BA-3 Amplifier illustrated build guide

wengtech, one chassis on the XS300 is for the power supply. The heatsinks on the power supply chassis are used for the constant current source mosfets which bias the output stage partially into single-ended class A even though it is a push pull output stage. Pretty neat trick. Similar to how the .5 and the .8 work but on a different scale. The constant current source are tied to the output stage.

6L6: do you think adding a little bit of feedback would make the amp stable enough to not need DC blocking caps? From the output to the JFET buffer with a resistor in line?
 
@wengtech -

tl;dr - It's always a really fun exercise to see how to get massive power. I know I like it. I think you've seen some very practical concerns raised. They're not insurmountable.... but...

To use the complimentary output stage as as a follower (big honkin' crippled F4) ...I think you'd need at least double the number of output pairs and a helluva set of heatsinks to reach what I'd consider an adequate setup for 300W into 8R at a reasonable Iq for performance.

You could go balanced if you wanted... but I was trying to slowly move toward some level of practicality vs. 300W of Class A ...

🙂
 
@wengtech: Since Jim brought up the objective performance of the output stage, I suggest you have a look this Papa article: https://firstwatt.com/pdf/art_leaving_class_a_2019_redux.pdf
hello Dennis, i still don't get it, is that mean biasing BA3 output stage into class AB? (is that even possible?)
i just scrolling and scrolling the article and came to the final conclusion by mr.Pass is that Class A amplifiers sound better than the alternatives😆

wengtech, one chassis on the XS300 is for the power supply. The heatsinks on the power supply chassis are used for the constant current source mosfets which bias the output stage partially into single-ended class A even though it is a push pull output stage. Pretty neat trick. Similar to how the .5 and the .8 work but on a different scale. The constant current source are tied to the output stage.

6L6: do you think adding a little bit of feedback would make the amp stable enough to not need DC blocking caps? From the output to the JFET buffer with a resistor in line?
stand alone chassis for ccs stage? cool...😎
hello Mr. @Nelson Pass would you shared the picture inside XS300 amp and those PSU?🤣😍🙏

Dang it! You stole my punchline...

I was waiting to see...

I know I don't have any pre-amp in my arsenal that will deliver 140Vpp (ish). I wanna see!

tl;dr - It's always a really fun exercise to see how to get massive power. I know I like it. I think you've seen some very practical concerns raised. They're not insurmountable.... but...

To use the complimentary output stage as as a follower (big honkin' crippled F4) ...I think you'd need at least double the number of output pairs and a helluva set of heatsinks to reach what I'd consider an adequate setup for 300W into 8R at a reasonable Iq for performance.

You could go balanced if you wanted... but I was trying to slowly move toward some level of practicality vs. 300W of Class A ...

🙂
well, tbh the 300W class A is not my major concern, it's just to make it more interesting and challenging to build.😛
my major concern is to drive this output stage with tube preamp, it just tickles my hands for doing that.
a friend of mine just lent me a tron seven line stage preamp to be tested with this ba3 os
so far i know it use 2x5687 tube with cathode follower design and has capacitor coupling at the output
the gain is about 15dB and Output Impedance <600 Ohms, that's what i know so far
is that enough juice to drive ba3 os? what do you think?
 
^ I love to build the amps too, but choosing the parameters is fun (for me) also. Maybe it will be for you too.

So maybe consider this... among a few things.

Let's assume that the pre-amp designer planned on max of 2Vrms input and added bit of headroom, so that the pre-amp would not voltage clip. You can also look up max voltage output in the specs. You know the gain.

Alternatively or side-by-side, you can look up the max Vp of whatever source you're using. Factor in the gain of the pre-amp assuming that the pre-amp will handle it well. Then you know the max Vp that will exit the pre-amp. You'll likely never hit that output, but it's nice to have the cushion. Clipping is very annoying (to me).

Either way... figure out a practical max Vp that will enter your new (not BA-3) complimentary output follower amplifier. Let's just call it a crippled F4 => CF4 for all intents and purposes.

You know that the CF4 has no voltage gain... so Vp in <=> Vp out or close enough for government work.

You know the load that you're most interested in.

Set your rails based on the Vp and do the calculations based on that. It will be your max power achievable into that load. There's no need to design an amp for 70Vp if it will never see 70Vp (IMO). There are too many things that could go (very) wrong... and you'd possibly and unknowingly make compromises that would deteriorate the actual performance of the amplifier in the name of a number. I'll guess that you could use a more 'normal' build and be thrilled. So many people ask "will it drive an F4". You're now in the boat of "Will it drive a MASSIVE F4?" 🙂

Boring stuff below... but it's how I work things through in my wee brain.

As an example, my "newly-remodeled", 12-pair CF4 build will initially run off ~32V rails instead of 24V like it was previously. Why? 1) That's the power supply I have built at the moment. 2) It's nice b/c my tube pre-amp has a max output of ~60Vp => 40Vrms. Wait! you say... then shouldn't you have higher rails? No. The gain of the pre-amp is 20dB => 10X. Max output of any SE source I have is 2Vrms, at 0dBFS, which is never hit. Digital clipping annoys me even more than analog clipping. So... the max Vrms out (for me) from that pre-amp would be 20Vrms => ~29Vp. Again, it's likely never to hit that, but ...

