It was a good day.
Sunday, November 16, 2008
Friday, October 24, 2008
Fluid Jets
Alright, this one isn't about amplifiers, but I recently got accused of having a "dead blog," so as I'm sitting around the lab waiting for our copper vapor laser (more on that later!) to cool, I thought I'd share some awesome images with the world of some micron-sized bubbles undergoing asymmetric oscillation and the creation of a fluid jet.
Excited yet? I figured not. We'll start at the beginning. Check out this propeller:
Looks pretty sweet, huh? Judging by the damage on the lower left, you might think someone had a little too much of Grandpa's cough syrup and run the boat upon the shore or something. Well, boating drunk can certainly damage a propeller, but this damage has a pretty characteristic look to the trained eye, called "cavitation" damage.
In order to keep this entry fun, I would like you now to turn back the clock to the 1800s. You're the captain of a mighty steamboat and you motor around the seven seas fearless. However, every few months your propellers need to be replaced because you see damage similar to that shown above. Why the hell is this propeller, which just spins in the water, always getting this same damage??? Now, please return to the present with the understanding that this unexplained damage was certainly a mystery for a period of time.
Fortunately, the process slowly began to be understood (first modeled by Lord Rayleigh). As a propeller turns, it creates pockets of negative pressure in the fluid. These negative pressure areas are strong enough to pull gas out of the liquid creating small bubbles. Since negative pressure was necessary to pull the gas out of solution, when the negative pressure is no longer present, the bubble will collapse. So you have this small bubble (let's say a few millimeters in diameter) and it collapses. You might, think "oh, big deal. a small bubble bursts underwater." HOWEVER, this little bubble collapses so fast it will make your head spin. In some situations the bubble can collapse so fast it generates light through a process called sonoluminescence (the generation of light through sound). Think about that for a second--when was the last time you saw something move in such a way that it forced light to be emitted from surrounding molecules? Yeah, not too often. Pretty cool stuff if you ask me.
Sonoluminescence does not occur in the situation with the propeller to my knowledge, though. However, the bubble does collapse at a very high velocity near the boundary. A typical growth and collapse cycle of a bubble near a boundary looks something like the following images (from Kodama et al.):
What you see above is a fluid jet that is formed as the bubble collapses asymmetrically. In this case, the jet penetrates a gel surface, and the bubble is on the order of 1 mm in diameter. This all occurs quite quickly and the above images have a frame rate of 100 kHz (100,000 frames per second). The speed of this so-called "fluid jet" is 66 meters / second. Consider this in terms of say .. your typical super soaker water gun. Imagine if you fired that thing from the end zone of a football field and 1 second later it had crossed the 50 yard line. When you think of it this way, you can see how these small jets can cause some major damage. Clearly, there are equations that describe this activity in terms of the momentum of the fluid, but c'mon .. this is a blog.
This brings me to the beautiful image I want to share, which is a micron-sized bubble undergoing the creation of a fluid jet near a gel boundary. Since this bubble is only a couple of microns in diameter it oscillates much faster than the one above. Not to mention it is smaller, so it's a little harder to image. Anyway, with a very good microscope and a 30 nsec laser for a flash, I was able to capture the following image, which is a fluid jet impinging through the bubble near a boundary.
These bubbles are being researched for applications in localized drug delivery, where they can be injected (since they are the size of blood cells). During appropriate insonation parameters, you can see how a small fluid jet is created, which would impinge on the wall of a living vessel, potentially increasing the permeability of that vessel so a drug could reach beyond the vasculature.
And there you have it: the reason I have not been building amplifiers and writing about it :)
Excited yet? I figured not. We'll start at the beginning. Check out this propeller:
Looks pretty sweet, huh? Judging by the damage on the lower left, you might think someone had a little too much of Grandpa's cough syrup and run the boat upon the shore or something. Well, boating drunk can certainly damage a propeller, but this damage has a pretty characteristic look to the trained eye, called "cavitation" damage.
In order to keep this entry fun, I would like you now to turn back the clock to the 1800s. You're the captain of a mighty steamboat and you motor around the seven seas fearless. However, every few months your propellers need to be replaced because you see damage similar to that shown above. Why the hell is this propeller, which just spins in the water, always getting this same damage??? Now, please return to the present with the understanding that this unexplained damage was certainly a mystery for a period of time.
