Bradley Labs Guitar Amplifiers

Builder’s Notes:

Article 1:  Iron Horse Design Notes (
Click Here

Article 2:  Musings on Speakers updated April 2010 (Click Here)

Article 3:  Musings on Guitar Amplifier Output Power 2012 (Click Here)

Iron Horse Design Notes 

  The input  is controlled by a Hi Lo switch. The Lo position is loaded with 1 Meg Ohm. The Hi position reduces this impedance to 140K Ohm and reduces the input to the pre amp by 6 dB, ( 1/2 )

Volume: A single control that allows both “Clean” amplification as well as “Crunch and Sustain” In operation the Volume control operates as a Pre Amp Gain control, used on any single channel amplifier up to about 80% CW. Going beyond 80% on to full CW continues to increase preamp gain, but now starts to reduce power amp gain, or Master Volume. The combination of the Guitar Volume control, which is an extension of the Pre Amp Gain control, the LO/HI input selection, the LoZ/HiZ Drive selection and the front panel Volume control with the Quench mode allows any level of sound and/or style one could accomplish with separate Gain and Master Volume potentiometers. See graph below for a graphical representation. 

Pre and Post Shelving Filters:   These selections control the frequency dependant  local feedback on the preamp Triode and the Post Stack Triode.  “Low” on both selections match the settings on most Guitar amplifiers built to this day. The “Mid” and “Hi” selections are new to the Guitarist. The lower value capacitors used in the Mid and Hi positions allow AC Filament Voltage to Cathode full wave rectification which causes Hum. Typically DC is used for the Filament Voltage however this is cumbersome and expensive. A DC bias placed on the Filament voltage reverse biases the rectification process and eliminates the hum. This is most likely the reason this selection was never allowed. Heavy bypassing of the Cathode resistor eliminates this phenomenon. Hence the hardwired “Lo” selection on all amplifiers. Drive: This selects either Hi or Lo impedance drive to the Tone Stack. HiZ is the standard drive impedance of typical Guitar Amplifiers. LoZ utilizes the LoMu section of the 12DW7 dual Triode as a Cathode Follower. This gives the Tone Stack a solid low impedance source to work from allowing less interaction of the Treble Middle and Bass controls with each other. Additionally the lower output impedance is not affected by the loading of the Tone stack which allows 6 dB ( X2 ) more signal to pass through. Post Coupling:  The Post coupling selection is a complex RC and CR network on either side of the "Gain" Triode. The two end selections "Low" and "Hi" are based on the Showman and the AC30. The "Mid" selection is placed at the geometrical mean.  Brite: This selects a small value capacitor that connects across the wiper to top terminal of the Pre amp Gain section of the Volume control. It effectively allows high frequencies to pass through with no attenuation at low volume settings. 
 The Character section of the amplifier caters to the personality of the speaker used.  Notes on the Character Section of the Iron Horse: Damping Factor, as used in audio amplifiers, is the ratio of speaker impedance to amplifier output impedance. Low output impedance amplifiers typical of Solid State design are in the 0.1 Ohm range, thus if an 8 Ohm speaker were driven by one of these amplifiers, it would have a damping factor of 80.  So what are the implications of Damping Factor? The simple answer is for accurate reproduction of Sound Pressure Level related to applied voltage, (sound output follows voltage input ), the best results are obtained with a high Damping Factor. How low can the DF go before we notice any ringing or delayed response in the speaker? You will notice a slight delay and ringing when DF is below 2. DF below 2 and into the 0.1 range produces interesting effects depending on the speaker and enclosure variations. Is a high DF what we want when we spend good money on a speaker that has Fat lows, Firm mids and Bright highs? The answer is probably no. We want the speaker to be part of the sound experience. This means we need high output impedance, low DF. Tube amplifiers do just that, especially the Pentode configuration. The Pentode is so good at supplying a high output impedance, that negative feedback is needed to reduce it. Simply put a DF greater than 5 will make all speakers sound about the same with minor variations due to SPL response. A speaker is nothing more than a linear Motor/Generator. An AC voltage applied to its voice coil terminals will produce a linear in and out motion. Conversely an external linear in and out motion will produce an output voltage at the speaker terminals. Inertia will keep an excited speaker cone moving in and out at its resonant frequency until it dies out. The Q of the speaker determines this time. High DF will dampen it immediately, low DF will let it ring.

