In a previous thread Tubelab explained to me that the tube, OT and speaker form a system that is greater than the sum of it's parts. The frequency spectrum impedance is effected by the characteristics of the individual speaker and is felt by the OT and the tube.
Does this include the mechanical physics of the speaker's Fs? If I use a sealed, ported or weighted passive radiator to change the Fs, is this also felt by the OT and the tube? Impedance changes with the Fs, does it not?
If that is so then for a truly, as perfect as possible, HiFi design we would have to design the entire system from tube, to OT and the speakers as well as the enclosure used?
Does this include the mechanical physics of the speaker's Fs? If I use a sealed, ported or weighted passive radiator to change the Fs, is this also felt by the OT and the tube? Impedance changes with the Fs, does it not?
If that is so then for a truly, as perfect as possible, HiFi design we would have to design the entire system from tube, to OT and the speakers as well as the enclosure used?
There's a way to cheat, which is something KEF does in their Xover circuitry. They compensate the impedance variations an amplifier would normally encounter with additional networks that dissipate energy, but dont contribute to the sound output.
Their theory is that if an unknown amplifier sees a purely resistive load (or as close to one as you can practically make it, considering a volume production product) chances are it will behave better - and that will be reflected in the sound of their speaker. "Easier to drive".
Some folks will disagree with that approach, even though it seems logical.
Their theory is that if an unknown amplifier sees a purely resistive load (or as close to one as you can practically make it, considering a volume production product) chances are it will behave better - and that will be reflected in the sound of their speaker. "Easier to drive".
Some folks will disagree with that approach, even though it seems logical.
Last edited:
Why would the tube care if the speaker bass-bump is at 33Hz or 76Hz? The tube does not know one frequency from another (up to MHz).
The OT should be selected to not fight the speaker bass-bump.
Read, read, and re-read chapter 3, "The Relationship Between The Power Output Stage and the Loudspeaker" in Radiotron 3rd {16MB PDF}
What I am asking is does the mechanical characteristics of the speaker's Fs change the impedance felt by the OT over the audio spectrum, or felt by the tube? If so the cabinet would effect the system. What you are saying is no. I have to read the pdf you linked.
The mechanical effect on impedance must have some effect. A 4 ohm speaker has all kinds of different impedance over the spectrum. If the tube and OT doesn't care then why does it matter what impedance the speaker is at all? 4,8,16 so what? That is just an RMS calculation. A 4 ohm speaker has many times that impedance over the entire spectrum. The OT or the tube has to feel that?
Last edited:
Of course, if a 4 Ohm speaker has impedance of 30 Ohms at resonance, the power stage will see more than it's nominal impedance load. How much more, will depend on ability of OPT to reflect it's secondary load. If OPT has high enough inductance, the output stage will see 30/4=7.5 times it's nominal load. Real life OPTs won't be able to do that, so the power stage will see the load equal to 2pifL, where f is speaker's resonance frequency and L is inductance of OPT's primary.
Interaction between amplifier and speaker may be quite complex if the amplifier is a voltage source, like most amplifiers are. Speaker "commands" such an amplifier to certain extent, in other words, all speaker's distortion, resonances, box colorations, etc., are fed into the amplifier and contribute to amplifier' s output signal. Only a current source amplifier, for example pentode with no feedback, is not affected by what speaker is doing.
Interaction between amplifier and speaker may be quite complex if the amplifier is a voltage source, like most amplifiers are. Speaker "commands" such an amplifier to certain extent, in other words, all speaker's distortion, resonances, box colorations, etc., are fed into the amplifier and contribute to amplifier' s output signal. Only a current source amplifier, for example pentode with no feedback, is not affected by what speaker is doing.
Last edited:
Blaxshep,
Oh, do not forget the woofer impedance dip that is nearly equal to its DCR, happening at 20 Hz, and again at perhaps 200 or 300Hz.
And then the woofer impedance rises again, due to the woofer voice coil inductive reactance.
Then there is the typical 2 way crossover, with the resonant impedance peak, perhaps at 2kHz or 3 kHz, the tweeter's lower impedance (nearly equal to its DCR) at perhaps 5kHz to 7 kHz, and finally the tweeter's inductive reactance that rises from 10kHz all the way to 20kHz.