Fortunately, the process slowly began to be understood (first modeled by Lord Rayleigh). As a propeller turns, it creates pockets of negative pressure in the fluid. These negative pressure areas are strong enough to pull gas out of the liquid creating small bubbles. Since negative pressure was necessary to pull the gas out of solution, when the negative pressure is no longer present, the bubble will collapse. So you have this small bubble (let's say a few millimeters in diameter) and it collapses. You might, think "oh, big deal. a small bubble bursts underwater." HOWEVER, this little bubble collapses so fast it will make your head spin. In some situations the bubble can collapse so fast it generates light through a process called sonoluminescence (the generation of light through sound). Think about that for a second--when was the last time you saw something move in such a way that it forced light to be emitted from surrounding molecules? Yeah, not too often. Pretty cool stuff if you ask me.
Sonoluminescence does not occur in the situation with the propeller to my knowledge, though. However, the bubble does collapse at a very high velocity near the boundary. A typical growth and collapse cycle of a bubble near a boundary looks something like the following images (from Kodama et al.):
What you see above is a fluid jet that is formed as the bubble collapses asymmetrically. In this case, the jet penetrates a gel surface, and the bubble is on the order of 1 mm in diameter. This all occurs quite quickly and the above images have a frame rate of 100 kHz (100,000 frames per second). The speed of this so-called "fluid jet" is 66 meters / second. Consider this in terms of say .. your typical super soaker water gun. Imagine if you fired that thing from the end zone of a football field and 1 second later it had crossed the 50 yard line. When you think of it this way, you can see how these small jets can cause some major damage. Clearly, there are equations that describe this activity in terms of the momentum of the fluid, but c'mon .. this is a blog.This brings me to the beautiful image I want to share, which is a micron-sized bubble undergoing the creation of a fluid jet near a gel boundary. Since this bubble is only a couple of microns in diameter it oscillates much faster than the one above. Not to mention it is smaller, so it's a little harder to image. Anyway, with a very good microscope and a 30 nsec laser for a flash, I was able to capture the following image, which is a fluid jet impinging through the bubble near a boundary.
These bubbles are being researched for applications in localized drug delivery, where they can be injected (since they are the size of blood cells). During appropriate insonation parameters, you can see how a small fluid jet is created, which would impinge on the wall of a living vessel, potentially increasing the permeability of that vessel so a drug could reach beyond the vasculature.
And there you have it: the reason I have not been building amplifiers and writing about it :)
Sunday, August 17, 2008
6D30 - An amp builder's amplifier
I finished the Weber 6D30 sometime in June after a month of working on it in the evenings. For those who want to cut to the mustard, here are some sound samples:
12AX7 Channel (Random country riff)
EF86 (random country riff)
EF86 (random country riff #2)
EF86 (palm-mute distortion)
(Recorded with an Audio-technica condenser through a cheap preamp in a terrible sounding room)
I'm not huge on sound samples, since it depends so much on recording quality and the player. Not sure if you can tell much other than that it sounds like an amplifier that breaks with a certain character. I think it's hard to comment on how it sounds as a stock amplifier, since any one person might do something different in the build. Below are some of the high and low points of the building process and amplifier.
The Parts
Anyone who has built an amplifier from scratch knows that half of how the amplifier turns out is based on choosing the right parts for your purpose. Depending on your goals, you will make different decisions here. If you're trying to clone a specific vintage amplifier, you might go out of your way to acquire "obsolete" carbon comp resistors that are widely known to change value at different temperatures, paying as much as $3 for a single Riken or Allen-Bradley new old-stock (see the selection at Angela Instruments). So what was my goals with the 6D30? Create an amplifier that I can A/B various amplifier designs for a learning what kinds of circuits suit various sounds I want my guitar to make. If you're wondering why I don't just go amp shopping to figure this out, I just want to comment that it's really difficult to find out if you like an amplifier by playing it at a music store where there are people playing instruments all around you and sales guy trying to pitch to you why a specific amplifier is good. Even if you're lucky enough to be at a chill music store with a good sales people who are not annoying, I still find it hard to decide whether I like an amplifier until I hear it for a week or two or with the band I'm playing in. So what parts did I want for this amplifier? New ones that were relatively inexpensive, but not cheap and crappy. Weber excels in this category, and I think all of the parts met or exceeded my standards, except for the pilot light and tubes. Web forums had warned about the cheap pilot light, and I think this is relatively insignificant, although I will definitely replace this eventually. On the other hand, I think I understand why Weber does not include cool NOS-tubes: people oftentimes already have a preference, and nice tubes are $$$. Not to mention, who wants to plug in a high-dollar quad of Mullards into a circuit being tested? The stock tubes sound just fine, but they're not amazing. No big deal. The cabinet below is a pretty good example of the quality of workmanship that goes into the amplifiers:
Shag carpet not included.
