Home |
Search |
Today's Posts |
#1
|
|||
|
|||
My first amp -- opinions?
Digital guy comes to tube land. I tried my first amp and like the sound. If anyone is interested: http://wosch.teratronik.com/amp/amp.html Maybe you RATs can point me to all those big blunders I made... I shamelessly posted a few questions onto that webpage and would be happy for a few hints and tricks. Keep glowing... -- Olav Wölfelschneider |
#2
|
|||
|
|||
Hi RATs!
Welcome! You did the most important thing right: Got it working and listened Happy Ears! Al Alan J. Marcy Phoenix, AZ PWC/mystic/Earhead |
#3
|
|||
|
|||
Hi Olav,
Maybe you RATs can point me to all those big blunders I made... Hmmm, let's see ... - you feed the screen (g2) from an OPT tab more positive (nearer to B+) than the plate. This is very unusual. Is this a drawing error or intentional? If intentional, what's the objective behind that? - the cathode resistors (R3, R7) of the ECC83 systems seem to be extremely high (probably since you wanted to reduce the gain of both stages). What plate current do both ECC83 systems draw (each)? At least for the gain/driver stages there probably are better alternatives than using high gain (mu=100) sections and then "destroying" the gain by such extreme plate/cathode resistor ratios. At least on first sight I see no point in using two such stages after another when a single one with a less mu type triode section would do instead. Or am I missing something? Tom -- If in doubt, mumble. |
#4
|
|||
|
|||
Tom Schlangen wrote: Hi Olav, Maybe you RATs can point me to all those big blunders I made... Hmmm, let's see ... - you feed the screen (g2) from an OPT tab more positive (nearer to B+) than the plate. This is very unusual. Is this a drawing error or intentional? If intentional, what's the objective behind that? This is usually called "Ultra-Linear" (UL), although originally UL referred to a specific tap ratio (43% if I remember correctly). It applies local feedback to the screen grid, and basically acts as a topology halfway between normal pentode operation, and triode operation (you could view triode operation as UL with 100% tap). - the cathode resistors (R3, R7) of the ECC83 systems seem to be extremely high (probably since you wanted to reduce the gain of both stages). What plate current do both ECC83 systems draw (each)? At least for the gain/driver stages there probably are better alternatives than using high gain (mu=100) sections and then "destroying" the gain by such extreme plate/cathode resistor ratios. At least on first sight I see no point in using two such stages after another when a single one with a less mu type triode section would do instead. There may well be lower total distortion using this approach, since each stage will have a lot of local feedback to linearize the response. Cheers, Fred -- +--------------------------------------------+ | Music: http://www3.telus.net/dogstarmusic/ | | Projects, Vacuum Tubes & other stuff: | | http://www.dogstar.dantimax.dk | +--------------------------------------------+ |
#5
|
|||
|
|||
Hi Fred,
This is usually called "Ultra-Linear" (UL), although originally UL referred to a specific tap ratio (43% if I remember correctly). Oh, Fred I, would simply call it pentode mode; there's not a little bit UL in this circuit ;-) Tom -- The first rule of magick is simple: Don't waste your time waving your hands and hoping, when a rock or club will do. |
#6
|
|||
|
|||
Hi Fred,
At least for the gain/driver stages there probably are better alternatives than using high gain (mu=100) sections and then "destroying" the gain by such extreme plate/cathode resistor ratios. At least on first sight I see no point in using two such stages after another when a single one with a less mu type triode section would do instead. There may well be lower total distortion using this approach, since each stage will have a lot of local feedback to linearize the response. This may be, but: A certain ammount of K2 distortion even may be desired to cancel out the distortion of the power stage. And probably the EL504 (a TV sweep circuit power tube) generates quite some distortion in pentode (not UL) mode. Without having listened to such a design of course, just from viewing the schematics, I still think that the gain of roughly 4 x 4 = 16 of both stages in that topology more easily and straight forward could be obtained from other circuits with less parts (e.g coupling caps) with more "fitting" tubes, than with two capacitor coupled grounded cathode stages using a high mu double triode like 12AX7/ECC83. I mean, a gain of 16 really *screams* for other tubes than a 12AX7/ECC83, and for a single stage with a medium mu triode (even without a cathode bypass cap). At least, this is a rather unconventional way of doing things, so maybe we both miss the point - lets see what the author had in mind if he cares to elaborate on this circuit and why he used this topology and what its benefits are. Tom -- Life: All in all a bad game, but graphics are really good. |
