Thanks for explaining which PCB is with LED on negative. This is my case. Thanks a lot. Will go to check it. Interesting why there is not polarity labeling on PCB for LED… it will make life so much easier for newbies like me
If the PCB is still unmounted, hold it up to a strong light and follow the traces. "Illuminating!" to make a pun.
As others have pointed out, the boards aren't mirrored and the rails flip between them. Seeing where the ground traces are helps learn the layout and how it matches to the schematic.
As others have pointed out, the boards aren't mirrored and the rails flip between them. Seeing where the ground traces are helps learn the layout and how it matches to the schematic.
I agree. I always thought that was an oversight.Interesting why there is not polarity labeling on PCB for LED… it will make life so much easier for newbies like me
Have fun with the rest of your build.
Cheers,
Dennis
This is exactly what I was doing for two evenings!The man who invented the USB port was buried today. The put him in the ground, pulled him out, flipped him over, and put him back in again.
Same with the LED.
Thanks for advice.
I turned the polarity of that misleading LED and have just finished setting up F6 and made some listenings with test speakers. Everything went smoothly. Good temperatures on MOSFETs, BIAS set up to 600 mV (+- 1 mV or less), about 0,0 mV offset (+- 0,5 mV or less) on both channels. By the way, is 60* C ok for Jfets? Haven’t seen any info about them, only about MOSFETs.
It seems to be, that everything is good. Very glad about that! Thank you Dennis and 6sX7 for your help with that LED.
Gonna build Troels Gravesen’s Discovery-861 speakers this week. Very excited to hear this pair with F6.
This is a work of beauty. Wow! I've built a F6, but now I want to modify it, following your lead.Finished this F6 recently; my first DIY amp. Reading schematics and holding a soldering iron are familiar things for me, but there's no doubt in my mind this project succeeded on the first try only due to the wealth of knowledge shared here!
Deluxe 4U case, PSU boards, amplifier boards with transformers and transistors all sourced from the store. Antek 150VA power transformers. Filter caps from Apex Jr and most everything else from Mouser. I built this amp stock save for the picomods and dual mono power supplies. Currently biased at 550mv. I found nothing tricky at all to implement the dual mono configuration; each PSU drives its own channel with each PSU board grounded through its own CL60 to the single star ground. This power supply configuration allows short wiring runs but due to my stacked PSU getting to the bias adjustments is tricky. I left the wires to all panels long so I may remove any heatsink or panel without disconnecting any wiring. Final biasing had the case stacked together like a house of cards; only bolting it together once that was complete
Layout of transformers and rectifier bridges. The transformer on right does not yet have all its wires dressed; finished state will be as the left side. You can see the all thread standoffs in place that will support the PSU boards above the transformers. Heatsinks and amplifier boards only mocked up here; heat shrink is still in place to protect from scratches.
View attachment 1111312
All PSU wiring completed and being tested. PSU boards mounted in place to standoffs using riser panels from the store. I wished to mount the PSU boards using the holes between the filer caps and therefore could not use a simple standoff as the transformers would be in the way. The riser panels mount to the all thread standoffs, and the PSU boards than mount to the riser with wiring from the rectifiers coming up through the the middle. Notice that the riser panels are notched to clear the caps on the amp boards and how access to the bias and offset pots is not convenient. Wiring of IEC terminal and power transformers is as all other Firstwatt amps with two CL60 to limit inrush and a safety cap across the line voltage. Difficult to make out in the picture but the standoffs nearest the line voltage terminal block each have pieces of neoprene tubing slid over them as extra insurance against abrading the power wires.
View attachment 1111316
Completed wiring. You can make out the CL60 for ground on each PSU board. The green wire from each heads to the star ground where the IEC ground and transformer static shields are also connected. Shielded coax for signal inputs; it and the line voltage input wiring again left long so the back panel can be unbolted and laid flat if desired.
View attachment 1111322
Initial biasing in progress before stacking the case together to "cook" for an hour. Final setting was 550mV with 0V offset.