32V rails (loaded) would be pretty appropriate. Whether I actually ever, ever hit that mark or not, I'd consider that a reasonably well-thought out chain. In a REALLY loud listening session with my least sensitive speakers... ~20Vp is max I've seen for a transient if I measured it properly. I also usually sit 5m away from my speakers or stand about ~7m away at the bench. So... I'm not really going deaf. It'd be fine if the music had any dynamic range, but... C'est la vie, I don't listen to audiophile recordings. Dynamic range is typically 7dB, but occasionally I dip my toe into the refined waters of some nice classical recordings. tl;dr - I've got more usable voltage than needed, but I'll never clip. I like that. With that said, am I still going to build a Wolverine with 71V rails... maybe.

I also am going to put my 300B amplifier(s) in front of my CF4 and see if I fall in love again. If so, I may come up with a slightly different plan.

Either way, here's the rub. I'll have to lower the Iq b/c I don't have massive heatsinks, and I was already in Burning Amp territory at 24V rails at the Iq I chose with the heatsinks they're attached to. That's the opposite direction for my 4ohm speakers hooked up at the moment and for distortion performance. I'd rather have the lower rail voltage / higher Iq. I will likely swap out toroids. Fortunately, I put together a PSU that makes it very easy to do that. I'm knowingly 'starting off on the wrong foot', but I've accepted the compromise to just check it out and make my life a little easier for the time being while I decide the final fate of the boards.

If love is in the air with the 300Bs ... then I have to noodle over a few things. The 300B amps have an input sensitivity of 3.9Vrms, and only have gain of 7.3dB => ~2.4X So... that's actually a really attractive option, since I don't need much gain in my system with most of my speakers at 'normal' listening levels. A buffer / passive attenuator with a 300B amp or two as my "pre-amp" is one of those really goofy things I'm dying to try. I tried the 'normal' F4 with an SS pre in front of the 300Bs, and it was pretty fantastic.

It's an adventure. Have fun along the way. Any way... that's a really really long way of saying that I personally don't give one iota about the power output of an amp initially. I start with the max voltage from my source and follow it through the chain adding gain where I want until it reaches the speaker. Conversely, I start at the speaker and work backwards based on SPL => voltage desired. The two meet in the middle somewhere.

Initially... I'll have max 2Vrms from the source... gain of 10x => 20Vrms

So... what's the power output into 8R of a 20Vrms signal (if I phrased that properly)? A 'paltry' - 50W.

Edited to add - I forgot... What's the power output at the highest peak I've measured in my system... 20Vp... => ~30W (rounding) assuming 8R... but... for me... it's closer to 4R or lower in the bass region so... closer to 50W. This is one of the things I'm trying to figure out when deciding how to configure a few things.

You may have seen (if you read the other wall of text) that I am planning on using some FE "modules" with the output stage also some time later. Then, I'll need to rethink it... again... but I think I've got a solid set of building blocks to play with.

This is all easy stuff for the resident pros, but I still have to plod through it. Standard caveats apply... I'm a dodo that's learning... so...

Enjoy!
Patrick

:cheers:
 
Last edited:
That's a very fun adventure👍
300B as preamp is also on the list😍
(I wonder how you hooked up 300B amp into CF4, isn't that 300B amp using output transformers design for speakers?)

There are still many possibilities that can be explored, increasing Iq lowering Vcc or vice versa, pp or se, and not to mention about those lateral mosfet (K1058/J162 families)

btw the tron seven max voltage output in the specs is 18v

I think i will start with what i have and stop calculating, just finish the project and connect any preamp/FE i have or borrow from friends😅 and see if it will drive this amp or not.

Thanks for sharing your builds.
:cheers:
 
  • Like
Reactions: Holzarbeiter
^ Thank you! I know it was a wall of text... but I do love this stuff. I hope you have twice as much fun as I have.

re: how to connect the F4 and use it for extra ooooooomph... Basically you just put a resistor across the output of the preceding amp. Make sure it's rated properly. Some people like certain types, and people have experimented with values based on their OTs and preferences.

This person was one of my inspirations for giving it a go...

https://www.diyaudio.com/community/...pass-f4-amplifier.234355/page-10#post-3864179

Have a blast and enjoy the tunes! You really can't go wrong.
 
Building a BA-3 with the SE output boards. I'm a little confused with the hookups between the bias board and the output board. These boards have pads marked OUT and S. 6L6's pics show the push-pull boards only have the OUT pads which are joined. The schematic for the SE boards shows the S joined, but there is no reference to OUT. Which ones should be joined?
 

Attachments

  • bias .jpg
    bias .jpg
    232.8 KB · Views: 92
What I would do is...

  • Connect the 5 pads on one side with V+ through V- between the boards exactly as they are marked and aligned from the bias board to the output board.
  • Connect one S pad from bias board to output board similarly
6 connections.

I would not worry about a guide for a different amp... See BA-1 thread.

I'd get double confirmation.
 
  • Thank You
Reactions: maltboy1
you need proper biasing mechanism, then you end with OS having both voltage and current gain, then you need FE or preamp, then you need to enclose NFB around everything to get where you need to get

what are the reasons of that idea of yours?

anyhow - one of ways is - take an inspiration from https://www.diyaudio.com/community/threads/new-stasis-front-end.363701/, build your own Stasis Mosfet OS, and inform us of your experience

:clown:
 
  • Like
Reactions: wengtech