#7
|
|||
|
|||
Olav Wölfelschneider wrote:
Digital guy comes to tube land. I tried my first amp and like the sound. If anyone is interested: http://wosch.teratronik.com/amp/amp.html You may want to add in parallel a cap across R12, the 1K ultralinear tap resistor. G2 needs a low impedance source to operate in ultralinear mode properly. Not sure if 1K is low enough or not. G2 will draw some plate current (which is varying according to what the control grid is doing) and this varying current will (in a nonlinear fashion) vary the voltage on g2 in addition to what the ultralinear tap and quiescent g2 current is doing. |
#8
|
|||
|
|||
"Tom Schlangen" wrote in message
... This is usually called "Ultra-Linear" (UL), although originally UL referred to a specific tap ratio (43% if I remember correctly). Oh, Fred I, would simply call it pentode mode; there's not a little bit UL in this circuit ;-) Hi Tom I've seen something like this done locally, selecting the taps on a similar cheap PA output transformer to make a P/P ultralinear tube transformer (a use for which it was _never_ intended). If you look closely at the schematic, the screen is fed at some part of the total primary impedance, but whether it's the usual ~43% is beyond my knowledge, or maths ability (I suspect it's a bit higher). It is, however, a basic UL circuit ..... Cheers David |
#9
|
|||
|
|||
Hi Olav,
regarding the question on your homepage about NFB and oscillation: Obviously you got positive feedback - so simply ground the other end of the OPT secondary winding and get the feedback voltage from the (now) not grounded end. Tom -- The first rule of magick is simple: Don't waste your time waving your hands and hoping, when a rock or club will do. |
#10
|
|||
|
|||
Hi Olav,
You give me way too much credit. I still have much theory to learn. Hey I am a beginner in tube DIY myself. Just don't be confused with all these fancy concepts out there to consider, but stay with common sense and stone-age rules like "best amplifyer is a wire with gain". I think most of tube circuit benefits (including good sound) are comming from the sheer simplicity of very few active elements in simple circuits. Think about why e.g a 6SN7 is considered to be the most linear active device in audio technology invented and known so far in universe :-) Where there is no fuss, you don't have to cope with correcting it. Keep it simple, understand it and learn from it. At least these directives worked for me :-) Anyway the driver stage was once from a historic pre-amp design where there was a tone-control network between the two stages. Of course, this ate away a lot of the gain, so they took two stages. Ah, now this makes sense. I built the tone network, didn't like it and removed it. Now it seems I have too much gain on my hands. So you have at least two options to improve on your circuit now: - Use NFB to reduce gain by intention to gain more over-all linearity and less distortion instead of gain (just try it, but don't misuse it apply too much to cure other design faults. 6-9 dB should be enough for a "common" two stage SE design using "common" valves and circuits. EL504 probably will need more than that to be tamed together with the OPT you use - go for less gain (which means go for the gain you really need) without NFB, using less parts, and so on. You have to try out to find the way you prefer sonically and technologically wise. As a beginner myself with only few amps built so far I prefer to just keep it simple, so I have a chance to really understand what is going on. I'm thinking about trying the SRPP stage from http://www.ndh.net/home/kboehm/kt88.htm instead. Ah, Mr. Boehm, from the same very little provider I have my own home at. The circuit shown at Mr. Boehms site is quite generic, but of course there is absolutely nothing wrong with that. Try it, measure it (as he suggest, with diffent tube types) and understand how it works! Be aware that you probably have to lift heaters so you don't exceed max. heater/cathode of the tube you intend to use. But, you should consider wether a SRPP circuit suits your intention _besides_ being know to be okay within certain design rules and margins in the first place. In my first project: http://www.ndh.net/home/schlangen/ro...ecl82pset.html I had an ECL82 triode section