dan@rna.UUCP (02/04/87)
I don't want to start a lot of controversy with this query, just some "straight" answers, if possible. I was reading in TAS (The Absolute Sound) their technical evaluation of the Jadis 200 tube Class A amplifier. In that article they brought forth the contention that solid state rectifiers (I believe they were refering to those in the power supply, correct me if I'm wrong) degrade the sound of the amplifier relative to the use of tube rectifiers. They also suggested that the Jadis benefited from the the lack of solid state power regulation. Now I'm (I hope) fairly open minded. So without getting into a lot of snide remarks (I could generate plenty), any sincere, shot-in-the-dark speculations as to why tube rectifiers would be preferable to suitably rated solid state diodes ? It can't be noise, tubes should be worse and it should all be filtered out by supply caps (1500mfd total via resistors). Perhaps its switching transients, but they, too, should be filtered out. Solid state should have lower voltage drop with any decent amount of current and larger current capacity. So I really don't have any good idea why... Solid state regulation versus no regulation, I could perhaps see. Fast, high current peaks should be more readily delivered with no regulation. I don't know if they (or anyone) claims that solid state regulation would be sonically worse than tube regulation. Here again, I can't see why...
straka@ihlpf.UUCP (02/06/87)
> the contention that solid state rectifiers (I believe they were refering to > those in the power supply, correct me if I'm wrong) degrade the sound of the > amplifier relative to the use of tube rectifiers. They also suggested that > Any ... > speculations as to why tube rectifiers would be preferable to suitably > rated solid state diodes ? I'll join you in getting flamed. This sounds totally ridiculous, perhaps even religious! Here we go again. I must admit that it sometimes gets entertaining. -- Rich Straka ihnp4!ihlpf!straka
ken@rochester.UUCP (02/06/87)
In my opinion, if one can afford to pay the kind of money asked for this kind of equipment and you like the sound, what does it matter what scientific explaination, real or imagined, one comes up with? Don't get me wrong, I'm not against scientific inquiry. But somebody who simply says "I like the sound of this better" is being more honest with himself than one who says "I bought this because it has 0.0004% fifth harmonic distortion as compared to the other one which has 0.001%". I enjoy listening to music a lot on my crufty old turntable, not even CD, and I wonder some days whether some audiophiles listen to the music or for the noise. (Just think, in the time it takes you to type in another diatribe against bipolar output stages, you could have heard another Corea piano solo :-).) Ken
dsi@unccvax.UUCP (02/06/87)
In article <598@rna.UUCP>, dan@rna.UUCP (Dan Ts'o) writes: > > I don't want to start a lot of controversy with this query, just > some "straight" answers, if possible. > > I was reading in TAS (The Absolute Sound) their technical evaluation > of the Jadis 200 tube Class A amplifier. In that article they brought forth > the contention that solid state rectifiers (I believe they were refering to > those in the power supply, correct me if I'm wrong) degrade the sound of the > amplifier relative to the use of tube rectifiers. They also suggested that > the Jadis benefited from the the lack of solid state power regulation. But of course! The typical Thevenin impedance of a vacuum tube rectifier varies with load, but is about 50 to 100 ohms. Furthermore, the vacuum tube rectified DC power supply cannot be modeled as simply a resistor and an ideal voltage source. The supply impedance varies with the current demand of the PA tubes in the amplifier (mostly); therefore, we conclude, the power supply + amplifier combination exhibits nonlinearity because the operating point of the PA tubes is being jerked around. Now, we all know that the current drawn by a class A amplifier is constant, right? However, the PLATE DISSIPATION IS NOT CONSTANT WITH POWER DELIVERED TO THE LOAD. (If you don't believe this, I'll invite you to roast marsmellows in the exhaust air stream of my Julian Betts AB10SG AM transmitter with the excitation turned off...) In fact the plate dissipation does not vary linearly with the square of the load current because: 1) The active devices are vacuum tubes 2) The driving point impedance looking into the plate-to-speaker matching transformer varies with frequency, amplitude, room acoustics, etc. 3) Plate dissipation increases at higher frequencies. (If you don't believe this, connect a 20 kHz sinewave generator to your amplifier and a suitable load, and dial in maximum power. Watch the plates of the tubes - or measure the airstream temperature). This is due primarily to core losses in the plate transformer. Why all this grumbling about plate dissipation? Well, in an ideal class A amplifier, the power is either being dissipated in the plates, or in the load, and the ratios are proportional to tube plate current or load current, squared, and respectively. In a nonideal class A amplifier using real world tubes, the supply current does vary somewhat with the power delivery to the load, particularly as the grid voltage is approaching cutoff of one of the tubes or as the control grid voltage of one of the tubes becomes nonzero. (You can achieve more output by drawing grid current, but this places unbelieveable demands on the driver amplifier). Thus, the small signal model of a class "A" vacuum tube amp fails for the golden ears. Thus, since the vacuum tube rectified power supply presents a nonlinear effective impedance, and the supply current varies with power delivered to the load, we have an additional active device introducing distortion into the recovered waveform. I would suspect two things: 1) As the tubes reach saturation, the vacuum tube rectifier will cause the peaks of the reproduced waveform to assume more second order distortion. The solid state rectification with or without regulation causes the tube at saturation to assume more odd harmonic distortion. 2) That the amplifier would sound mellower with age (now, what we want are 20 year old 5R4GY's out of Grandpa's TV set, right?) due to aforementioned characteristics. Vacuum tube regulation is notoriously disgusting. Any decent zener diode can blow off such wonderful vacuum tube reference sources as the 0A2. Any decent pass transistor can supply much tighter control (in the right circuit) than pass tubes such as the 4CX250D and the 6080. There are probably some golden ears out there who use this stuff, but why bother. I don't deny that the vacuum tube amplifier sounds different, and there is a certain romance associated with glowing filaments in a darkened room. However, I propose a test. Take two wirewound resistors, rip out the 5U4 (or whatever) in your tube amplifier, and replace the tube with the wirewound resistors in series with the silicon diodes you will replace them with. Better still, find some positive temperature coefficient resistors with a reasonable thermal lag. Then, do some listening tests. Phase II: Replace the standard thermionically heated vacuum tube rectifier with some mercury vapour rectifiers (e.g. 866A's) and do some tests (the voltage drop across mercury vapour tubes is virtually constant with supply current). Phase III: Place an oscilloscope lead at the center tap of the plate transformer of your class A amplifier if you believe that such devices are ideal... Pmmmmmph. David Anthony DataSpan, Inc.
