Tube amplifier - pls help me design it.

[madriaanse]

I would like to build a super high quality tube amp for my monitoring system - something good enough to drive a set of Wilson Watt/Puppies. I have been browsing around various tube amp DIY web pages, but haven't really found anything I like. I have the following design criteria:
1: high quality transformer balanced line input (Jensen, Lundahl)
2: fixed gain - wide open - no pots.
3: minimalist, purist signal path using commonly available tubes
4: super high quality speaker output transformer (who's "the best" in this field?)

Before I put one line on paper I'd love to hear your thoughts or pointers in the right direction!!

Thanks in advance!!

Martin

 

[macwanj]

I suggest you check out www.diyaudio.com; theres an entire section dedicated to tube amps. I pretty much learnt all my tube amp theory on that forum. What are your speakers rated at in terms of effeciency? Do you want something clean or are you looking for more even order harmonic distortion? Do you want something snappy with god transient response (like a PP Class A or AB) or do you want something smooth (like a triode SE). Some more information might help folks in pointing you in the right direction. IMHO there is no absloute best transformer, cap; the key is in trying them out and deciding which ones you like. A good place to start with transformers is Hammond and Edcor. If you have a higher budget, theres Tamura and MagneQuest.....

 

[madriaanse]

Hello Macwanj,

Thanks for your reply!! you wrote:

What are your speakers rated at in terms of effeciency?

92 dB @ 1 watt

Do you want something clean or are you looking for more even order harmonic distortion? Do you want something snappy with god transient response (like a PP Class A or AB) or do you want something smooth (like a triode SE).

Since this will be for my "big" speakers - a quasi-mastering setup - I would like a very clean amp with excellent transient response.

Some more information might help folks in pointing you in the right direction.

Ok, let me know if there are any other specifics I should mention. With what I've written above, do you have any amp designs you'd recommend as a starting point? I found this design interesting: http://www.lundahl.se/claus_b_se.html It is probably a bit underpowered for my needs, but I like the idea.

Thanks again!!

Martin

 

[macwanj]

Hey Martin,

Funny, I was going to suggest that exact design to you. I personally love the sound of the EL34 in triode mode. You know, you could mod that design to give you as much as 15W of power as a pentode whilst still retaining the flavor of a triode. This is done by getting a SE output transformer with a ultra linear tap and conecting to the screen (a form of selective feedback I think...). Alternatively you could look for an EL34 or EL84 push pull amp like the Mullard schematic I've attached...

On a side note, you should be aware that tube amplifers tend to compress program material; also having an ultra clean amp defeats the purpose of using a tube amplifier as most of the warmth that we perceive is even order distortion. Also a tube amp will never have as good of a transient response as a solid state....

Once again I strongly reccommend going to the diyaudio forums; you'll find so many project there...

Hope this helps....

Have fun and be safe! Always remember when testing live circuits; put your left hand in your pocket and never forget to drain psu caps!

 

[alexc]

If you don't mind a pre-existing design,
I suggest the modified Dynaco MkIII KTM88 based 60W Ultalinear PP monoblocks (from Triode Electronics)

Extremely popular for the last 40years. As fundamental, simple and purist as it's possible to be!

In fact, that's what I am doing as soon as our dollar improves a little 

You can leave off the volume pot and add transformer balancing.

Am planning on running them in 'passive bi-amping' arrangement with my
exisiting stereo tube amp (9W SE single 6L6) to my Tannoy 8" concentrics.

I was using just the stereo tube amp for monitoring but decided *I need more POWER*.
Actually, at 9W/channel, it was pretty loud! I just want more tube amps around me  ;D

The difference from the previous 200W mosfet amp and the 9W SE was pretty impressive.

And if you live in a cool climate, these are doubly good. They warm up the room nicely

I am now literally converted 

 

[schmidlin]

You brought up a concept I have been noodling with for a while:
If you have balanced in, why not stay balanced all the way thru the amp?  Something like a Gates StaLevel (no side chain, of course) or a McIntosh after the first (PI) stage.  You get the idea.  It shouldn't be too hard, and may be the most stable design going.  Two preamp tubes and a pair of power pentodes would do the it.  Neg feedback may be tricky, tho, but not necessary.

All you EEs out there: any reason not to try this? (it may have been done, for all I know)

 

[macwanj]

You brought up a concept I have been noodling with for a while:
If you have balanced in, why not stay balanced all the way thru the amp?  Something like a Gates StaLevel (no side chain, of course) or a McIntosh after the first (PI) stage.  You get the idea.  It shouldn't be too hard, and may be the most stable design going.  Two preamp tubes and a pair of power pentodes would do the it.  Neg feedback may be tricky, tho, but not necessary.

All you EEs out there: any reason not to try this? (it may have been done, for all I know)

Yes, it has been done, a good example is Manley's Snapper monoblocks: http://www.manleylabs.com/containerpages/snapper.html

One disadvantage to this approach is that a custom transformer is required; atleast if you're going to implement a push pull topology for higher output power. This can be expensive...

 

[mcs]

You brought up a concept I have been noodling with for a while:
If you have balanced in, why not stay balanced all the way thru the amp?

No problem doing that.

Two preamp tubes and a pair of power pentodes would do the it.  Neg feedback may be tricky, tho, but not necessary.