The Layout
The panel design and circuit board was way cramped on this amplifier, although I liked the separation of the power supply and tone circuit. I visited Bryce Gonzales (local Sacramento amplifier expert) at BG Electric to help me iron out some oscillation issues. As you can see in the photo below, he pretty much moved the entire tone circuit of the 12AX7 from the board up to the pots and removed the effects loop:
He explained these moves showing me how the signal wires passed near the DC voltage wires. Oh, of course: the magnetic field generated around the signal wire will be inducted in loops of neighboring DC wires creating an oscillating power supply to the tubes. After his help, the amp stopped making non-harmonic noises and sounded fuller. Nice!
The Speakers
Uhmm .. Freaking amazing.
The Future and other notes
I think the amplifier is perfect for my goal of having a lot of things in one place that I can A/B. The 6D30 is the ultimate amplifier as far as having many options inside of a single box goes. HI/LO gain settings, a cathode-fixed bias switch, load impedance switching, built-in bias meter, .. the list goes on. As far as becoming familiar with various sounds goes, this amplifier is an amp builder's dream. I think as it stands, I would compare it most to a Mesa amplifier due to the bells and whistles. Among amplifiers I've played, it has a super fast response where the attack of all notes is well-pronounced, quick, and loud. The highs produced by this amplifier are decidedly Vox-like but with more grittiness.
Basically, I think the amplifier as built would work well for a guitarist who likes a Vox tone at high volumes. I really like the EF86 channel, but I am probably going to tone down the 12AX7 channel a bit more. My main complaints on this channel would be that it always seems to have too many highs that get a little "ice-pick" like at times. This would be a good lead tone, but I play rhythm/harmonic guitar, so this sound is not very useful to me. No worries, I will change out some capacitors, reduce the gain (maybe sub in a 12AT7 ??), and dial it in.
The big question for me with this amplifier is volume. I love playing it, but it's too loud. I currently use an attenuator at half volume or so, and I basically always play this amplifier on cathode mode. I am wondering about redesigning the output stage for 2 tubes to avoid dragging an attenuator around. And that friends, is the fun of this amplifier: you finish it, and all the parts are there to mod it into something else :)
12AX7 Channel (Random country riff)
EF86 (random country riff)
EF86 (random country riff #2)
EF86 (palm-mute distortion)
(Recorded with an Audio-technica condenser through a cheap preamp in a terrible sounding room)
I'm not huge on sound samples, since it depends so much on recording quality and the player. Not sure if you can tell much other than that it sounds like an amplifier that breaks with a certain character. I think it's hard to comment on how it sounds as a stock amplifier, since any one person might do something different in the build. Below are some of the high and low points of the building process and amplifier.
The Parts
Anyone who has built an amplifier from scratch knows that half of how the amplifier turns out is based on choosing the right parts for your purpose. Depending on your goals, you will make different decisions here. If you're trying to clone a specific vintage amplifier, you might go out of your way to acquire "obsolete" carbon comp resistors that are widely known to change value at different temperatures, paying as much as $3 for a single Riken or Allen-Bradley new old-stock (see the selection at Angela Instruments). So what was my goals with the 6D30? Create an amplifier that I can A/B various amplifier designs for a learning what kinds of circuits suit various sounds I want my guitar to make. If you're wondering why I don't just go amp shopping to figure this out, I just want to comment that it's really difficult to find out if you like an amplifier by playing it at a music store where there are people playing instruments all around you and sales guy trying to pitch to you why a specific amplifier is good. Even if you're lucky enough to be at a chill music store with a good sales people who are not annoying, I still find it hard to decide whether I like an amplifier until I hear it for a week or two or with the band I'm playing in. So what parts did I want for this amplifier? New ones that were relatively inexpensive, but not cheap and crappy. Weber excels in this category, and I think all of the parts met or exceeded my standards, except for the pilot light and tubes. Web forums had warned about the cheap pilot light, and I think this is relatively insignificant, although I will definitely replace this eventually. On the other hand, I think I understand why Weber does not include cool NOS-tubes: people oftentimes already have a preference, and nice tubes are $$$. Not to mention, who wants to plug in a high-dollar quad of Mullards into a circuit being tested? The stock tubes sound just fine, but they're not amazing. No big deal. The cabinet below is a pretty good example of the quality of workmanship that goes into the amplifiers:
Shag carpet not included.The Layout
The panel design and circuit board was way cramped on this amplifier, although I liked the separation of the power supply and tone circuit. I visited Bryce Gonzales (local Sacramento amplifier expert) at BG Electric to help me iron out some oscillation issues. As you can see in the photo below, he pretty much moved the entire tone circuit of the 12AX7 from the board up to the pots and removed the effects loop:
He explained these moves showing me how the signal wires passed near the DC voltage wires. Oh, of course: the magnetic field generated around the signal wire will be inducted in loops of neighboring DC wires creating an oscillating power supply to the tubes. After his help, the amp stopped making non-harmonic noises and sounded fuller. Nice!