shop@uwmcsd1.UUCP (02/07/87)
> > the contention that solid state rectifiers (I believe they were refering to > > those in the power supply, correct me if I'm wrong) degrade the sound of the > > amplifier relative to the use of tube rectifiers. They also suggested that > > > Any ... > > speculations as to why tube rectifiers would be preferable to suitably > > rated solid state diodes ? > > I'll join you in getting flamed. > This sounds totally ridiculous, perhaps even religious! > Here we go again. I must admit that it sometimes gets entertaining. > > -- > Rich Straka ihnp4!ihlpf!straka One major problem using solid state diodes in my power amps is that there is a large current surge on turnon and unless properly ventilated (I have the amps tilted up using shot glasses) the diodes will melt down and short. - Tom tjk@csd4.milw.wisc.edu -- Thomas Krueger ...ihnp4!uwmcsd1!uwmcsd4!tjk or University of Wisconsin Milwaukee tjk@csd4.milw.wisc.edu Computing Services, Electronics Shop 3200 N. Cramer St. (414) 963-5172 Milwaukee Wi 53211
ken@rochester.UUCP (02/08/87)
|One major problem using solid state diodes in my power amps is that there |is a large current surge on turnon and unless properly ventilated (I have |the amps tilted up using shot glasses) the diodes will melt down and short. Excessive turnon transients will blow out your diodes right away. In designing a power supply, one calculates the transient current on the assumption that the capacitors are a dead short and that the maximum voltage is applied. The current is limited by the total impedance (secondary + reflected primary). The maximum transient current is given in power diode specs. If the impedance is not enough, then one adds a little resistance in series with the diodes. If the diodes melt later that is due to inadequate heat removal. Correct solution, wrong cause blamed. A different issue is soft turn on. Some power supply regulators ramp up the voltage gently so that the output stages don't cause plops in the speakers. Ken
henry@utzoo.UUCP (Henry Spencer) (02/08/87)
> the contention that solid state rectifiers (I believe they were refering to > those in the power supply, correct me if I'm wrong) degrade the sound of the > amplifier relative to the use of tube rectifiers... They may perhaps be thinking of the use of switching power supplies, which are invariably solid-state, versus linear supplies. Switching supplies do have a rather noisier output, less suited for precision analog hardware like stereo amplifiers. But there's no need to resort to tubes to build a linear supply, since semiconductor ones work fine. I can't think of anything else that would account for this curious claim, except the various forms of religious fervor that pervade high-end audio. (I recall an interesting article which pointed out that the single most cost-effective way to improve the sound from your stereo is to have your ears cleaned by a doctor. I wonder how many of the audio fanatics have bothered to do that?) Mind you, I am not an expert on them there strange analog thingies (notably those built with discrete components: op-amps I understand, but transistors are still inscrutable all too often...); give me digital any day. -- Legalize Henry Spencer @ U of Toronto Zoology freedom! {allegra,ihnp4,decvax,pyramid}!utzoo!henry
caf@omen.UUCP (02/09/87)
In article <615@unccvax.UUCP> dsi@unccvax.UUCP (Dataspan Inc) writes:
: Now, we all know that the current drawn by a class A amplifier is
:constant, right? However, the PLATE DISSIPATION IS NOT CONSTANT WITH
:POWER DELIVERED TO THE LOAD. (If you don't believe this, I'll invite you
:to roast marsmellows in the exhaust air stream of my Julian Betts AB10SG
:AM transmitter with the excitation turned off...) In fact the plate
:dissipation does not vary linearly with the square of the load current because:
No wonder you can cook marshmallows that way. Most AM transmitters use class
C for the finals, not class A. If you remove excitation from a typical tube
type final amp with "self bias" developed be rectification of the grid drive,
you will indeed toast marshmallows and whatever else is handy.
I would also suspect the mudulator output tubes are run class B or AB, not
class A. There may not be a class A high power stage in that transmitter.
wam@cdx39.UUCP (02/09/87)
[verbiage about thermionic vs. solid state rectifiers affecting sound...] You could make a case for tube diodes in an instrument amp, where the artist will bend it to its limits. When pushed, the high resistance of thermionic diodes will drag down the supply voltage in an interesting fashion, not easily duplicated by solid-state gadgets. At home, listening to recorded or broadcast programs, I don't think you want to set it so loud that the amp clips (also different with tubes than x-istors). Neither do you want the audio peaks to drain the power supply of its vital working fluid-- still, I suppose tubes wilt more gracefully than silicon... -- Bill MacLeod telephone: 617/364-2000x7520 Email: ...{cthulhu,inmet,harvax,mit-eddie,mot[bos],rclex}!cdx39!wam Smail: Codex Corporation; Mailstop C1-65; 20 Cabot Blvd; Mansfield MA 02048 send two boxtops and one thin dime for standard non-disclosure agreement...
dsi@unccvax.UUCP (02/09/87)
In article <471@omen.UUCP>, caf@omen.UUCP (Chuck Forsberg WA7KGX) writes: > In article <615@unccvax.UUCP> dsi@unccvax.UUCP (Dataspan Inc) writes: > : Now, we all know that the current drawn by a class A amplifier is > :constant, right? However, the PLATE DISSIPATION IS NOT CONSTANT WITH > :POWER DELIVERED TO THE LOAD. (If you don't believe this, I'll invite you > :to roast marsmellows in the exhaust air stream of my Julian Betts AB10SG > :AM transmitter with the excitation turned off...) In fact the plate > :dissipation does not vary linearly with the square of the load current because: > > No wonder you can cook marshmallows that way. Most AM transmitters use class > C for the finals, not class A. If you remove excitation from a typical tube > type final amp with "self bias" developed be rectification of the grid drive, > you will indeed toast marshmallows and whatever else is handy. > > I would also suspect the mudulator output tubes are run class B or AB, not > class A. There may not be a class A high power stage in that transmitter. Actually, it is class "AB" at the present time. We have modified it so that it will operate with low level modulation. There are high level mod tubes in a Julian Betts AB10SG. The reason for this is so that we could operate the transmitter at 4700 watts critical hours and 500 watts postsunset. Nevertheless, the Betts final stage can be operated and biased class "A" if one is so inclined. The transmitter was originally screen modulated class "C" but this blows weeds.... David Anthony DataSpan, Inc.
cgs@umd5.UUCP (02/09/87)
>>> Any speculations as to why tube rectifiers would be preferable >>> to suitably rated solid state diodes ? > >One major problem using solid state diodes in my power amps is that there >is a large current surge on turnon and unless properly ventilated (I have >the amps tilted up using shot glasses) the diodes will melt down and short. >Thomas Krueger ...ihnp4!uwmcsd1!uwmcsd4!tjk or >University of Wisconsin Milwaukee tjk@csd4.milw.wisc.edu Tom, As you should well know, the current surge at turn-on is due to the fact that the large power supply caps are charging up, not due to anything inherent in the diodes themselves. The diodes themselves have a sufficient thermal mass so that the junction temperature does not rise anywhere near the point of damage during the cap charging at turn-on. While the peak power of the current surge may be large, the duty cycle is very small, the net effect is that what you've stated is a non-problem. However, lack of ventilation during continuous operation of the equipment can elevate the temperatures inside the enclosure, hence causing junction temperatures to soar. Excessive junction temperatures will cause permanent damage. Also, I would expect any well designed equipment to operate without damage in a well ventilated environment that you would find very uncomfortable -- about 40 deg. Celsius (104 deg. Fahrenheit), 90% Rel. Humidity (non- condensing). -- --==---==---==-- .. The jaws that bite, the claws that catch! .. ARPA: cgs@umd5.UMD.EDU BITNET: cgs%umd5@umd2 UUCP: ..!seismo!umd5.umd.edu!cgs
muller@sdcc7.UUCP (02/09/87)
Until recently most of the power supplies I saw in audio equipment were far from what one would call high quality. It was only recently that the equipment manufacturers started to actually include reasonable supplies. In terms of the quality of the supply you have to look at it's stability under load (it has to feed loads whose impedence rapidly changes with time) and the quality of its output. Most supplies have a fair residual noise content (this include both series pass and switching supplies) as does the majority of batteries. A good power supply will have much lower noise levels than any available battery. Switching supplies which do not have carefully designed filtering and shielding will pass harmonics of its switching rate onto it's output. Series pass supplies (either have one of more transistors or tubes... yes tubes are used in very high voltage supplies) will in general be a lot quieter, but since they regulate by dumping execess output through heat they are a lot larger and consume a lot more power. Really what you have in a good power supply is a closed loop feedback system not one of those full wave brigde rectifier systems were filtering is done by the addition of coils and capacitors. Of course all the design features problems etc of feedback systems apply. All supplies have the following basic ideas: --------------------- ------------------- --------- ---------- | raw supply source |->| raw supply system|->|regulator|-->|protection|-->out | ac 120v for example| ------------------- ---------- ---------- | --------------------- ^ | | ----------- ---------- | | --->| reference|->|comparison|----- ---------- | | supply | | circuit |<------|load sense|<---------- ----------- ---------- ---------- Depending on the type of supply (series pass, switching etc) the exact contents of the boxes will vary (and where physically the sense is done is also important). A good supply should not blow up on power up, in fact one of the ways power supply manufactures use to test supplies is to test its ability to recover under high switched rate SHORT circuit loads, good supplies will recover to rated load in milliseconds. Most of the supplies have current and overvoltage protection for the device it is driving. Current protection limits the current drawn and overvoltage (aka crowbar circuits) limit the voltage in case of circuit failures. How anyone could HEAR tubes in the audio output after this mess is beyond me. What they could hear is a bad supply, but TUBES???? Keith Muller University of California muller@sdcsvax.ucsd.edu
