Post by colinf on Jun 4, 2017 10:08:05 GMT 12
upload full resolution photosIn this photo you can see the two triode sections. The cathode (with heater filament inside it) is the small silver coloured tube in the middle. The anode is the grey box with flanges surrounding it, and holes in the side. The control grid is wire that has been wound horizontally around two copper supporting posts next to the cathode. These posts can be seen next to the cathode tube on the white mica insulators. The mica insulators hold the whole triode construction together and hold it in place inside the glass envelope.
AMR-iFi R&D
|
Post by Graham on Jun 4, 2017 10:23:54 GMT 12
Hi Colin
Thanks for your input, yes fully understood. One of my important jobs when I was a National Service Manager was attending technical training courses in NZ and overseas to learn about new technology and then present this usually complex information in simple to understand language and explanations for our technicians. Both you and Owen have so far managed to achieve the same outcome !!!!! Keep up the good work.
|
Post by colinf on Jun 4, 2017 12:16:45 GMT 12
Ok great! Now for the 6sn7: 150 volts on the anode and -3v on the grid results in 10mA cathode current (Ik). k is short for kathode, the German word for cathode. I is from French, Intensite de courant, or current intensity in English. Dynamically, for Vp=150v, Vg1 = -2v results in Ik= 13mA; and Vg1= -4v results in Ik=7mA. Still following ok?
AMR-iFi R&D
|
Post by Owen Y on Jun 4, 2017 13:32:47 GMT 12
Thanks colin. I've built quite a few preamps & 'Single Ended' amps, but only one Push-Pull amp. And never messed with KT88s (6550s, yes) - so bear with me.... Looking at the Beam Tetrode diagram above, the valve is illustrated elementally quite clearly - but, as colin describes, it is actually built concentrically (like most valves): (The famous RCA drawing of a beam tetrode.) A modern KT88.
|
Post by Owen Y on Jun 4, 2017 13:57:18 GMT 12
Before we dissect those abbreviations for Gryffles....looking again at the diagrams of the various valves: In particular, we are interested in the KT88 - referring to the 'Pentode' diagram, things to note: - the Control Grid (g1) is usually always 'biased -ve volts (in relation to the Cathode). - the Screen Grid (g2) is given a +ve voltage, so that it helps attract electrons from the Cathode which accelerate through it to the Anode. You can appreciate that the Screen's +ve voltage is generally lower than Anode voltage, in order for this to happen. And that the Screen thus is effective in (greatly) increasing current flow & amplification as a result. - the Suppressor Grid (g3), as said, is there to shield the Screen from electrons bouncing off the Anode. It is usually connected to the Cathode - ie. 0 volts. This low voltage tends to repel these electrons back to the Anode. This Suppressor Grid is instrumental in making Pentodes able to 'swing much more voltage than Tetrodes without the aforementioned problems, capable of more power (output) & more efficient valves (with less input volts). So, Pentodes are pretty impressive, huh?
|
Post by Owen Y on Jun 4, 2017 14:23:39 GMT 12
I know you've been desperately wondering (jez kiddin) why the KT88 was called a 'Kinkless Tetrode' (Beam Tetrode)? The Suppressor Grid above, also solves a problem with Tetrodes in that when the Anode handles large signals, the voltage 'swing' can send electrons back to the Screen instead - this causes a drop or 'kink' in the nice smooth current delivery of the valve (Tetrodes specifically). Not good. PS. Beam Tetrodes have lower 3rd Harmonic Distortion than Pentodes - maybe one reason why some prefer the (warmer) sound of KT88s, KT66s, 6550s - over say EL34s?
|
Post by Owen Y on Jun 4, 2017 14:45:01 GMT 12
Speaking of signal voltage swings, we should properly grasp that music electrical signal in an amp is an AC (Alternating) electrical signal - swinging +ve & -ve (around 0 volts) like a sinewave: Studying the basic Triode diagram again, you will see: - A small signal (sine wave) at the Grid (input) & a larger (anplified) signal out (coming off the Anode). (Output can also be taken off the Cathode, which can be useful, but it will be un-amplified - but 'Cathode Followers' are another chapter ) - You will have spotted immediately that, whereas the input sinewave is swinging -ve-to-+ve, the output signal is swinging +ve-to--ve - ie. the output has been mirrored or 'phase' reversed. All valve amp stages do this (except for the above-mentioned Cathode Followers). This is because when the Grid incoming signal swings negative, the Grid becomes more -ve & the electrons flowing off the Cathode are repelled (ie. throttled, Graham ) & the Anode becomes more +ve. OTOH when the Grid swings +ve, more (-ve) electrons are attracted off the Cathode & the Anode becomes more -ve.
