Home |
Search |
Today's Posts |
#1
Posted to rec.audio.tubes
|
|||
|
|||
What makes a good curve?
Its been said here many times that plate curves for a particular tube
look good. So what do good curves look like? How do you tell them from not so good curves? Adam |
#2
Posted to rec.audio.tubes
|
|||
|
|||
What makes a good curve?
Adam Stouffer wrote: Its been said here many times that plate curves for a particular tube look good. So what do good curves look like? How do you tell them from not so good curves? Adam The lines are fairly straight, and the distance between them is fairly linear. i.e. for a triode curve if you look at the line for -2V, it is about an equal distance from -1V as it is from -3V, all along the length of the line. What this means is that in an audio application, the tube will have fairly low harmonic distortion if it's biased properly. Try to look up "ideal triode" and you will find the ideal curves for a perfect triode which does not exist, and you can compare it to the real devices. |
#3
Posted to rec.audio.tubes
|
|||
|
|||
What makes a good curve?
On Apr 6, 6:13*pm, Adam Stouffer wrote:
Its been said here many times that plate curves for a particular tube look good. So what do good curves look like? How do you tell them from not so good curves? Adam 1. Good plate curves are straight for the maximum length above a short bottom curvature. That ensures linearity and predictability in the amplification. 2. Good plate curves are equidistant between negative bias lines. That makes the amplification linear. 3. For an SE amp good plate curves are nearer vertical than diagonal. This gives the greatest signal swing that is still linear, and thus greater gain without paying a price in distortion. Working visually on the loadlines is my preferred approach to new tube amp design (as Patrick's is arithmetical and algebraic), so there is a much discussion of loadlines on my netsite http://members.lycos.co.uk/fiultra/T...mp%20INDEX.htm but already expert may go directly to the illustrations page http://members.lycos.co.uk/fiultra/KISS%20190.htm On that page they will find individually taken curves for a pair of my WE300B, which indubitably have good curves: WE300BindivSPEC.jpg An illustration to show how powerful the curves are as a tool to tell you what will happen inside your amp if you do x, y or z: 112KISScurves.jpg An illustration showing how I derived the operating points for the 300B in The KISS Amp: 113KISS300Boperatingpoints.jpg Ditto for the 417A drivers: T39-Ultrafi-417A-Q.jpg The circuit is here if you want to see how I turned all the operating points into the specification of the components in the amp: http://members.lycos.co.uk/fiultra/T...trafi-crct.jpg The same page http://members.lycos.co.uk/fiultra/KISS%20190.htm also holds frequency response curves for several of my amps which you can by now surely find for yourself. They show how flat a response can be when you start with tubes with "good curves" and apply good design principles. All this is discussed at: http://members.lycos.co.uk/fiultra/T...mp%20INDEX.htm and in particular in chapters 112 to 115. HTH. Andre Jute Visit Jute on Amps at http://members.lycos.co.uk/fiultra/ "wonderfully well written and reasoned information for the tube audio constructor" John Broskie TubeCAD & GlassWare "an unbelievably comprehensive web site containing vital gems of wisdom" Stuart Perry Hi-Fi News & Record Review |
#4
Posted to rec.audio.tubes
|
|||
|
|||
What makes a good curve?