View attachment 1111405
How I was able to easily mount the front panel. Notice nut driver with extension and allen bit attached. This made what would have been a tedious operation a breeze by zipping all four screws right in! Also a good shot of final wiring. Coiled twisted pair goes to LED indicator in face plate. Also a good shot showing notches in riser panel to clear caps on amp board.
View attachment 1111324
Time to solder the input coax to the amp boards and close up rear panel.
View attachment 1111406
Good overall internal shot. There's about 1/4" clearance between the filter caps and top cover when fitted.
View attachment 1111407
Big sexy beast at just over 40 pounds! Handles are rose gold kitchen cabinet pieces from Amazon. LED displays are holes drilled through front panel and then filled with hot glue. LED's are then hot glued onto back of faceplate so glue acts as a diffusor. Only thing I will change at this point is to up the resistance to the display LED's as I find this too bright. I may add an additional power switch to the front panel in the future but at same time I tend to leave my amps on 24 hrs a day so not a necessity.
View attachment 1111327
Gratuitous amp porn shot
View attachment 1111330
This amp has floored me! It drives a pair of DIY Jim Holtz/Curt Campbell Bordeaux speakers and I was concerned I'd be lacking power as this amp replaces a PAIR of Carvers that were each bridged to over 200W as monoblocks. My speakers are setup to accommodate bi-amping and if the F6 lacked grunt the back up plan was to run it through the mids and ribbons only and leave the bass to something else. I was gobsmacked when I first installed the F6! I can only describe it as fantastic sounding and cannot believe how deep the bass goes and how well controlled it is. The midrange and ribbons are also signing like they never have and these speakers now show me what they are truly capable of. This "little" amp beats the crap out of what was there before and can take on all comers; regardless of power. After adjusting speaker levels in my processor I can still push the volume to uncomfortable levels if desired with absolutely zero noise. This project has exceeded my expectations in every way thanks to all who've shared along the way. Enjoy the music!
A couple of questions:
Could you provide the model # of the Antek transformers you used?
What are the values of the Apex Jr caps, I can't quite read them in the photos.
I used Mark's low voltage start and assume I can fit it in here.
I've been wondering what my next project should be and thanks to you, now I know.
It is ingenious the way you laid it all out. Congrats!
I didn't know that Antek Made a 150va 18v transformer. I know there is a 200va 18v with the measurements on the website. That would work nicely assuming the measurements still work out. If you use the typical 25v-35v 15000uf caps, they should be smaller by quite a lot. However, you could even do 22,000 with the soft start if you wanted to.
I like to step up the voltage to 35v in case you decide to repurpose the power supplies to something like a BA-3 at some point or something like that. They don't cost a lot more.
I like to step up the voltage to 35v in case you decide to repurpose the power supplies to something like a BA-3 at some point or something like that. They don't cost a lot more.
Mikerodrig27 is correct; my memory was poor. The power transformers are 200VA 18V+18V; Antek model AS-2218. The Apex Jr caps are 15,000 mfd each and rated for 63V @ 105 degrees C. They should last pretty much indefinitely, though they are taller than most.
I'm recently learning about the F6 and have been reading up on the various elements of the the circuit. The Pass lecture gave me my first peak into what is going on with the various stages, the bias circuit obviously being key. But now I'm scratching my head about some basic power calculations regarding the stepdown to bias voltage from the PSU. If you operate with a 550mV bias at 1.3A and that is being supplied via 23V.... is that not a ~30 watt loss to the stepdown? I must be missing something obvious, otherwise wouldn't it make sense to have a bias dedicated supply that is not having to scrap 98% of the voltage on tap?
Edit: Now that I have posted this it dawns on me that the 1.3A must on the other end, or not directly coming from the stepdown but rather part of the output.
Edit: Now that I have posted this it dawns on me that the 1.3A must on the other end, or not directly coming from the stepdown but rather part of the output.
Last edited:
One thing that might help is not to conflate '550mV bias' with anything. It's not 550mV bias. It's simply 1A3 bias. It helps me to always add the key word... bias current. It's not bias voltage. It's an error I made frequently until it was pointed out to me, and one I still see often.