|
Post by Owen Y on Jun 4, 2017 16:16:14 GMT 12
Finally, in answer to Gryffles' question, the valve data sheet (let's take the KT88 maximum ratings),.... Vh - heater voltage Ih - heater current Va - Anode voltage (referenced to Cathode). Vg2 - Screen voltage ( ditto ). Va,g2 - Anode/Screen voltage for 'Triode' connection (g2 is connected to Anode when in quasi-Triode configuration). Pa - Anode 'dissipation' in watts (this is the max. amount of power that the Anode can dissipate as heat - before thermal runaway & self destruction!) Pg2 - Screen dissipation max. Pa+g2 - Anode + Screen dissipation max. Ik - Cathode current max. Vh-k - Heater to Cathode voltage (if this voltage difference is high, current can 'leak' from the heater to the Cathode, causing noise pickup - eg. hum. Rg-k (cathode bias) - the max,. value of 'grid resistor' (Grid to Cathode to 'bias' the valve when 'Cathode/Self-Biased'.) Too high can be potentially thermally dangerous for an output valve. Rg-k (fixed bias) - as above but this resistance must be lower for 'fixed biased' amps. The design choice of Cathode/Self-bias or Fixed Bias (a subject in itself) would have been made in Graham's Yarland amp - at this stage we think it is Cathode Biased because we can see from the datasheet that Fixed Bias produces over twice the power output (100wpc). Note - voltages are DC voltages.
|
Post by Owen Y on Jun 4, 2017 16:31:28 GMT 12
I think it's worthwhile now looking at a typical schematic for a KT88 Push-Pull amp - Graham's Yarland might be similar to this (but with 12AX7 & 6SN7 input & phase-splitter/driver valves): (Also we think our Yarland is Cathode-biased, not Fixed-biased.)
|
Post by Graham on Jun 4, 2017 17:10:44 GMT 12
Whew ! Getting into some heavy stuff now eh Gryffles. I need a cuppa tea and a lay down after digesting that lot.
|
Pundit
Post by Gryffles on Jun 4, 2017 18:03:36 GMT 12
Yes, some of it is over my head but I get most of it.
I'll need to re-read it when I'm not so tired.
Just back from a cold days trout fishing!
|
Post by Owen Y on Jun 4, 2017 18:11:34 GMT 12
Whew ! Getting into some heavy stuff now eh Gryffles. I need a cuppa tea and a lay down after digesting that lot. Me too.... PS. Do you notice how the OPT is connected to the KT88s?
|
Post by Owen Y on Jun 5, 2017 9:44:00 GMT 12
I've updated the KT88PP schematic above to indicate 12AX7 & 6SN7 valves. I'll leave the characteristics of the 6SN7 to colin's expertise & we'll continue to examine the KT88 output stage.... If you look at Graham's photos (the 4th one in particular), you can just about see 5 wires from the KT88s disappearing into the OPT. This corresponds to the schematic above. (The OPT 'primary' winding is connected to the amplifier & the 'secondary' winding is connected to the lspkrs.) I've annotated Graham's 2nd photo:
|
Post by Owen Y on Jun 5, 2017 10:19:38 GMT 12
KT88s: You see that the Anode of ea KT88 is connected to opposite ends of the OPT primary winding & thus the KT88s operate in Push-Pull fashion to 'swing' the signal voltage. You can see also that 470 volts 'HT' (High Tension) supply is connected to the 'Centre Tap' of the Primary winding - to supply the KT88s with current. In this way, the OPT Primary (which has many turns of wire around an iron core, like a giant resistor) acts also as a nice solid high impedance AC load for the tubes - maybe 5,000 - 7,000 ohms for 2 x KT88s. You'll see also that the OPT secondary is illustrated as smaller, ie. with fewer windings & thus the high valve impedance is 'stepped down' to 4/8/16 ohms spkr impedance. (The opposite of a Step Up Transformer / SUT.) This amp schematic shows also a widely used output configuration called 'Ultra-Linear', where additional connections are provided on the primary (notes added) in order to 'feed back' a portion of the output of the Anode to the Screen of the Beam Tetrode/Pentode. This 'Negative Feedback' scheme lowers distortion & is commonly used for Beam Tetrode/Pentode amps. See - en.wikipedia.org/wiki/Ultra-linear
|
Post by Owen Y on Jun 5, 2017 10:40:37 GMT 12
Possibly a good juncture to discuss anything, rather than me just raving on (Besides, I may be reaching the extent of my knowledge!)
|
Post by Graham on Jun 5, 2017 11:08:16 GMT 12
Hi Owen
The explanation provided by you and Colin is excellent. Several spec sheets for this particular amp refer to the ability to switch the feedback on or off. This must be an optional fitment as mine has no such switch. Do I presume it runs with permanent feedback? There must be several versions or configurations as some have the on/off switch on the front, and some do not have the remote volume control that mine has. I also note that some owners have upgraded to KT120 valves for a very slight increase in output power.