The Speakers
Uhmm .. Freaking amazing.
The Future and other notes
I think the amplifier is perfect for my goal of having a lot of things in one place that I can A/B. The 6D30 is the ultimate amplifier as far as having many options inside of a single box goes. HI/LO gain settings, a cathode-fixed bias switch, load impedance switching, built-in bias meter, .. the list goes on. As far as becoming familiar with various sounds goes, this amplifier is an amp builder's dream. I think as it stands, I would compare it most to a Mesa amplifier due to the bells and whistles. Among amplifiers I've played, it has a super fast response where the attack of all notes is well-pronounced, quick, and loud. The highs produced by this amplifier are decidedly Vox-like but with more grittiness.
Basically, I think the amplifier as built would work well for a guitarist who likes a Vox tone at high volumes. I really like the EF86 channel, but I am probably going to tone down the 12AX7 channel a bit more. My main complaints on this channel would be that it always seems to have too many highs that get a little "ice-pick" like at times. This would be a good lead tone, but I play rhythm/harmonic guitar, so this sound is not very useful to me. No worries, I will change out some capacitors, reduce the gain (maybe sub in a 12AT7 ??), and dial it in.
The big question for me with this amplifier is volume. I love playing it, but it's too loud. I currently use an attenuator at half volume or so, and I basically always play this amplifier on cathode mode. I am wondering about redesigning the output stage for 2 tubes to avoid dragging an attenuator around. And that friends, is the fun of this amplifier: you finish it, and all the parts are there to mod it into something else :)
Friday, July 18, 2008
To live in a bike city
The other day a friend of mine who just moved over near Chapel Hill, NC asked me to send him some pictures of some of bike-friendly Davis for a bike planning committee there. Although Davis is a pretty expensive place to live, it is nice island in commuter-drive, suburbia-ridden America. The following pictures were taken on a 10-minute ride across town. If I were seeking things out, I would have also photographed the Dave Pelz Bike Overpass. Bikes save more than just gas--they save space and enforce good city planning.
Thursday, June 26, 2008
Adding a Master Volume to the Traynor Bassmaster
(Usual tube amp disclaimer: These amplifiers contain voltages that can kill you if you do not understand what you're doing, so please understand what you're doing if you decide to attempt anything described on this page.)
Today, I want to discuss the addition of a master volume control to the Traynor YBA-1. When I first heard about amplifiers that lacked a master volume, I was actually pretty confused because I thought this meant the amplifier had a fixed gain and no volume control. Fortunately, this is not the case, and non-master volume amplifiers have only a single volume control, and it is located in the preamp stage of the amplifier. Below is the general signal flow from guitar to speaker in one of these setups:
So here you can control the volume of the output by just one switch. Some people at this point might be thinking: now why would you need TWO volume pots for a device that has a single output? This is a reasonable thought, since a single volume control can control the output. However, each tube has a slightly different "flavor" of potentially desirable distortion. With one knob, the guitar player can certainly create distortion but distortion from the tubes will only occur at a specific volume. Your amp will need to be quite loud to create nice juicy distortion, potentially annoying all of your neighbors/roommates/bandmates. Master volume helps with this problem by pulling out some of the power right before the signal reaches the output stage. This allows only the preamp (and Phase Inverter) stages of the amplifier to be run "all out," while some of the signal is drained to ground prior to entering the output stage, thereby allowing a quieter sound from the speaker yet maintaining the distortion. The distortion characteristics of so-called "preamp distortion" certainly differ from "all out" distortion where the output tubes and speakers are involved as well.