Just look at a magazine test review for a loudspeaker impedance curve.
To answer you question:
Whatever the loudspeaker impedance is at any given frequency, is "felt" by Both the output transformer, and Also by the output tube.
If that is not true, we either have an infinite loss transformer, or we are violating the conservation of energy theory (just a theory?).
"All Generalizations Have Exceptions"
Then just relax, and sit back, and enjoy listening to the music.
Oh, do not forget the woofer impedance dip that is nearly equal to its DCR, happening at 20 Hz, and again at perhaps 200 or 300Hz.
And then the woofer impedance rises again, due to the woofer voice coil inductive reactance.
Then there is the typical 2 way crossover, with the resonant impedance peak, perhaps at 2kHz or 3 kHz, the tweeter's lower impedance (nearly equal to its DCR) at perhaps 5kHz to 7 kHz, and finally the tweeter's inductive reactance that rises from 10kHz all the way to 20kHz.
Just look at a magazine test review for a loudspeaker impedance curve.
To answer you question:
Whatever the loudspeaker impedance is at any given frequency, is "felt" by Both the output transformer, and Also by the output tube.
If that is not true, we either have an infinite loss transformer, or we are violating the conservation of energy theory (just a theory?).
"All Generalizations Have Exceptions"
Then just relax, and sit back, and enjoy listening to the music.
Last edited:
Yep your bass bump in impedance is caused by the mechanical mass of the cone and drive and the elasticity of the suspension. If the cone moves it generates an emf. Try moving the cone manually with meter across the terminals.
Yes, exactly. A speaker's nominal impedance better needs to be interpreted as the minimal impedance that an amplifier has to cope with.
Best regards!
Best regards!
I always wonder how a speaker driver could arrange with a new cabinet.
It took several weeks, in some cases months to optimize for both to perform on the best level. Those people who change speaker designs after auditioning two days miss most of this process. And even thats not enough of the mystery, a speaker relates to the room in which it is playing because that, too gives a feedback to the speaker. Is it a fully damped room or one able to reflex energies from the walls back into the room?
In every new room, the speakers need some weeks to settle, to perform best. And then it happens, there is some click- moment when the sound snaps in. Everything works.
But that new level only happens when a level of quality is achieve that enables the speakers to let them shine, to show those qualities.
With most of the equipment outside that will not show in such a clear way, because they are far below that level of refinement and too many mistakes are being involved in the design of the speakers and the whole audio chain of electronic amplified music.
But when you experience that level and the tricks that could happen, its like magic.
ALL what is happening in the speaker is directly reflected via the transformer(s) into the amps means to the tube.
We are talking about very complex processes, which couldn't be described in detail by the school physics models of trivia.
It took several weeks, in some cases months to optimize for both to perform on the best level. Those people who change speaker designs after auditioning two days miss most of this process. And even thats not enough of the mystery, a speaker relates to the room in which it is playing because that, too gives a feedback to the speaker. Is it a fully damped room or one able to reflex energies from the walls back into the room?
In every new room, the speakers need some weeks to settle, to perform best. And then it happens, there is some click- moment when the sound snaps in. Everything works.
But that new level only happens when a level of quality is achieve that enables the speakers to let them shine, to show those qualities.
With most of the equipment outside that will not show in such a clear way, because they are far below that level of refinement and too many mistakes are being involved in the design of the speakers and the whole audio chain of electronic amplified music.
But when you experience that level and the tricks that could happen, its like magic.
ALL what is happening in the speaker is directly reflected via the transformer(s) into the amps means to the tube.
We are talking about very complex processes, which couldn't be described in detail by the school physics models of trivia.
Last edited:
Are you worried that your speaker will be resonating at the Fs 100% of the time as a song is playing? Where will those frequencies come from?
First off, My comments in the 100 year old amp thread were directed mostly to guitar amp applications since the question was asked about a nearly 70 year old Fender 5C1. In this case the amp usually is looking into a single speaker or multiple speakers of the same type. The signal source is one to six tones, their harmonics, the IMD products generated in the amplifier, and any low frequency dynamic (string movement) effects that make it through the amp. Things get more complex in a multi way HiFi speaker cabinet.