Do you mean this lecture from BAF? https://firstwatt.com/pdf/art_f6_baf.pdf
I also can't reconcile your 550mV against a 1A3 bias current with the circuit in post #1. When you site a voltage measurement, it's extremely helpful to site where in the circuit it's measured / calculated and the schematic. Generally the bias current is determined by measuring the voltage drop across either of the source resistors and using Ohm's law to calculate the current through them. 550mV across R2 is a target I've seen posted. Did you / are you planning on changing the values of the source resistors and/or did I misunderstand? Is that the circuit you're referencing?
550mV across R1 => 0A982 (550mV / 0R56)
550mV across R2 => 1A17 (550mV / 0R47)
If you wanted a bias current of 1A3 you would need 0V611 across R2 and 0V728 across R1 (Assuming DC offset of 0V00 and assuming we're talking about the same circuit)
What do you mean by 'step-down' and 'other end'? If you show the math you used, it may be easier to determine what you're asking.
Are you looking to calculate the rough 'power' available at a certain load in Watts vs. heat 'dissipation' in Watts? You said you were doing power calculations. Is the difference between the two what you're calling step down? Are you wondering how you get from dual rail +- 23V to a measurement of mV across the source resistors? Again, I may be way off base in understanding what you're trying to determine / understand. If so, apologies.
All disclaimers still apply... I think I've got this correct, but I may not. It helps me to learn alongside, so I try. Wiser minds will likely zero in on exactly the situation much faster than I.
Do you mean this lecture from BAF? https://firstwatt.com/pdf/art_f6_baf.pdf
I also can't reconcile your 550mV against a 1A3 bias current with the circuit in post #1. When you site a voltage measurement, it's extremely helpful to site where in the circuit it's measured / calculated and the schematic. Generally the bias current is determined by measuring the voltage drop across either of the source resistors and using Ohm's law to calculate the current through them. 550mV across R2 is a target I've seen posted. Did you / are you planning on changing the values of the source resistors and/or did I misunderstand? Is that the circuit you're referencing?
550mV across R1 => 0A982 (550mV / 0R56)
550mV across R2 => 1A17 (550mV / 0R47)
If you wanted a bias current of 1A3 you would need 0V611 across R2 and 0V728 across R1 (Assuming DC offset of 0V00 and assuming we're talking about the same circuit)
What do you mean by 'step-down' and 'other end'? If you show the math you used, it may be easier to determine what you're asking.
Are you looking to calculate the rough 'power' available at a certain load in Watts vs. heat 'dissipation' in Watts? You said you were doing power calculations. Is the difference between the two what you're calling step down? Are you wondering how you get from dual rail +- 23V to a measurement of mV across the source resistors? Again, I may be way off base in understanding what you're trying to determine / understand. If so, apologies.
All disclaimers still apply... I think I've got this correct, but I may not. It helps me to learn alongside, so I try. Wiser minds will likely zero in on exactly the situation much faster than I.
As Patrick says, the bias current is being measured as a voltage drop across a resistor in the circuit - the voltage drop (in this case 550mV) across the 0.47ohm resistor shows that the current is about 1.2A .
The voltage is not dropping from the 23V rail to 0.55V, merely that current is being drawn across the resistor, with a small drop of course, and measuring that drop (in our case 550mV) is a convienent place to use Ohm’s law and determine the bias.
The voltage is not dropping from the 23V rail to 0.55V, merely that current is being drawn across the resistor, with a small drop of course, and measuring that drop (in our case 550mV) is a convienent place to use Ohm’s law and determine the bias.
Ah! this:
"The voltage is not dropping from the 23V rail to 0.55V, merely that current is being drawn across the resistor, with a small drop of course, and measuring that drop (in our case 550mV) is a convienent place to use Ohm’s law and determine the bias."