Common Master Volume Method
On a push-pull output amplifier, a typical place to put the Master Volume is right after the phase inverter (PI). The 6D30 I just built has the master volume there, and I will implement the MV in simulation below in the Push-Pull EL84 example from Duncan Amps using LTSpice (free simulation software). If you're thinking about learning circuits, I can't say enough times how nice it is to have a good simulator handy. I think if you can read schematics that the tutorial here should be enough to follow along with the master volume implementation I will show. I start with the Duncan Amps push-pull EL84. A picture of the full schematic is shown below:
I have circled the individual stages of the amplifier as a guide to understand what's going on. Working from left to right, the RED box indicates the input. The next green box is the preamp stage. In this case, it adds about 30 dB of gain. The next stage (yellow) is the power output stage, followed by what I consider the "actuator stage." The actuator stage is basically where the signal goes from being an electrical current to a sound. The top area consists of the power supply. Since the master volume resides at the power output stage, let's focus on that. It is cut out below:
Basically, the resistance between the two points on the left side will determine the amount of signal passed to the 6CA7 tubes. Right now, the resistance between A and B is infinite. If you choose a large valued resistor and make it variable, it can function as a master volume here. A 1 MegaOhm resistance seems pretty common as a value that functions as an approximation to infinity in this position. Let's see how this works:
The gain at 1 kHz BEFORE the addition of R14 is 53 dB, and with the addition of R14, and after the addition, it is still approximately 53 dB. Therefore we have successfully added something, that does not affect our circuit (at least as far as gain is concerned). Dialing R14 down to from 1000k to 10k (two decades), the gain at 1 kHz is reduced to 50 dB (-3 dB => half). This trend continues at 1k where the gain is reduced to 38 dB. Given the very wide range of values required for this effect to take place, it is worth mentioning that you would only want to use an audio-tapered potentiometer in this position for it's logarithmic response (smooth over wider range). So there you have it, one method for implementing master volume!
Another MV Method Implemented On Traynor YBA-1
On the Traynor YBA-1, I implemented a slightly different MV control. The output section looked a little weird to me, so I opted for tapping some of the voltage off to ground before it made it to the output section. If you are interested in YBA-1 schematics, I have posted them here. The schematic for this mod is shown below:
The area on the left is the output from the preamp/phase inverter stages, and the 1 Meg potentiometer is added in sequence. Physically achieving this on the stock amplifier was not a terrible challenge and required no drilling, since I had an extra hole on the chassis where I had removed the 'ground switch.' As far as I could tell, the two-pronged plug with a ground switch was basically just a device for sending 120VAC from your arm to whatever was grounded. People always say to convert 2-prong tube amps to 3-pronged with proper grounding. Well, this amplifier made me a believer. A photo of the modded amplifier below:
The bright yellow wire that is heat-shrinked to the dull yellow wire running from the front panel to the back shows where the mod is connected. Basically, it works pretty well, although it is a little annoying that the MV is at the back of the amp. Ah well, there is no room on the front regardless.
Recap
Okay, today we have gone through two methods of implementing master volume on tube amplifiers that lack this feature. As always, these new 'features' do not come for free. The first method mentioned adds a HP filter that changes with the volume level, and you might find a similar effect with the second method, depending on the tone circuit in the preceding stage. You might try experimenting around with these MV methods in LTSpice to choose the one you want to implement if you're thinking about it. Or just plug them in and see which one you like better. Another MV method I did not talk about but will probably implement on the Traynor at some point is one I've heard referred to as the Ken Fischer #3. From what I understand, this is a master volume that drains the gain from all stages of the amplifier at the same time via a "ganged" potentiometer. This is a pretty interesting idea that I believe would change the overdrive characteristics by a good bit.
Today, I want to discuss the addition of a master volume control to the Traynor YBA-1. When I first heard about amplifiers that lacked a master volume, I was actually pretty confused because I thought this meant the amplifier had a fixed gain and no volume control. Fortunately, this is not the case, and non-master volume amplifiers have only a single volume control, and it is located in the preamp stage of the amplifier. Below is the general signal flow from guitar to speaker in one of these setups:
So here you can control the volume of the output by just one switch. Some people at this point might be thinking: now why would you need TWO volume pots for a device that has a single output? This is a reasonable thought, since a single volume control can control the output. However, each tube has a slightly different "flavor" of potentially desirable distortion. With one knob, the guitar player can certainly create distortion but distortion from the tubes will only occur at a specific volume. Your amp will need to be quite loud to create nice juicy distortion, potentially annoying all of your neighbors/roommates/bandmates. Master volume helps with this problem by pulling out some of the power right before the signal reaches the output stage. This allows only the preamp (and Phase Inverter) stages of the amplifier to be run "all out," while some of the signal is drained to ground prior to entering the output stage, thereby allowing a quieter sound from the speaker yet maintaining the distortion. The distortion characteristics of so-called "preamp distortion" certainly differ from "all out" distortion where the output tubes and speakers are involved as well.