The Fs spec given in the speaker spec sheet is it's "free air" resonant frequency. It will change in surroundings other than free air. The amplitude of the impedance peak will change as well as its frequency. This peak is created when the push created by the voice coil coincides with the natural swing of the cone in its suspension. Change that with any damping or external air movement, as occurs inside a cabinet, and everything changes.
If the cabinet is a closed box with a port, and the box resonance is significantly different from the speakers resonance, then the entire system may have two impedance peaks, or one wider peak. This can be seen in Bass Box or WIN ISD simulations.
Most speaker testing and building assumes that the load presented to the amp is purely resistive. It's is agreed that the speaker's impedance changes with frequency, but that is still assumed to be resistive, IE my "8 ohm" guitar speaker is "20 ohms" at resonance. The load is always assumed to be resistive, meaning that the applied voltage, and the current drawn by the load are in phase. This is not the case, even when there are no reactive crossover components in the system.
Real speakers do have an inductance and capacitance associated with their voice coils. They also have physical properties associated with movement of the coil and cone, IE it is easier to move the cone at it's natural resonant frequency than at another frequency, and that movement is restricted by several things like the surround and spider.
The coil IS moving in a magnetic field, so it does generate a considerable electric current, which is fed back into the amplifier. This current is rarely in phase with the drive signal that the amp is feeding to the speaker, making the speaker's apparent impedance reactive, either as a capacitor, inductor, or usually both.
That EMF that get's fed back into the amp may get into the GNFB loop creating a disturbance, but that is already a controversial subject that I'm not going to touch. Most of my amps do not use GNFB.
Add some inductors and capacitors as found in the crossover networks of a modern multi-way HiFi speaker and you have a very complex load that varies from capacitive to inductive over the speaker's frequency range.
All speaker testing currently done assumes a continuous tone. Dynamics are rarely considered.
30 years ago I went to college at age 37 to get an engineering degree at my company's expense. The project that I designed for my "analog design project" was a multi channel car stereo amp with a big LF amp feeding a 15 inch sub in a large wood box. The power supply was a DIY switcher using an SG3525 chip, 4 large mosfets and a DIY step up transformer. All older surplus parts. It worked well, I graduated and left school....one day the amp smoked, yet kept on playing. Autopsy revealed blown power supply caps. I replaced them, but within a few months the same thing happened again, on EXACTLY the SAME PART in the SAME SONG, ZZ Top's Sharp Dressed Man, turned up to the edge of clipping.
The subwoofer power amp itself was a DIY clone of a SWTPC Universal Tiger, modified with bigger parts to be pretty much blow proof. It was a relatively old simple design that I had built dozens of in the 70's and 80's for home, car, and guitar amp duties. All of the individual channel boards were similar modified Plastic Tigers. I still had plenty of built boards, so I used them.
I set out to find out why this happened. What I discovered led me down a nearly year long rabbit hole that just kept getting deeper. After a year I had more questions than answers, and finally called it a draw and gave up.
Imagine this scenario. The bass guitar is pushing that heavy 15 inch speaker cone back and forth about half an inch. At the instant it's at maximum velocity moving in one direction a huge drum transient tries to quickly reverse it's direction. What is the instantaneous dynamic impedance of that woofer? I was convinced at the time that for a millisecond or two it could actually go negative, the speaker is stuffing current back into the amp faster than the amp can fight back. Now I'm not so sure, but it can definitely be lower than the DCR of the coil.
It can obviously be felt as far back as the power supply. In that case modern low ESR capacitors capable of handling the large current pulses demanded by my less than 4 ohm subwoofer fixed the "smoking amp" problem. They also fattened up the bass to where the plywood subwoofer box needed help, but that was another problem.
Even in this old solid state situation the speaker is a large part of the total equation. It always will be.
The Fs spec given in the speaker spec sheet is it's "free air" resonant frequency. It will change in surroundings other than free air. The amplitude of the impedance peak will change as well as its frequency. This peak is created when the push created by the voice coil coincides with the natural swing of the cone in its suspension. Change that with any damping or external air movement, as occurs inside a cabinet, and everything changes.
If the cabinet is a closed box with a port, and the box resonance is significantly different from the speakers resonance, then the entire system may have two impedance peaks, or one wider peak. This can be seen in Bass Box or WIN ISD simulations.
