Initial 'DQ5' testing

[rsumperl]

Greetings everyone,
I breadboarded a P-P 6DQ5 amplifier with a 6SN7 voltage amp. I coupled with a 1:3 transformer. The output transformer was a 15W 8K:4 transformer.
I used 500vdc on the plate and 175vdc on the screen of the 6DQ5 and 175vdc on the plate of the 6SN7. I adjusted the bias for an idle current of 32ma/tube on the 6DQ5. So, considering the output transformer is not even close to what the 6DQ5s wants to see, the amp was LOUD. Real LOUD. I ordered 3.3K 100W transformers from Edcor. We'll see how things change with them.

Ray

[Arnulf]

Try paralleling two 4R speakers until those transfiormers arrive. This will give you a better idea of what you should expect with them.

Post the schematic when you finish your design, I'm sure others would be happy to see it and perhaps offer some suggestions.

[tubelab.com]

OK, it is Sunday and for the first time since July 4 I am not at work. So....time to make 'em glow. I had guessed that the 6DQ5 would make about 100 watts at 550 volts and a 3300 ohm load without melting. Speculation on the bias current was 30 to 40 mA.

So I wired a pair of 6DQ5's into the red board, set the screen supply at 150, the plate supply at 550, and adjusted the bias current for 30 mA. I applied enough signal to get 1 watt and tweaked the bias current and the balance for minimum distortion. 0.37% distortion comes at 33 mA and more or less current raises the distortion although it remains under 0.5% from 27 to 50 mA.

I then applied more crank. As the input signal went up, so does the output power and the distortion. 50 watts makes 1.89%. 100 watts gives 3.27%. I usually state amplifier power as the power at which the distortion crosses 5% which comes at 107 watts. Crank it till square waves appear reads 129 watts although this is not an accurate number.

The tubes show a faint glow if left at 100 watts for several minutes in a dark room. Since the power goals were met at 550 volts, I see no reason to go higher since more glow would appear, or the bias current would need to be reduced.

[rsumperl]

Hey George,
Thanks for the info. I'll reduce screen to 150vdc and see what happens. So my question for you is at the 100w level you were getting 3.27% distortion, is that with or without feedback and if it is without, how much lower would the distortion be and what is the price on power.

Thanks,
Ray

[tubelab.com]

The red board uses both local (Schade) feedback and Global Negative Feedback. I am using the local feedback which is a resistor from the plates of the output tubes to the plates of the driver tubes. This lowers the output impedance without sucking the life out of the sound. I am not using the GNFB on this board since all of my listening tests have sounded better without it despite slightly higher measured numbers. Applying negative feedback will not reduce the available power output. It will reduce the total amplifier gain so that you will need more input voltage for a given power output.

Regarding the Kepco power supply, I have one and it is working, but it has been banished to the warehouse since the big HP came to town. Too many power supplies, not enough bench space. I don't know if I will get there before leaving town next weekend.

Got to crank up some more tubes before dinner. I should have more info tonight but I will post to the red board thread.

[rsumperl]

Thanks for your help George! Last question, you posted some distortion numbers. At what point does it become noticable? Is 3.27% noticable?


Thanks,
Ray

[Arnulf]

Perhaps you should have a look at Variable Harmonizer, especially the pages beyond 2nd. Not all THDs are created equal, mr. Jefferson said

[tubelab.com]

At what point does it become noticable? Is 3.27%.......Not all THDs are created equal

As most people know not all DISTORTIONS are created equal. Unfortunately most THD meters including my multi-killobuck (when new) HP do not measure THD. Most meters measure the level of the fundamental tone, then engage a sharp notch filter to remove it, measure what's left, and compute the total distortion. This includes THD, IMD, hum, noise and anything else that is not the applied tone. An FFT analyzer running on a PC will provide a visual display of all of the frequency components in the output spectrum of the unit under test. Mine is non functional at this time.

Lets review a bit about distortion:

First there is "THD" Total HARMONIC Distortion. THD is the sum of all of the harmonics created in the unit under test, measured and referenced to the total output. All harmonics do not sound equally displeasing.

