"The TV sweep tubes brought out the supressor grid to a seperate pin so that they could be biased slightly positive to eliminate Barkhausen oscillation. This voltage is specified in the data sheets. For most tubes the spec is 0 volts minimum, and 30 volts maximum with a few tubes rated for 50 volts."
The negative suppressor grid will cause more off axis electrons to return to the screen grid instead of making it to the plate. This will make for more rounded corners on the plate characteristics, and a hotter running screen grid. Neg. voltage on the suppressor does not "get it out of the way". Check the 6LE8 datasheet to see plate curves for negative suppressor voltages.
http://www.mif.pg.gda.pl/homepages/frank/sheets/135/6/6LE8.pdf
Don
The negative suppressor grid will cause more off axis electrons to return to the screen grid instead of making it to the plate. This will make for more rounded corners on the plate characteristics, and a hotter running screen grid. Neg. voltage on the suppressor does not "get it out of the way". Check the 6LE8 datasheet to see plate curves for negative suppressor voltages.
http://www.mif.pg.gda.pl/homepages/frank/sheets/135/6/6LE8.pdf
Don
The 6LE8 is a special purpose tube, so its characteristics may not apply to a sweep tube. As such sweep tubes are beam tubes and the previously mentioned output tube is a true pentode, so they probably don't behave the same either. I think that it's worth the twist of a power supply knob the next time I have a P-P experiment going.
Michael Koster said:
The mere fact that you choose Triode over the alternatives already does that.
BTW, triodes needs a lot more voltage swing into the Grid and needs to be watched.
Requires that the speaker system have a decent alignment (and reasonably flat Z non-reactive). If vented I'd have something as close as possible to a Bessel and if closed box then reasonably low Qc. Since these amps are basically a hybrid sitting somewhere between voltage and current send amps, imagine a current send amp with infinite output Z into a loudspeaker. It shouldn't work as in theory the Qe part of the equation gets blown to infinity as well. But in the real world it still works, just ask Nelson Pass who has done some significant things with his F1 current amp. Unity Coupling was actually coined, AFAIK, by Lynn Olson. At least I learned it from him. The idea has also been called 'hard amps' which I understand is a reference to their immunity to back-EMF effects. Every little cone (and other) mechanical resonance will have a mirrored current going out of the back door and into the amp. If you wander about it, if using a decent SS amp with ideal voltage characteristics (output Z ten or more times lower than speaker Z), then put a series resistor (make it decent quality) of similar Z to the speaker and listen. Mind you, your max output before clipping will be -6dB or as if you only had a quarter of the power. You now have simulated Unity Coupling. Try it, it may surprise you, and indeed suprise your SS amp that it only has to deal with a fraction of the 'many happy returns' the speaker sends back. It will also tell you how well the speakers cope.
Menno did tests on the audibility of low permeability. If you take the transformer and inject very low signals with extremely low noise, then you can see waveform changes. The problem is similar to that of DACs where linearity decreases with level and that has been measured. Put it differently, input an audio signal at very low level and then amplify afterward to hear the contribution of the tx and it does sound grainy. That is lack of resolution.
The problem is simply this, we can make small signal transformers with high permeability but less headroom. Make transformers BIG and it becomes much more of a problem dealing with very low but clearly audible levels. You can use amorphous cores but that will halve the power rating compared to silicon steel. But bias silicon steel and you have full power rating and high perm, maybe even better than expensive cores.
Does no harm whatsoever. You have to find an 'input' and drive it from a high Z so as not to alter the functioning of the amp's circuit. The amp treats it as if it is simply another small signal to process. The most difficult place to inject is actually directly into the Primary. Some has suggested that the bias signal has some conditioning effects on the components. I have an open mind.
Joe R.
tubelab.com said:
If the tx is say -1dB @ 20KHz (representaive of my latest examples), don't worry. It will work just fine with 50-55KHz bias. I may have given the impression that it needs 100KHz bandwidth but in reality not so. It's the voltage swing on the Primary that gives the benefit, not what emerges from the Secondary.
