[dBs]

If you'll allow, I will do a brief history. If you don't want to read any of this, feel free to skip right to the part that says PLANS.

HISTORY
I just graduated this last year with a bachelors in EE. Unfortunately, my senior project was less than successful. I will fail, but that doesn't mean that I have to like it. I wont let it dissuade me from pushing through, learning, and progressing. I decided that I wanted to design, from the ground up, a DIY open-source project amp as a way to "give back" to the audio community. I had asked for help in form of a mentor on here and received no reply. Being of thick head, I wasn't going to let that stop me, it simply meant that I would have to teach myself. Unfortunately, tubes are a technology of yesteryear and college looks at them as outdated (which they are for most purposes, lets be honest) so I learned nothing of them in college.

I bought some books, attempted some light reading but found the turmoil of setting up my new "working world" life more time consuming than I anticipated. Now life has slowed down enough to allow me to do the research I wanted to do. I have researched relentlessly the last three weeks. I think I am ready to start.

PLANS
There are already a million different amps out there already for the open-source DIY community, even in the tube genre. Being the kind of person who doesn't like to tackle a challenge the same way as those before me, I would like to take a slightly different approach and no idea is really too out there.

I am looking to keep the parts list no higher than the $300 parts range so it can be afforded by anyone. At this point I would like it to be a single ended push-pull (SEPP) topology. I will be attempting to avoid output transformers (OTL) for their cost and potential influence on the sound. This will likely limit my output impedance (Zo) for lower impedance headphones, but the P-P should help with that a bit. I would like the output tubes to be triodes. I will be driving them as linearly as possible while keeping the bias current as low as possible. I don't want there to be a chance of them going into class AB operation.

The input stage I am still debating on and will think about a bit more this week. I am considering pulling one of the P-P tubes grids from the plate of the input tube and the other from the cathode of the input tube. I would prefer a triode in this stage. I am hoping to avoid the added cost and complexity of a differential pair.

I am leaning toward chokes instead of cathode resistors (Rk) if I can find them cheaply enough. I am debating bypass capacitors but I wont know if I will use them or not until I've started building (I will design for both possibilities).

I am looking to avoid negative feedback (NFB) if I can help it, which I believe I will be able to do. I like how it lowers the harmonic distortion level but I don't like how it mirrors it to higher orders.

I am probably half way through my research of power supplies. I like the idea of using a tube rectifier but I am not sure how that will affect cost and complexity. I suspect I will be forced into solid state. Chokes in the Pi filters would be nice but I think would prove too expensive.

As this project progresses, I will be keeping this up to date with schematics, math, thoughts, questions, etc. until a final project is realized. At this point we will see what happens regarding the possibility of a kit or group buy for parts if anyone is actually still interested at that point. Even if this never garners any other attention, at least I would have vindicated myself from my senior project, hahaha.

YOUR HELP
Where I would like the communities help is while college taught me the theory, they didn't teach me real world implementation. I could spend at least a month intently researching the various tube types and the trade offs associated with them but seeing as this is intended as a "giving back" and I want there to be interest in the project as well as the result, I think asking you all is the perfect solution to this.

What tube/s would you like to see used? Are there any that are horribly under appreciated and deserve their spot in the limelight? Tube rolling is always nice but what tubes offer the best flexibility? What about rectifier tubes? What type of tube offers the best sonic flexibility? Etc.

Also, I haven't researched yet the different capacitor types and brands. The same applies to chokes, resistors, transformers, etc. I was planning on doing this anyway, but if time can be saved by tapping into the cumulative knowledge of this highly educated community I certainly wouldn't turn it down.

So, thoughts?

UPDATES
I forgot to mention the most basic parts! I plan this amp to be for headphones. The Zo is naturally the lower the better, but until the output tube/s have been decided upon and whether or not there will be output capacitors or bypass capacitors, it cannot be determined yet. A basic potentiometer based volume control that will be put to ground in parallel with another resistor to keep the pot itself out of the signal path. Only one headphone output at this point to keep costs down. A basic 2 input (+), 1 output headphone amp.

It is a bit of an "out of thin air" number, but I will be trying for a Zo of around 50 Ohms. I believe that could be a realistic value. Output power I am not as certain on. My aim is quality to that end I need to run my tubes as linearly as possible so less than 1 watt but more than 400 mWatts. I want to be able to drive Hd650's without difficulty and with enough confidence to run 600 Ohm DT880/990s. P-P running in lower power operation should hopefully enable running lower impedance headphones like the K701s. I think I would find it difficult to run Denons or Grados but I will try my best.

 

[pabbi1]

If you have not already, I would read Pete Millett's DIY Audio pages - it is full of real world implementation, as well as discussions on tubes.

To keep glass costs down, I would suggest Russian tubes - 6n1p, 6n6p and 6h23, all plentiful and cheap.

Otherwise, best of luck.

 

[Leny]

Hello dBs, I admire your enthusiasm. However before progressing you need to take a step back and decide exactly what you are going to design. For example…

Is it going to be for dynamic headphones, electrostatic headphones, etc.?
If dynamic, what impedances must it drive, and to what level?
Is it to have a volume control or reply upon a volume control built into the source?
How many inputs?
Etc,. etc., etc..

So, my recommendation is to write a short statement defining exactly what it is to be. Then write an objective list of target performance, features and facilities.

Only then can you start assessing alternative ways to achieve your objectives. Likewise remember that cost is closely associated with technical feasibility, so you may need to vary your budget a bit.

Good luck.

 

[dBs]

Quote:

Originally Posted by pabbi1 /img/forum/go_quote.gif If you have not already, I would read Pete Millett's DIY Audio pages - it is full of real world implementation, as well as discussions on tubes. To keep glass costs down, I would suggest Russian tubes - 6n1p, 6n6p and 6h23, all plentiful and cheap. Otherwise, best of luck.

Thanks, I will look into those tubes. The site I'm sure will also prove useful, thank you.

Quote:

Originally Posted by Leny /img/forum/go_quote.gif Hello dBs, I admire your enthusiasm. However before progressing you need to take a step back and decide exactly what you are going to design. For example… Is it going to be for dynamic headphones, electrostatic headphones, etc.? If dynamic, what impedances must it drive, and to what level? Is it to have a volume control or reply upon a volume control built into the source? How many inputs? Etc,. etc., etc.. So, my recommendation is to write a short statement defining exactly what it is to be. Then write an objective list of target performance, features and facilities. Only then can you start assessing alternative ways to achieve your objectives. Likewise remember that cost is closely associated with technical feasibility, so you may need to vary your budget a bit. Good luck.

I have it up there but I suppose I didn't state it explicitly but this would be for headphones. I haven't thought about preamp capabilities or anything yet but at this point I would probably not involve any as it would up the complexity and cost.

As far as the number and types of inputs, there isn't much to that. I don't plan to incorporate any DAC functionality, so I could leave it at RCA type plugs. As far as Zo, I haven't decided on that yet as that will be determined in part by the type of tubes I go with and whether or not I end up incorporating any bypass capacitors or output capacitors. Naturally the lower the better, but until I have a set of tubes decided upon, there wont be much of a way to determine what Zo I will be aiming towards.

I do plan on incorporating a simple POT (probably use a paralleled variation implementation to remove the pot itself from the signal path. I haven't researched any sort of delay circuit yet but that is up for consideration.

Cost is a factor but I suppose not entirely set in stone. If some leeway is needed than that is fine as ultimately the DIY'er can go with a cheaper capacitor or resistors if they felt the cost prohibitive in default configuration. I definitely understand that my developmental costs will go above the amps cost, that is to be expected. I don't even own an oscilloscope or frequency generator yet XD. This will be a piece by piece endeavor, as it will be personally funded.

Features are basic, no real frills: I will at least create it with two inputs in mind simply so a person could very easily add in parallel any additional inputs they may desire, basic potentiometer volume control, one headphone port. Basic 2 in, 1 out =).


Someone PMed me saying that there is an over-saturation of this style of amp already and that I should think about creating a DIY DHT SET amp that doesn't hum. Any thoughts on this idea? It's early so P-P can be "voted" out.

 

[nikongod]

Quote:

Originally Posted by dBs /img/forum/go_quote.gif Someone PMed me saying that there is an over-saturation of this style of amp already and that I should think about creating a DIY DHT SET amp that doesn't hum. Any thoughts on this idea?

There is not an oversaturation of what you described, indeed it is quite unusual. Unfortunately it is already the basis of a DIY project (Cavalli Bijou). The idea to run a SEPP stage off of a long tail pair is neat and I can tell you it sounds awesome with 6sn7.

