrate . Conclusions are wrong completely.
Crown Reference studio amp 20000damping factor
So...
speaker cables ARE real!
dude DUUDE! How about getting a dual voice coil (Coil HA, like my name) Uhum... Dual Voice Coil speaker and revisit "SPEAKER DAMPING FACTOR" again. One driver coil to move the speaker and the other to monitor it's movements or (Better) ATTACH VARYING RESISTANCE ACROSS IT TO DIAL IN DAMPING !?!
>> Fakk .... A Year Ago ? No one will ever see this. SO , getting a good scope soon ...Maybe I'll make my first video... THANX JOHN A_T
My pioneer vsa 910 has 100 damping factor and 0.005% thd
I've had this Yamaha CP2000 since 2004... Damping = 200 ...450W RMS X 450W RMS@8Ohm > 650WX650W RMS@4Ohm >2000W RMS @ 4ohm BRIDGED< 1%THD 20Hz-50KHz +|- .5 DB...#) what, WHAT...? Did you say something ? Tinnitus y' know... Hope talking about it doesn't let out the magic smoke........ THANX AGAIN, John A_T
Nice explanation, but please, there's no t in across.
Robert Langford - None of us came here for an English lesson.
There's a "Ross" in every crowd listening intently for the chance to pounce; it inflates their little ego.
This make sense if all your going to do is tap your speaker with a pencil all day, but hooked up to a subwoofer with high back EMF is real world and those back EMF WILL enter the feedback circuit more or less.
The whole point here is that the amplifier's low impedance kills the back EMF.
First ma dude