Re: A Bodger's Tale
Posted: Wed Mar 27, 2019 8:21 pm
If your amp sounds good to you, don't spend all your waking hours wondering if it could sound even better, just enjoy the music. All too often I have fallen prey to that bug and believe me, it is debilitating. Yes, tweaking can be rewarding, but if your amp is constantly in bits, you're not going to get much out of this hobby.
Now then, valve amps aren't rocket science; actually they are more complicated than that. OK, so their circuits might be a lot simpler than a computer, but they work towards outputting a signal that affects both the senses and the emotions. The sound of a valve amp, or any amp for that matter contains a complex blend of objective and subjective elements that's very difficult to unravel, even if you have an electrical engineering degree and a lab full of test equipment.
Sometimes, you nail it as I did a few years ago with a pair of monoblock SET amplifiers based on a pair of new old stock, Sylvania 6B4G direct heated triodes. The 2A3 valve runs on 2.5V filaments. The 6B4G is the same valve but has 6V filaments and it is far cheaper to by than new old stock 2A3 valves.
At the time I built this pair of monoblocks, the Audio Talk crew were well into driving direct heated power triodes with the high transconductance pentodes that were once common as video amplifiers in old TV sets. These valves such as the 12GN7, EF184, C3g etc are (apart from the German C3g) cheap and plentiful as new old stock, can run high currents up to 30mA without breaking sweat and with a low value anode resistor, they put out a decaying spectrum of even order distortion harmonics, rather than the odd order harmonics that result when pentodes are used conventionally in ultra high gain mode. They make excellent two stage DHT amplifiers a realistic proposition.
Pentodes have five electrodes. The two extra electrodes are called the screen grid and the suppressor grid. The screen grid does what it says, and "screens" the anode from the grid, effectively preventing the grid from "seeing" the anode. This has an important effect on the Miller capacitance of the valve itself. If the grid can't see the anode electrically, then it can't see any capacitance either. This has the effect of greatly increasing the bandwidth of the device, up into the MHz region; waaay out of the audio band, so slewing distortion at audio frequencies is banished, which should result in far better top end response from the amp, provided the output transformer is capable of showing the difference.
The other electrode is the suppressor grid, which suppresses secondrary electron emission from the anode, basically by throwing any stray electrons emitted by the hot anode (which is only supposed to collect electrons) back to where they came from. It does this by interposing itself between the screen grid and the anode and being at an earthy potential with respect to the said anode. Normally this earthy potential is achieved by internally connecting the supressor to the valve's cathode, although, in some pentodes such as the EL34 the suppressor is brought out to its own terminal. There is a danger of it attracting electrons on their way perfectly legitimately to the anode, so to avoid this, the suppressor grid has as coarse a pitch as practically possible and it is arranged so that its wires are aligned exactly with those of the screen grid. In this way, the supressor doesn't 'get in the way.'
In order for a pentode to work at its best, the screen grid needs to be at a fixed potential relative to the anode, as its voltage has a huge effect on the operation of valve in comparison to that of the anode. To this end it can be locked by a zener diode, or much more prettily be held by a gas discharge tube. These lovely devices, are filled with an inert gas and glow orange or purple according to what gas they contain. They are usually available in 75V, 108V and 150V versions and can be strung together in series to drop higher voltages eg a regulated 300V can be obtained by stringing two 150V tubes together.
So I built this 6B4G amp, with a 12GN7 pentode driver, with 108V on the screen, regulated by a 108V glow discharge tube per channel. Also I did something to each power supply that I had never done before....and the improvement in the music quality was not subtle.........
Now then, valve amps aren't rocket science; actually they are more complicated than that. OK, so their circuits might be a lot simpler than a computer, but they work towards outputting a signal that affects both the senses and the emotions. The sound of a valve amp, or any amp for that matter contains a complex blend of objective and subjective elements that's very difficult to unravel, even if you have an electrical engineering degree and a lab full of test equipment.
Sometimes, you nail it as I did a few years ago with a pair of monoblock SET amplifiers based on a pair of new old stock, Sylvania 6B4G direct heated triodes. The 2A3 valve runs on 2.5V filaments. The 6B4G is the same valve but has 6V filaments and it is far cheaper to by than new old stock 2A3 valves.
At the time I built this pair of monoblocks, the Audio Talk crew were well into driving direct heated power triodes with the high transconductance pentodes that were once common as video amplifiers in old TV sets. These valves such as the 12GN7, EF184, C3g etc are (apart from the German C3g) cheap and plentiful as new old stock, can run high currents up to 30mA without breaking sweat and with a low value anode resistor, they put out a decaying spectrum of even order distortion harmonics, rather than the odd order harmonics that result when pentodes are used conventionally in ultra high gain mode. They make excellent two stage DHT amplifiers a realistic proposition.
Pentodes have five electrodes. The two extra electrodes are called the screen grid and the suppressor grid. The screen grid does what it says, and "screens" the anode from the grid, effectively preventing the grid from "seeing" the anode. This has an important effect on the Miller capacitance of the valve itself. If the grid can't see the anode electrically, then it can't see any capacitance either. This has the effect of greatly increasing the bandwidth of the device, up into the MHz region; waaay out of the audio band, so slewing distortion at audio frequencies is banished, which should result in far better top end response from the amp, provided the output transformer is capable of showing the difference.
The other electrode is the suppressor grid, which suppresses secondrary electron emission from the anode, basically by throwing any stray electrons emitted by the hot anode (which is only supposed to collect electrons) back to where they came from. It does this by interposing itself between the screen grid and the anode and being at an earthy potential with respect to the said anode. Normally this earthy potential is achieved by internally connecting the supressor to the valve's cathode, although, in some pentodes such as the EL34 the suppressor is brought out to its own terminal. There is a danger of it attracting electrons on their way perfectly legitimately to the anode, so to avoid this, the suppressor grid has as coarse a pitch as practically possible and it is arranged so that its wires are aligned exactly with those of the screen grid. In this way, the supressor doesn't 'get in the way.'
In order for a pentode to work at its best, the screen grid needs to be at a fixed potential relative to the anode, as its voltage has a huge effect on the operation of valve in comparison to that of the anode. To this end it can be locked by a zener diode, or much more prettily be held by a gas discharge tube. These lovely devices, are filled with an inert gas and glow orange or purple according to what gas they contain. They are usually available in 75V, 108V and 150V versions and can be strung together in series to drop higher voltages eg a regulated 300V can be obtained by stringing two 150V tubes together.
So I built this 6B4G amp, with a 12GN7 pentode driver, with 108V on the screen, regulated by a 108V glow discharge tube per channel. Also I did something to each power supply that I had never done before....and the improvement in the music quality was not subtle.........