I have come to realize that advertising plays a huge part in the equation of mass buying. Television has become a way to shape and control audience behavior and purchasing.
The sad fact is that marketing geniuses often craft lies to sell whatever product they are paid to promote. That's certainly true about amps, and many other things.
Ebay and the internet, however, are changing everything, because everyone has access to information and a worldwide marketplace.
For instance, on Ebay we now see tube amps made specifically for iPods and mp3 players.
The sad fact is that marketing geniuses often craft lies to sell whatever product they are paid to promote. That's certainly true about amps, and many other things.
Ebay and the internet, however, are changing everything, because everyone has access to information and a worldwide marketplace.
For instance, on Ebay we now see tube amps made specifically for iPods and mp3 players.
Did I see someone ask what is the topology for this amp.
Well lets see:
Class AB1 35W from EL34s - Maybe Ultralinear but probably just Pentode Mode. I think they would have made a point of saying so if it was Ultralinear.
2 x 2SN7 and 1 x 6SL7 for front end for 2 channels.
So:
The 6SL7 is split between channels (one triode section each channel) as a common cathode amp, with a lot of negative feedback applied at the cathode to get Zout of the pentode Mode Output down to something reasonable.
There is a 6SN7 for each channel - almost certainly paraphase splitter/driver.
In other words it is a so so, ho hum, design straight out of the 1960's - I expect zero stereo imaging to speak of and no real dynamics (because of the feedback levels), and bad channel separation although it would probably still leave the modern Class D SS rubbish for dead.
All the tubes are too close to each other - the thing is going to run damn hot.
There are better choices of cheap Chinese Amps around. Pick a physically larger one so that its easier to rebuild.
Cheers,
Ian
Well lets see:
Class AB1 35W from EL34s - Maybe Ultralinear but probably just Pentode Mode. I think they would have made a point of saying so if it was Ultralinear.
2 x 2SN7 and 1 x 6SL7 for front end for 2 channels.
So:
The 6SL7 is split between channels (one triode section each channel) as a common cathode amp, with a lot of negative feedback applied at the cathode to get Zout of the pentode Mode Output down to something reasonable.
There is a 6SN7 for each channel - almost certainly paraphase splitter/driver.
In other words it is a so so, ho hum, design straight out of the 1960's - I expect zero stereo imaging to speak of and no real dynamics (because of the feedback levels), and bad channel separation although it would probably still leave the modern Class D SS rubbish for dead.
All the tubes are too close to each other - the thing is going to run damn hot.
There are better choices of cheap Chinese Amps around. Pick a physically larger one so that its easier to rebuild.
Cheers,
Ian
Well, I'm just outing myself as a twenty something guy into good audio.
It's not totally a lost cause, I have a number a friends that are pushing me to build them gear, I had one friend, my own age, not long ago asking me where he could get better speakers because he'd heard something he'd never heard before on my system (and I was only running Soundcraft near-field monitors, no audiophile stuff).
I'm the biggest cheapskate in the world, but I know good sound when I hear it, so do a few of my friends. And none of us have (quite) hit 30.
Of course a bunch of my friends simply can't hear any difference between their boom tizz system from k-mart and mine, but I'm sure it has ever been thus.
Don't give up!
It's not totally a lost cause, I have a number a friends that are pushing me to build them gear, I had one friend, my own age, not long ago asking me where he could get better speakers because he'd heard something he'd never heard before on my system (and I was only running Soundcraft near-field monitors, no audiophile stuff).
I'm the biggest cheapskate in the world, but I know good sound when I hear it, so do a few of my friends. And none of us have (quite) hit 30.
Of course a bunch of my friends simply can't hear any difference between their boom tizz system from k-mart and mine, but I'm sure it has ever been thus.
Don't give up!
I'm glad to hear it! I have communicated with a few college students who have discovered the world of vacuum tube audio. We need more young people to keep this hobby alive, and maybe invent something new. Otherwise we will be building those " so so, ho hum, design straight out of the 1960's" forever. I work in a building with at least 1000 engineers. There are 3 of us who play with tubes (no 20 somethings) and I am the only one actively designing new circuitry.
Todays engineers have modeling tools and processing power at their disposal that wasn't even dreamed of in the 1960's. Now if we could get some better vacuum tube models.......
