[triod750]

No I probably won't! I'm very happy with 5687, 7044, and E182CC.

I like them better; 5687 and E182CC. Haven't tried 7044.

 

[L0rdGwyn]

I like them better; 5687 and E182CC. Haven't tried 7044.

I'm glad you like them. Can't say if the comparison will be the same seeing as they are in different circuits.

 

[triod750]

I'm glad you like them. Can't say if the comparison will be the same seeing as they are in different circuits.

Same, same but different...
As always.

 

[g0ldl10n]

Glad I could help @Nayyte .

I went from cap in the signal path battery bias back to battery in the signal path battery bias. No doubt battery in the signal path is better, dramatically. I thought I was getting over my capacitorphobia but I think it's here to stay.

So, I finally started my own DIY build which the circuit is currently sitting on some plywood - probably what you would consider a basic SET OTC amp currently with a pair of 6J5s (cathode bias) driving a pair of 6CK4s (fixed bias via a couple 9V batteries for each) which I have compared to cathode bias and its night and day - fixed battery bias is much better sounding than cathode bias.

Now I must ask, what type of battery do you use and is that how this amp is still being biased? Do you have any means of adjusting its voltage reference? Right now, I have a couple 'amazonbasic' 9V batteries in this circuit for each power tube which I find pretty comical when I think about how simple it is but how amazingly well it works/sounds compared to cathode biasing.

I am now about 100% sure any DIY amp I make will always use a form of fixed bias on at least the power tubes - this circuit is the first time I have had the pleasure of experiencing fixed bias and you could say I am a bit head over heels for the SQ compared to auto bias.

Edit: Oh, and forgot to ask, do you use any form of negative feedback at all? I have done a lot of experimenting with it and have found that a very small amount of negative feedback from the OPT output tap back to the plate of the 6J5s seems to work very well .. just a very small amount though.

 

[L0rdGwyn]

So, I finally started my own DIY build which the circuit is currently sitting on some plywood - probably what you would consider a basic SET OTC amp currently with a pair of 6J5s (cathode bias) driving a pair of 6CK4s (fixed bias via a couple 9V batteries for each) which I have compared to cathode bias and its night and day - fixed battery bias is much better sounding than cathode bias.

Now I must ask, what type of battery do you use and is that how this amp is still being biased? Do you have any means of adjusting its voltage reference? Right now, I have a couple 'amazonbasic' 9V batteries in this circuit for each power tube which I find pretty comical when I think about how simple it is but how amazingly well it works/sounds compared to cathode biasing.

I am now about 100% sure any DIY amp I make will always use a form of fixed bias on at least the power tubes - this circuit is the first time I have had the pleasure of experiencing fixed bias and you could say I am a bit head over heels for the SQ compared to auto bias.

Edit: Oh, and forgot to ask, do you use any form of negative feedback at all? I have done a lot of experimenting with it and have found that a very small amount of negative feedback from the OPT output tap back to the plate of the 6J5s seems to work very well .. just a very small amount though.

Yes, fixed bias is generally going to be better than cathode bias as it eliminates a cathode bypass cap in the output stage. But essentially no commercial headphone amplifiers use fixed bias because it is more complex, expensive, and prone to failures.

My HY69 amp is currently cathode biased, but I am making plans to rebuild it with a fixed bias circuit, Simpson milliammeters, and a separate chassis for the power supply. So it will be a three chassis design :O

The input stage of the HY69 amp is battery biased, not the output stage. Generally you want some sort of adjustable bias supply for the output stage of a fixed bias design so you can adjust the bias current of the output tubes as they age, and also ensure the bias current is appropriate for the output transformers being used. That means a negative voltage supply and a potentiometer to adjust the grid voltage, in addition to some sort of method to measure the bias current. That can be a milliammeter, or you can just use a sense resistor in the cathode and measure the voltage across it to check the current. With a battery, you are going to be at the mercy of tube-to-tube variations.

In a fixed bias amplifier, you also have to be very careful that the bias supply does not fail, and if it does, the amp should be properly fused to protect against it. If one of your batteries were to die or short and the grid voltage of the power tube went to ground (0V), the tube will conduct its maximum possible current, not so great for your output transformer primary windings. In my fixed bias amp, I have fuses for the primary of the mains transformer (slow blow), the secondary of the B+ winding (fast blow), and a fast blow fuse in series with the primary winding of each output transformer. This way, if the bias supply fails for any reason and the tubes run over their intended bias current, the fuse blows and the output transformers aren't damaged. It's also important to include some protection resistors for a bias pot so if the wiper goes open, the grid will be biased at the full negative voltage of the bias supply (meaning low operating current).