Keep the health man. It reminds the rest of us what really matters.
Joe R.
Not necessarily, I've gotten more power out of triode connected KT88s than I got in ultra-linear by driving them with mosfets. However, I did destroy the tubes in testing. I'm in the process of coming up with an instrumentation setup so that I can measure grid dissipation so that I can best tune the output impedance of the driver so as to get the most out of the triodes without destroying them.
tubelab.com said:
Since we are actually referring to EL34, the point is simply this, we have an option where to connect whereas the usual other suspects don't give one. Indeed it has been suggested tying it to the Screen which was my first inclination, but Bill Perkins felt otherwise - but I think he does something different now?
Joe R.
Bob Carver was doing demonstrations back in the early 80's with a black box that he connected between one of his big SS amps and the speaker. The box was said to make a SS amp sound like a tube amp. He later revealed that the black box contained a big resistor and nothing more. This was part of one of his demonstrations designed to convince the world that the future was made of silicon. I wound up buying a Carver M400 and a Phase Linear 4000. It was extremely loud, but I lost interest within a few years and they are stuffed in a box somewhere.
Within a few years Carver figured out that there was money to be made in tube amps and the Silver Seven came out.
I had BobCarves later or last(?) SS amp with switchmode supply, or whatever it was(some kind of classD maybe?)
It did have an exstra output with a resistor
I was relustant to use it as it somehow didnt seem right to me
It did sound different with the resistor
Not softer really, just a bit mellow and closed
Still VERY far from any tubeamp I have heard
I was told that the output with the resistor could be improved considerably with small mods, like a better resistor
Didnt last long with me though
Maybe output transformers would have been much better
It did have an exstra output with a resistor
I was relustant to use it as it somehow didnt seem right to me
It did sound different with the resistor
Not softer really, just a bit mellow and closed
Still VERY far from any tubeamp I have heard
I was told that the output with the resistor could be improved considerably with small mods, like a better resistor
Didnt last long with me though
Maybe output transformers would have been much better
Joe Rasmussen said:
Hmmm - what if the HF bias is actually 20-30 kHz worth of noise (moved up to the top of the OPT passband), instead of a single tone? That way, less trouble with intermodulation from digital, television, and various switching supplies (which are everywhere these days). Also, less chance of intermodulation with a high-order harmonic in the music itself, or a poorly-filtered artifact from the CD player.
I can't claim credit for the "Unity Coupling" idea. I met a Western Electric old-timer in the Seattle area that told me that the WE/ERPI/Altec movie theaters were set up with a low-power amp in the projection booth, a 600-ohm balanced transformer-coupled interface to the main power amps which were physically close to the speakers behind the screen, and the output Z of the amps set up to be similar to the load Z of the speaker system (Altec A1, A2, A4, or A5 or earlier). Wente and Thuras knew a few things about resonance control and optimized coupling between system elements, thus, the selection of 2nd-order 500 Hz crossovers and horns that liked a damping factor of unity.
I don't how true this is, but it sounds plausible. This conversation is what led me to the research on the Thirties-vintage WE "Harmonic Balancer" that was forgotten after the Williamson swept everything else aside in 1947. I find it interesting that the company that invented feedback in 1927 reserved it for their SE amp (the 91A) and used only the Harmonic Balancer (and not overall loop feedback) for their top-of-the-line PP amps.
(Since the complete theater sound system was leased, not sold, from WE/ERPI, there was never a "retail price" on the 86, 91, and 92 family of amplifiers. Americans old enough to remember the Bell Telephone System probably also remember that the telephones were never for sale, and were simply part of the monthly bill. No plug-in RJ-11 sockets back then! The Bell technicians would threaten to remove service if you so much as touched the insides of their phones or any of their wiring. It took a Supreme Court decision in the early Seventies to force Bell to open their system to non-Bell devices like answering machines.)