Yes it has. I worked at a stereo store in the early 1970's. Our best selling system was a best seller only because the sales people got a commission bonus for every one they sold. I had to fix them, they broke a lot! There were Fisher systems for the same money that worked pretty good and sounded far better. Even in 1971, we sold no tube equipment.
tubelab.com said:
I highly doubt that there is anything "new" out there, waiting to be discovered. We've been at this now for over 100 years, with God only knows how many tens of millions of man-years and $$$$$ from all the R&D depatments of the military, corporate, and academic worlds. we've tried just about every concievable topology: common cathode, grid, and plate configs, as well as all the stacked topologies: cascode, SRPP, Mu stages of all sorts, White cathode followers, &c, ad infinitum. All sorts of other possible topologies that incorporate feedback, positive and negative. There are even the "backwards" topologies that input the signal between plate and cathode and take out the signal from the grid/cathode circuit. None of this is anything really new.
Do we really need anything novel here? A good implementation beats hell out of something "wierd" every time. The proliferation of complexity seems to be more related to relieving the atechnological end user from the responsibility of managing his equipment (hence the preference for cathode bias over the better performing fixed bias) and/or to make it easier to move product out the door with minimal testing, even if that complexity compromises the sonics. (Then, again, when has mass production for a mass market ever not settled for the lowest common denominator?) A basic design implemented well certainly gets the job done, even if it requires more testing and tweaking than would be practical for mass production. What's more important is that we have far better materials: better dielectrics for capacitors that didn't exist in the 1930's, and better xfmr core materials. Modern materials can make a 1930's era design perform far better than it could at the time it was designed.
While I can say I "design new equipment", it really isn't "new". What I did was adapt circuitry more commonly associated with precision equipment to an audio application, borrowing pre-existing ideas from RF and SS practice. So it's not really "new".
Sure there is. I work as an engineer in a high tech facility. We dream up "the next big thing" and then invent the technology for it. New tech is being invented every day. Granted absolutely NONE of it is being developed for vacuum tube audio, but that doesn't mean that WE can't find a home for some of this new tech in our amps.
50 years ago we didn't have a way to drive a tube into A2 with A1 distortion levels. Now we have mosfets and CCS IC's. Screen drive is now easy, and allows for some efficient class B amps. All amps used to have vacuum rectifiers, then came mercury vapor, and now FREDs and high voltage Schottkys.
Some of this technology has found its way into our amps, and some of them sound nice. Improvements will continue. Now we have DSP's and SMPS technology. These may find use in vacuum tube amplifiers someday. That day may come sooner than later due to efficiency and power factor requirements that may be legislated upon us.
I have been building vacuum tube amps for over 40 years. I have built just about every configuration that has been published over the past 100 years. Most of my amp designs can be traced to an "old school" design. Most also have modern "sand" sprinkled in where improvements can be made. I, for one am not content to build the same old stuff. I will continue to experiment.
Maybe it is just the engineer in me, but I have a (actually two) tube amplifier under construction that has a SMPS controlled by a DSP chip. Why would I do this? Because I can build a class A triode amp with OVER 50% efficiency, and the audio signal path is still all tube! Doing this requires a power supply that can readjust its output voltage 200000 times a second. Four years ago I designed one for a mobile radio. It doubled the efficiency in the RF power amp. Now I am trying the same thing with tubes. Most of this only exists in the simulation world but reality (hardware) is unfolding slowly.
I've built some tube circuits with SMPS, but for car audio. My question is why/how are you using a DSP chip? Usually there are just PWM chips meant for SMPS that only cost a couple dollars and no programming necessary. Also, how are you raising the efficiency of a class A using a SMPS? Of course SMPS regulation is more efficient than linear regulated, but it's still less efficient than a non-regulated mains based, linear power supply.
Just curious to get a little more info.
Just curious to get a little more info.
I will post more info once I have everything working, which may be a while, but the short story:
An amplifier needs enough voltage headroom to pass the loudest signals without clipping. 99% of the time the amplifier is making far less than full power. At very low power all of the voltage across the tube in a class A amp is being dissipated as heat (the current is constant). If this voltage can be adjusted on the fly to be just enough to avoid clipping at the required power level at any instant in the complex audio signal, a large increase in efficiency can be achieved. Other power supply "modulation" techniques are under investigation.