Now you might see why commercial amplifier makers might not want to sell these amps tubes and transformers can be damaged if you are not careful.

I sometimes use feedback, when necessary or if will benefit the circuit. My 801A amplifier uses some negative feedback, around 11dB, from the output tube plate to the input tube cathode, with the feedback loop buffered by a MOSFET. This is due to the 801A being a relatively high plate resistance output tube - using NFB lower the output impedance and allows for the use of better output transformers (5K primary instead of say, 11K or 14K). So no, I'm not averse to using feedback if it helps the design.

 

[g0ldl10n]

Yes, fixed bias is generally going to be better than cathode bias as it eliminates a cathode bypass cap in the output stage. But essentially no commercial headphone amplifiers use fixed bias because it is more complex, expensive, and prone to failures.

My HY69 amp is currently cathode biased, but I am making plans to rebuild it with a fixed bias circuit, Simpson milliammeters, and a separate chassis for the power supply. So it will be a three chassis design :O

The input stage of the HY69 amp is battery biased, not the output stage. Generally you want some sort of adjustable bias supply for the output stage of a fixed bias design so you can adjust the bias current of the output tubes as they age, and also ensure the bias current is appropriate for the output transformers being used. That means a negative voltage supply and a potentiometer to adjust the grid voltage, in addition to some sort of method to measure the bias current. That can be a milliammeter, or you can just use a sense resistor in the cathode and measure the voltage across it to check the current. With a battery, you are going to be at the mercy of tube-to-tube variations.

In a fixed bias amplifier, you also have to be very careful that the bias supply does not fail, and if it does, the amp should be properly fused to protect against it. If one of your batteries were to die or short and the grid voltage of the power tube went to ground (0V), the tube will conduct its maximum possible current, not so great for your output transformer primary windings. In my fixed bias amp, I have fuses for the primary of the mains transformer (slow blow), the secondary of the B+ winding (fast blow), and a fast blow fuse in series with the primary winding of each output transformer. This way, if the bias supply fails for any reason and the tubes run over their intended bias current, the fuse blows and the output transformers aren't damaged. It's also important to include some protection resistors for a bias pot so if the wiper goes open, the grid will be biased at the full negative voltage of the bias supply (meaning low operating current).

Now you might see why commercial amplifier makers might not want to sell these amps tubes and transformers can be damaged if you are not careful.

I sometimes use feedback, when necessary or if will benefit the circuit. My 801A amplifier uses some negative feedback, around 11dB, from the output tube plate to the input tube cathode, with the feedback loop buffered by a MOSFET. This is due to the 801A being a relatively high plate resistance output tube - using NFB lower the output impedance and allows for the use of better output transformers (5K primary instead of say, 11K or 14K). So no, I'm not averse to using feedback if it helps the design.

Thanks for all the pointers with fixed bias, I will definitely employ them. The current battery only bias came about due to wanting to try and hear what fixed bias had to offer, and a drawer of some 9V batteries and figured it was easy enough to use them to give it a try. Very happy I did.. I am a bit floored how good it sounds compared to cathode bias, its great!

I definitely plan on making the bias adjustable in its final form due to the exact reason you stated, and also to be able to bias other tubes all together, but will now be adding a couple more safety features that I may not have!

Something else I wanted to ask you considering your experience.. right now, I unfortunately have no real way of taking any meaningful measurements of the circuit outside of a standard handheld DMM. I saw in a previous post you mentioned the Analog Discovery, is that something you would recommend for a general DIYer, or something(s) else for that purpose?

 

[L0rdGwyn]

Thanks for all the pointers with fixed bias, I will definitely employ them. The current battery only bias came about due to wanting to try and hear what fixed bias had to offer, and a drawer of some 9V batteries and figured it was easy enough to use them to give it a try. Very happy I did.. I am a bit floored how good it sounds compared to cathode bias, its great!

I definitely plan on making the bias adjustable in its final form due to the exact reason you stated, and also to be able to bias other tubes all together, but will now be adding a couple more safety features that I may not have!

Something else I wanted to ask you considering your experience.. right now, I unfortunately have no real way of taking any meaningful measurements of the circuit outside of a standard handheld DMM. I saw in a previous post you mentioned the Analog Discovery, is that something you would recommend for a general DIYer, or something(s) else for that purpose?

Most people haven't heard a fixed bias amp, it is the best way to design a SET output stage. Aegis is cathode bias, but given it is a cathode follower power stage, gets many of the same benefits of fixed bias, i.e., no bypass capacitor necessary. That's why it sounds good. And without the headaches of a fixed bias circuit.