@ItsAllInMyHead yes that is the lecture I was referring to. As far as where I was coming from on the power requirement, 6L6 zeroed in on what I was getting at. I need to read up more on how the measurements are taken is probably a good next step. From a very basic standpoint I was just looking at the rail voltage and getting to thirty watts based on bias current. AND I was misconstruing the meaning of the voltage measurements being mentioned elsewhere in the threads. The "other end" remark was my assumption that 1.2A is not seen directly at the bias circuitry (which was my original impression) hence it must be showing up at the output.
I'm still somewhat confused. What I think I understand is that the bias is a steady/constant source of DC injected into the AC signal because reasons (for the FETS and facilitating class A). NP mentioned in the lecture that you could supply the bias using a 1.5V battery. So, I see Papa mention running the bias from 1.5V, and I see the bias current averaging 1.3A. Based on these observations, is it safe to say that the bias circuit is a sub 2 watt DC source injected into the signal in order to properly drive the gate? But that can't be correct, because I'm fairly sure 1.3A is not actually going through the AC secondary and into the gate. So realistically the circuit must be 1.5V or less and some minimal current more acceptable to the fine magnet wire of the secondary and of the small resistors used.
Thanks gents!
"The voltage is not dropping from the 23V rail to 0.55V, merely that current is being drawn across the resistor, with a small drop of course, and measuring that drop (in our case 550mV) is a convienent place to use Ohm’s law and determine the bias."
@ItsAllInMyHead yes that is the lecture I was referring to. As far as where I was coming from on the power requirement, 6L6 zeroed in on what I was getting at. I need to read up more on how the measurements are taken is probably a good next step. From a very basic standpoint I was just looking at the rail voltage and getting to thirty watts based on bias current. AND I was misconstruing the meaning of the voltage measurements being mentioned elsewhere in the threads. The "other end" remark was my assumption that 1.2A is not seen directly at the bias circuitry (which was my original impression) hence it must be showing up at the output.
I'm still somewhat confused. What I think I understand is that the bias is a steady/constant source of DC injected into the AC signal because reasons (for the FETS and facilitating class A). NP mentioned in the lecture that you could supply the bias using a 1.5V battery. So, I see Papa mention running the bias from 1.5V, and I see the bias current averaging 1.3A. Based on these observations, is it safe to say that the bias circuit is a sub 2 watt DC source injected into the signal in order to properly drive the gate? But that can't be correct, because I'm fairly sure 1.3A is not actually going through the AC secondary and into the gate. So realistically the circuit must be 1.5V or less and some minimal current more acceptable to the fine magnet wire of the secondary and of the small resistors used.
Thanks gents!
Last edited:
Thanks for asking the question and to @ItsAllInMyHead and @6L6 for responding.
This conversation helped test and solidify my understanding of some concepts.
Learning and understanding electronics is a long journey I like to compare to getting to scratch at golf where some pick it up faster than others, but you have to learn a LOT of things along the way. And you have to be willing to go back from time to time and revise your understanding from almost (pardon the pun) ground zero. With golf, I've been trying for 30 years and shoot low 80s, but my son wins tournaments and has a handicap below par. I shall break 80 ere I leave this mortal shell!
This conversation helped test and solidify my understanding of some concepts.
Learning and understanding electronics is a long journey I like to compare to getting to scratch at golf where some pick it up faster than others, but you have to learn a LOT of things along the way. And you have to be willing to go back from time to time and revise your understanding from almost (pardon the pun) ground zero. With golf, I've been trying for 30 years and shoot low 80s, but my son wins tournaments and has a handicap below par. I shall break 80 ere I leave this mortal shell!
@Haze Head That's a lot... but IMO, you're asking wonderful questions. I have no idea where your level of understand lies, but I have a few things I can offer.
I'd advise first (many might advise differently) on taking a few basic electronics courses and/or doing some reading on the basics. That'll help a ton with terminology and understanding.
You get to 30W by using the voltage swing available at the output and the load. See the "power equation". The amplifier must also be able to allow enough current to deliver that power without releasing magic smoke (or just shutting off). The same amplifier will likely have different power delivery into 16, 8, 4, and 2 ohm loads. Many people also include a distortion figure they're comfortable with and/or site the power available just prior to clipping the output signal. They may also provide some useful frequency response data. An amp that is has a nice "flat" response at 1W may have much lower bandwidth and or distort much more at higher power delivery. That may sound complex, but each part is important. Many people want even more information, but ... 30W by itself is relatively meaningless.
