g0ldl10n
1000+ Head-Fier
I want to start off this thread with a bit of background in regards to my personal DIY journey. Since about 2012 I have been what people call an "audiophile", but didn't get into tube amps until middle of 2022. Well, needless to say I got hooked on the notorious "tube sound" and now am a proud hoarder of tubes (it is a bit ridiculous tbh, 
). Since I was late to the game so many of the famed tubes were already really expensive so I decided to take a different route, search for unknowns or tubes with very little information on them besides their datasheets - this ofc required adapters, a LOT of adapters which got expensive - so in order to curb adapter cost I began making them myself which to date, probably have over 100 different adapters I have made for myself.
Anyways, since I started making adapters I always had the desire to learn to design/build my own circuit/amp and I finally took the dive into the studies Feb/March of this year (2024) - for reference, my background in electrical engineering is null, I have none at all so that was my starting point, literally step 1.
I grabbed a couple books from Merlin Blencowe (thanks for the book recommendations @L0rdGwyn ), one which focused on the amplifying circuit and the other focused on the power supply. These books built my foundation of knowledge, but I also picked up a lot of other things from diyaudio.
That said, I probably learned the most from hands on experience with my prototype circuit, because when you have no formal background in electrical engineering so much of what you read in books or online, no matter how well it is explained can still be hard to grasp and it isn't fully realized/understood until you put it into practice in the real world.
Lastly, this circuit which is pretty much finalized is not any sort of copy/clone I found online - ofc, it has many similarities of any tube amp, but it's quite a bit different especially in regards to the output topology, and the output transformers being used, which are 10VA Antek toroidal PTs repurposed as OPTs, which can be done due to the parafeed output which keeps DC off of the OPTs allowing us to use non air-gapped transformers. The sound quality benefit between using air-gapped and non-air-gapped transformers in parafeed is huge from what my ears can hear - ofc, you can find a select few OPTs designed specifically for parafeed output (or probably have them custom made), but they are quite expensive, much much much more than Antek 10VA toroidal transformers cost - I am sure they sound good, but I will hold off on those as I am inclined to believe any SQ benefits wouldn't be worth the upcharge, at least not to me - but if anyone is willing to buy me a pair I will be happy to test them out for you!
So, from knowing literally nothing of electrical engineering in Feb/Mar 2024 to fully building, testing and completing a stable and good sounding circuit within the span of several months, I would say I am making pretty good time. The circuit itself has been working without issues for the past 3 months on top of my desk and piece of plywood.
It is a complete rats nest for the one main reason, it is the circuit which I tested an ungodly amount of different things in, again, starting from knowing nothing - so, all of my real world testing of circuit building has been there.
Moving on to the chassis, this I saw as another daunting task - I either needed to teach myself some sort of CAD program so I could design a chassis and have it precut by a 3rd party, or get myself a drill press + needed bits and buy blank chassis and do all of the cutting/design "in-house". Well, I decided to go the drill press route - found some nice aluminum chassis' online and got to work.
Okay, enough with the preamble, here's the build
---------------------------------------------------------------------------
THE AMPLIFYING CIRCUIT:
-SET parafeed HP amp
-Cascoded CCS (source) load for input stage set at 5mA
-Cascoded CCS (gyrator - constant voltage at anode, rather than current) load - voltage reference for the gyrator is 50% of B+ via simple resistive voltage divider (IME, dropping anywhere between 40-50% of B+ thru the gyrator gives good results)
-Input valve(s): x2 6J5 or x1 6SN7 - or any tubes which can operate efficiently at 5mA of plate current (6829, 6KN8, 6414, EBC33, to name a few)
-Output valves: Sockets wired for 6L6 and EL34 - gyrator loaded output allows much more freedom to tube rolling - basically can roll anything that will bias below its max dissipation of power per datasheets
-Currently I have been stuck on the amazing sound from the Svetlana 6P13S, great sounding valve imo
-Output is simple 1/4" TRS connected to a 3 way rotary dial to select between low, medium, and high impedance HPs
B+~440V ~495V
---------------------------------------------------------------------------
THE POWER SUPPLY:
-Antek AS-2T400 - 200VA 400V
-400V dual secondaries wired in parallel which allows up to 400mA of current
-This PT has dual 6.3V secondaries for our filaments, both supporting up to 4A - one of these windings are used for input/output stages, and the other winding is used for the rectifier, which is of the 6.3V half-indirectly heated variety, EYY13.
CHANGE: I have decided to use a completely separate filament transformer for the EYY13 rectifier and paralleling the two separate 6V windings on the main PT instead, although, both do work. The reason for this change is so that I can take the VDC+ output from the CT of the separate filament transformer instead of from the rectifier tube itself. In my own testing, this is far superior in regards to keeping the amp quiet.