Feedback would be very necessary. Pentodes are terrible without feedback. If you made the amp with triodes, you wouldn't need feedback.

it may have been done, for all I know

It has.

Best regards,

Mikkel C. Simonsen

 

[analag]

Here is a design I was playing around with. In cases like this, I like to wind my own output transformer on C-Core for total electrical and physical symmetry.

 

[schmidlin]

That is totally nutz.    Part of me thinks you are just bragging, but I know you  better.  All comments in respect.  I'm getting most of it, but what is that cluster off the cathodes of the 6SL7?

And what is your overall goal?? Low THD?  Any prototypes?

 

[analag]

The cluster hanging off the cathodes is there to present an extremely high impedance value or as high as I can make it without resorting to more complicated measures. It's set to hold the cathodes at 2mA. With theoretically matched triodes such as those in a simulator, the out of phase signals are closely balanced, before it goes to the second stage.  Yes I made this one a few years back as (2) mono blocks on the same chassis for better stereo separation.
Something I like to do is to use individual bias pots. As you tweak them you can actually hear when you achieve balance in the output tubes. Dramatic reduction in noise. May need some small value grid stopper between the CF and the 6L6's. The cf is there to facilitate the use of smaller coupling caps not class AB2 drive potential.

 

[analag]

http://www.plitron.com/shopping/specs/2100CFBH.pdf

I would try an output from these guys if I were in the market for such a thing.

 

[kevinc]

Here's a schematic of an input stage to which you can attach your own PP output stage.  I've used it in this form and with cathode current sources to promote true differential amplifier action and it sounds excellent.  Other tubes can be used (with care to the consequences) and the plate choke can be replaced by plate resistors (have to adjust PS dropping resistors to do this) and interstage transformer can be replaced with plate resistors and coupling capacitors (have to adjust PS dropping resistors to do this) to save money.  The results are still very good.  The key to preserving the good qualities of this circuit is to retain the input phase splitting transformer and the differential stages.  I prefer not to use overall feedback.  The input can be connected for unbalanced (as shown), balanced, or both.

http://www.kandkaudio.com/images/PPAmpInput.pdf

If you don't need too much gain or power look at this http://www.kandkaudio.com/poweramplifier.html.

 

[madriaanse]

Wow, thanks everyone for the great suggestions and ideas!! The Dynaco MK3 looks like a convenient and quality solution: http://www.triodeel.com/dynamk3.gif

I would like to add a Jensen JT-11P4-1-1 on the input: http://www.jensen-transformers.com/datashts/11p41.pdf
Just follow the Jensen data sheet (20K on secondary) straight into the Mk3's 6AN8, right?

Thanks Again!!

M.

 

[alexc]

The 6AN8 (pentode and triode combination) is often replaced with 2 triodes in the Mk3.
There are a couple of modified 'driver boards' offered out there.

I would first go with a modified one such as offered in the base Triode kit.

But yes - the traffo would unbalance and connect to grid of first preamp tube.

But these amps normally have a sensitivity of around 1Vrms for rated output.
(the Mk4 is 1.3Vrms for rated output).

So - if you are driving from your desk Control Room Output, you need to
consider the level of output drive. Particularly if you plan to use no volume control.

The Jensen is 1:1.4 so feeding it with a CRM output hotter then 1V is probably going to
give you a limited range on your CRM pot and possibly overdrive the amp.

What's your feed level ? I think you may want to have 1:1 but there's not much in it.

The impedance needs to be considered also. What are you driving it from? 10K or 600ohm?

Perhaps someone can give some advice? The Jensen seems to be around 11K:3.4K or so.
Probably OK for a CRM output

 

[madriaanse]

Thanks for the valid points. I will be feeding the amp from a Coleman Audio M3PH MK2 Passive Control Room Monitor, being fed by a Lavry 4496.

 

[alexc]

Sounds good - I imagine the Lavry  won't be putting out a crazy hot signal, like some units can
and the controller should give a good range of volume control.

Good luck with it.

 

[Igor]

Just my $0.02...

You wrote, "I would like a very clean amp with excellent transient response."

Hmmm... If you going to make single-ended amp, it is easier to find
suitable trunny. But, most of these designs suffering from nice, but...
a way distorted bass and low damping factor.
From the other side, it is hard to get absolutely symmetric slew rate in SE design.

For P-P, good transients means absolutely symmetric designed trunnys.
Core material means a lot.
For input, I'd prefer 1:1 (or 1:1.5) on high grade nickel, like 8414 UI-31 from Magmet.
Allmost ideal core for this purpose.
For output, 2 chokes on HI-B C-core preferred. Sectionizing/watts is up to your taste.

I listened to some very high quality tube amps, did some for my home use and sold some as well.
Try next experiment....
Take very fast clean transistor amp in class A prefereably, which suits good to your speakers.
Simple switch or relay, to commutate input/output...Equalize levels within 0.1 db with measurement mic on stand
and same test signal on CD or converters (suggestion is -18db, 1 or 2 kHz).

Then listen....

Tube amps almost allways sounding very pleasant to ears, but good fast transistor amp
can have more details and better bass on same level. Less low-midrange coloration as well
(no trunny's no coloration

Without too much getting into audiophoollery debris..
All this is just IMHO. You may have different taste and prefer other kind of sound.
Hope it can be helpfull.

 

[Sorr]

I like this amp
http://www.drtube.com/schematics/radford/sta100.gif