1W, 8R, 0.005% THD+N, < - 1dB 20Hz to 20kHz
30W, 8R, 1% THD+N, 1kHz, < -3dB 20Hz to 20kHz
Above may be more meaningful to some. Note, I made that up for illustration, and it has nothing to do with the F6.
Exercise - Try to calculate the voltage (Vpp and Vrms) along with the current that would flow through an "8 ohm" loudspeaker at 30W. Might be an interesting way to start understanding power.
Bias.... Hard term. To me... It's tricky because "Class A bias", "bias current", "biasing a device" "reverse biasing a component" can refer to any number of things. Take two examples within this lecture: the choice around how to "bias" the JFETs in the front-end is a wholly different matter than choosing the quiescent current (or bias current) of the output devices and how to set the bias current.
I think I see where the confusion may have come from. Setting the bias with the battery... AHA!
"Now we have a circuit where the voltage V2 determines the bias current of the amplifier and V1 determines the output DC offset, and they are independently adjustable...You can use anything to make the bias voltages – you can use a battery, in fact a 1.5V battery is just about the right voltage for that."
In the diagram on page 4 - V1 and V2 go to the gates of the MOSFETs. The way I like to think of it, and a way I've seen taught, is that the gate is the knob for the spigot and the current running through the drain and source is the water flow. You have to have enough of a potential difference (and in the correct polarity) to open the spigot => "turn on" the output devices. V1 and V2 supply voltage to the gates of Q1 and Q2 ... when a point is reached where the potential difference between the gate and the source is 'high enough' the tap opens. Then, you'll get flow from drain to source (directly from the rails) through Q1 and Q2. That allows the rail voltage sources (+V1 and -V1) to enter the game "straight through" the output devices and to the output through the source resistors. That's why to measure the current through Q1 and Q2, you measure the voltage across the source resistors. The current through the source resistors = current from drain to source of Q1 and Q2.
You may see the term Vgs_threshold tossed about a bit. That's the voltage to 'turn them on'. If you keep 'turning the knob' (increasing the potential difference between the gate and the source), then more water / current flows. That's until you get to the current you want or the device goes up in smoke. That is how you set the current (bias => quiescent current) through the output devices and why you measure it at the source resistors. You bias both devices (Q1 and Q2) until the current flowing through them is equal => no DC offset.
That 'constant' flow of current (along with the rail voltages and the voltage drop through the output devices among a few other things) ultimately relates to the output power of the amp, the Class A bias of the amp, and some choices around "making those devices work hard" so they're operating in a better region and/or in a way that produces a sonic affect we like. Nelson talks in one lecture about tired puppies, I recall. It helped me understand a bit. I still have a VERY limited understanding.
Onto the 1V5 vs. +- 23V rails.
That 1V5 is not directly a part of determining the output power of the amp. The flow of water helps determine the output power of the amp, but the 1V5 is just a tool to turn on the tap. It's also important to note that this voltage does not affect the signal... it's not "a steady/constant source of DC injected into the AC signal". It just turns on the tap and controls the flow through the tap.
Edited to add - It's also important to note that it's not necessarily 1V5. In the "real" amp, the choices of the resistor values + Zeners (or other mods) along with the setting of P1 and P2 ultimately determine the voltage.
There's obviously no battery in the circuit, so where do we get the 1V5 (not in all cases) ... and maybe why did some people have such a tough time reaching the intended bias of the amplifier, thus different Zener values and/or mods?
See on page 5 that the rails are now 'connected' to the gates through an LED, a few resistors and a trimmer? There's your bias supply vs. a battery.