Common Master Volume Method
On a push-pull output amplifier, a typical place to put the Master Volume is right after the phase inverter (PI). The 6D30 I just built has the master volume there, and I will implement the MV in simulation below in the Push-Pull EL84 example from Duncan Amps using LTSpice (free simulation software). If you're thinking about learning circuits, I can't say enough times how nice it is to have a good simulator handy. I think if you can read schematics that the tutorial here should be enough to follow along with the master volume implementation I will show. I start with the Duncan Amps push-pull EL84. A picture of the full schematic is shown below:
I have circled the individual stages of the amplifier as a guide to understand what's going on. Working from left to right, the RED box indicates the input. The next green box is the preamp stage. In this case, it adds about 30 dB of gain. The next stage (yellow) is the power output stage, followed by what I consider the "actuator stage." The actuator stage is basically where the signal goes from being an electrical current to a sound. The top area consists of the power supply. Since the master volume resides at the power output stage, let's focus on that. It is cut out below:
Basically, the resistance between the two points on the left side will determine the amount of signal passed to the 6CA7 tubes. Right now, the resistance between A and B is infinite. If you choose a large valued resistor and make it variable, it can function as a master volume here. A 1 MegaOhm resistance seems pretty common as a value that functions as an approximation to infinity in this position. Let's see how this works:
The gain at 1 kHz BEFORE the addition of R14 is 53 dB, and with the addition of R14, and after the addition, it is still approximately 53 dB. Therefore we have successfully added something, that does not affect our circuit (at least as far as gain is concerned). Dialing R14 down to from 1000k to 10k (two decades), the gain at 1 kHz is reduced to 50 dB (-3 dB => half). This trend continues at 1k where the gain is reduced to 38 dB. Given the very wide range of values required for this effect to take place, it is worth mentioning that you would only want to use an audio-tapered potentiometer in this position for it's logarithmic response (smooth over wider range). So there you have it, one method for implementing master volume!Another MV Method Implemented On Traynor YBA-1
On the Traynor YBA-1, I implemented a slightly different MV control. The output section looked a little weird to me, so I opted for tapping some of the voltage off to ground before it made it to the output section. If you are interested in YBA-1 schematics, I have posted them here. The schematic for this mod is shown below:
The area on the left is the output from the preamp/phase inverter stages, and the 1 Meg potentiometer is added in sequence. Physically achieving this on the stock amplifier was not a terrible challenge and required no drilling, since I had an extra hole on the chassis where I had removed the 'ground switch.' As far as I could tell, the two-pronged plug with a ground switch was basically just a device for sending 120VAC from your arm to whatever was grounded. People always say to convert 2-prong tube amps to 3-pronged with proper grounding. Well, this amplifier made me a believer. A photo of the modded amplifier below:
The bright yellow wire that is heat-shrinked to the dull yellow wire running from the front panel to the back shows where the mod is connected. Basically, it works pretty well, although it is a little annoying that the MV is at the back of the amp. Ah well, there is no room on the front regardless.Recap
Okay, today we have gone through two methods of implementing master volume on tube amplifiers that lack this feature. As always, these new 'features' do not come for free. The first method mentioned adds a HP filter that changes with the volume level, and you might find a similar effect with the second method, depending on the tone circuit in the preceding stage. You might try experimenting around with these MV methods in LTSpice to choose the one you want to implement if you're thinking about it. Or just plug them in and see which one you like better. Another MV method I did not talk about but will probably implement on the Traynor at some point is one I've heard referred to as the Ken Fischer #3. From what I understand, this is a master volume that drains the gain from all stages of the amplifier at the same time via a "ganged" potentiometer. This is a pretty interesting idea that I believe would change the overdrive characteristics by a good bit.
Sunday, June 8, 2008
Examination of a 6SN7 Preamp Circuit
Today, I will simulate a cathode-biased amplifier using 6SN7 tube model and compare the simulation to the data sheet. Using the data sheet acquired from the Tube Data Sheet Locater, we can see the schematic for a cathode-biased amplifier circuit and various values for resistances in the circuit. This example is derived from the excellent tube HOWTO at Boozhound Labs. A similar circuit analysis will be simulated here with a test for the accuracy of the onset of distortion in the 6SN7 tube in the simulation.
Input Capacitor
Viewing the circuit in the above chart, you can see Esig, which is the input signal for the circuit. There is initially a capacitor in the circuit for the purpose of coupling the signal to a prior stage or instrument. In case this has not been mentioned previously, the purpose of a capacitor in this position is two-fold:
1) Block the DC voltage from a prior stage
2) Pass the AC signal to the amplification stage
Input Resistor
So the signal travels from a source (such as a guitar or record player) through this capacitor. The next thing in the chain is the input resistor Rg1. The basic requirement for this resistor is that it be large enough to dominate the voltage divider formed by the source impedance. As you can see, we can choose any value in the 3rd column for this resistor, ranging from .1 to 10 MOhms. Following the example on Boozhound, I am choosing 100k. This is a pretty typical value for a record or guitar amplifier. Now the voltage across Rg1 is our input signal, which exists at the grid of the 6SN7.