Yeah the Analog Discovery is a good option, can take basic measurements with it. I haven't messed around with the distortion measurements much, not sure of their accuracy. Most signal generators are not designed to generate low distortion signals, like a DAC is. But it can do frequency response, square wave, power output, etc.

 

[g0ldl10n]

Most people haven't heard a fixed bias amp, it is the best way to design a SET output stage. Aegis is cathode bias, but given it is a cathode follower power stage, gets many of the same benefits of fixed bias, i.e., no bypass capacitor necessary. That's why it sounds good. And without the headaches of a fixed bias circuit.

Yeah the Analog Discovery is a good option, can take basic measurements with it. I haven't messed around with the distortion measurements much, not sure of their accuracy. Most signal generators are not designed to generate low distortion signals, like a DAC is. But it can do frequency response, square wave, power output, etc.

Going to have to do some digging on cathode follower power stage.. I know one of the books you recommended to me awhile ago which I did end up buying (now I need to order the one focused on power supply design, which is likely even more important to learn more about) has a good section on this, although, so much information in that book that has been key to me getting to this point its hard to take it all in, I will definitely circle back to that section.

All of my tube hunting/curiosity has now turned into amp building.. this will be, and already is very addicting

With the mediocre parts I have used in this build, I am stunned by the SQ I am getting - I can now only imagine how good it can actually get, hence the addictive part.

I am sure I will be in and out of this thread with more questions as they arise, and once I get the prototype circuit looking a bit more organized, as it currently looks like a birds nest from so much testing/experimenting, ill share some photos.

The chassis bit is the next thing I need to figure out, and also need to start putting money aside for the Analog Discovery.. so much to do and so little time!

 

[L0rdGwyn]

Going to have to do some digging on cathode follower power stage.. I know one of the books you recommended to me awhile ago which I did end up buying (now I need to order the one focused on power supply design, which is likely even more important to learn more about) has a good section on this, although, so much information in that book that has been key to me getting to this point its hard to take it all in, I will definitely circle back to that section.

All of my tube hunting/curiosity has now turned into amp building.. this will be, and already is very addicting

With the mediocre parts I have used in this build, I am stunned by the SQ I am getting - I can now only imagine how good it can actually get, hence the addictive part.

I am sure I will be in and out of this thread with more questions as they arise, and once I get the prototype circuit looking a bit more organized, as it currently looks like a birds nest from so much testing/experimenting, ill share some photos.

The chassis bit is the next thing I need to figure out, and also need to start putting money aside for the Analog Discovery.. so much to do and so little time!

Blencowe's book does have a section on cathode followers, but not cathode follower power stages, for that you'd have to look on diyAudio or TubeCAD. The power supply book is very good and goes over bias supplies pretty extensively.

Amp building addiction is much more serious than just tube addiction, the possibilities are literally endless, and even more expensive! The lust becomes more about transformers than for tubes, at least for me.

Jones' Building Valve Amplifiers is a good read for laying out an amplifier chassis. And Blencowe's article on grounding schemes. Both have big noise floor implications. Also, you must have a safety earth connection in your amplifier. Jones' book will discuss that.

 

[g0ldl10n]

Blencowe's book does have a section on cathode followers, but not cathode follower power stages, for that you'd have to look on diyAudio or TubeCAD. The power supply book is very good and goes over bias supplies pretty extensively.

Amp building addiction is much more serious than just tube addiction, the possibilities are literally endless, and even more expensive! The lust becomes more about transformers than for tubes, at least for me.

Jones' Building Valve Amplifiers is a good read for laying out an amplifier chassis. And Blencowe's article on grounding schemes. Both have big noise floor implications. Also, you must have a safety earth connection in your amplifier. Jones' book will discuss that.

Ah man, I hope its not more expensive, that would be quite a lot of money lol.. will have to start selling some tubes!

 

[smodtactical]

Took a little longer than I thought, but it always does. The 2A3 amp is done.

Here is the circuit.



And the final build with RCA 5687 and RCA 2A3.







The amp has excellent measurements. Distortion is very low.

1mW into 32ohm 0.036% THD
1mW into 120ohm 0.0066% THD
1mW into 300ohm 0.020% THD
1W into 32ohm 0.36% THD
1W into 120ohm 0.35%
1W into 300ohm 0.21%

Bandwidth is excellent as well.







Listening for the first time, this may be the most transparent amp I have made or heard. Amazingly clear and dynamic with the AC, huge soundstage. Probably the most impressive thing I've noticed so far: given its detail, clarity, and precise imaging, you can listen at very low volumes and not feel you are missing any of the detail. Oh and it's dead silent on high impedance setting

Any chance of posting a guide on how to build this just like you did with Aegis? I know its expensive but would appreciate if you could consider it one day. Thanks brother.