What if it took more ooomph to turn on the tap (Higher Vgs_threshold)? What if the ratio of current that flowed between the drain and source was different depending on the voltage differential? What if both those factors are typical to different output devices? Then, you may need to adjust the biasing mechanism. Notice that in the schematic in Post #1 we have Z1 and Z2 instead of the LEDs in the lecture. Notice that the kits come with two values of Zeners. 2Pico did some super-cool work around 'going back' to the LEDs (among other things). All that just to show that different output devices are... well.. different, and we need to adjust the biasing mechanisms accordingly. The "bias voltage" => Vgs is not injected into the signal.
Just tossing out a few things that have helped me along in understanding this amazing amplifier and a few others. May not help you at all, but hey... Also a bit looooooooong. I reorganized it a bit, but it's still a ramble. Jim can come back and make it MUCH more concise.
Nelson, ZM, Jim, Dennis, Ben, and many, many others that likely deserve mention, have taught that each of these "little chunks" assembles into one very, very cool Lego set to play with. I still can barely figure out what each individual piece does, but it sure is fun. If I've explained it in a way that helps anyone... then it helps me solidify how I've learned it... and/or people will tell me just how wrong I am... then I learn more.
I'd advise first (many might advise differently) on taking a few basic electronics courses and/or doing some reading on the basics. That'll help a ton with terminology and understanding.
You get to 30W by using the voltage swing available at the output and the load. See the "power equation". The amplifier must also be able to allow enough current to deliver that power without releasing magic smoke (or just shutting off). The same amplifier will likely have different power delivery into 16, 8, 4, and 2 ohm loads. Many people also include a distortion figure they're comfortable with and/or site the power available just prior to clipping the output signal. They may also provide some useful frequency response data. An amp that is has a nice "flat" response at 1W may have much lower bandwidth and or distort much more at higher power delivery. That may sound complex, but each part is important. Many people want even more information, but ... 30W by itself is relatively meaningless.
1W, 8R, 0.005% THD+N, < - 1dB 20Hz to 20kHz
30W, 8R, 1% THD+N, 1kHz, < -3dB 20Hz to 20kHz
Above may be more meaningful to some. Note, I made that up for illustration, and it has nothing to do with the F6.
Exercise - Try to calculate the voltage (Vpp and Vrms) along with the current that would flow through an "8 ohm" loudspeaker at 30W. Might be an interesting way to start understanding power.
Bias.... Hard term. To me... It's tricky because "Class A bias", "bias current", "biasing a device" "reverse biasing a component" can refer to any number of things. Take two examples within this lecture: the choice around how to "bias" the JFETs in the front-end is a wholly different matter than choosing the quiescent current (or bias current) of the output devices and how to set the bias current.
I think I see where the confusion may have come from. Setting the bias with the battery... AHA!
"Now we have a circuit where the voltage V2 determines the bias current of the amplifier and V1 determines the output DC offset, and they are independently adjustable...You can use anything to make the bias voltages – you can use a battery, in fact a 1.5V battery is just about the right voltage for that."
In the diagram on page 4 - V1 and V2 go to the gates of the MOSFETs. The way I like to think of it, and a way I've seen taught, is that the gate is the knob for the spigot and the current running through the drain and source is the water flow. You have to have enough of a potential difference (and in the correct polarity) to open the spigot => "turn on" the output devices. V1 and V2 supply voltage to the gates of Q1 and Q2 ... when a point is reached where the potential difference between the gate and the source is 'high enough' the tap opens. Then, you'll get flow from drain to source (directly from the rails) through Q1 and Q2. That allows the rail voltage sources (+V1 and -V1) to enter the game "straight through" the output devices and to the output through the source resistors. That's why to measure the current through Q1 and Q2, you measure the voltage across the source resistors. The current through the source resistors = current from drain to source of Q1 and Q2.
You may see the term Vgs_threshold tossed about a bit. That's the voltage to 'turn them on'. If you keep 'turning the knob' (increasing the potential difference between the gate and the source), then more water / current flows. That's until you get to the current you want or the device goes up in smoke.