Load Resistor
The next decision we must make is based on what we will be driving with this tube. Again, following the Boozhound example, we will be driving the input to a subsequent stage of the circuit. In this case, we are driving a 240kOhm resistance. So far we have
Rg1 = 100k
Rs = 240k
Therefore we direct our attention to the only row where Rg1 and Rs have these values (row 2). Here you can see we have a few options as far as the plate voltage goes: 90V, 180V, and 300V. You can see by comparing the columns for Eo at each voltage that the plate voltage determines how much output the amplifier can drive. For instance, looking at the row we have chosen to work with based on the input and load resistance (again, row 2), if Ebb=90V, the maximum output from this amplification stage is 16 volts RMS before the onset of distortion. Increasing Ebb to 180V allows the tube to produce 33 Vrms before the onset of distortion. I have chosen 180V. After decisions for input resistance, output load, and plate voltage are made, you pretty much just need to match the values to the parts in the schematics and simulate. I have made this file for LTSpice and stored it here. A drawing of the file is here:
Accuracy of Simulation Compared to Data Sheet
According to the first table, we expect the amplifier to have a voltage gain of 15 with the onset of distortion occurring when the amplifier tries to produce a signal with an RMS value of 33V. Using an input signal of 1V for normalization, we can get the following plot to test the transfer function of the equation:
Reading this plot, you can see that the gain is approximately 24 dB. “Argh,” I hear you moan in response to having to compare dB to a simple multiplicative gain. Honestly, I completely agree with you—this is annoying. However, decibels are nice for describing amplification because numbers like 1500X as loud start sounding a little silly. And even though this stage is only 15X the original signal, it will be fed into another stage that will multiply that signal by more, so the overall gain gets out of hand pretty quick. Back to the question at hand of whether the simulation is accurate, we must answer whether 24 dB gain is the same as 15X the original signal. The equation for the output gain in dB is
# dB = 20 * log(Vout/Vin)
Solving for dB, using Vout/Vin = 15 yields 10^1.2, which is approximately 15.8. Therefore, the simulation is pretty true to the data sheet. While we’re at it, I think it would be instructive to discuss the onset of distortion. Distortion is an important topic for both Hi-Fi tube amplifiers and guitar amplifiers, since you generally want to avoid distortion in the former and sometimes you want to create it in the latter. Viewing the data sheet you can see that the amplifier will experience 5% harmonic distortion when trying to create an output with an RMS-voltage of 33V. This time, I will use the transient response capabilities of LTSpice to simulate the scenario where the amplification creates a 33V output signal. I will use a 1 kHz signal for the simulation. From the frequency response, you can see the 1 kHz signal will be amplified by about 22 dB (12.6X the original signal using the above equation). Using the gain and definition of RMS voltage, you can find that a voltage of 3.6V will cause the gain stage to try to drive approximately 33V RMS, thereby introducing 5% harmonic distortion. This is shown below.
You can see a small amount of asymmetry in the negative cycles of the waveform, but this looks pretty nice, so presumably 5% harmonic distortion, although probably not an acceptable parameter for audiophiles with good ears, is not visible to the eye at least. To test if we are near a threshold point, I am running the simulation again with a 6V input signal at 1 kHz (shown below):
Here you can see the characteristic onset of distortion in the negative cycle of the waveform. This is all pretty cool, if I do say so myself, since it seems the simulation and data sheet match fairly well.
Cathode Bypass Gain Switch
I will do an analysis a bit later regarding how this sounds, but for now, I would like to discuss a simple method of adding a gain switch to this preamp stage. This method of adding gain is very simple. All I am going to do is add a switch to the cathode capacitor (Capacitor C1 in the circuit). The tube’s amplification abilities are related to the voltage at the cathode. There is always an AC signal at the cathode related to the input signal due to the nature of the tube. The role of C1, called the cathode bypass capacitor, is to filter this signal so that the node at the cathode is mostly zero. Removal of this capacitor allows the AC signal to exist at the cathode, thereby creating fluctuations in the cathode voltage. These fluctuations result in a reduced gain. The frequency response of the amp without the capacitor is shown below:
Wow! We have reduced the gain from 24 dB to 19 dB. You will also note that this capacitor was acting as a HP filter when in the circuit, since the low frequency response is much more flat. Two things can be learned from this: 1) you can put a switch to connect or disconnect the capacitor C1 to create a +5 dB gain stage and 2) if you want to maintain a flat frequency response, you should be careful in choosing your cathode bypass capacitors, since they are HP filters.