jeffw@midas.UUCP (02/10/87)
In article <1663@uwmcsd1.UUCP> shop@uwmcsd1.UUCP (Thomas Krueger) writes: >One major problem using solid state diodes in my power amps is that there >is a large current surge on turnon and unless properly ventilated (I have >the amps tilted up using shot glasses) the diodes will melt down and short. Since the time constant associated with the turn-on surge in any power supply I have ever seen (and that's quite a few) is at least an order of magnitude less than the thermal time constants of common power semiconductor packages, it's hard for me to imagine how ventilation could have much to do with it, unless the design was marginal in the first place. Which then is a problem with the design, not solid-state rectifiers in general. Could you elaborate a bit? Jeff Winslow
henry@utzoo.UUCP (Henry Spencer) (02/10/87)
> ...somebody > who simply says "I like the sound of this better" is being more honest > with himself than one who says "I bought this because it has 0.0004% > fifth harmonic distortion as compared to the other one which has 0.001%". Almost certainly, the right explanation for tube fanaticism is that its adherents simply like the type of distortion that vacuum tubes introduce. Our electronics man (who grew up with vacuum tubes) noticed my previous posting and chatted a bit about the horrors of trying to make vacuum tubes amplify signals cleanly. His parting comment: "linear, they're not!". -- Legalize Henry Spencer @ U of Toronto Zoology freedom! {allegra,ihnp4,decvax,pyramid}!utzoo!henry
rdp@teddy.UUCP (02/10/87)
In article <615@unccvax.UUCP> dsi@unccvax.UUCP (Dataspan Inc) writes: ->In article <598@rna.UUCP>, dan@rna.UUCP (Dan Ts'o) writes: ->> ->> I don't want to start a lot of controversy with this query, just ->> some "straight" answers, if possible. ->> ->> I was reading in TAS (The Absolute Sound) their technical evaluation ->> of the Jadis 200 tube Class A amplifier. In that article they brought forth ->> the contention that solid state rectifiers (I believe they were refering to ->> those in the power supply, correct me if I'm wrong) degrade the sound of the ->> amplifier relative to the use of tube rectifiers. They also suggested that ->> the Jadis benefited from the the lack of solid state power regulation. -> -> But of course! The typical Thevenin impedance of a vacuum tube rectifier ->varies with load, but is about 50 to 100 ohms. Furthermore, the vacuum tube ->rectified DC power supply cannot be modeled as simply a resistor and an ->ideal voltage source. The supply impedance varies with the current demand ->of the PA tubes in the amplifier (mostly); therefore, we conclude, the power ->supply + amplifier combination exhibits nonlinearity because the operating ->point of the PA tubes is being jerked around. -> -> .... -> -> In fact the plate dissipation does not vary linearly with the square -> of the load current because: -> -> 1) The active devices are vacuum tubes -> 2) The driving point impedance looking into the plate-to-speaker -> matching transformer varies with frequency, amplitude, room -> acoustics, etc. -> ->Pmmmmmph. ->David Anthony Pmmmmmph, thyself, Mr. Anthony. I (and many others) would be most interested (and possibly amused) by your explanation of how room acoustics affect plate dissipation. Since we are talking about Thevenin equivalents and all that, what is the Thevenin equivalent circuit of the loudspeaker acting as a generator of electrical signals caused by "room acoustics". Well, let's see, the equivalent radiation resistance of a driver reflected back through to the driver terminals is a few milliohms at best, and given the fact that most dynamic louspeakers have a series DC resistance of anywhere from 3.5 to 7 (a typical range) I suspect that the effect of room acoustics is pretty much non-existant. Unless of course your talking about the wind generated by somebody's phenominal subwoofers cooling the tubes enough to change their operating points :-) Dick Pierce
shop@uwmcsd1.UUCP (02/11/87)
> In article <1663@uwmcsd1.UUCP> shop@uwmcsd1.UUCP (Thomas Krueger) writes: > > >One major problem using solid state diodes in my power amps is that there > >is a large current surge on turnon and unless properly ventilated (I have > >the amps tilted up using shot glasses) the diodes will melt down and short. > > Since the time constant associated with the turn-on surge in any power supply > I have ever seen (and that's quite a few) is at least an order of magnitude > less than the thermal time constants of common power semiconductor packages, > it's hard for me to imagine how ventilation could have much to do with it, > unless the design was marginal in the first place. Which then is a problem > with the design, not solid-state rectifiers in general. Could you elaborate > a bit? > > Jeff Winslow Certainly. The amp is a Dyna Mk III. I added a large filter choke, about 10 HY's in series with the Dyna choke, which is in basically a pi filter. To get the added space, I took out the 5AR4 (which I have also been told is famous for heater-to-cathode shorts, but that's another story), and replaced with 1000PIV/2.5A diodes which were bypassed with .005MFD/3KV ceramic capacitors in an effort to keep the turnon surge away from the diodes. These diodes were placed under the chassis on the inside. If I would leave the amp on for a long time (like three days) enough heat would build up to the point where either a diode would melt down and short spontaneously, or more rarely (why? I dunno) on turnon. I seriously doubt that the design in this case was marginal, although it looks like the ventilation certainly was. The chassis was never designed to be a welcome environment for solid state devices. - Tom (btw, mail to csd4, not csd1) -- Thomas Krueger ...ihnp4!uwmcsd1!uwmcsd4!tjk or University of Wisconsin Milwaukee tjk@csd4.milw.wisc.edu Computing Services, Electronics Shop 3200 N. Cramer St. (414) 963-5172 Milwaukee Wi 53211