On the topic of the individual who PM'd you: Have you researched this individuals recent threads and general opinions? Would you follow this person into a project based on their previous track record. If the individual has more than about 200 posts the find started threads button should be quite helpful. fewer than 150, read everything they have to say before you believe a word of it.
Quote:

It's early so P-P can be "voted" out.

Why should push-pull be voted out? There are plenty of reasons to say that its a great topology to look at if you have a budget. Here are a few of the better ones:

1. The amp has better PSR. A simpler PS can be used, saving money.

1. Cheap PP transformers generally have better bandwidth than cheap SE transformers.

1. With the goal of creating something new, there is not much to say for SE. There are tons of neat things to try with PP that have not been used in ages.

Technically the last one is not really new but nobody has seen it in long enough that you can take credit.

I should add that a parafeed SE amp offers many advantages to a budget minded amp as well. Another great topology to look into.

 

[Leny]

Quote:

Originally Posted by dBs /img/forum/go_quote.gif As far as Zo, I haven't decided on that yet as that will be determined in part by the type of tubes I go with…

Yes and no; there is a better way...

First decide what you are trying to achieve, then look for alternative ways to get there.

So first research the Zo you need, and document it. THEN look for ways to achieve it.

It may be that as the project evolves you find that it is difficult to achieve your target, and you may have to accept a ‘deviation’ from specification. However in order to increase your chances of success you should always start with a specified target, not with a decision on the ingredients.

So what output impedance would you like as a target?
Must this be the same for all impedances of headphones?


Quote:

Originally Posted by nikongod /img/forum/go_quote.gif I should add that a parafeed SE amp offers many advantages to a budget minded amp as well. Another great topology to look into.

Seconded.


Keep it up,
L.

 

[dBs]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif There is not an oversaturation of what you described, indeed it is quite unusual. Unfortunately it is already the basis of a DIY project (Cavalli Bijou). The idea to run a SEPP stage off of a long tail pair is neat and I can tell you it sounds awesome with 6sn7.

I was hoping to avoid a differential pair if I could help it to keep costs down. I suppose it could be done without significant increased cost if the tube was a twin triode, maybe.

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif On the topic of the individual who PM'd you: Have you researched this individuals recent threads and general opinions? Would you follow this person into a project based on their previous track record. If the individual has more than about 200 posts the find started threads button should be quite helpful. fewer than 150, read everything they have to say before you believe a word of it.

I trust this individual. Ultimately, while this project is for myself, it is also for everyone else as well. I want people to be interested in this project as well. I want people to actually want to build this thing so catering to the tastes of people here is important to that end =)

I've done my research pretty well on P-P and basic SET. I know the trade offs and advantages well =) I'm hoping to do something a bit unique though, and I haven't seen a lot of SEPP in the cheaper DIY genre.

I did consider "bringing back the retro" with a circlotron but to implement them correctly you really need to bring in an output transformer and that would cripple the cost.

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif I should add that a parafeed SE amp offers many advantages to a budget minded amp as well. Another great topology to look into.

I have not researched these very extensively yet as I was concerned with the OPT cost. I thought I would incorporate more complex and expensive options in later designs (I don't intend this to be the only thing I design =)). If this is what the customer wants though, I haven't steadfastly written anything off.

Quote:

Originally Posted by Leny /img/forum/go_quote.gif First decide what you are trying to achieve, then look for alternative ways to get there.

I'm just concerned of painting myself into a corner. If someone comes up with a really intriguing idea--to me and any community members who take interest in this idea--but I can't give it a shot because I set up prerequisites, then I'm not doing this project any favors. I kind of like the idea of experimenting along the way =D

You are right though, I should have set something up a bit more definitive. I will include some more specific thoughts into the UPDATE on the OP.

Quote:

Originally Posted by Leny /img/forum/go_quote.gif Must this [Zo] be the same for all impedances of headphones?

I didn't entirely understand what you asked here, could you rephrase it?

 

[Leny]

Quote:

Originally Posted by dBs /img/forum/go_quote.gif I didn't entirely understand what you asked here, could you rephrase it?

Sorry for not being clear.

Transformer coupled amps with switchable taps (for matching to different impedance headphones) change their output impedance from tap to tap according to the reflected impedance of the tube.