I will wax philosophical for a moment here and would love any input from more experienced builders regarding the topic of cathode bypass. From the above analysis, it seems like this is one place where the goals of a guitar amplifier and Hi-Fi amp might be at odds. Generally, it seems like Hi-Fi amps would want to avoid bypass caps as much as possible to maintain signal fidelity, while guitar amplifiers might want to use the capacitors to achieve higher gain, earlier onset of preamp distortion, etc.. Any thoughts?
Okay amp lovers. Today we have learned
1) How to pick parts for building a cathode-biased preamplifier stage using the data sheet for a 6SN7 tube
2) How to simulate said circuit in LTSpice
3) Made comparisons between data sheet and simulation for amplification parameters and distortion onset
4) Discussed the use of cathode bypass capacitors as a gain switch
That is all for now. Now, go huff some solder fumes, grill some bratwurst, have a beer, or go biking. Personally, I am gonna do all four.
Thursday, May 15, 2008
Preamp stage of 6D30 simulated
I've been doing a lot of wiring on the amp and realizing some of the subtleties of circuit design, as well as the practical aspects of fitting all of this stuff into a combo cabinet. The final sound is really where the proof lies, but elegance can be recognized anywhere. In the spirit of learning about each stage of the amp, I will try to simulate as many of them as possible. Below is the schematic for the first preamp stage of the 6D30. I have employed here, the generic triode model from Duncan Amps.
In case you are interested, the file for the schematic is here.
The image plots the input signal and the voltage across the 500k resistor on the output. 300 mV is a reasonable approximation to a guitar signal. As you can see, this is a pretty big amplification from a small signal. This is why the wiring from the input jack should be both SHORT and SHIELDED. Due to the low signal level at the input, a smaller amount of noise can drive your signal to noise ratio (SNR) very low fast. Then every stage downstream amplifies your noisy signal resulting in .. loud noise mucking up your sweet guitar tone! :(
The other thing worth noting in this circuit is the high-pass filter that couples this stage to the next stage. This consists of C2 and R6 and is a typical RC circuit. The -3 dB (half-amplitude) point for this circuit is ~600 Hz (found by using 1/(2*pi*R*C)). Why am I mentioning this? The fundamental frequency range for guitar is from 80 Hz to 1 kHz. Since 600 Hz and below are attenuated by more than half in the first stage of this circuit, the guitar will certainly be affected. Frequencies above 1 kHz account for the "timbre" of a guitar, so none of that is lost. If channel 1 has too many highs, this might not be a bad place to throw in some more low frequencies by increasing the value of the C2, and dropping the -3 dB to more like 400 Hz.
As a note to anyone building this amp, the resistor on the layout says 1M, but the package contains a 500K resistor, and 500K is what the schematics say. I have noticed one should trust the schematic more than anything with the kit. The wiring layout is really confusing to me, since it seems to be mirrored and upside down. Maybe I'm just not thinking about it right.
In case you are interested, the file for the schematic is here.The above circuit seems innocent enough: two inputs running into the circuit via the ubiquitous 68k resistor and separate 12AX7 stages mixed at the output with a gain of ~30 (15 dB). Thus, given a 300 mV sinusoid, the voltage across the 500kOhm resistor will be ~8.7V sinusoid. This can be seen below:
The image plots the input signal and the voltage across the 500k resistor on the output. 300 mV is a reasonable approximation to a guitar signal. As you can see, this is a pretty big amplification from a small signal. This is why the wiring from the input jack should be both SHORT and SHIELDED. Due to the low signal level at the input, a smaller amount of noise can drive your signal to noise ratio (SNR) very low fast. Then every stage downstream amplifies your noisy signal resulting in .. loud noise mucking up your sweet guitar tone! :(
The other thing worth noting in this circuit is the high-pass filter that couples this stage to the next stage. This consists of C2 and R6 and is a typical RC circuit. The -3 dB (half-amplitude) point for this circuit is ~600 Hz (found by using 1/(2*pi*R*C)). Why am I mentioning this? The fundamental frequency range for guitar is from 80 Hz to 1 kHz. Since 600 Hz and below are attenuated by more than half in the first stage of this circuit, the guitar will certainly be affected. Frequencies above 1 kHz account for the "timbre" of a guitar, so none of that is lost. If channel 1 has too many highs, this might not be a bad place to throw in some more low frequencies by increasing the value of the C2, and dropping the -3 dB to more like 400 Hz.
As a note to anyone building this amp, the resistor on the layout says 1M, but the package contains a 500K resistor, and 500K is what the schematics say. I have noticed one should trust the schematic more than anything with the kit. The wiring layout is really confusing to me, since it seems to be mirrored and upside down. Maybe I'm just not thinking about it right.
Subscribe to:
Posts (Atom)