rep@genrad.UUCP (02/11/87)
In article <1669@uwmcsd1.UUCP> shop@uwmcsd1.UUCP (Thomas Krueger) writes: > >Certainly. The amp is a Dyna Mk III. I added a large filter choke, about >10 HY's in series with the Dyna choke, which is in basically a pi filter. >To get the added space, I took out the 5AR4 ...... and >replaced with 1000PIV/2.5A diodes which were bypassed with .005MFD/3KV >ceramic capacitors in an effort to keep the turnon surge away from the >diodes.If I would leave the amp on for a long time (like three days) >enough heat would >build up to the point where either a diode would melt >down and short spontaneously, or more rarely (why? I dunno) on turnon. I'm not familiar with the Dyna Mk III and you didn't mention the power supply voltage or the full secondary voltage of the transformer. Assuming the power supply was designed for let's say 350 volts output under full load with vacuum tube rectifiers, the transformer would probably be rated at something like 400 volts RMS each side of center tap (further making the reasonable assumtion that a vacuum tube design would use a full-wave-center-tap type rectifier circuit). This would give you a PIV requirement of about 1130 volts ( 400 * 2 * sqrt(2) ). Perhaps you're really having a breakdown voltage problem, aggravated by the high temperature, not a current surge problem. It might be worth trying higher PIV rectifiers. Is the rest of the amplifier happy with the increased voltage resulting in the change to silicon rectifiers? It was sometimes necessary to take steps to reduce the voltage when conversions like this were made. You can use NTC (negative temp. coeff.) elements to limit inrush currents or take the approach that my old Heath 70 Watt (dual 6550) power amps used which was a weird element that they called a surgistor. It had a bimetallic thermostat element adjacent to a small heating element connected in parallel with it; this assembly was in series with the transformer primary. When the amp was turned on, the resistance of the heating element limited the current and heated the bimetallic element which shorted out the heating element thus delivering full voltage to the transformer. Apparently the bimetallic element had enough resistance so that it's self-heating kept it in the turned-on state. No -- I'm not seriously suggesting the use of one of these strange devices, I just thought it was an amusing approach to the problem. Pete Peterson {decvax,linus,wjh12,mit-eddie,masscomp}!genrad!rep
dsi@unccvax.UUCP (02/11/87)
In article <3748@teddy.UUCP>, rdp@teddy.UUCP (Richard D. Pierce) writes: > In article <615@unccvax.UUCP> dsi@unccvax.UUCP (Dataspan Inc) writes: > ->In article <598@rna.UUCP>, dan@rna.UUCP (Dan Ts'o) writes: > ->> > ->> <question about vacuum tube v. solid state rectifiers> > -> But of course! The typical Thevenin impedance of a vacuum tube rectifier > ->varies with load, but is about 50 to 100 ohms. Furthermore, the vacuum tube > ->rectified DC power supply cannot be modeled as simply a resistor and an > ->ideal voltage source. The supply impedance varies with the current demand > ->of the PA tubes in the amplifier (mostly); therefore, we conclude, the power > ->supply + amplifier combination exhibits nonlinearity because the operating > ->point of the PA tubes is being jerked around. > -> > -> .... > -> > -> In fact the plate dissipation does not vary linearly with the square > -> of the load current because: > -> > -> 1) The active devices are vacuum tubes > -> 2) The driving point impedance looking into the plate-to-speaker > -> matching transformer varies with frequency, amplitude, room > -> acoustics, etc. > -> > ->Pmmmmmph. > ->David Anthony > > Pmmmmmph, thyself, Mr. Anthony. I (and many others) would be most interested > (and possibly amused) by your explanation of how room acoustics affect > plate dissipation. Very simple. Measure the terminal impedance characteristics of a loud- speaker as a function of frequency in, say, an empty, unfurnished bedroom, and measure the same speaker in an anechoic chamber. Suffice it to say that, should the speaker resonance and room resonance correspond, there would be a very high impedance pole (magnitude) at the resonance frequency (in the empty room case). The plates of the vacuum tubes operate into a load impedance. This load impedance exists for the case where the secondary of the plate transformer is terminated into an essentially conjugate match. Now, we all know that the loudspeaker doesn't present even close to a conjugate match to the plate transformer, and furthermore, doesn't even present the same terminal characteristics to the plate transformer from room to room of the speaker-plate transformer pair. Furthermore, every EE freshman knows that the impedance seen at the primary of a theoretical transformer is related to the impedance presented to the secondary by the following relation: Z1 = Z2 * (N1/N2)**2 where Z1 and Z2 are the impedances and N1/N2 is the primary to secondary turns ratio. Now, we have our real-world speaker in two different rooms connected to the same amplifier. We're using a class "A" amplifier, so the current drawn by the amplifier is essentially constant. Either the current is being dissipated in the plate of the tube, or is being delivered to the load. The load impedance seen is the transformer primary with respect to the PA tube. If the load impedance of the transformer changes with room acoustics (and it obstensibly does) then the transformer current must also change with room acoustics. Thus, the instantaneous plate dissipation must also change.... For example, take your typical 8417 beam power tube operating class "A" with plte voltage of 300 volts, plate current of 100 mA, and a load impedance of 16,000 ohms. The speaker, whatever load is nominally 8 ohms. This requires a turns ratio of sqrt(16000/8) or 44.72:1. Now, it is not unusual to find a speaker which varies from, say, 40 ohms to 3 ohms which is nominally labeled 8 ohms. Now, assuming our 8417 is connected to an "ideal" transformer, its load resistance varies from 80 k ohms (in the 40 ohm load case) to 6k ohms (in the 3 ohm case). It sure seems to me that if the AC load impedance can be made to vary over a 13:1 range as a function of frequency, a 2:1 or even 5:1 change at specific frequencies can be made with the same speaker from room to room. This is pretty damn drastic where vacuum