Connecting the terminals of a speaker to an AC voltmeter and thumping the cone with your finger will produce a few volts. This voltage goes back into the amplifier and is damped when high DF present. If a recording of a finger thump is played through a speaker with an amplifier with a high DF, the speaker cone will simply move in and out to reproduce the recording with no ringing of its own. If an amplifier with a low DF plays that same thump, the speaker cone will move in and out and then ring at its natural resonant frequency. This is the personality we are controlling with the Feedback control. There are other configurations that a tube can have that produce lower output impedance, namely Triode and Ultra Linear. These two alternate configurations have much lower output impedance than the Pentode configuration, but no where near the output impedance of a solid state amplifier. Less feedback (DF) is needed in these modes as there is inherent local negative feedback developed in the tube itself. The Feedback adjustment is used to give varying amounts of negative feedback to dampen the speaker ringing ( DF adjustment ). Maximum clockwise rotation is Minimum feedback ( Minimum DF ). This convention was used so all controls appear to increase volume when rotated clockwise. The range of DF is approximately 0.1 to 2 depending on tube configuration. The Presence adjustment is used to apply frequency selective feedback (DF). Maximum clockwise rotation is decreasing feedback (DF) at higher frequencies which translates to higher frequencies are louder than lower frequencies. Output impedance as described above is not the same as output impedance selection described below. These selections match the tube output voltage/current excursions to accomplish maximum power transfer to the load, (speaker). The output transformer translates this optimum impedance by using different turn-ratios from the fixed primary winding to the selectable secondary windings.  Impedance Selection:  As an example, using a cab with a 4 Ohm Speaker impedance, Optimum power transfer will be made at the 4 Ohm selection. When using 4 tubes (4AB1) use the 8 Ohm selection to get the full 30 Watts. See explanation in FAQ’s. Of course we are talking about a 4 Ohm fixed resistive load, A speaker is at its characteristic impedance at DC and around 300 Hz. The impedance at the speaker's self resonance (~100 Hz for most guitar speakers), is in the 100 to 200 Ohm region. The impedance starts rising linearly from about 300 Hz reaching 50 Ohms at 20 KHz.   Referring to Figure 1,  Speaker Impedance/Sound Pressure Level vs. Frequency , there really is no 4, 8 or 16 Ohm impedance across all frequencies, If you want maximum volume at higher frequencies greater than 1 KHz put the Horse in the 16 Ohm position. It will be approximately twice as loud above 1 KHz. At the speaker resonance, (the peak at 100 Hz) very little power is needed, so the sound will appear loud and Boomy at this frequency. This is very close to the 83 Hz on a standard tuned guitar E string. You can tell a speaker’s resonance by simply thumping the speaker cone with your index finger. The note you hear is the self resonant frequency. Although little power is required to drive the speaker at the resonant frequency, the speaker cone size starts to limit the efficiency of the transfer of energy into the air. The SPL portion of the graph shows the output sound falling off just above the resonant frequency and continuing as the frequency decreases further. Figure 1:
 Bias:  The Self selection allows the grid bias to be developed across a Cathode resistor. The bias voltage developed across this resistor is effectively subtracted from the available plate voltage thus reducing the available output power. The cathode resistor is bypassed with a capacitor to prevent local feedback which would reduce the gain of the tube. This resistor capacitor arrangement will cause a “Sagging” of the output under high volume attacks due to the voltage increasing across this resistor due to additional Cathode current. The Fixed position “Grounds” the Cathodes and supplies a negative voltage to the grids. All available plate voltage and current is used to develop the output  signal. Rectifier: This selects a 5AR4 Rectifier Tube or a Solid State Diode pair. More output tube plate voltage is available with the SS Diodes. The voltage also has a lower output impedance which holds up under high volume attacks. The Tube selection is less efficient as well as higher in output impedance. This produces less output power capability as well as a “Sagging” of the output voltage under high volume attacks. Old tube amps had this phenomenon. This is also a convenient place to allow the Standby function as a selection between the two. Iron Horse Block Diagram  



