The measurements were taken under identical conditions within the same minute. The speaker and microphones didn't move. No settings were changed other than the amplifier and the speaker cable. I did the measurements several times. It's clear that it is the amplifier and the results are consistent. I never had these results with my 8 ohm nominal speakers on the tube amp. I agree that I should measure the amplifiers but I am not experienced with doing that. I am, however, very experienced with speaker measurements so I am confident in my results there.
I agree but my measurements were done within the same minute. Nothing changed except the amplifier and speaker cable. I have measured both sides of the amplifier, plus a Class a/b and Class D amplifier. The Class D and Class a/b amp measure the same through the speaker. The tube amp struggles. I also measured a different speaker where the tweeter circuit has a rising impedance (due to a single resistor rather than an L-Pad for attenuation) and the tube amp increases the treble over this range. It really just points to poor performance from an old design/ tube amp. Time to move over to Class D.
I am experienced and fairly accomplished with speakers and crossovers. But I am a bumbling fool with electronics in an amplifier. I could not be trusted to do this!!! Ha! I will gladly sell this amplifier to someone who will use it and love it. I don't want my amplifier changing the frequency response of my speakers.
Such measurements only gives a consistent answer, when each "measured" parts are the same.
Some amplifier behaves frantically to (rising) capacitive load, so if the speaker cable differs, this HF difference may occurs.
The speaker cable is not causing a 2db shelf above 4Khz at the exact spot where the impedance drops to 4 ohms. The culprit is clear. It is the tube amp's inability to drive the load properly. These speakers have measured really well on 2 different amplifiers and cables (Class D, Class A/B). Then the tube amp can't handle it.
How about seeing if you can fix your tube amplifier to work better with speakers of many different impedances.
A complete and accurate schematic of your amplifier is worth 1000 Words.
Add some Global Negative feedback? Perhaps, or perhaps not.
A better output transformer, with about 1.5 to 2.0 times higher primary impedance, and less leakage inductance, plus less distributed capacitance? Perhaps. But that will have less power, just more Finesse to the sound, and be less speaker dependent.
(I do not use global negative feedback on my amplifiers).
Your output stage does not have Global negative feedback, or other output stage negative feedback, Right?
What are the output tubes? Triode; Ultra Linear Pentode/Beam power; Pentode/Beam power modes?
You are at the right Forum, and the right thread to see how little, or how much might be required to modify the amplifier to get less sensitivity to different loudspeaker impedance variations.
You will need a non-inductive power load resistor, so you can test the amplifier by itself. Two 8 Ohm resistors will allow you to load the amplifier with 8 Ohms, 4 Ohms, and . . . 16 Ohms (most loudspeaker woofers are at least 16 Ohms at one or more bass frequencies).
Your woofer is more than 30 Ohms at the peak that is below the port frequency, and more than 30 Ohms at the peak that is above the port frequency.
Have Fun!
A complete and accurate schematic of your amplifier is worth 1000 Words.
Add some Global Negative feedback? Perhaps, or perhaps not.
A better output transformer, with about 1.5 to 2.0 times higher primary impedance, and less leakage inductance, plus less distributed capacitance? Perhaps. But that will have less power, just more Finesse to the sound, and be less speaker dependent.
(I do not use global negative feedback on my amplifiers).
Your output stage does not have Global negative feedback, or other output stage negative feedback, Right?
What are the output tubes? Triode; Ultra Linear Pentode/Beam power; Pentode/Beam power modes?
You are at the right Forum, and the right thread to see how little, or how much might be required to modify the amplifier to get less sensitivity to different loudspeaker impedance variations.
You will need a non-inductive power load resistor, so you can test the amplifier by itself. Two 8 Ohm resistors will allow you to load the amplifier with 8 Ohms, 4 Ohms, and . . . 16 Ohms (most loudspeaker woofers are at least 16 Ohms at one or more bass frequencies).
Your woofer is more than 30 Ohms at the peak that is below the port frequency, and more than 30 Ohms at the peak that is above the port frequency.
Have Fun!
Last edited:
I am in no way qualified or motivated to modify a product that is this inferior when I can easily get one that works properly. I can't answer your questions about this amplifier because I really don't know. I also don't want to waste my time with a product like this. The an amplifier should not change the frequency response on a speaker with a load that is this easy to drive. These speakers sound absolutely fantastic with the Class D or Class A/B amp. Goodbye tube amp. Never again.
VTA120 uses good transformers and is a classical pushpull design BUT with quite low NFB. Replacing 7.5k resistors with 2k
will make it close to dynaco MkIII performace.
I believe the issue here could very well be the interaction between the amplifier and the loudspeaker load.
Amplifier output impedance can have a strong effect on the full system's frequency response. These changes in system frequency response are often not taken fully into account.
Measuring the amplifier's output voltage vs frequency alone into a resistive load will provide a good reference point to start with and needs to be done first.
This measurement alone will however tells us nothing about the amplifier's frequency response into a real loudspeaker load and so it's effect on the full system's frequency response.
After measuring the amplifier's output voltage into a constant resistance like 4 ohms, secondly one needs to remeasure the amplifier's output voltage loaded with the loudspeaker to be used to see changes in amplifer frequency response in the system.
The difference between the two curves will be the amplifier's contribution to system errors due to the loudspeaker's loading of the amplifier.
This error in system frequency response from the amplifier's output impedance can be quite large as pointed out by EdGr.
There is a third effect on system frequency response that amplifier output impedance can have. This is due to changes in the loudspeaker's Qt at low frequencies caused by the added series resistance from the amplifier's output impedance.
The variation in low frequency Qt will effect both the loudspeaker frequency response near Fb and can have a strong effect on the loudspeaker's transient response near Fb. This effect is of couses very dependent on the loudspeaker design.
The measurement data that "ThatSoundsGood" provided does seems to show about a 1.5dB gain at 45Hz and about a 2dB loss at 65Hz between the two amplifiers. This would be consistent with increased Qt for a higher tube amplifier output impedance.
Long ago when all amplifiers were tube based many loudspeaker designer's took great care to designing loudspeakers with near constant input impedance to reduce errors in system response caused by amplifier loading. The introduction of solid state amplifier's in the 1960 with near zero output impedance however caused almost all loudspeaker designers to place a low priority on loudspeaker constant impedance. So most modern loudspeaker show strong variation in input impedance VS frequency.
This makes amplifier/loudspeaker interaction a very real issue in many of todays systems when non-zero output impedance amplifiers (mostly tube amplifiers) are used.