For example it might be 10 Ohms on the low impedance setting and 30 ohms on the high impedance setting. Their designers believe this to be fine, and I wondered if this is something that you had considered, and whether you believed it important or not.

However, I have now re-read your posts and see that you would prefer no output transformer, so my question does not apply anyway!!!


But then again, if you do decide to consider an output transformer, take a look at the Hammond 119DA which is about US $25 at Antique Electronic Supply. It's performance seems good for the price. It is used in the Hagtech Castanet and the M.A.D. amps.

 

[dBs]

Quote:

Originally Posted by Leny /img/forum/go_quote.gif But then again, if you do decide to consider an output transformer, take a look at the Hammond 119DA which is about US $25 at Antique Electronic Supply. It's performance seems good for the price. It is used in the Hagtech Castanet and the M.A.D. amps.

That is a good piece of information and far cheaper than I have seen unto this point. That will be worth looking into.

If I was going to do an OPT design, I would have done tapped, absolutely =)

 

[regal]

Quote:

Originally Posted by dBs /img/forum/go_quote.gif It is a bit of an "out of thin air" number, but I will be trying for a Zo of around 50 Ohms. I believe that could be a realistic value. .

How do you plan on Zo=50ohms with a P-P OTL and little NFB? You're looking at a lot of output tubes and a huge power supply. Unless you are talking about a Futterman type design? An OPT is much less costly.

I'm personally not inerested another OTL when they are all over eBay and from China for cheaper (just one example for $213 http://www.head-fi.org/forums/f5/ell...lifier-420767/than one can build and as mentioned for DIY the Bijou is tough competetion. The problem with OTL is the big cap at the output makes them all sound like that big cap to a certain degree. But I applaud you for making a group project, great DIY spirit.

As far as parafeed, looks at dsavitsk's page, he has done a lot of good work with parafeed spud amps.

 

[burgunder]

I think it would be great if you and the community would design an OPT amp, but I will suggest another desing than parafeed as it seems there might be a kit avaiable with the L'espressivo design at least the amp has had it's own section at DIYforums for some time. But this is only speculation, so a parafeed could still be OK.

 

[nikongod]

Quote:

Originally Posted by dBs /img/forum/go_quote.gif I've done my research pretty well on P-P and basic SET. I know the trade offs and advantages well =) I'm hoping to do something a bit unique though, and I haven't seen a lot of SEPP in the cheaper DIY genre.

Despite the fact that people know no limits on how much to spend on parts for it a bijou can be built fairly inexpensively.

Quote:

I did consider "bringing back the retro" with a circlotron but to implement them correctly you really need to bring in an output transformer and that would cripple the cost.

Edcor makes push pull output transformers for as little as $18. The bandwidth is kind of poopy, but the ones for $25 look much nicer.
Quote:

Originally Posted by regal /img/forum/go_quote.gif How do you plan on Zo=50ohms with a P-P OTL and little NFB? You're looking at a lot of output tubes and a huge power supply. Unless you are talking about a Futterman type design? An OPT is much less costly.

Quote:

Originally Posted by runeight /img/forum/go_quote.gif the Zo of the Bijou is about 55R with no NFB.

I dont understand what makes the futerman circuit anything more than a white cathode follower with a bias scheme that circumvents older patents.

 

[dBs]

Ok, I like what I'm seeing, a good exchange of interest now!

I will look into parafeed and see what the good/bad/trade-offs/major considerations are.

Keep it up!

Don't forget to give me ideas on tubes. That's one area I could spend months learning about and only be scratching the surface. Any tubes that you find are unappreciated, over hyped, just plain bad, or exceptional?

 

[nikongod]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif I dont understand what makes the futerman circuit anything more than a white cathode follower with a bias scheme that circumvents older patents.

I just goggled exactly what a futerman circuit means, thinking forever that it meant the output stage of the Bijou. A futerman circuit is an OTL amplifier with a positive feedback loop nested in a global negative feeback loop.

 

[TimJo]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif I just goggled exactly what a futerman circuit means, thinking forever that it meant the output stage of the Bijou. A futerman circuit is an OTL amplifier with a positive feedback loop nested in a global negative feeback loop.

So the positive feedback is coming from the cathode current of the phase splitter in that design, correct?