Musings on Speakers 

  (Be sure to read the update at the very end of this article!)
 My musings will be limited to traditional paper cone voice coil speakers as used in Guitar Amplifiers. These speakers are intended to cover the entire audio range. Traditional Hi Fidelity speaker cabinets have 3 separate speakers covering the Bass Mid and Treble ranges. For example the Woofer in one of these arrangements is only called on to reproduce the frequencies between 20 and 300 Hz, All other frequencies are bypassed to the other two speakers.  The design of the Woofer allows great freedom in design. There are no modes or standing waves present on the cone, so any type of material can be used. Additionally, the cone and voice coil assembly are designed to have enormous linear travel in order to move a lot of air, in some cases an inch of travel can be accommodated. This makes the Woofer very inefficient, as the voice coil must remain in the magnetic flux path over its entire excursion to reduce distortion.  Additionally a typical woofer must go two octaves lower in frequency than a standard guitar speaker. This needs at least 4 times the travel to reproduce the lowest frequency. A guitar speaker on the other hand only has to move a tenth of an inch or so to get the same output sound pressure level at 80 Hz, the lowest frequency a standard tuned guitar can produce.

So let’s start with the magnet. There are several types of magnets available i.e., Alnico, Neodymium, Ceramic. The difference is primarily cost and weight. For a given material the only criteria that is important to the operation of the speaker is Flux Density measured in Tesla. There are other implications that come into play with the physical properties of the magnets that show up as different sound responses. The Ceramic magnet is most efficiently constructed in the “pancake” form factor. This causes the pole pieces that direct the magnetic flux density around the voice coil to have air pockets in the assembly.  These pockets have a resonant frequency that gives rise to accentuated output at certain frequencies and diminish output at other frequencies. A closed dust cap over the voice coil can diminish these to a large extent. This forces the cooling convection air currents to be forced out the back of the speaker assembly.

A speaker with a grilled dust cover will allow the resonance to “voice” the speaker response. See a picture of a grilled dust cap below.  Next to consider is Power Handling. Power handling has many factors involved starting with the Magnetic Flux Density. More Power needs more Flux Density. This is because larger wire is needed on the voice coil which in turn opens up the gap between the pole pieces.  More flux density is needed to maintain the same concentration. Additionally the voice coil moves further, so a different winding configuration is used. See the over hung and under hung examples used to keep the winding in the magnetic flux path. This means a bigger magnet or a more powerful magnetic material. Ceramic magnets are the least efficient so they will weigh the most and cost the least due to ease of manufacturing.  The next thing to consider is the Voice Coil. This is nothing more than insulated wire wound on a form. The form is centered in the magnetic flux path and creates the in and out motion of the speaker cone. There is very little heat conduction from this coil as it is wound on a very light weight form made of high temperature Polyimide material. 97% of the power applied to a speaker is dissipated as heat due to the poor efficiency of the cone to air energy transfer.  The wire is either Aluminum or Copper. The trade off is weight. Aluminum is lighter, but has more resistance per turn causing it to self heat more than copper. The lighter weight means less mass to drag along with the speaker cone, so the efficiency goes up a slight amount. The heat can only dissipate by convection. To get air convection maximized, the voice coil diameter must be made larger. Heavier wire is also used due to increased current. These two increases cause more mass that has to be moved, so the efficiency usually suffers as you increase power capability. This also translates into responsiveness of the speaker to electrical stimulus.

Lower wattage speakers using small diameter voice coils have the most responsive sound of all the single speaker designs. A good example is the classic Celestion Blue. This Alnico magnet speaker is rated at 15 Watts, has a 1.75 inch voice coil, and weighs 9.3 pounds. On the other extreme is a 350 Watt Peavey Black Widow which uses a Ceramic magnet and has a 4 inch voice coil, it weighs a hefty 16 pounds. Needless to say the Black Widow is a bit sluggish.  What else should we consider when dealing with power handling? Referring to the data sheet on a particular speaker will reveal a term labeled Xmax. This is the maximum displacement the cone can move before the voice coil goes 10% out of the magnetic flux path. Any more current or power applied to the speaker will result in no more movement or at best limited movement due to residual flux paths. This translates into distortion. The speaker “clips” just as an amplifier would when overdriven.  Additional current or power goes up in heat, and to a limited extent voice coil movement. The voice coil can over shoot the Xmax limit due to inertia of the mass of the cone assembly. If the voice coil goes approximately 25% over the Xmax limit it will hit the rear pole piece used to complete the flux path of the magnet. This is a destructive event and non recoverable voice coil damage results. The voice coil form will be deformed at the bottom end and will rub the center magnet pole piece. Distortion at all volume levels is the result. Some speaker manufacturers will provide this specification as Xlim.