tube amplifiers are concerned. THE SLOPE OF THE TUBE LOAD LINE IS CORRECT FOR ONE SET AND ONLY ONE SET OF BIAS AND OPERATING CONSTANTS. IF THE SLOPE OF THE LOAD LINE CHANGES, THE OPERATING POINT IS NOT CORRECT. Not only are the distortion characteristics of the amplifier changed, but the power delievered to the load changes (of course), and hence, the plate dissipation. The slope of the load line is the AC impedance of the plate transformer looking into the load. Now, for the small signal model, all of the aforementioned effects may be de minimis. But for the amplifier delivering its full maximum sinewave (voltage, current, whatever) the effect of load impedance on the operating characteristics does become significant. I did not say that room acoustics were "the" factor. I merely said that room acoustics were "a" factor. The determination of amplifier performance as a function of load impedance seen by the amplifier is an area for scholarly research. Perhaps it can be determined that changes in plate dissipation as a function of room "load" is insignificant. The change in operating point characteristics for a vacuum tube amplifier with a matching transformer, even for purely resistive loads, does nevertheless exist as a function of the load impedance. David Anthony DataSpan, Inc.
rfg@hound.UUCP (02/11/87)
Ive been away from net.audio for a monthy or so because the feed was disrupted when the net was reorganized. Today's is the first really sizeable set I've received this year. And what do I find? Another insane controversy has arisen. Oh, Joy, Oh, Joy! Just like the good old days of nut.audio!!! How I have missed it!! And it's just what I needed to cure the depression brought on by reading a national audio mag last night that recommended buying a special audio Power Cord to improve the sound of a "dream system." "Don't knock it," they said, "Buy it and listen if it doesn't improve the sound." THe dream system only cost $30,000.00. Anyone with $30,000.00 to spend on an audio system (should, shouldn't) spend any of it on an audiophile power cord. Choose one. What difference does it make anyway? Well, one difference it makes is that when the poor slob with $277.95 to spend starts worrying about the power cord (or the design of his rectifiers) it becomes a really sad situation and all you greedy little audio dealers and two-bit manufacturers can (should) go to hell for it. It is a fact that compared to vacuum tubes, solid state silicon rectifiers are one of the few nearly perfect devices. I am sure that it is possible to screw-up sound quality with a carefully misdesigned power supply, or one with some components failed, but silicon rectifiers do make it harder to accomplish that end. Now, let's go back to a discussion of burning candles next to the turntable to improve disc reproduction. THAT is something I've actually done and seen work. But it all depends on what the candle is made out of, and where can you buy pure beeswax today? ...There is a REAL subject for discussion. (I will admit, however, that no one ever made a more beautiful rectifier than a mercury vapor tube. Combined with just slightly gassy 2A3's the result was ... ineffable. Who needed the music?) Cheers! Onward, nut - er - wreck.audio! -- "It's the thought, if any, that counts!" Dick Grantges hound!rfg
newton2@topaz.berkeley.edu.UUCP (02/12/87)
Sender:Doug Maisel Thank God for the return of hound!! To Dick Pierce: Lighten up, bud. Anthony's a righteous knowledgeable non-bullshit-mode dude. One of the few (actually, one of the two..., heh, heh-- OK, it was a joke! a JOKE!) Doug Maisel 56 Panoramic Way Berkeley, CA 94704 (415) 848-5247
ornitz@kodak.UUCP (02/12/87)
In article <1669@uwmcsd1.UUCP> shop@uwmcsd1.UUCP (Thomas Krueger) writes: > >Certainly. The amp is a Dyna Mk III. I added a large filter choke, about >10 HY's in series with the Dyna choke, which is in basically a pi filter. >To get the added space, I took out the 5AR4 ...... and >replaced with 1000PIV/2.5A diodes which were bypassed with .005MFD/3KV >ceramic capacitors in an effort to keep the turnon surge away from the >diodes.If I would leave the amp on for a long time (like three days) >enough heat would >build up to the point where either a diode would melt >down and short spontaneously, or more rarely (why? I dunno) on turnon. To add to some of my previous comments, Thomas: If your transformer secondary is 800 volts center tapped which is likely, your diode PIV rating should be more like 3 KV with a capacitor input filter and adequate safety margin. The capacitors across the diodes are for transient protection, not surge protection. Also remember that many of the small 1000 PIV, 2.5A diodes on the market (HEP170, etc.) have a 2.5A steady state current rating but their surge ratings are no higher than 1A diodes. In a capacitor input filter, the surge rating overrides the DC rating. Try using three 1N5408 diodes in series for each leg of the rectifier. Parallel the 1N5408's with a 0.001 uF cap and a 470K 1 W resistor across each diode. You also might move the choke to make a dual section L filter. This will put less stress on the diodes and give you better ripple rejection. Barry
larry@kitty.UUCP (02/12/87)
In article <1945@hound.UUCP>, rfg@hound.UUCP (R.GRANTGES) writes: > ... > (I will admit, however, that no one ever made a more beautiful rectifier > than a mercury vapor tube. Combined with just slightly gassy 2A3's the > result was ... ineffable. Who needed the music?) Don't forget about tungar tubes. A "distributed" power supply using tungar tubes as rectifiers would not only light the room with a quaint orange ambience, but would eliminate having to use those nasty, unpredictable, and unreliable silicon diodes. Perhaps I will offer an upgrade kit for amplifiers that presently have silicon diodes, and advertise it with the slogan: "Turn your favorite mogul-socket floor lamp into the ultimate audiophile power supply..." Another possibility to eliminate silicon diodes is to use copper oxide rectifiers - made with oxygen-free copper, of course... > Cheers! Onward, nut - er - wreck.audio! Ditto. <> Larry Lippman @ Recognition Research Corp., Clarence, New York <> UUCP: {allegra|boulder|decvax|nike|rocksanne|watmath}!sunybcs!kitty!larry <> VOICE: 716/688-1231 {hplabs|ihnp4|mtune|seismo|utzoo}!/ <> FAX: 716/741-9635 {G1,G2,G3 modes} "Have you hugged your cat today?"