Adam Stouffer wrote: Its been said here many times that plate curves for a particular tube look good. So what do good curves look like? How do you tell them from not so good curves? Adam Triode curves look the nicest, and what you see on the Ia vs Ea data sheet is the plot of Ia ves Ea for fised values of Eg. The curve is just a well drawn scribble unless you know what the curve means. But as Ea is raised from zero, Ia begins to increase at a rate of a constant x the cube root of a number squared, or something to the power of 2/3. Don't laugh, I'm serious. Thermionic diodes display this character, but there is only one curve for them because thay have no grid. With a grid and increasing grid bias voltage, the curves plotted for each value of Eg1 appear as if stepped across the page towards the right. A good set of triode curves includes these "Ra" lines being very close the same shape for changing Eg1 bias, and they are very evenly spaced if you measure along a horizontal line above 0.0 mA of anode current. A good set of triode curves means that if the gris swings say +/-5Vdc the anode voltage change along a horizontal line is maybe -100Vdc and + 98V in a typcal case of a linear triode in circuit with a load of more than 20 x Ra. Tangents drawn to the curves at any Ea/Ia point can be extended to see ehat Ia change you can get for a change of Ea. This gives you the Ra or dynamic anode resistance of the triode for that particular Ea/Ia point, The slope of the tangent line used to calculate Ra graphically changes with the Ia. And so Ra also changes dynamically, because the Ea/Ia ratio vareis as something rasied to the power of 2/3. Hence there is distortion in triodes. But the Ra as calculated for a given triode for a fixed value of Ia will be nearly equal over a wide Ea range in a nice set of curves. Pentodes and beam tetrodes are somewhat different in that for a fixed Eg1 voltage, Ia rises very rapidily with a slight rise in Ea, then at some point, the Ia rise becomes extremely slight for a huge rise in Ea. Their Ra can be graphically determined with tangent lines, but the Ra variation can rather a lot, and Ra can be so high the Ra is considered almost an infinite resistance, ie, above 1Mohm, and a nearly horizontal graph line. Pentodes and tetrodes have a bigger change in shape no matter which way you look at them. so they never look as well as a triode. However, a pentode or beam tetrode can be connected with NFB and taught to behave to give about the same gain and Ra as a triode, and were you to test such an arrangeent, you might end up with a better looking set of curves than a triode with similar Ra and Ia and gain. That is about the right amount of information to perhaps completely bamboozle nearly all the group. More descriptions of things worth knowing about triodes can be read after going to http://www.turneraudio.com.au/education+diy.html Also, please read the Radiotron Designer's Handbook, 4th Ed, 1955. Beg borrow or steal any books on basic electrtonics written before 1960. Patrick Turner. |
#5
Posted to rec.audio.tubes
|
|||
|
|||
What makes a good curve?
"maxhifi" wrote in message ... : : : Adam Stouffer wrote: : : Its been said here many times that plate curves for a particular tube : look good. So what do good curves look like? How do you tell them from : not so good curves? : : Adam : : The lines are fairly straight, and the distance between them is fairly : linear. i.e. for a triode curve if you look at the line for -2V, it is : about an equal distance from -1V as it is from -3V, all along the length : of the line. : : What this means is that in an audio application, the tube will have fairly : low harmonic distortion if it's biased properly. : : Try to look up "ideal triode" and you will find the ideal curves for a : perfect triode which does not exist, and you can compare it to the real : devices. : ........................................ here's a practical example (mind you, good curves are in the eye of the beholder, eh) of a tube with just a lean 0.5 W Pamax, but capable of some amazing curves: look up the (Philips) EBC81 datasheet. suppose we have 380V B+ available. using a 680K plate R, ignoring load, you'll want it biased at 0.25 mA, 210V, -3.1Vg1k if you draw the loadline on a printed sheet, you can see that a +/- 1V input signal is amplified pretty blamelessly, swinging 55V both ways, Av=55. as you increase the input signal, however, the distance between subsequent gridlines increases on positive going inputs, towards 0Vg1k, decreases for negative going inputs: 2H distortion (and that's unloaded, so it'll be worse when in circuit). however, were you to use a 1.6 mA ccs, with a cap coupled to a subsequent 680K load, at 210V, -2Vg1k, you have a stunningly lineair, 6SN7 class, but with Av=67 stage, that is near 100Vrms out @ (est.) 1 % THD. (move that earlier loadline upwards to go from 70V, 1.8mA to 345V 1.4mA and you see the very equidistant gridline intersections) Rudy |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
What makes a good room? | Pro Audio | |||
Frequency curve (response) good disco-sound system... | Pro Audio | |||
What makes a mix buss sound good? | Pro Audio | |||
What makes a good linear triode? | Vacuum Tubes | |||
who makes good 8" subwoofers now a days | Car Audio |