Most speaker testing and building assumes that the load presented to the amp is purely resistive. It's is agreed that the speaker's impedance changes with frequency, but that is still assumed to be resistive, IE my "8 ohm" guitar speaker is "20 ohms" at resonance. The load is always assumed to be resistive, meaning that the applied voltage, and the current drawn by the load are in phase. This is not the case, even when there are no reactive crossover components in the system.
Real speakers do have an inductance and capacitance associated with their voice coils. They also have physical properties associated with movement of the coil and cone, IE it is easier to move the cone at it's natural resonant frequency than at another frequency, and that movement is restricted by several things like the surround and spider.
The coil IS moving in a magnetic field, so it does generate a considerable electric current, which is fed back into the amplifier. This current is rarely in phase with the drive signal that the amp is feeding to the speaker, making the speaker's apparent impedance reactive, either as a capacitor, inductor, or usually both.
That EMF that get's fed back into the amp may get into the GNFB loop creating a disturbance, but that is already a controversial subject that I'm not going to touch. Most of my amps do not use GNFB.
Add some inductors and capacitors as found in the crossover networks of a modern multi-way HiFi speaker and you have a very complex load that varies from capacitive to inductive over the speaker's frequency range.
All speaker testing currently done assumes a continuous tone. Dynamics are rarely considered.
30 years ago I went to college at age 37 to get an engineering degree at my company's expense. The project that I designed for my "analog design project" was a multi channel car stereo amp with a big LF amp feeding a 15 inch sub in a large wood box. The power supply was a DIY switcher using an SG3525 chip, 4 large mosfets and a DIY step up transformer. All older surplus parts. It worked well, I graduated and left school....one day the amp smoked, yet kept on playing. Autopsy revealed blown power supply caps. I replaced them, but within a few months the same thing happened again, on EXACTLY the SAME PART in the SAME SONG, ZZ Top's Sharp Dressed Man, turned up to the edge of clipping.
The subwoofer power amp itself was a DIY clone of a SWTPC Universal Tiger, modified with bigger parts to be pretty much blow proof. It was a relatively old simple design that I had built dozens of in the 70's and 80's for home, car, and guitar amp duties. All of the individual channel boards were similar modified Plastic Tigers. I still had plenty of built boards, so I used them.
I set out to find out why this happened. What I discovered led me down a nearly year long rabbit hole that just kept getting deeper. After a year I had more questions than answers, and finally called it a draw and gave up.
Imagine this scenario. The bass guitar is pushing that heavy 15 inch speaker cone back and forth about half an inch. At the instant it's at maximum velocity moving in one direction a huge drum transient tries to quickly reverse it's direction. What is the instantaneous dynamic impedance of that woofer? I was convinced at the time that for a millisecond or two it could actually go negative, the speaker is stuffing current back into the amp faster than the amp can fight back. Now I'm not so sure, but it can definitely be lower than the DCR of the coil.
It can obviously be felt as far back as the power supply. In that case modern low ESR capacitors capable of handling the large current pulses demanded by my less than 4 ohm subwoofer fixed the "smoking amp" problem. They also fattened up the bass to where the plywood subwoofer box needed help, but that was another problem.
Even in this old solid state situation the speaker is a large part of the total equation. It always will be.
In the case of a guitar amp the resonance of a guitar speaker is often within the guitar's frequency range, and playing loud at or near resonance is a real thing. I have measured peak plate voltages of over 2 KV in a guitar amp driven to clipping whose B+ voltage was 420 volts. It is something to consider when laying out PC boards and wiring up tube sockets.
TL, you must have some terrible nightmares.... fears of being struck by lightning.... I'm glad I have simple needs in this hobby.
Have a look at winISD - it will gives you plots of impedances with different ported and unported designs.
The solution is very simple: Don't use PCB in tube amps. Its unecessary and it flattens the sound and makes it sound thin compared to good (I'm writing of knowledge in this field) executed free wiring with the parts mounted on boards and using turret lugs.
PCB was invented for mass market production with automated installation and soldering of the parts. And, of course, its way cheaper and faster to build a PCB based amp.
But it will always be second best. In terms of achievable sound.