rdp@teddy.UUCP (02/12/87)
In article <625@unccvax.UUCP> dsi@unccvax.UUCP (Dataspan Inc) writes: ->> ->> Pmmmmmph, thyself, Mr. Anthony. I (and many others) would be most interested ->> (and possibly amused) by your explanation of how room acoustics affect ->> plate dissipation. -> -> Very simple. Measure the terminal impedance characteristics of a loud- ->speaker as a function of frequency in, say, an empty, unfurnished bedroom, ->and measure the same speaker in an anechoic chamber. Suffice it to say that, ->should the speaker resonance and room resonance correspond, there would be ->a very high impedance pole (magnitude) at the resonance frequency (in the ->empty room case). -> OK, I routinely do this (dozens of time a week, in fact). I can state categorically that unless the room we are in is comparable in size to the enclosure volume, the room acoustics have NO effect on the terminal impedance of the loudspeaker. None. (at least as far as my impedance measuring equipment is concernned, certainly far more accurate than any concept of plate load is concerned). You missed the point I was making. The effect of acoustics on the terminal impedance of the louspeaker is non-existant simply because the effective Thevenin equivalent of that generator is so small, and there is a whole series of effective series impedances that are so much greater. The effect can be ignored to any approximation that even remotely deals in reality. ->we all know that the loudspeaker doesn't present even close to a conjugate ->match to the plate transformer, and furthermore, doesn't even present the ->same terminal characteristics to the plate transformer from room to room ->of the speaker-plate transformer pair. -> No, we all don't know that. Those of us who have measured the impedance effects don't know that. You have presented no evidence to that effect. I would suggest you try measuring either the effective terminal impedance or (even) the plate dissipation with changes in room acoustics. I have absolutely no argument with your contention, however, that plate dissipation is load impedance sensitive, just that load impedance is insensitive of room acoustics. -> Now, we have our real-world speaker in two different rooms connected ->to the same amplifier. We're using a class "A" amplifier, so the current ->drawn by the amplifier is essentially constant. Either the current is ->being dissipated in the plate of the tube, or is being delivered to the ->load. -> -> The load impedance seen is the transformer primary with respect to ->the PA tube. If the load impedance of the transformer changes with room ->acoustics (and it obstensibly does) then the transformer current must also ->change with room acoustics. Thus, the instantaneous plate dissipation ->must also change.... -> Again, without either and analysis or actual data to support you, I must strongly dispute the claimed effects of room acoustics. > -> For example, take your typical 8417 beam power tube operating class ->"A" with plte voltage of 300 volts, plate current of 100 mA, and a load ->impedance of 16,000 ohms. The speaker, whatever load is nominally 8 ->ohms. This requires a turns ratio of sqrt(16000/8) or 44.72:1. Now, it ->is not unusual to find a speaker which varies from, say, 40 ohms to 3 ->ohms which is nominally labeled 8 ohms. Now, assuming our 8417 is connected ->to an "ideal" transformer, its load resistance varies from 80 k ohms ->(in the 40 ohm load case) to 6k ohms (in the 3 ohm case). It sure seems ->to me that if the AC load impedance can be made to vary over a 13:1 range ->as a function of frequency, a 2:1 or even 5:1 change at specific frequencies ->can be made with the same speaker from room to room. This is pretty damn ->drastic where vacuum tube amplifiers are concerned. -> I do not in any way dispute your analysis on this section, in fact, it is a pretty damning case against using tube amplifiers to drive speakers, period. I have a couple of nitpicks, however. I have measured the real impedance of quite literally hundreds of loudspeakers in all conditions from anechoic to highly reverberent to extremely resonant environments. The effects of room acoustics you claim are neither apparent in the measurements, nor in the analysis of the phenomenon itself. Again, to restate my point, The effective radiation load presented by the room is such a riduculously small part of the total impedance as to be unmeasurable. Look at the radiation impedance of a 10 inch woofer at 100 Hz as reflected back through the voice coil. It is a couple of milliohms, at best. This compared to such large effects as the simple series DC resistance (6 ohms or so) is negligable (go look it up for yourself in texts such as Leo Beranek's "Acoustics", McGraw-Hill, et al) Secondly, typical loudspeaker impedances vary in a range of about 5 to 1. 13 to one is a vary bizarre exception. > I did not say that room acoustics were "the" factor. I merely said >that room acoustics were "a" factor. The determination of amplifier >performance as a function of load impedance seen by the amplifier is an >area for scholarly research. Perhaps it can be determined that changes >in plate dissipation as a function of room "load" is insignificant. I contend that the effects are no factor at all. Dick Pierce
ornitz@kodak.UUCP (02/13/87)