Also I am happy to pay a fee for the plans. Thanks.

 

[L0rdGwyn]

Any chance of posting a guide on how to build this just like you did with Aegis? I know its expensive but would appreciate if you could consider it one day. Thanks brother.

Also I am happy to pay a fee for the plans. Thanks.

Sorry, but I don't plan to release that amp as a project. It's not likely I'll do another Aegis-like release, it was a huge amount of work which is still producing more work for me. That circuit is also more complicated to build, more risky, expensive, more difficult to source parts, and there is some proprietary circuitry of a friend's that I cannot willingly give away schematics for.

 

[g0ldl10n]

Amp building addiction is much more serious than just tube addiction, the possibilities are literally endless, and even more expensive! The lust becomes more about transformers than for tubes, at least for me.

In regards to OPTs, I am currently using a pair from one of the highest quality markets around, Aliexpress .. ofc, this was done for the sole purpose of budget in mind considering it is my first design and build, which was/is to get my understanding of the fundamental SET circuit solidified to some degree, which I would then build my hands on experience from - basically a starting point so I didn't see a very good reason to shell out a lot of cash on expensive parts, who knows, I could've failed, and or caused the magic smoke to appear from them, right?

Well, something very much to my surprise happened with this current "Plywood Special", it sounds very good - better than any amp I currently own quite easily (not that I own anything that is considered TOTL).. this however, I was not expecting, wasn't even something I was really attempting to achieve considering the parts in use, which brings me to my questions.

All things being equal the OPTs obviously have a very large impact on final sound quality and frequency response, and I fear you're right about transformer lust possibility, but I am not made of money (unfortunately!!) so I need to choose wisely here - but since this current circuit sounds this good with these OPTs ($90 for the pair shipped!) I can only imagine what it might/could sound like with some others - from your experience, what would you consider very good "budget" OPTs for SE operation? Say, no more than $200/ea ($400 pair), but preferably $300 or less for a pair?

If money wasn't an object, which SE OPTs would you recommend?

Also, I have got to ask, have you ever used any OPTs sourced from Aliexpress? If so, what was you're experience with them?

 

[AnyDave]

It's not likely I'll do another Aegis-like release, it was a huge amount of work which is still producing more work for me.

I've said it before, but I've released DIY projects in the past and nothing I did even remotely approached the amount of work that must've gone into the documentation you put out for the Aegis. I'm a bit surprised you did it even once. I think everybody who built the Aegis owes you thanks. Not only for the work, but for the continued support.

 

[L0rdGwyn]

In regards to OPTs, I am currently using a pair from one of the highest quality markets around, Aliexpress .. ofc, this was done for the sole purpose of budget in mind considering it is my first design and build, which was/is to get my understanding of the fundamental SET circuit solidified to some degree, which I would then build my hands on experience from - basically a starting point so I didn't see a very good reason to shell out a lot of cash on expensive parts, who knows, I could've failed, and or caused the magic smoke to appear from them, right?

Well, something very much to my surprise happened with this current "Plywood Special", it sounds very good - better than any amp I currently own quite easily (not that I own anything that is considered TOTL).. this however, I was not expecting, wasn't even something I was really attempting to achieve considering the parts in use, which brings me to my questions.

All things being equal the OPTs obviously have a very large impact on final sound quality and frequency response, and I fear you're right about transformer lust possibility, but I am not made of money (unfortunately!!) so I need to choose wisely here - but since this current circuit sounds this good with these OPTs ($90 for the pair shipped!) I can only imagine what it might/could sound like with some others - from your experience, what would you consider very good "budget" OPTs for SE operation? Say, no more than $200/ea ($400 pair), but preferably $300 or less for a pair?

If money wasn't an object, which SE OPTs would you recommend?

Also, I have got to ask, have you ever used any OPTs sourced from Aliexpress? If so, what was you're experience with them?

I don't have a ton of experience with budget transformers to be honest. TubeLab did an article some years back about various budget options.

https://tubelab.com/articles/component-testing/budget-output-transformers/

Probably Edcor would be a good choice for a budget headphone transformer. Unfortunately a lot of transformer winders don't make headphones specific output transformers. You can use speaker transformers but they are a little less than ideal. Lundahl is a good mid-range option.

If you want to go big, custom transformers for Monolith Magnetics are near the top with their Finemet amorphous core line. The Audio Feast also makes a Finemet core headphone transformer for a pretty reasonable price.

I like the 49% nickel core transformers from Intact Audio, but they are pretty pricey at over $1,000 a pair and Dave is not always available to wind them.