It looks like Altec switched over to more conventional PP pentode-with-feedback amplifiers in the Fifties for the Cinerama, VistaVision and Todd-AO 70mm widescreen stereophonic theaters, while retaining the A1, A2, A4, and A5 family of speakers (515 woofers and 288 compression drivers with multicell horns).
Impedance matching was the way to get less power loss that means less needs of power amplification that leads to less distortions and wider dynamic range. However, it was not perfectly matched on all frequency band, also it was not linear, so a negative feedback was used instead to make the life easier.
"Every new thing is a well forgotten old one"
"Every new thing is a well forgotten old one"
The term Unity Coupling here is confusing, since McIntosh called their xfmrs by that. Don't you mean the same thing as Critical Damping as used in the RDH4? Which implies impedance matching the amplifier output to the speaker. Lots of amps had adjustments of positive current feedback and neg. voltage feedback (ElectroVoice comes to mind.) to achieve critical damping of the speakers.
Don
Don
I was born in 1949, so I have memories of the Fifties as a kid that was gonzo for hifi. The terminology was pretty loose, and advertising became dominant as the decade went on and the Space Age got its start. It really went Big Time when stereo came out in the movie theaters from 1953 through 1958, on prerecorded stereo tape in 1956-57, on LPs in 1958, and on FM tuners in 1962.
Power ratings got more and more fictitious, especially after the notorious IHFM ratings crept in during the Sixties, eventually drawing the attention of the Fair Trade Commission in the early Seventies. Advertisers could get away with claiming pretty much anything - it was pretty loose. So anybody could claim Unity Coupling, or Z-Matic, and it would mean whatever the advertiser would claim.
Power ratings got more and more fictitious, especially after the notorious IHFM ratings crept in during the Sixties, eventually drawing the attention of the Fair Trade Commission in the early Seventies. Advertisers could get away with claiming pretty much anything - it was pretty loose. So anybody could claim Unity Coupling, or Z-Matic, and it would mean whatever the advertiser would claim.
"Advertisers could get away with claiming pretty much anything - it was pretty loose. So anybody could claim Unity Coupling, or Z-Matic, and it would mean whatever the advertiser would claim."
That's exactly why I stopped buying audio equipment a long time ago, manufacturers lost all credibility. And they sure haven't gotten it back with 500 Watt computer speakers lately.
Don
That's exactly why I stopped buying audio equipment a long time ago, manufacturers lost all credibility. And they sure haven't gotten it back with 500 Watt computer speakers lately.
Don
It's interesting to re-read the advertising literature of the Fifties from a modern perspective. Even Altec told tall tales on occasion, but in the absence of Theile/Small modeling techniques, you have to cut them a little slack. But when comparing the all-too-similar schematics of most Golden Age electronics, and the absurd advertising claims, well ... even Marantz and McIntosh stretched it at times. All of the manufacturers, including the ones at the top of the market, were selling watts-per-dollar and the lowest-THD numbers.
It took the advent of really cheap transistor watts to change the focus of the high-end market - and even then, there is still only a very small market for high-quality transistor amps with modest (60-watt or less) power. By contrast, building a high-quality 60-watt tube amp is not an exercise in cheap watts by any stretch of the imagination, much less if you want Class A or DHT triodes.
Joe, thanks for the interesting schematic and the commentary that followed. If I recall right, that's a Mullard phase inverter, right? Any opinions on that phase inverter versus others? I also like the detail of the pull-down on the suppressor grids, nice touch!
It took the advent of really cheap transistor watts to change the focus of the high-end market - and even then, there is still only a very small market for high-quality transistor amps with modest (60-watt or less) power. By contrast, building a high-quality 60-watt tube amp is not an exercise in cheap watts by any stretch of the imagination, much less if you want Class A or DHT triodes.