The next thing to discuss is cone material and damping. The cone will move in and out linearly over its entire surface when the excitation frequency is low. This is referred to as the low end or Bass frequency spectrum.  The webbing around the outside of the cone at the top or largest diameter has many functions, first it must keep the cone centered, second it has to be flexible to allow the cone to reach its Xmax excursions, and lastly it has to dampen the free resonate frequency of the cone mass. When frequency is increased, the standing waves on the cone can cause ripples along the face of the cone. These ripples deform the cone with a sinusoidal pattern. See picture below for 3 different frequencies.  The excursion can be much larger than the Xmax of the voice coil, but are only present on the cone material. The type of material used for the cone as well as reinforcement rings can limit the ripples due to global damping. See speaker cone picture below for example of reinforcement rings. Stiff paper or plastic will have the most ripple excursions; where as treated paper will have the most damping and least ripple effects. These ripples translate into peaks and valleys of output sound or Sound Pressure Level for a given drive level.

So where does this leave us when trying to select a speaker? First will be power handling. Lower power speakers are more responsive and more efficient. Use the maximum output power of your amplifier as a guideline so you don’t damage the speaker. Over driven amplifiers can produce output power in excess of twice the rated value. Choose a speaker that is rated at twice your amplifier’s rated output power. If you choose to use two speakers, you can match your amplifier’s rated output power.  If you want your amplifier to be the major distortion element in the chain choose a speaker or speakers that are rated 4 times your amplifier output power rating. Next are you looking for “clean” undistorted sound or an “over driven” distorted sound?  Clean sounds are most readily available with closed back enclosures and flat SPL curves on the data sheet.

Live Speaker Testing:
Matt and I had occasion to audience a pair of Jensen Blackbirds which use an Alnico magnet and have a very flat frequency response and a pair of Jensen MOD12-70 which use a Ceramic magnet and have a 5 dB dip at 1.5 KHz and a 6 dB increase in output from 2 to 5 KHz. (See frequency response below).  

Both sets of speakers were in closed back enclosures and had a cab impedance of 8 Ohms. We both liked the MOD12-70 setup by a large margin. We played our own style for the test. Matt played his usual overdrive material, and I played my usual Venture, Sultans of Swing material. In both cases we were blown away with the cheaper speaker with the “wild” frequency response. (Read the update below!  Have our opinions changed???  Keep Reading!!!)

The Horse does bring out the personality of the speaker with its “Character” section, so the sky is the limit when selecting a speaker. The picture of the dust cap below is of the MOD12-70. The choice of speaker impedance should match your amplifier output impedance specifications. For the Iron Horse use 8 Ohms, as this will allow full power in the 1 and 4 tube selections by dropping the Horse output impedance to 4 Ohms when using 1 Tube and increasing the Horse output impedance to 16 Ohms when using 4 tubes.  Please refer to the pictures below for clarity.  I apologize for not acknowledging the original creator of the speaker diagrams, since I have long lost their original source.  The Dust Cap picture is my own.  Impedance/SPL graph was provided by Jensen. 




Update on Speaker Musings: 06/09/10 I was interested in the note listed on the Blackbird speaker data sheet, “ Rated power measured with 2 hours test with pink noise signal, 6 dB crest factor, loudspeaker mounted on enclosure”. This coupled with several YouTube postings on speaker break in techniques, got me thinking about why the expensive $250 Blackbirds were out done by the relatively less expensive $70 MOD12-70 speakers during our listening comparison tests.

I decided to purchase a Pink Noise Generator and give the new speakers a break in. I was able to purchase a Model PNG-100 from Mystic Marvels on eBay for $70. This generator is based on a amplified filter weighted Zener Diode noise source. It has 230mv RMS output from 20 Hz to 20 KHz. I inserted a 20 dB attenuator at its output to drop the level enough to drive the input of a 65 Watt Twin Reverb Normal channel. I purchased this Twin about two years ago and was not impressed with its sound. Nothing like my 65 I grew up with and unfortunately lost in my life’s journey. I attributed it to the speakers, vowing to replace them if I ever found a speaker I liked. It came loaded with a pair of 50 Watt Jensen C12Ns. It’s been sitting mostly ignored since.