Most SMPS chips are designed to make a fixed regulated voltage. They are not designed to be "modulated". Chips designed for PFC circuits are designed to be modulated at a 100 or 120 Hz rate. The open loop gain - bandwidth product of the analog circuitry is not high enough to be useful for audio frequency modulation.
There are specialized DSP chips from Microchip and Freescale that are intended for motor control and even SMPS use. Experiments with audio frequency modulation are not quite ready for prime time yet. Phase shift issues due to latency in the A/D and poor programming ability.
It is also possible to build the entire SMPS with discrete components. I have found white papers in the academic world showing modulated converters at the 50 watt level having a modulation bandwidth of 20 MHz for CDMA base stations. Audio should be simple. Google "supply modulation" or "envelope tracking" for more info.
An amplifier needs enough voltage headroom to pass the loudest signals without clipping. 99% of the time the amplifier is making far less than full power. At very low power all of the voltage across the tube in a class A amp is being dissipated as heat (the current is constant). If this voltage can be adjusted on the fly to be just enough to avoid clipping at the required power level at any instant in the complex audio signal, a large increase in efficiency can be achieved. Other power supply "modulation" techniques are under investigation.
Most SMPS chips are designed to make a fixed regulated voltage. They are not designed to be "modulated". Chips designed for PFC circuits are designed to be modulated at a 100 or 120 Hz rate. The open loop gain - bandwidth product of the analog circuitry is not high enough to be useful for audio frequency modulation.
There are specialized DSP chips from Microchip and Freescale that are intended for motor control and even SMPS use. Experiments with audio frequency modulation are not quite ready for prime time yet. Phase shift issues due to latency in the A/D and poor programming ability.
It is also possible to build the entire SMPS with discrete components. I have found white papers in the academic world showing modulated converters at the 50 watt level having a modulation bandwidth of 20 MHz for CDMA base stations. Audio should be simple. Google "supply modulation" or "envelope tracking" for more info.
tubelab.com said:
Tubelab,
I have heard others speak of such ideas elsewhere and I guess I don`t get the reason for it . Correct me if I am mistaken but I`ll try to draw an analogy to explain how I think what you are speaking of will work.
Imagine a painter hand holding his paint roller against a wall he wishes to paint in vertical strokes. The painter represents our class-A vacuum tube. The roller can be the loudspeaker voicecoil. Now let`s imagine he is standing on a platform placed on the forks of a forklift that can move up and down in sync with the painter`s hand. The forklift platform represents the power supply to the vacuum tube. Thinking about audio the forklift can do most of the
heavy lifting and basic LF waveshape , the painter`s delicate hand can do the high frequency detail and touch up . Both the forklift (power supply) and the painter`s hand (output tube) are modulated by the incoming audio signal.
Sorry about mixing metaphors. Am seeing the basic mechanism correctly?
That's a good idea. I've had a similar idea of adjusting a fixed bias output stage to give just enough bias to meet the requirement of the signal by tracking the signal amplitude. Would probably require rectifying the signal then controlling some active device. I'll probably never find the time with all the other projects to try this though.
You may want to try a PWM chip using one of the compensation op amps on it. Feed the signal to it and adjust the gain. It will vary the pulse width with the signal. Which will raise and lower the supply rail.
Anyhow, sounds interesting. Keep us posted!
You may want to try a PWM chip using one of the compensation op amps on it. Feed the signal to it and adjust the gain. It will vary the pulse width with the signal. Which will raise and lower the supply rail.
Anyhow, sounds interesting. Keep us posted!
This is just about right. I look at it this way. There is a pattern traced on the wall that the painter is trying to follow (the audio signal) The fork lift can "see" this pattern and is fast and agile enough to follow it (the power supply is modulated at the audio rate and has a modulation bandwidth of at least 30 KHz). The fork lift is "digital" and can move with a very high efficiency (SMPS converter with 90% efficiency). Because the fork lift is "digital" it can move to one of only 1024 places (10 bit A/D).
The painter stands on the forks and follows the pattern with the roller. If the fork lift was perfect in following the pattern, the painter would have zero work to do, he would just hold the roller in one spot (zero dissipation in the output tube). Since we said that the fork lift was not perfect, the painter has a small amount of work to do. We will also feed the painter well enough to motivate him to be able to stoop down low and reach up high if he has to (headroom for the output tube so that the effects from the fork lift are never noticed, and tubes don't work very well without voltage).