Joe, thanks for the interesting schematic and the commentary that followed. If I recall right, that's a Mullard phase inverter, right? Any opinions on that phase inverter versus others? I also like the detail of the pull-down on the suppressor grids, nice touch!
Some thoughts ,insights and a suggestion
I have followed the advices of the good people here and studied some of the plans of tube amplifiers , specifically those of Tubelab and NK.
NKs design is very appealing to me. However, it uses some proprietary transformer and has not been built by anyone else.
Tubelab's minitron is very interesting combination of old and new but has low output power,and again is a prototype with no follow up, as tubeblab himself is preoccupied with his day job and other projects.
I am not skilled enough to deal with any of the tube or HV parts by myself.
It occurred to me that these and other designs are actually a one or few persons, one time endeavors, and are thus prone to being less than the optimum they could have been.
I work in a telecommunication company. We produce cards that are several orders of magnitude more complex than the above designs (No disrespect is either implied or intended in any way, I have the highest respect to both Tubelab and NK), yet we had only single re-spin of the cards and even that was minor. The major differences are that this is a team efforts and we have strong simulation and verification tools.
On the other hand there are many skilled, knowledgeable and experienced people here, that their collaborated knowledge could produce wonders.
So my suggestion is to follow the "open source" community paradigm and produce some "oped design" projects of tube amplifier designs that would be a shared work of all of the good people in this forum.
This would enable to collectively produce deigns that would be probably better than any single one of us alone.
Each can contribute in his own way. I could easily layout low voltage SS PCBs. With some guidance, I could probably do the HV part. Others will probably have more meaningful contribution than that.
Your comments are appreciated.
I have followed the advices of the good people here and studied some of the plans of tube amplifiers , specifically those of Tubelab and NK.
NKs design is very appealing to me. However, it uses some proprietary transformer and has not been built by anyone else.
Tubelab's minitron is very interesting combination of old and new but has low output power,and again is a prototype with no follow up, as tubeblab himself is preoccupied with his day job and other projects.
I am not skilled enough to deal with any of the tube or HV parts by myself.
It occurred to me that these and other designs are actually a one or few persons, one time endeavors, and are thus prone to being less than the optimum they could have been.
I work in a telecommunication company. We produce cards that are several orders of magnitude more complex than the above designs (No disrespect is either implied or intended in any way, I have the highest respect to both Tubelab and NK), yet we had only single re-spin of the cards and even that was minor. The major differences are that this is a team efforts and we have strong simulation and verification tools.
On the other hand there are many skilled, knowledgeable and experienced people here, that their collaborated knowledge could produce wonders.
So my suggestion is to follow the "open source" community paradigm and produce some "oped design" projects of tube amplifier designs that would be a shared work of all of the good people in this forum.
This would enable to collectively produce deigns that would be probably better than any single one of us alone.
Each can contribute in his own way. I could easily layout low voltage SS PCBs. With some guidance, I could probably do the HV part. Others will probably have more meaningful contribution than that.
Your comments are appreciated.
Lynn Olson said:
I wondered about that too if that would prove problematic, but as you know we do a lot of work on digital players and use relatively minimal filtering and have not had any trouble whatsoever. It seems 50KHz, more than twice 20KHz, at the level we have decided upon is adequate. But a noise generator fed through a bandpass filter should of course also work and if problems had revealed itself... but so far so good. BTW, Allen and I discussed his Sony VC24 player conversions, they have a low level HF component when in SACD mode, so when hooked up to tube amps and playing SACDs, users should be getting some benefit on the side without knowing.
I come back to the curious comparison with DACs and poor low level linearity. Kind of dithering for output txs.
Generally the safe level is about 100-200mV RMS and that translates to usually around 10V p/p on the Primary.
The scale on the left is that of a Fletcher-Munson curve. Generally in real cases end up better than plateauing at -70dB and perm remains very healthy indeed at all times.
Joe R.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.