-EYY13 is a full wave half-indirectly heated rectifier
Rectification via hybrid full bridge using EYY13
EYY13 requires 2.5A for its heaters
400AC supply, it can rectify up to 350mA of current
---------------------------------------------------------------------------
MAIN COMPONENT HIGHLIGHTS
-EIZZ 'EZ-1308A' bakelite sockets / gold plated contacts, which are very hefty as well - picked these because the socket pins/contacts on the underside are made for P2P wiring/soldering, instead of their other nice aftermarket sockets which have the pins - this will make wiring them in P2P fashion so much easier and more secure
-EIZZ Stepped Attenuator, 25k, 24 steps (although the EIZZ attenuators aren't too expensive, I have tested some very inexpensive SMD stepped attenuators which sound very good - definitely eye opening - but, the build quality of the EIZZ is much better)
-Jupiter 0.47uF coupling caps (I have tested mass prod / new prod caps in this position, and many of them are very good, and IMHO, to the point where paying the upcharge on audiophile caps makes very little sense to me now)
-'NOS' Sprague 10uF Capacitor 400V DC + 1uF MBGP PIO Soviet cap (paralleled) serving as parafeed coupling cap
-10VA Antek toroidal PT repurposed as OPTs (specifics below)
High Impedance HPs (200-400 ohm) - AN-0115 - 10VA 15V
Medium Impedance HPs (80-150 ohm) - AN-0109 - 10VA 9V
Low Impedance HPs (20-80 ohm) - AN-0106 - 10VA 6V
---------------------------------------------------------------------------
TESTING EQUIPMENT
Quant Asylum QA403 with DIY dummy loads
Standard DMM
---------------------------------------------------------------------------
CURRENT PROGRESS
First, here are a couple measurements taken from the QA403 (granted, this circuit is right next to a monitor and laptop, so the noise measurements I would think could get better once built within the chassis, and or, measured away from these other devices).
That said, I am not 100% sure if the noise measurements are that good either, maybe someone else can tell me.
Frequency Response of both channels
Left channel noise at ~1mW into 300ohm load (right channel is very close to the same)
I will spare pictures of the current rats nest, because it look's terrible
- if anyone really wants to see it, let me know and I will share a picture, but ofc, it will look nothing like that once built within the chassis.
There may be one small hiccup to this current chassis design which another friendly DIY'er here pointed out to me @carlman14 , but won't know for sure until its all built inside the chassis, which is the EYY13 rectifier may be placed too close to the 6SN7 input socket and induce some hum, but again, I won't know if that is an issue until the circuit is in - I did buy a few of these chassis', so if needed I can start over.
Here is a progression of the chassis being built out
FYI: I do not consider myself an expert in any way shape or form - I haven't even been doing this for a year yet, so if anyone more knowledgeable sees something that is very wrong, please tell me
More to come, including a schematic of the amplifying circuit.
). Since I was late to the game so many of the famed tubes were already really expensive so I decided to take a different route, search for unknowns or tubes with very little information on them besides their datasheets - this ofc required adapters, a LOT of adapters which got expensive - so in order to curb adapter cost I began making them myself which to date, probably have over 100 different adapters I have made for myself.Anyways, since I started making adapters I always had the desire to learn to design/build my own circuit/amp and I finally took the dive into the studies Feb/March of this year (2024) - for reference, my background in electrical engineering is null, I have none at all so that was my starting point, literally step 1.
I grabbed a couple books from Merlin Blencowe (thanks for the book recommendations @L0rdGwyn ), one which focused on the amplifying circuit and the other focused on the power supply. These books built my foundation of knowledge, but I also picked up a lot of other things from diyaudio.
That said, I probably learned the most from hands on experience with my prototype circuit, because when you have no formal background in electrical engineering so much of what you read in books or online, no matter how well it is explained can still be hard to grasp and it isn't fully realized/understood until you put it into practice in the real world.
Lastly, this circuit which is pretty much finalized is not any sort of copy/clone I found online - ofc, it has many similarities of any tube amp, but it's quite a bit different especially in regards to the output topology, and the output transformers being used, which are 10VA Antek toroidal PTs repurposed as OPTs, which can be done due to the parafeed output which keeps DC off of the OPTs allowing us to use non air-gapped transformers. The sound quality benefit between using air-gapped and non-air-gapped transformers in parafeed is huge from what my ears can hear - ofc, you can find a select few OPTs designed specifically for parafeed output (or probably have them custom made), but they are quite expensive, much much much more than Antek 10VA toroidal transformers cost - I am sure they sound good, but I will hold off on those as I am inclined to believe any SQ benefits wouldn't be worth the upcharge, at least not to me - but if anyone is willing to buy me a pair I will be happy to test them out for you!
So, from knowing literally nothing of electrical engineering in Feb/Mar 2024 to fully building, testing and completing a stable and good sounding circuit within the span of several months, I would say I am making pretty good time. The circuit itself has been working without issues for the past 3 months on top of my desk and piece of plywood.