The fork lift does all of the heavy lifting, but is very efficient at doing so (the modulated SMPS can be 90% efficient). The painter is very linear, but terribly inneficient, so we let him do maybe 10% of the work (class A triode amp at 5 or 10 watts in a 100 watt amp).
Using this analogy we could let the fork lift represent a 100 watt amp at 90% efficiency, and the painter be a 10 watt amp at 20% efficiency. The result is still a 100 watt amp because ideally the painter is doing nothing, but you still have to feed him. To get my 100 watts out, I am putting 111 watts into the fork lift and 50 watts into the painter. This is 161 watts into the system for 100 watts out giving an efficiency of 62.5%.
No laws of physics are broken here. This is just a new way of approaching an old problem. There is indeed something "new" out there and I am exploring it. This could open up the possibility for 10 watt SET amplifiers that you could hold in your hand, and 100 watt SET's that you could actually build. Is this added complexity worth it. In most cases no. The average casual tube amp builder will never go here, but why ignore the possibilities.
Why use a DSP? I have built the modulated SMPS for mobile radio. I know that I could build it with discrete parts. I have tried feeding an audio signal into an off the shelf SMPS chip. The response is far from flat. Look up the GBW product (or the gain and phase curves) of the on chip amps. They lose half their gain by 10KHz. Besides that, there is $15000 reasons for using DSP:
http://www.circuitcellar.com/microchip2007/
Isn't this very similar to the Sunfire "Linear Tracking Downconverter" ? Basically a tracking device that allows the supply to stay a step (or two) ahead of the output device's power requirements .
Not that there is anything wrong with this, but it is only "new" in the respect that it hasn't been applied to tubes.
Someone mentioned that a "Fixed" bias design is superior in performance to a "Cathode" biased system. How so ?
...................Blake
Not that there is anything wrong with this, but it is only "new" in the respect that it hasn't been applied to tubes.
Someone mentioned that a "Fixed" bias design is superior in performance to a "Cathode" biased system. How so ?
...................Blake
I am not familiar with the Sunfire amp, but the principles could be similar. I was under the assumption that the Sunfire subwoofer amp was a totally digital design. It is entirely feasible to build a bipolar SMPS and use it for an amplifier. In fact while testing the one I built for a mobile radio I tested its use as an AM radio transmitter. I could get over 100 watts out of it at 500 KHz.
I have one of the original Carver M400 "Magnetic Field Amplifiers". There is nothing magnetic about it but it used multiple sets of output transistors fed from multiple supply rails. The appropriate pair of output devices were used as needed according to the instantaneous output power demands. It was very ingenious for its time, and could play very loudly without getting hot. Just don't use it for a guitar amp.
That amp was designed before SMPS technology was perfected. The "Magnetic Field" power supply was an early SMPS that used a triac to switch at 60 Hz. The switching could be heard in the speakers. Some of my audio buddies did a group buy and got 10 of these when they first came out. Mine is the only one that still works. I would assume that Mr. Carver has perfected the tracking SMPS in the 25 years since the M400 came out.
I don't know if the Sunfire or any other existing amp wraps a modulated power supply around a class A audio amp or not. I know that it is now fairly common in the linear transmitter world. Whether or not this is "new", I am going to try it in the vacuum tube domain. It may work, or it may smoke and blow up. Either way I am going down a road that I don't think has been heavilly traveled before. Again that is just the engineer in me. I have "discovered" a thousand things that don't work! (that has been said before)
George,
Aren't you just describing a Class H amp?
Wikipedia definition of amp classes - scroll down for Class H
Aren't you just describing a Class H amp?
Wikipedia definition of amp classes - scroll down for Class H
Yes, I am. That is where this all started. Think about how you would apply Class H technology to a traditional tube amp. It is not trivial. I have some ideas and a few simulations. There are still some serious obstacles to overcome, and I won't have time for any more experiments for a few weeks (2 week road trip).
I don't want to be guilty of hijacking this thread any further, so I will explain it to you in Dayton. I will start a new thread when I have some real results to post.
I don't want to be guilty of hijacking this thread any further, so I will explain it to you in Dayton. I will start a new thread when I have some real results to post.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.