The Ultimate Tube DIY Amp
Sep 11, 2006 at 1:47 AM Thread Starter Post #1 of 31

Nerull

100+ Head-Fier
Joined
Jun 14, 2005
Posts
344
Likes
11
Greetings, hehe.

I'm looking to make an amp that surpasses my wildest expectations, but this time, I want to look into tubes. (Although the Beta22 will replace my M^3 probably next year.)

The simplicity in design is remarkable, but I'm not sure which route to go. There is the DynaTube, but that doesn't look like it's complete and whatnot. So I guess I'd have to ask someone if it's okay to use that design (boards are done right, no problems with the schematic, etc...). I guess I'm just a little weary on it.

Another design under consideration is the Aikido headphone amp, but there isn't too much information on how they sound, so again, I'm skeptical on whether it'd be worth building (considering I want the ultimate amp).

Are there any other designs that can be considered ultimate? I really have one of the best sources I can possibly get (without extreme measures like DSPs and swapping out the output stage of the CD player, which I haven't finished yet...), but anyway, I want to make a monster. We're talking audio nirvana. If there isn't a design, I guess I'll have to read up on how to build one, but still, I'm sure there's something out there that is incredible. Price really isn't a concern, despite me being a college student, so fire away. If there's a schematic that's the ultimate amp, I'd also be willing to make a PCB design for it.

~Tom
 
Sep 11, 2006 at 1:56 AM Post #2 of 31
for the aikido: have you looked at diyaudio or AA (tubes diy)? I found the design there and people seems to have only praise for the design. from my readings, John Broskie is a respected tube designer and this respect wasn't achieve through BS claims.

of course in head-fi, the aikido isn't famous. probably because it wasn't designed by the usual head-fi/headwize suspects (kgilmore, amb, cavalli and etc)

btw, you seem to like to use "ultimate" quite a lot. it's a subjective word when it comes to audio.
smily_headphones1.gif
but I think with the aikido you would get a great amp. is it the ultimate? who knows? unless you have extensively used all the headphone amps (diy or commercial) out there, how can you say it's the ultimate?

I must've spent around $700 USD in this amp already. and will probably need to spend another $30 on bits and pieces.
 
Sep 11, 2006 at 3:29 AM Post #3 of 31
Aikido X2

Great bang for the buck and sounds great.

-R-
 
Sep 11, 2006 at 6:10 AM Post #4 of 31
There's two topologies I'd consider.

1) Single high-gm triode w/ output transformer.

2) Push-pull pair of triodes w/ output transformer. Can possibly be a diff. pair.


I have #1 right now, with an ECC99, CCS loads for AC-coupled output transformers (called parafeed). It sounds great with my HD600's. The advantages are simplicity, extremely good control of AC current loops, and only a single capacitor (4 parts in all actually) in the current path. The disadvantage like all single ended designs is higher 2nd order harmonic distortion due to the dynamic impedance (tube has higher source impedance as it turns off, lower as it turns on).

The second option is basically never used in headphones but there's no reason it couldn't be. The advantage is essentially total elimination of capacitors from the signal path, and much more flat, symmetric dynamic source impedance. The disadvantage is the extreme cost of good transformers that can handle the DC standing current and have good balance, and slightly more complicated AC currents in the power supply. You also need to provide phase splitting. In headphones this can practically be done by making it into a differential pair (CCS on the cathodes), but that puts the two tubes in series rather than parallel, which makes the source impedance not so flat, but still symmetrical.

I'd strongly recommend staying away from output transformerless types, as with 32 ohm phones you'd be talking about parallel tubes or cathode followers or SRPP-type stages which have AC current loops that are much harder to control. It also essentially guarantees being stuck with capacitors in the signal path as well.

In all designs remember that the power supply is as important as the amp, you can't always hide capacitors or other nonlinear elements from the AC current paths by drawing them on a second page.

In my SE amp I use parafeed so that high quality amorphous type transformers can be used, and a CCS-fed shunt regulated supply with CCS loads for the tubes provides absolute isolation from the PSU caps. You can see pics & schematics here: http://s58.photobucket.com/albums/g241/gridstop/ I used autotransformer wiring because then commonly available 10K:600 transformers could be used as 7.5K:300, but this requires 4-pin headphones. For 32 ohm cans you could use MQ's TL-404 autotransformer (again requiring 4 pin phones) or one of the lundahl 14.4:1's would be good and allow 3 pin connectors. But like any SE amp, you would need cathode bypass capacitors to allow 3-pin connectors, while a 4-pin setup allows elimination of the bypass cap. Most headphones bring all 4 wires to the connector so it doesn't require a complete recable anyways, just a new connector on your phones.
 
Sep 11, 2006 at 9:29 PM Post #7 of 31
Quote:

Originally Posted by regal
If you do go P-T-P. I lot of people have been waiting 3 months for a board and they already started collecting money for round two. The designer basically quit the scene do to dishonorable activity of the board maker.


... and some of the customers are dishonorable in their attempts to spread vicious rumors. Despite the problems of delivery from China, I wouldn't hesitate to order one of his boards: they are state-of-the-art, two-sided, ground plane, silkscreened flexible parts sizes with several configurations possible.

You also fail to mention that the two other primary designers haven't gone anywhere, neither have the prototype team members. I have my own reasons for suspecting what happened, as I'm sure many others do, too. Nevertheless, I dare say none of us know the whole story, but you would do well to follow up with him as he promised on the 12th - and he will send a new board - rather than spreading the issue over here.
 