That is how you set the current (bias => quiescent current) through the output devices and why you measure it at the source resistors. You bias both devices (Q1 and Q2) until the current flowing through them is equal => no DC offset.That 'constant' flow of current (along with the rail voltages and the voltage drop through the output devices among a few other things) ultimately relates to the output power of the amp, the Class A bias of the amp, and some choices around "making those devices work hard" so they're operating in a better region and/or in a way that produces a sonic affect we like. Nelson talks in one lecture about tired puppies, I recall. It helped me understand a bit. I still have a VERY limited understanding.
Onto the 1V5 vs. +- 23V rails.
That 1V5 is not directly a part of determining the output power of the amp. The flow of water helps determine the output power of the amp, but the 1V5 is just a tool to turn on the tap. It's also important to note that this voltage does not affect the signal... it's not "a steady/constant source of DC injected into the AC signal". It just turns on the tap and controls the flow through the tap.
Edited to add - It's also important to note that it's not necessarily 1V5. In the "real" amp, the choices of the resistor values + Zeners (or other mods) along with the setting of P1 and P2 ultimately determine the voltage.
There's obviously no battery in the circuit, so where do we get the 1V5 (not in all cases) ... and maybe why did some people have such a tough time reaching the intended bias of the amplifier, thus different Zener values and/or mods?
See on page 5 that the rails are now 'connected' to the gates through an LED, a few resistors and a trimmer? There's your bias supply vs. a battery.
What if it took more ooomph to turn on the tap (Higher Vgs_threshold)? What if the ratio of current that flowed between the drain and source was different depending on the voltage differential? What if both those factors are typical to different output devices? Then, you may need to adjust the biasing mechanism. Notice that in the schematic in Post #1 we have Z1 and Z2 instead of the LEDs in the lecture. Notice that the kits come with two values of Zeners. 2Pico did some super-cool work around 'going back' to the LEDs (among other things). All that just to show that different output devices are... well.. different, and we need to adjust the biasing mechanisms accordingly. The "bias voltage" => Vgs is not injected into the signal.
Just tossing out a few things that have helped me along in understanding this amazing amplifier and a few others. May not help you at all, but hey... Also a bit looooooooong. I reorganized it a bit, but it's still a ramble. Jim can come back and make it MUCH more concise.
Nelson, ZM, Jim, Dennis, Ben, and many, many others that likely deserve mention, have taught that each of these "little chunks" assembles into one very, very cool Lego set to play with. I still can barely figure out what each individual piece does, but it sure is fun. If I've explained it in a way that helps anyone... then it helps me solidify how I've learned it... and/or people will tell me just how wrong I am... then I learn more.
Last edited:
Very useful info @ItsAllInMyHead, I'd say I probably knew about half and learned from the other half, so good rant As far as taking classes I have actually, but I don't really put much weight on that because I never retain anything I don't do regularly.
Since the nature of the bias voltage is now sorted out I come to another interesting problem. If the aforementioned battery sourced bias were to be employed, how would it tie in, and will this negate the need for the capacitors? In NP's early simplified schematic he referenced the bias to the source pins without the caps; he only then adds caps when the bias circuit reference is moved to ground.
Big picture here is what I'm thinking about. I want to attempt an R-core made with Arnon7 electrical steel. An R-core F6 with single primary, as NP originally envisioned. I like the idea of chopping out the bias circuit for an externally sourced voltage. If the caps can go thanks to the external bias voltage even better. Lastly, I'm looking at using a dual channel mosfet SMD. That combination of mods, on paper, appears to take the F6 in a very minimal direction, in terms of parts count and connectivity.
I think this guy looks pretty interesting. I'm practically assuming it would need to be water cooled, CPU style, perhaps on both sides of the chip. I did bounce the key spec off of several other FETs, several in stock SiCs included, and it holds up really well. I'm thinking you may get good match potential on a single die IC. I see it does have the extra diode symbol, dunno if that means anything here:
As far as taking classes I have actually, but I don't really put much weight on that because I never retain anything I don't do regularly.Since the nature of the bias voltage is now sorted out I come to another interesting problem. If the aforementioned battery sourced bias were to be employed, how would it tie in, and will this negate the need for the capacitors? In NP's early simplified schematic he referenced the bias to the source pins without the caps; he only then adds caps when the bias circuit reference is moved to ground.