to spare per channel, so I could have used a SRPP circuit instead of the simple grounded cathode stage I used in the end. Or I could have built a grounded cathode stage with a cathode follower (but what for? no hard load to drive ...) or a cascode or a constant current draw pair, or what else one could do with two similar triode sections. I simply stayed with the single triode section because it worked nicely and I did understand its operation at that moment. Not need for "fancy" stuff like SRPP there in this circuit at that moment, which I wasn't able to understand at that time anyway. Some months later and after trying other things (like SRPP, cascodes, or simple paralleling of triode sections) on the bench, I designed and built the second amp: http://www.ndh.net/home/schlangen/ro...t/el34set.html and although being eager to apply all what I have learned from the first amp and all experiments thereafter, I stayed with simplicity and just paralleled two triode driver sections because this approach just does and _delivered_ what was needed. And the resulting amp simply sounds gorgous. This doesn't keep me from experimenting - just the opposite is true. But for things intended not for experimenting but for daily use I just keep it simple, reliable and proven, and concentrate on the really important "other" issues, like electrical safety, and so on. using an ECC83? As said already, ECC83 is a very high amplification factor tube. ECC83 definitely is very suitable for its intended purpose, but that is just the point ... if you don't need this high gain, there are better (e.g more linear) candidates. As for the tube selection... since I'm just beginning my stock of tubes is limited. I've got several EL504 (which got me started with tubes) a few low grade ECC83 from EI/Jugoslavia and some russians. I wouldn't worry about brands at the moment. Just primarily go for a correct electrical design, and after that start tube brand rolling and after that just use what you like yourself most sonically wise (in a certain circuit). Then think about why this tube of the same type sounds other than that other tube in the same circuit and draw your conclusions, and use what you like, ignoring what other folks say. At that moment you probably are already one step above those folks who say that Telefunken, Amperex buggle boys or whatelse is fashionable at the moment is "the best brand" without knowing what they are talking about. Just remember that a tube circuit not only consists of a tube, less than a certain tube *band*. So, just find out yourself about certain brands and circuits and so on, but before that built a solid base of knowledge and understanding to derive your opinion from. I tried something with two of the russians, but they where microphonic like hell... (6S2P & 6SH49P-DR) So you probably snipped your fingernail against some of them and heared a more ore less loud click out of your speakers compared to other brands. Okay, but the question is, if this really matters: Do you snip your fingers against the tubes in the amp you are listening *music* with? Do you have 100dB efficiency horns that shake the ECC83 input tubes when driven by your EL504 output stage so hard that there are audible interferences? Do you have a very low signal level stage (e.g phono inputs) with these tubes, were microphonics of input tube really is an issue? It is good that you consider such things, but don't let them deflect you from what *really* matters for the intended purpose you design the actual circuit for. Right now the amp is disassembled, as I am working on a chassis. That is the hard and no-fun stuff, if one doesn't have appropiate tools and doesn't want to fall back to generic off-the-shelf chassis. Maybe my homepage can give you some hints for chassis construction; there are (short) sub-chapters with pictures how these chassis are made. Welcome to the wonderful world of tube! Tom -- MS-DOS is the worst text adventure game I have ever played: Poor vocabulary, weak parser and boring storyline. |
#11
|
|||
|
|||
Hi Olav,
I wrote: - go for less gain (which means go for the gain you really need) without NFB, using less parts, and so on. E.g by dropping one of the two gain stages and optimize the remaining one :-) Tom -- Anyone who considers protocol unimportant has never dealt with a cat. - R. Heinlein |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Opinions Needed on New DIY Radio Site | Tech | |||
Wharfdale Opinions | Audio Opinions | |||
Opinions on Sub | Car Audio | |||
Opinions on a digital audio workstation? | Pro Audio | |||
sub $2000 rackmount mixer for project studio -- mic pre opinions | Pro Audio |