Sep 12, 2006 at 12:43 AM Post #9 of 31
Quote:

Originally Posted by tomb
... and some of the customers are dishonorable in their attempts to spread vicious rumors. Despite the problems of delivery from China, I wouldn't hesitate to order one of his boards: they are state-of-the-art, two-sided, ground plane, silkscreened flexible parts sizes with several configurations possible.

You also fail to mention that the two other primary designers haven't gone anywhere, neither have the prototype team members. I have my own reasons for suspecting what happened, as I'm sure many others do, too. Nevertheless, I dare say none of us know the whole story, but you would do well to follow up with him as he promised on the 12th - and he will send a new board - rather than spreading the issue over here.



Well said, tomb. Although I waited a long time for my SOHA board, it finally arrived and it was a very good quality board. I had no problems with the built at all.
 
Sep 12, 2006 at 1:12 AM Post #10 of 31
Nerull, I think you need to specify what you expect from and what price range. I am not sure what you really meant by "ultimate". Something like this will be ultimate enough for you?
wink.gif


http://www.headphone.com/products/he...oshino-v20.php

There's a lot of good amps out there. Some commercial and some DIY stuffs. Find yourself what fits your liking. If you decided to go for DIY route, I think there's two way of doing DIY. First way is to start to build from simple amp to more complicated ones step by step. This is good way because it will give you some sort of idea about how things work. Second way is to buy/make the one you want. You may not able to learn anyting but you will get what you want if you look harder to find the one. So, search this forum first to see what people said about amps.
 
Sep 12, 2006 at 2:22 AM Post #12 of 31
I guess I should really rephrase this, as I guess I wasn't very clear in state of mind last night. I have a lot of DIY experience (1 DIY cd player, 2 M^3 amps, and a bunch of modded CD players), so that isn't a problem, and I don't want to buy a commercial amp (this is the DIY forum, yes?).
So I'm definately looking into Dr. Gilmore's design, the DynaTube.

Knowing Dr. Gilmore's designs and goals in his designs, I'm going with his. But there's one things that concerns me.
Will the DynaTube work in the last current state as posted by Dr. Gilmore here:

http://gilmore.chem.northwestern.edu/bamaslama4.zip

And here:

http://gilmore.chem.northwestern.edu/bamaslama2.jpg


If so, then I'll get boards of it made and make it. If not, what flaws can be seen in the design?
Again, sorry if it was a bit vague, I have done quite a bit of searching near and far and wondering if anything compared to Dr. Gilmore's beast.

~Tom
 
Sep 12, 2006 at 2:37 AM Post #13 of 31
I looked at the dynatube before going with the aikido. from my readings, no one has built the dynatube since mr. gilmore hasn't completed the testing/design. Or at least I got the impression that the design wasn't final.
==================
1 thing to note about 6AS7 family is that the triodes are highly mismatched.

6AS7 has been used in Atmasphere (10K+ amps). So they may have a way of making the tubes matched through the design or maybe they matched the tubes as close as possible using a tester.
==========================
anyway looking at the tube section, it's simple enough for a p2p. less than 10 components! but I think the servo section is quite complicated and needs a proper pcb.
 
Sep 12, 2006 at 4:55 AM Post #14 of 31
Quote:

Originally Posted by jarthel
I looked at the dynatube before going with the aikido. from my readings, no one has built the dynatube since mr. gilmore hasn't completed the testing/design. Or at least I got the impression that the design wasn't final.
==================
1 thing to note about 6AS7 family is that the triodes are highly mismatched.

6AS7 has been used in Atmasphere (10K+ amps). So they may have a way of making the tubes matched through the design or maybe they matched the tubes as close as possible using a tester.
==========================
anyway looking at the tube section, it's simple enough for a p2p. less than 10 components! but I think the servo section is quite complicated and needs a proper pcb.



Jarthel,
I am considering building a tube amp, and the Aikido is on the list of possibilities, was wondering if you built the octal or 9 pin version?
Sounds like you like it.

Tom,
I am also looking for a tube amp to build. Right now, I am leaning towards building a BCJ, as described at Headwize. I looked at the Dynatube too, but did not want to for the reason given, not sure if anyone has built it other than Kevin.
Here is a thread from headwize with a "ranking" of some current can amp designs
http://headwize.com/ubb/showpage.php...Cavalli;Jones;
And here is one of the threads about the bcj
http://headwize.com/ubb/showpage.php...srch=bcj;mkii;

Randy
 
Sep 12, 2006 at 5:19 AM Post #15 of 31
Quote:

Originally Posted by randytsuch
Jarthel,
I am considering building a tube amp, and the Aikido is on the list of possibilities, was wondering if you built the octal or 9 pin version?
Sounds like you like it.


Randy



I am building the aikido. All the parts hasn't arrived. I am going all out on this one. tube regulation and DC heating. little extras as well like B+ power-on delay and etc.

as for the sound, I do not for the reason mentioned above. But as I said, I did some research on the aikido in AA and diyaudio and have heard only good things about it. As you know, Mr. Broskie isn't popular in headwize/head-fi so his designs are ignored/unknown to the community. but he is a respected tube designer from what I have read.

If I have to buy the resistors again, I will use Mills non-inductive wirewounds. pricy but from my readings, it seems to be worth it. I am using Kiwame at the moment.
 

Users who are viewing this thread

Back
Top