I think a rough approximation may be a little like this:
Woofer in a closed box, from very low frequency to very high frequency:
DCR; Inductive, Resonant, Capacitive, DCR, Inductive. (not quite that simple, because there is always the DCR term there, and at or near Resonance there are other factors affecting the Q.
Woofer in an open baffle, from very low frequency to very high frequency:
DCR; Inductive, Resonant, Capacitive, DCR, Inductive. (not quite that simple, because there is always the DCR term there, and at or near Resonance there are other factors affecting the Q.
Woofer in a ported box, from very low frequency to very high frequency:
DCR; Inductive, Resonant, Capacitive, DCR, Inductive, Resonant, Capacitive, DCR, Inductive. (not quite that simple, because there is always the DCR term there, and at or near Resonance there are other factors affecting the Q.
Woofer in a closed box, from very low frequency to very high frequency:
DCR; Inductive, Resonant, Capacitive, DCR, Inductive. (not quite that simple, because there is always the DCR term there, and at or near Resonance there are other factors affecting the Q.
Woofer in an open baffle, from very low frequency to very high frequency:
DCR; Inductive, Resonant, Capacitive, DCR, Inductive. (not quite that simple, because there is always the DCR term there, and at or near Resonance there are other factors affecting the Q.
Woofer in a ported box, from very low frequency to very high frequency:
DCR; Inductive, Resonant, Capacitive, DCR, Inductive, Resonant, Capacitive, DCR, Inductive. (not quite that simple, because there is always the DCR term there, and at or near Resonance there are other factors affecting the Q.
Not only that, but also don't overlook the effects of humidity and whether or not the floors are hardwood w/wo rugs or carpet and where the speakers are located... corners, stands, walls, windows treatments, and if you are following the matching requirments of Jopple (or maybe it's Joppel).
I spent the first 62 years of my life in South Florida where severe lightning storms were a daily occurrence for the entire summer, and most of the rest of the year. I have been shocked by nearby lightning hits numerous times, but knocked unconscious only once. Been there, done that...post#10 here:
Shocking experience
I started building tube amps for guitar and HiFi applications at a very young age, and was selling then before I got to high school. All were PTP designs. I worked in consumer electronics repair starting at age 16, and took over the service department of a large electronics store at age 18. I have seen far too many fried amps, and began looking into why this happened.
I test every amp I build whether it is PC board based, or PTP with an electric guitar preamp cranked to 11 and the line voltage above rated max. Why, because someone somewhere WILL do something like that, and I want to find that failure mode, and fix it before they do.
41 years of designing "mission critical" radio equipment has taught me that. Note, "mission critical" equipment is deemed so, because someone's life may depend on it working correctly under less than optimum conditions. Most of my career was spent designing circuits, devices, and technologies for police, fire, and other public safety radio systems.
That may be your opinion, which is also shared with many in the field of tube amps. It is not mine. I also have knowledge in this field, and have designed PC boards for applications up to 2 GHz which contain RF, audio, DSP, and high speed digital circuitry and would not exist without a very well thought out and carefully designed and executed PCB, cell phones. After leaving the consumer electronics world in 1972, I became a design engineer at Motorola.
PC board was invented for several purposes and uniformity of design was the primary design driver. The early PCB's of the 1950's were all hand stuffed. Automated assembly came later. Granted the PCB's from the early days were pretty bad. The paper based materials used had all sorts of problems with heat and moisture, and fell apart within a few years. Today's FR4 materials are very good and with proper technuques can be used up to 4 or 5 GHz. A competent designer should not find any obstacles to PCB use at audio frequencies. Someone who can't create a clean hum free design with PTP, will have enen less success with a PCB since the designer is constrained to basically two dimensions....very little use of the "Z" dimension.
There is no valid reason that a well designed PCB can't sound just as good as a well designed PTP version of the same amplifier. The difference is that once the board is properly working, every one will sound exactly the same if the same parts are used in it's assembly. There are plenty of good and bad examples of both construction techniques.
I have been selling tube amp PCB's for 15 years. Plenty of the amps built 15 years ago are still around, and working fine. Even an inexperienced builder can make a hum free amp because the star grounding system is built into the board at design time.
#1, That explains it. #2, You're gonna break the knob off some day if you keep twisting it past the stop point.