It is a complete rats nest for the one main reason, it is the circuit which I tested an ungodly amount of different things in, again, starting from knowing nothing - so, all of my real world testing of circuit building has been there.
Moving on to the chassis, this I saw as another daunting task - I either needed to teach myself some sort of CAD program so I could design a chassis and have it precut by a 3rd party, or get myself a drill press + needed bits and buy blank chassis and do all of the cutting/design "in-house". Well, I decided to go the drill press route - found some nice aluminum chassis' online and got to work.
Okay, enough with the preamble, here's the build
---------------------------------------------------------------------------
THE AMPLIFYING CIRCUIT:
-SET parafeed HP amp
-Cascoded CCS (source) load for input stage set at 5mA
-Cascoded CCS (gyrator - constant voltage at anode, rather than current) load - voltage reference for the gyrator is 50% of B+ via simple resistive voltage divider (IME, dropping anywhere between 40-50% of B+ thru the gyrator gives good results)
-Input valve(s): x2 6J5 or x1 6SN7 - or any tubes which can operate efficiently at 5mA of plate current (6829, 6KN8, 6414, EBC33, to name a few)
-Output valves: Sockets wired for 6L6 and EL34 - gyrator loaded output allows much more freedom to tube rolling - basically can roll anything that will bias below its max dissipation of power per datasheets
-Currently I have been stuck on the amazing sound from the Svetlana 6P13S, great sounding valve imo
-Output is simple 1/4" TRS connected to a 3 way rotary dial to select between low, medium, and high impedance HPs
B+
---------------------------------------------------------------------------
THE POWER SUPPLY:
-Antek AS-2T400 - 200VA 400V
-400V dual secondaries wired in parallel which allows up to 400mA of current
-
CHANGE: I have decided to use a completely separate filament transformer for the EYY13 rectifier and paralleling the two separate 6V windings on the main PT instead, although, both do work. The reason for this change is so that I can take the VDC+ output from the CT of the separate filament transformer instead of from the rectifier tube itself. In my own testing, this is far superior in regards to keeping the amp quiet.
-EYY13 is a full wave half-indirectly heated rectifier
Rectification via hybrid full bridge using EYY13
EYY13 requires 2.5A for its heaters
400AC supply, it can rectify up to 350mA of current
---------------------------------------------------------------------------
MAIN COMPONENT HIGHLIGHTS
-EIZZ 'EZ-1308A' bakelite sockets / gold plated contacts, which are very hefty as well - picked these because the socket pins/contacts on the underside are made for P2P wiring/soldering, instead of their other nice aftermarket sockets which have the pins - this will make wiring them in P2P fashion so much easier and more secure
-EIZZ Stepped Attenuator, 25k, 24 steps (although the EIZZ attenuators aren't too expensive, I have tested some very inexpensive SMD stepped attenuators which sound very good - definitely eye opening - but, the build quality of the EIZZ is much better)
-Jupiter 0.47uF coupling caps (I have tested mass prod / new prod caps in this position, and many of them are very good, and IMHO, to the point where paying the upcharge on audiophile caps makes very little sense to me now)
-'NOS' Sprague 10uF Capacitor 400V DC + 1uF MBGP PIO Soviet cap (paralleled) serving as parafeed coupling cap
-10VA Antek toroidal PT repurposed as OPTs (specifics below)
High Impedance HPs (200-400 ohm) - AN-0115 - 10VA 15V
Medium Impedance HPs (80-150 ohm) - AN-0109 - 10VA 9V
Low Impedance HPs (20-80 ohm) - AN-0106 - 10VA 6V
---------------------------------------------------------------------------
TESTING EQUIPMENT
Quant Asylum QA403 with DIY dummy loads
Standard DMM
---------------------------------------------------------------------------
CURRENT PROGRESS
First, here are a couple measurements taken from the QA403 (granted, this circuit is right next to a monitor and laptop, so the noise measurements I would think could get better once built within the chassis, and or, measured away from these other devices).
That said, I am not 100% sure if the noise measurements are that good either, maybe someone else can tell me.
Frequency Response of both channels
Left channel noise at ~1mW into 300ohm load (right channel is very close to the same)
I will spare pictures of the current rats nest, because it look's terrible
- if anyone really wants to see it, let me know and I will share a picture, but ofc, it will look nothing like that once built within the chassis.There may be one small hiccup to this current chassis design which another friendly DIY'er here pointed out to me @carlman14 , but won't know for sure until its all built inside the chassis, which is the EYY13 rectifier may be placed too close to the 6SN7 input socket and induce some hum, but again, I won't know if that is an issue until the circuit is in - I did buy a few of these chassis', so if needed I can start over.
Here is a progression of the chassis being built out
FYI: I do not consider myself an expert in any way shape or form - I haven't even been doing this for a year yet, so if anyone more knowledgeable sees something that is very wrong, please tell me
More to come, including a schematic of the amplifying circuit.
Last edited:
