I've given this a bit of thought recently because of claims that people make about amplifiers 'sounding' different. One thing that comes to my mind is the effect of amplifier compensation. It seems that applying that at the voltage gain stage tends to use up the 'headroom' of the differential amp. I'm guessing that is why distortion rises at higher frequencies. It seems then since this lowers feedback at higher frequencies that the output impedance would necessarily rise too. Hence the 'different' sound of amplifiers. While I'm on the subject it seems to me that since bipolar transistors have large swings in current with minor changes in voltage, they would give a lower output impedance since they don't need the voltage feedback thru the loop like an FET. BTW, a most excellent and informative video. Signed, an audiophile from the 70's. Also, I'd welcome any feedback. :-) KC9WVR
I once worked on a guitar amp that presented with a strange distortion problem. The input to the output stage was clean, but the output signal had an unusual distortion component clearly visible on a scope. It turned out to be a cone resonance coming from the speaker at certain frequencies that was reflected back through the output transformer. It looked at first like the problem originated in the output stage, but it was generated by the speaker itself. The cone had softened with age. And the guitar amp had no NFB around the output transformer, so there was not much the amplifier could do to damp this speaker cone resonance. You could feel the speaker cone doing this with your hand like a ripple on the surface. A new speaker of the same type made the problem disappear. This is also one reason why there's a certain dynamic synergy between tube guitar amps and particular speakers.
Excellent explanation video of how the damping factor affects the loudspeaker behaviour and what its contribution is in the audio chain. Some really obvious facts are laid out that sometimes audiophiles and audio gear enthusiasts forget. It is good to get back to basics and make sure we got these covered with a proper grounding in the subject.
Superior damping is one of two reasons why eliminating the passive filter, and its typically substantial impedance, greatly improves signal tracking. The other is that the filter itself is a resonant circuit. Those are two a several compelling reasons to execute the filter with a digital signal processor ahead of amplification.
I have constructed bass modules with dual voice coil drivers. The second voice coil is used, with a shorting resister, to dampen the driver. Works quite well to clean up the response.
Excellent episode, thank you! As a side note though, moving mass increase (Mms) lowers the damping, not the other way around. Damping of fundamental resonance is the product of the motor strength related damping (Qe) and suspension related damipng (Qm). In both of corresponding equations, the moving mass is in the numerator.
I work in cinema and when wiring surround speakers, we try to avoid any series strings because it will crush damping factor, which results in a muddy/sloppy bass. If you wire two 8Ω speakers in series, each speaker has, effectively, an 8Ω resistor in series with it. You also tend to get a less efficient system as the power output of a solid-state amplifier into 16Ω is going to be less than into 4Ω (if you wired in parallel). Most modern commercial amplifiers will work down to 2Ω (typically with an increase in some THD). So, for 8Ω speakers, I'll wire 4-5 in parallel before introducing a series-parallel string (some amps, most notably, the QSC DCA 1222 will work down to 1.6Ω). For a 6 speaker string, I'll wire 3 in parallel, another group of 3 in parallel and then series the two groups. That only puts a 2.7Ω impedance in series with any speaker. Not ideal but for surround speakers, certainly workable and leaves you with 5.3Ω to the amplifier, which is in its sweet spot of power/THD. Naturally, one can add more amplifier channels to avoid the series but few exhibitors would want to spend the extra money. I might add...that is an interesting driver you have there. Front vented with a whizzer cone "tweeter."
An 8ohm nominal impedance woofer in a an enclosure will be a much more demanding load for a class AB amplifier than this. You can approximate actual power dissipation of output devices by multiplying impedance modulus at a chosen frequency by the cosine of the phase angle and then squaring the result. You are basically trying to source or sink large current with the output voltage approaching 0.
Tony, you would make a great Physics professor or an electrical electrical engineering professor teaching a couple classes of physics. this stuff is fun
Negative Feedback affects the Gain , Bandwith and Distortion of an Amp. More Feedback reduces the Gain and Distortion but increases the Bandwith. Less Feedbach increases the Gain and Distortion but reduces the Bandwidth. Given the small change in signal level when you adjusted the "Damping" Control tells me only a small change was made to to the Feedback . Makes me wonder if this a "Hi-Fi feature" 😄 However you didn't test to see if the "Damping" control affected the Freq Response of the Negative Feedbacb loop. If it does I'd think the "Damping" control was being used to modify the overal Tonal qualities of the Amp.
Awesome video. There was a LOT of information explained here!!! I understood a small percentage of this. But this small % of learned information was significant!!!! I understand some things much better!!! Now if I could understand all of this I’d be so happy.
Very good informative video. Thanks Tony. (Like the intro part very much too) Also remember the old trick of connecting two speaker drivers together then pushing in one cone and seeing the other react/move out?
Nicely done Tony. Very informative. I was hoping you'd show the effect of an 8 ohm resistor across that speaker. It would have demonstrated that almost any damping factor greater than 1 at the voice coil is adequate.
Thanks for another very informative video, Tony! It just happens that I ran into this damping factor effect reather by accident when I decided to put suppresion diodes on the output of an amplifier I am modifying and improving. I only added the diodes just because I have seen them used in some old amp designs and I wanted to see what happens but to my very pleasent surprize it made my speakers sound way better and I believe louder too. Before this any beat in the music at high volume would make the speaker pop out very much but now they keep the cone in check and I can increase the volume without much mechanical distorsions from the speakers at the volume I used to listen to and it seems that the bass and vocals are a lot more clean and clear. I am actually blown away how well the speakers perform now, even though they are some old RFT cabinets with cheap QTX woofers in them.
Kenwood "Sigma Drive": Basic M1 power amplifier (intent: reduce the problems associated with overly long speaker wires) Wiring at the back of the amplifier: 4 terminals: 2.5mm primary wire (+/-) from output to speaker... second 1.5mm (+/-) wire to "Sigma drive" sense terminals
@@mackfisher4487 Thank you. I repair welders and generators for a living and once in a great while I get one that is in showroom condition, but needs a bit of TLC. I wish everyone would take better care of all of their equipment. All the best to you and yours.
The true measurement of damping factor (internal resistance of the amplifier) in solid state amplifiers is done with the negative feedback connected to the voltage amplifier output and not to the current amplifier output (Speaker output). For DC stability of the amplifier one should add a DC servo amplifier to obtain the offset adjustment. In my experience the best results are obtained with amplifiers that are designed in this way.
In the circuit diagram, the damping factor switch has two sets of selectors. One set changes the feedback gain and the other set places resistors in the GND return path of the speakers. The resistors placed in series with GND return are 0 ohm (Normal), 0.02 ohm (Medium) the final one is the 0.02 ohm and 0.11 ohm (comprising of two 0.22 ohm, 2 watt) resistor combined in series (0.13 ohm) (Soft). The gain change is most likely used to correct the loss of amplitude by placing these resistors in series i.e. volt drop across GND return resistors. For 0 ohm series resistor, gain is 29.9dB, for 0.02 ohm series resistor gain is 31.5dB and for 0.13 ohm resistor the gain is 35.9dB. Not sure exactly how the switch it connected, but is also seems like the volt drops across the GND return resistors could be part of the feedback. If so, the more current passing through these GND return resistors give higher volt drops adding to the feedback signal and killing the forward drive since it is negative feedback.
@xraytonyb Tony, another great video and excellent explanation and demonstration. What's your take on damping headphones in solid state amps that use drop resistors off the output and where the resistors are about 10x the impedence of the headphones? Does feedback take care of that effectively too?
What would be a great natural follow up to this would be a comparison between an active and a passive system. The great mystery though would be why some might prefer a passive system (with the same exact drivers).
@@Stelios.Posantzis no one of my audiophile friends use actives They are often changing amps and cables and make endless comparisons With actives you just need a preamp and some XLRs and its done The game stops too early
A design came to my mind. Single stage amplifier with positive and negative rails. Adjust the no signal point such that the collector of the transistor is zero volts. This way you connect the speaker directly to the amplifier without a dc blocking cap.
That's why damping factor is not as important then output impedance and resistance of a speaker cable is These two have much more contribution to how a system sounds like
When you showed the coil I thought it was a Pork Pie, not sure if you have them where you live as the pastry is a little different it's called 'hot water crust pastry' and as you expect its made with hot water and fat, usually Lard but if your posh you can use beef dripping. There ! you taught me something and I hope you have learnt a bit too !
46:40 I have a couple of Adcom 545 mk II amplifiers that have the warm feedback mod done to them. What is it? It is a second resistor with a bypass capacitor parallel across it, installed in series with the stock feedback resistor. The gain increased and I am sure the damping factor dropped some, but the sound is more tube like and less fatiguing to listen to.
I enjoyed the video, and the topic, but, using your method, I think you were doing a dis-service to the general technique of measuring rated damping factors. From everything have seen, it is measured at an amp's max rated output as, the output transistors are biased on more than at the levels you are measuring at, and will have less of a Vce drop ( along with a lower internal C-E resistance) than at reduced power, so, the measurements will be different.
34:00 In the late 1980's Kenwood made amplifiers that had extra sensor wires called "Sigma Drive" that were small gauge that ran all the way from the amp to the speakers. Kenwood KA-9X was a 120 wpc integrated amplifier which used class G amplifier, split dual rail power supply, (30v/70v) and had a damping factor of 1000 at 50 Hz. They quit doing that because people would get the wires crossed up and damage the amp.
@@zulumax1 it's ike the kelvin sensing in a 4 wire multimeter setup. I don't think it's -ve fb in and of itself, but simply moves the speaker cables inside the existing - ve fb loop to compensate for any voltage drop on those cables. Of course, any non-linear effects in those cables would already have been tweaked out by swapping them end for end 😅
@@123xqp more of a sales gimmick. Probably a placebo effect fooling oneself more than anything you might hear. I will have to see if schematic is available on hifiengine
I notice you are using the green 10Mhz GPSDO and distribution amp. All of my test equipment have different 10Mhz reference input voltage requirements. So my question is, did you use any attenuators between the distribution amp and your individual pieces of test gear or just or just hook them up direct? I'm hesitant to hook mine up without attenuators for fear of blowing an input. Thanks, Dave
45:02 i am now curios if the minimum impedance selector does something similar (referring to the option on some amps to select 8, 6, 4 ohm speakers). i remember reading someplace that at some point manufacturers in an effort to get compliant w/ regulations (overheating, fire hazard, etc) implemented the 4ohm speaker switch and what is does is limit the power/current on the transistors so that amps run cooler (but also arguably weaker) when used w/ 4ohm loads. apologies if i am mixing up things, i am not educated at all in this field :)
Greetings, I'm looking to purchase a class d power amplifier for bass duty use and it's claims a damping factor of 5000 Approx 2.5kw @8ohms 4kw @4ohmms 6kw /channel @ 2ohms Is this 5k damping factor possible?
Enjoyed; great session! Lots to mull over. Will probably rerun a couple times. I understand why you say it's not an important "spec" to you. "Variation, the spice of life." Thanks, Professor! Oh, love the Sanwa.