Graham
|
Post by Owen Y on Jun 5, 2017 11:57:26 GMT 12
Hi Graham - Ultra-Linear, yes I would expect so, because as far as I know, there are no disadvantages of using UL config for Beam Tetrode/Pentodes (aside from more costly & trickier to manuf OPTs). The only way to check is to look carefully at the KT88 wiring, in particular g2 (pin 4) will be connected to the OPT, not to the HT supply. Is the power output specified 40wpc? The biasing arrangement is of interest too, as Cathode Bias (Self Bias) is normally self-regulating, as the name says. However - (i) in PP the 2 valves may not be identically matched, ie. unbalanced (ii) different valves will draw different current - so your amp appears to have provision for balancing and/or adjusting bias (idling) current. The alternative method or biasing an amp is Fixed Bias, where instead of just having a resistor to set the Grid g1 -ve in relation to the Cathode, a -ve PSU provides a fixed -ve voltage to the Grid. This has the advantage of (i) greater output power (ii) deletion of the Cathode Bypass capacitor - these do not appear to be visible in your amp.
|
Pundit
Post by Gryffles on Jun 5, 2017 20:42:13 GMT 12
Still all interesting for me Graham, Owen & Colin
So I have another question
When you say "the Control Grid (g1) is usually always 'biased -ve volts (in relation to the Cathode)."
What actually causes the Control Grid to become biased negatively?
Hope that makes sense!
|
Post by Owen Y on Jun 5, 2017 23:02:19 GMT 12
So I have another question When you say "the Control Grid (g1) is usually always 'biased -ve volts (in relation to the Cathode)." What actually causes the Control Grid to become biased negatively? Thanks, good question! There are a few ways, but 2 common ways: (We can refer back to the basic Triode valve diagram from earlier, showing the sinewave signals....) (i) One way is to simply apply a -ve voltage to the Grid (as indicated above) - but this requires a small PSU (or battery supply) to generate the -ve DC volts. (Fixed Bias) (ii) The more common way is to insert a Cathode Resistor (Rk) between the cathode & Ground - & the quiescent (idle) current running down through the valve will result in a +ve voltage developed across the Cathode R (due to Ohm's Law V = I x R). Thus the Grid becomes -ve in relation to the Cathode. (Self Bias, Cathode Bias, Auto Bias.) (The Cathode R normally has a 'Bypass Capacitor' in parallel, but that's another story )
|
Post by colinf on Jun 6, 2017 19:47:44 GMT 12
Owen's diagram above, with the resistor in series with the cathode, is the most common circuit in the world. Easy to see why, it is very convenient to implement. All you do is put a resistor in series with the cathode and have the control grid connected directly to the audio source. The resistor in the cathode results in a very small amount of negative feedback which reduces the gain of the triode. A 12ax7 has an endemic gain of 100. Theoretically if you put 0.2 volts of audio signal into the grid, the output would be 20 volts, which is a lot. Thus the 12ax7 is considered a high gain valve. In a practical circuit like the one above though, the negative feedback, and the resistor load on the anode conspire to make it have a gain of (depending on the resistor values chosen) 60 to 70 times. A 6sn7 has a theoretical gain of about 20, but is capable of over 10 times the current drive of a 12ax7 which makes it a handy choice as a driver tube for output tubes like the KT88.
AMR-iFi R&D
|
Pundit
Post by Gryffles on Jun 6, 2017 21:01:03 GMT 12
Great. That makes sense to me Owen Y & colinfThanks for the explanations.
|
Post by Owen Y on Jun 7, 2017 17:27:45 GMT 12
Just referring back to an earlier point...... - You will have spotted immediately that, whereas the input sinewave is swinging -ve-to-+ve, the output signal is swinging +ve-to--ve - ie. the output has been mirrored or 'phase' reversed. All valve amp stages do this (except for the above-mentioned Cathode Followers). If you look at the KT88 amp schematic, you'll see that it has 3 stages - 12AX7 -> 6SN7 -> KT88. So, it would appear here that the output of this amp would be phase-inverted (in relation to the input). You would expect that the amp designer and/or OPT maker would address this at the OPT connections, but I always check correct-phase myself Graham, by swapping +/- lspkr connections at the back of the amp - listening for good stereo spread, soundstage width & lateral placement & focus in particular. When phase is inverted, typically, the sound is 'clumped ' in the middle, sometimes bass can appear stronger but confined very 'central'. (With an oscilloscope of course, this can be checked on the bench.)
|
Post by colinf on Jun 8, 2017 11:08:13 GMT 12
The absolute phase on the shown circuit diagram should be correct, although it isn't specified how the output transformer windings are oriented. (Usually marked with a dot for winding beginning, but not shown here.) From input to output: input to 12ax7 normal phase; input to top 6sn7 driver half is inverted; input to top kt88 is normal phase; input to top half of transformer is inverted; the inverted output of the transformer is grounded on the secondary which means the output is normal phase.
AMR-iFi R&D
|
Post by Owen Y on Jun 8, 2017 12:50:41 GMT 12
As said, looks inverted to me, but I would always check - either by listening and/or definitely on the bench if you have a scope (or using one of these clever devices... www.parts-express.com/galaxy-audio-cricket-polarity-test-set--245-261 ). Most audio transformers I've seen have windings start/finish mirrored input/output & the easiest way to invert is at the connections. Traditionally, transformer leadout wires would be usefully colour-coded, eg: Red (B+), Blue (plate, finish lead of prim), Green (sec. 'hot' end of voice coil), Black (return).
|