Big picture here is what I'm thinking about. I want to attempt an R-core made with Arnon7 electrical steel. An R-core F6 with single primary, as NP originally envisioned. I like the idea of chopping out the bias circuit for an externally sourced voltage. If the caps can go thanks to the external bias voltage even better. Lastly, I'm looking at using a dual channel mosfet SMD. That combination of mods, on paper, appears to take the F6 in a very minimal direction, in terms of parts count and connectivity.
I think this guy looks pretty interesting. I'm practically assuming it would need to be water cooled, CPU style, perhaps on both sides of the chip. I did bounce the key spec off of several other FETs, several in stock SiCs included, and it holds up really well. I'm thinking you may get good match potential on a single die IC. I see it does have the extra diode symbol, dunno if that means anything here:
Last edited:
Same here re: use it or lose it. Everyone has their own style.Very useful info @ItsAllInMyHead, I'd say I probably knew about half and learned from the other half, so good rant
Think of this as my poor attempt to learn alongside you. Admittedly, I haven't the foggiest of clues around some of it, but maybe we can refine some of the questions / ideas to the point where it's worth it for folks with real knowledge to chime in knowing that you've put some major effort into it. Me, I just like thinking about the concept. Don't want to hijack your thoughts.It's now up to you to think about the simplified schematic / block diagrams against the schematic with "extra stuff to actually make it work well / stable"
Which ones? Show your thoughts on a diagram. What do you think each of those caps do in the circuit? If you're unsure, think, research, and ask here. I like to think of caps fitting two main purposes. Am I filtering or blocking 'something' and/or am I storing some energy for a useful purpose.
Why might that be?
WOW! I haven't the foggiest re: building transformers. That would be super-neat!
For stability? Simplicity? Separate PCBs to make your life easier? Why? Nifty idea... many amps have "separate" PCBs for bias voltages / mechanisms. Some of them even run on different PSUs altogether and/or different power transformer secondaries. I think ZM's come up with a bunch of daughterboards for his various neato ideas for biasing some "tough" critters that like to run away. Edited to add, if you really are thinking of a cell-type battery, my opinion is that it had better have a VERY linear discharge and you'd better know exactly where that cliff is and when to change it out. Either that or a wall-wart or 'something'.
Simplicity / lowering parts count (if the parts are uneccessary is a great goal, IMO)
Comparing / contrasting the use of that specific chip in an F6-type output vs. the "bog standards" or even some others that have been proposed is waaaaaay outside my skills. I look at it, and it makes my head hurt. I'll watch for others ideas, but if I come up with anything, I'll throw it into the wind. Great to think about.
Note - Was typing before you put up the diagram of your simplified idea. I think I follow. Very cool! Looking forward to see how this all shapes up. DIY at its best.
Last edited:
@ItsAllInMyHead
About the role of the caps, I don't have a good guess just yet. And what ever I come up with will be a guess lol.
Glad you like the little sketch. It would be really cool looking I think, as it's very close to the original simplified circuit --but all of the requisite functions are there, just sort of off the board. Like you mention with ZM's boards. As to the rational, I'm just playing and this particular amp is super cool to play around with owing to the plentiful details provided in the lecture.
I'll see If I can get a decent R core figured out first. But I'm not sure if it will even be feasible to cool a dual channel FET, so my breath isn't being held on that deal. I'm trying three different things in the sketch and if I can make at least one of them work I'll be happy.
About the role of the caps, I don't have a good guess just yet. And what ever I come up with will be a guess lol.
Glad you like the little sketch. It would be really cool looking I think, as it's very close to the original simplified circuit --but all of the requisite functions are there, just sort of off the board. Like you mention with ZM's boards. As to the rational, I'm just playing and this particular amp is super cool to play around with owing to the plentiful details provided in the lecture.
I'll see If I can get a decent R core figured out first. But I'm not sure if it will even be feasible to cool a dual channel FET, so my breath isn't being held on that deal. I'm trying three different things in the sketch and if I can make at least one of them work I'll be happy.