To perform the speaker break in, I attached the Pink Noise Generator, and turned the Bass and Mid to max, the treble and brite to min. I wanted speaker excursions not loud noise. I turned up the volume until the speakers were rumbling like Niagara Falls. The Bass frequency was much lower than any standard guitar could produce. I looked in the back at the speakers and could see they were visibly moving, air was blowing out the back of the cabinet and the walls were shaking. I then closed the door and let it go for 2 hours, occasionally taking a peek to see if any thing was burning or smelled funny. Needless to say it was noisy. All in the name of science I told my wife. She went shopping after about 5 minutes of Niagara Falls.  The test was completed. I felt the back of the speakers. They were both warm and had a faint smell of burnt Lacquer.

I tried out the amp the next day and was pleasantly surprised with the sound. There was more bottom and the speakers had more sensitivity. I had to turn the volume down a notch as the amp produced more sound for my original volume setting. Next were the MOD12-70 speakers. I plugged them into my Twin and went through the same process. Once again the same thing:  more Bottom and better sensitivity. The Blackbirds were over at Matt’s house, so I couldn’t give them a run. Probably a good thing as my wife just got home from shopping. I talked to Matt about my findings and we set up a time when I could break in the Blackbirds. Matt had been using the cab at his band practice and at least one gig, so he was pretty well dialed into their performance. He still wanted to get a pair of MOD12-70 when he got some spare change. I gave them a 2 hour break in just like the others and returned them to Matt. They are now his favorites. Next up is Matt’s stock speakers in his Super Twin. I’ll let him run the test at his house this time, I’m afraid I pushed my wife’s patience to the limit! I’ll also let him comment on the results… 

Notes from Matt about the speakers:
  Don and I spent a lot of time comparing the Blackbirds and the Mod-70s back to back.  Both of us were totally convinced that the Mod-70s were the better sounding speakers.  Well, both of us were completely blown away after the break-in.  The Blackbirds came alive.  The true warmth of AlNiCo speakers were blooming with every chord and riff.  Much to my chagrin, Don's charts and graphs once again equated to something that rocked!    

I took our Cabinet with the Blackbirds to band practice the other night and all the guys were floored with the tone, just warm, round and full.  I was able to dial in the Horse with a killer clean sound just on the edge of break-up and played my Strat all night.  I have a fairly complex pedal board set-up and gave the Horse and the Blackbirds all kinds of sounds with positively beautiful results.  My drummer commented that it was the best guitar sound he has ever heard!  We even did some recording and during the playback we all agreed the tone was phenomenal.  The takeaway here is that you absolutely cannot overlook the speaker component of the signal chain.  A world-class amp deserves world class speakers!

Update June 2012:  Alnico vs. Ceramic speakers


Here is a scientific comparison of the two using Energy Product and Coercivity as a comparison.  Energy Product (EP) is the magnetic force the Voice Coil works against when a current is applied. The more EP the less power needed to move the cone, or the less magnetic material needed for the same cone movement. Coercivity (Hc) is the measure of resistance of a ferromagnetic material to becoming demagnetized. If the magnet gets demagnetized, the cone cannot move when current is applied.  As the EP gets smaller due to demagnetization, the input power must increase to maintain the same cone movement or loudness.  The table shows the differences between the two. Additionally, Neodymium Iron Boron (Nd-Fe-B) is also shown for comparison.

Material    EP(MGOe)   Hc(kOe)            Notes

Alnico 5         5.5      0.6     Typical grade used in Speakers

Ceramic        4.0      2.9     Typical grade used in Speakers

Nd-Fe-B     52.0     11.0     Light weight high power Speaker


Ceramic is slightly less efficient than Alnico 5

Alnico 5 is almost 5 times more likely to demagnetize than Ceramic

Nd-Fe-B Superior magnet in both categories

Does Alnico saturate and soften the cone excursion when high current is applied compared to Ceramic? This is the eternal debate, It is not a simple comparison of values in a chart. Speaker mechanical design, especially in the magnetic path surrounding the Voice Coil play a major role in the sound of Speaker. The simple under hung design can produce voice coil saturation effects due to the voice coil moving out of the magnetic path. Isn't this the effect Alnico is touted to have over Ceramic when it is over driven ? Is it a coincidence that the early speakers used this design. The low power rating of these speakers produce smaller speaker excursions allowing a smaller magnetic area. The later higher power Ceramic designs started using the overhung technique. This caused the efficiency to go down due to the Voice Coil being partially out of the magnetic path, but allowed more Voice Coil movement. The Power Wars of the 70's as well as the cost difference between Alnico and Ceramic all pushed speaker design into our present situation.