In article <3761@teddy.UUCP> rdp@teddy.UUCP (Richard D. Pierce) writes an excellent rebuttal to <625@unccvax.UUCP> dsi@unccvax.UUCP (Dataspan Inc). I would like to point out two areas in their discussion that may be overlooked by Net readers. Pierce says: >I do not in any way dispute your analysis on this section, in fact, it is >a pretty damning case against using tube amplifiers to drive speakers, period. The problem here is that the output transformer transforms the impedances by the square of the turns ratio. Without the transformer, only the direct impedance variation is involved. This still effects transistor output stages so the "damning case" is against the transformer method of impedance matching, not tube amplifiers per se. Many solid-state audio amplifiers approach zero ohms output impedance. With the case of a stiff voltage source, the output power will be inversely proportional to the impedance. Therefore there will still be an effect. My other point is a minor one with Mr. Anthony. The load presented by a speaker has a fair degree of reactance associated with it. In his discussion he mentions the changing slope of the load lines as the impedance changes. In an actual audio amplifier with a reactive load (whether a transformer is used or not), the load lines are elliptical making the analysis much more difficult when determining power output and plate (collector) dissipation. Good reference books in this area are now scarce unless your library stocks some of the classic texts of the 1940's. Barry L. Ornitz
dsi@unccvax.UUCP (02/13/87)
In article <3761@teddy.UUCP>, rdp@teddy.UUCP (Richard D. Pierce) writes: > In article <625@unccvax.UUCP> dsi@unccvax.UUCP (Dataspan Inc) writes: > ->> > ->> Pmmmmmph, thyself, Mr. Anthony. I (and many others) would be most interested > ->> (and possibly amused) by your explanation of how room acoustics affect > ->> plate dissipation. > -> > > OK, I routinely do this (dozens of time a week, in fact). I can state > categorically that unless the room we are in is comparable in size to the > enclosure volume, the room acoustics have NO effect on the terminal > impedance of the loudspeaker. None. (at least as far as my impedance > measuring equipment is concernned, certainly far more accurate than any > concept of plate load is concerned). Thank you for at least breaking one of my longest held misconceptions about acoustics and the speaker-room system (went and did some reading in the reference work cited by rdp). I was under the impression that the terminal characteristics of a loudspeaker varied pretty drastically depending on the acoustical "load" presented to the speaker. My assumption was totally wrong, evidently (please forgive me, I am used to antennas and such where the output devices are extremely environment sensitive at times, and even a 20 % change in load impedance can cook a set of tubes in a hurry). > ->(in the 40 ohm load case) to 6k ohms (in the 3 ohm case). It sure seems > ->to me that if the AC load impedance can be made to vary over a 13:1 range > ->as a function of frequency, a 2:1 or even 5:1 change at specific frequencies > ->can be made with the same speaker from room to room. This is pretty damn > ->drastic where vacuum tube amplifiers are concerned. > -> > > I do not in any way dispute your analysis on this section, in fact, it is > a pretty damning case against using tube amplifiers to drive speakers, period. > I have a couple of nitpicks, however. > > > Secondly, typical loudspeaker impedances vary in a range of about 5 to 1. > 13 to one is a vary bizarre exception. I agree (went back and read aformentioned reference). However, my BIC Formula 2 loudspeakers do, in fact, vary over a range from 3 to 40 ohms. Thank you for exposing my ignorance ... really. Your rebuttal is a strong case for at least exposing EE types to some acoustics in their training. Think I'll go back to video and RF for the time being... David Anthony DataSpan, Inc.
sjc@mips.UUCP (02/14/87)
>>> In fact the plate dissipation does not vary linearly with the square >>> of the load current because... >>> 2) The driving point impedance looking into the plate-to-speaker >>> matching transformer varies with frequency, amplitude, room >>> acoustics, etc. >>> Pmmmmmph. >>> David Anthony >> >> Pmmmmmph, thyself, Mr. Anthony. I (and many others) would be most interested >> (and possibly amused) by your explanation of how room acoustics affect >> plate dissipation. > > Very simple. Measure the terminal impedance characteristics of a loud- > speaker as a function of frequency in, say, an empty, unfurnished bedroom, > and measure the same speaker in an anechoic chamber. Suffice it to say that, > should the speaker resonance and room resonance correspond, there would be > a very high impedance pole (magnitude) at the resonance frequency... > > Furthermore, every EE freshman knows... Oh, to be a freshman again. Anyway, dramatic things do not happen just because two resonances coincide; only if the Q is high enough can you even detect a resonance. From the speaker's point of view, the Q of the room resonance is pretty small. After all, in most speakers, only a small percentage of the power driving the speaker makes useful sound; most of it heats the voice coil, compresses air inside the enclosure, and so on. Thus, most of the electrical load presented by the speaker is due to the voice coil impedance, the enclosure characteristics, and so on. This power dissipation largely damps out the effect of room resonance on the load presented to the amplifier. -- ...decwrl!mips!sjc Steve Correll