The second harmonic is a tone exactly one octave above the original tone. If the amp is fed "A above middle C" (A4) as played on the piano (440 Hz) the second harmonic will still be an A. So will the 4th and all of the even harmonics. Most agree that a reasonable amount of even harmonics are tolerable and even pleasing. After all look at all the SE amps.

The third harmonic usually falls very close to a note that sounds pleasing when played with the original tone. The third harmonic of A4 is 1320 Hz this is very close to E6 (1318.5 Hz) and an A and a E form a pleasing chord. So some 3rd harmonic within reason is usually not objectional. The 5th harmonic does not fall close to any musical note at all thus sounds dischordant to those used to the 12 tone musical scale. Ditto the higher odd harmonics.

There is also InterMoDulation distortion. No amplifier is perfectly linear. Theory tells us that whenever two signals are applied to a non linear device 4 (or more) signals will be present in the output. The two original signals, plus the mathematical sum and differences of their two frequencies. Apply a 440 Hz piano tone and 55 Hz thump from a bass guitar (both "A" notes) and you will get 385 Hz and 495 Hz. These are usually dischordant.

THD and IMD usually occur together so you will get the IMD mixing products of the input tones AND all of their harmonics too. We all know that hum and noise sound bad. Yes, the hum will IMD with the input tones too. There are also many more types of distortion, including some that we don't even know how to measure yet.

There is also a lot of psychoacoustics at work here. If you are cricitally listening to a piece of simple music at a low volume level it is easy to notice some simple nuances including distortion. If you are in the mosh pit at a Metallica concert (yes I have been there) can you tell the difference between 5% and 10 % distortion? I don't think so. SO...

If you have a very good system and you were listening at a reasonable level could you hear 3% of pure second harmonic? Maybe, maybe not. You might notice it especailly if you always listened to a SS system with .005% distortion. Could you hear even 1% of pure 5th harmonic at the 1 watt level. BIG TIME! Could you hear the difference between 1% and 3% assuming a reasonable harmonic spectrum? At 1 watt maybe. At 100 watts. NO your ears are saturated and probably generating even more distortion.

Music has a lot of dynamic range. The differences between the loudest and the quietest parts can be 50 to 80 db. When designing RF amplifiers for cellular and other communication systems, we speak of the peak to average ratio. Most music has a peak to average ratio of at least 20 db. Simply put, this means that if you took your 100 watt amp and turned it up until it just clipped the ocasional peaks, the AVERAGE power being produced will be 1 watt or less. What does this mean? As stated in the name of a high end solid state amp it is the FIRST WATT that counts. It MUST be clean. If the amp puts out 3 or 5 or 10 percent or even clips a bit at 100 watts who cares AS LONG AS IT RECOVERS QUICKLY and gracefully.

BTW, I measured 0.37% at 1 watt and even tweaked for minimum distortion at this level. Now you know why.

[rsumperl]

George, what do you think about driving the DQ5's with the circuit used in the thread "6L6GC AB2 AMP"? When you drove the DQ5's to 100W, were you in AB1 or 2?

Thanks
Ray

[DualTriode]

Originally Posted by tubelab.com

Music has a lot of dynamic range. The differences between the loudest and the quietest parts can be 50 to 80 db. When designing RF amplifiers for cellular and other communication systems, we speak of the peak to average ratio. Most music has a peak to average ratio of at least 20 db. Simply put, this means that if you took your 100 watt amp and turned it up until it just clipped the ocasional peaks, the AVERAGE power being produced will be 1 watt or less. What does this mean? As stated in the name of a high end solid state amp it is the FIRST WATT that counts. It MUST be clean. If the amp puts out 3 or 5 or 10 percent or even clips a bit at 100 watts who cares AS LONG AS IT RECOVERS QUICKLY and gracefully.

BTW, I measured 0.37% at 1 watt and even tweaked for minimum distortion at this level. Now you know why.

Hello George,
Thank you for the distortions and average level to peaks insights. Do you have any light to shine on the compression / loudness wars for current CD’s and vinyl vs vinyl back in the day?
DT
All just for fun!