Musings on Guitar Amplifier Output Power - April 2012 


How much output power do you need ?

How can you make a reasonable decision

Should you consider Solid State or Tube?


Let's take an engineering approach to the subject. Here are some facts about hearing along with some corresponding values in Decibels. The Decibel is a non linear Logarithmic unit used for the purpose of covering large variations of linear values. The value of 0 dB is usually the reference or starting point.

Looking at these statements may give us an idea of the range of dB we can expect.  If the values  below were given in linear terms, they would represent an amplitude range of 1,000,000 to 1.


 0 dB  Weakest sound heard, Mosquito buzzing 3 feet away.

30 dB  Whisper 6 feet away

60-65 dB Normal conversation 3 feet away.

95 dB Jackhammer 500 feet away.

110 dB Power Saw 3 feet away,

120 dB 50W Amplifier slight distortion.


Also looking at these characteristics of the Human Ear relating to changes in dB levels,


1 dB Imperceptible Change.

3 dB Barely Perceptible Change.

5 dB Clearly Noticeable Change

10 dB About Twice as Loud.

20 dB About Four Times as Loud.


And finally the relationship between Output power and dB, referenced to 1 mW.


 0 dB .001 Watts

30 dB    1 Watt

31 dB  1.3 Watts

33 dB    2 Watts

35 dB    3 Watts

40 dB   10 Watts

41 dB   13 Watts

43 dB   20 Watts

45 dB   32 Watts

50 dB  100 Watts

51 dB  126 Watts

53 dB  200 Watts

55 dB  316 Watts

60 dB 1000 Watts


This tells us we can just notice the difference between 1 and 2 Watts, or 10 and 20 Watts or 100 and 200 Watts. It also tells us we wouldn't notice the difference between 1 and 1.3 Watts or 10 and 13 Watts or 100 and 126 Watts. And finally 1000 Watts would only sound twice as loud as 100 Watts.


The role Psychoacoustics play in sound perception.   Psychoacoustics is the scientific study of sound perception. As seen above, the non intuitive thought of upgrading your 100W rig to 1000W and it only sounding twice as loud says something is going on here that needs explaining.   Rather than going into any detail, a good explanation is given on Wikipedia under Psychoacoustics. Some highlights taken away from the readings are that considerable preprocessing is done in the middle Ear before sending along the neural stimuli to the Brain. As in any preprocessing, data is reduced and assumptions are hard coded into the processing. This preprocessing was mostly intended to enhance our ability to hear each other communicate as well as detect danger.  Some of the non-intuitive aspects are, the Ear will replace missing fundamental frequencies if most of the harmonics are present. Two high frequency tones close

together will produce a low frequency phantom tone at the difference frequency.  This is caused by intermodulation distortion. If an undistorted pure sin wave is heard, the Ear will start adding harmonics to it as it gets louder, consequently if distortion is added to a lower level pure tone the Ear will perceive it as being louder than the pure tone even though no additional power was added. The type of distortion is important as well. Clipping distortion causes odd harmonics, very annoying, whereas even harmonics are pleasing as they represent Octaves. 

Solid State amplifiers are primarily clippers, You get pure sound up to the rated power, then clipping begins. The slightest clipping is heard quite noticeably. Tube amplifiers clip as well, but it is a more rounded softer clipping, not noticed until full limiting is reached. Additionally, the Ear was fooled earlier at lower volumes due to the amplifier producing even harmonics before it reached clipping.

Let's get back to our original questions, If you are seduced into the power availability of a Solid State amplifier, you will be disappointed with a 100 Watt unit, as it will clip at moderate playing levels. Consider at least 300 Watts. Tube amplifiers with 10 to 20 Watts will be sufficient for lead when cranked into moderate to heavy distortion:  40 to 80 Watts will be sufficient for clean or rhythm. Note the Ear will barely perceive a difference between 40 and 80 Watts. The type of output tube will also play into how loud an amplifier sounds for the same output power due to the way it distorts.

Finally, the speaker plays a role in the distortion as well as what we perceive as loudness.  Speakers are very inefficient electrical to sound converters transferring at most 3% of the power presented to them as sound pressure level. A speaker with a 3 dB higher SPL reading will represent a sound that would take twice the power to produce using the original speaker. Interestingly our Ear will barely perceive a difference.