Tweet
During Myles' visit to Genesis, he listened mostly to my records through my phonostage. The transparency of the phono caught his ear, and he asked me to say something about the design of the Genesis Platinum Phono.
I am primarily a loudspeaker designer, and this story begins the same way all my electronics designs do. I was dissatisfied with the existing phono stages that I owned and decided that I needed to design my own in order to show off the transparency of Genesis loudspeakers. I started off with designing and laying out the most sophisticated cartridge loading system that I could devise as the phono stages I liked all had sophisticated loading systems (some of them were even remote controlled).
A couple of months into this, I realized that I hadn't answered the question "WHY?" Why "load" the cartridge output (input to the phono stage) in the first place? It turns out that loading is needed to reduce an ultrasonic resonance peak due to the electrical interaction among the cartridge, the tonearm cable, and the input of the phono stage. This resonance peak is at a frequency that is not audible, but would overload the delicate electronics that amplify the tiny cartridge signal.
What's more, loading is actually very bad for music! The typical moving coil cartridge generates a miniscule signal. 0.5mV is typical, and some high-quality cartridges can go as low as 0.2mV. Loading dumps part of this signal to ground. The higher the loading (lower value resistance), the more of the current generated by the cartridge gets shunted to ground.
Without loading, the music is more powerful and dynamic. But the high resonant peak swamps the electronics of the phono stage and makes things sound harsh and hard. You don't actually hear the distortion, but you experience the detriment to the music. High loading reduces detail and resolution, but makes a "rounder and warmer" sound.
Once I realized this, I set out to solve the original problem – the resonance peak.
One characteristic of LP playback is the RIAA de-emphasis curve (Recording Industry Association of America’s standard equalization curve – pre-emphasis on recording, de-emphasis on playback) required to amplify the signal from the cartridge. The output of the cartridge is so low that the output voltage needs to be amplified by over 1,000 times (>60dB of gain) to get to the level required by the preamp. However, because of the RIAA curve, it needs to be amplified over 10,000 times (>80dB of gain) at 20Hz. But, it only needs to be amplified by 100 times (>40dB of gain) at 20kHz.
Hence, there is a 40dB (100 times) difference in gain between 20Hz and 20kHz. With the RIAA de-emphasis curve, the higher the frequency, the lower the gain wanted. That being the case, and the detrimental resonant peak being 30dB at ultrasonic frequencies, could we design an amplifier that doesn't amplify the resonant peak?
The typical phono stage is designed as a 3-step process:
Why not just design a variable gain amplifier that amplifies the bass frequencies more than the high frequencies, and match this variable gain to the RIAA de-emphasis curve? Then, make sure that the gain at the frequencies where the bad resonant peak of the tonearm cable can come in is low enough that it does not present a problem?
That is exactly what we did with the Genesis Phono - a single high gain stage where 20Hz has a gain of 80dB, 1kHz has a gain of 60dB, and 20kHz has a gain of 40dB with the variable gain structure matching the RIAA de-emphasis curve. 200kHz and above has a gain of less than 10dB, resulting in the resonant peak no longer being an issue (>20dB less than the midrange and well within the operating parameters of the high gain stage).
In this case, simple is really better – and following my mantra of “As simple as possible, but no simpler” worked once again.
I am primarily a loudspeaker designer, and this story begins the same way all my electronics designs do. I was dissatisfied with the existing phono stages that I owned and decided that I needed to design my own in order to show off the transparency of Genesis loudspeakers. I started off with designing and laying out the most sophisticated cartridge loading system that I could devise as the phono stages I liked all had sophisticated loading systems (some of them were even remote controlled).
A couple of months into this, I realized that I hadn't answered the question "WHY?" Why "load" the cartridge output (input to the phono stage) in the first place? It turns out that loading is needed to reduce an ultrasonic resonance peak due to the electrical interaction among the cartridge, the tonearm cable, and the input of the phono stage. This resonance peak is at a frequency that is not audible, but would overload the delicate electronics that amplify the tiny cartridge signal.
What's more, loading is actually very bad for music! The typical moving coil cartridge generates a miniscule signal. 0.5mV is typical, and some high-quality cartridges can go as low as 0.2mV. Loading dumps part of this signal to ground. The higher the loading (lower value resistance), the more of the current generated by the cartridge gets shunted to ground.
Without loading, the music is more powerful and dynamic. But the high resonant peak swamps the electronics of the phono stage and makes things sound harsh and hard. You don't actually hear the distortion, but you experience the detriment to the music. High loading reduces detail and resolution, but makes a "rounder and warmer" sound.
Once I realized this, I set out to solve the original problem – the resonance peak.
One characteristic of LP playback is the RIAA de-emphasis curve (Recording Industry Association of America’s standard equalization curve – pre-emphasis on recording, de-emphasis on playback) required to amplify the signal from the cartridge. The output of the cartridge is so low that the output voltage needs to be amplified by over 1,000 times (>60dB of gain) to get to the level required by the preamp. However, because of the RIAA curve, it needs to be amplified over 10,000 times (>80dB of gain) at 20Hz. But, it only needs to be amplified by 100 times (>40dB of gain) at 20kHz.
Hence, there is a 40dB (100 times) difference in gain between 20Hz and 20kHz. With the RIAA de-emphasis curve, the higher the frequency, the lower the gain wanted. That being the case, and the detrimental resonant peak being 30dB at ultrasonic frequencies, could we design an amplifier that doesn't amplify the resonant peak?
The typical phono stage is designed as a 3-step process:
- amplify (by 80dB or more) --> RIAA de-emphasis (filter by up to 40dB at 20kHz) --> amplify some more or just output buffer
Why not just design a variable gain amplifier that amplifies the bass frequencies more than the high frequencies, and match this variable gain to the RIAA de-emphasis curve? Then, make sure that the gain at the frequencies where the bad resonant peak of the tonearm cable can come in is low enough that it does not present a problem?
That is exactly what we did with the Genesis Phono - a single high gain stage where 20Hz has a gain of 80dB, 1kHz has a gain of 60dB, and 20kHz has a gain of 40dB with the variable gain structure matching the RIAA de-emphasis curve. 200kHz and above has a gain of less than 10dB, resulting in the resonant peak no longer being an issue (>20dB less than the midrange and well within the operating parameters of the high gain stage).
In this case, simple is really better – and following my mantra of “As simple as possible, but no simpler” worked once again.
Comment