300B - Push Pull or PSE to drive a commercial speaker

[arnoldc]

After auditioning dozens of speakers, and finally settling on a DIY speaker, for my near-cube room (4m x 4m x 3.5m H), I found a commercial speaker that conquered the room. It is the LSA1 from The LSA Group. It sounds very good, and is currently being driven by my vintage Harman Kardon 430.

However, I wanted a tube amplifier to ultimately drive this, and after selling my Anthem Amp 1, I'm left with flea powered amplifiers that definitely does not have the power.

So I wanted to embark on another project and to be more sensible, I will use a pair of 300B tubes.

The question is, should it be PSE or push pull? I have not heard any of the topology and does not know advantages of one from the other.

Can anybody enlighten me?

I also wanted to salvage the Lundahl LL1660S/10mA that I used in my last 300B to be used as either interstage or phase-splitter. The questions surrounding this would be, should I use it with single ended to push pull, or push pull to push pull (should I opt for push pull finals).

The other goal is to get as much power output as possible, and I've been playing with TubeCad that can do PSE, and I can only get 18W out of the 300B pair with around 200Vpp, with positive grid. Is there a way to do some 20W or more if I use push-pull?

Please bear with me as I'm a total newbie when it comes to push pull or PSE stuff.

 

[GK]

The other goal is to get as much power output as possible


Well that rules out class A and PSE.

 

[arnoldc]

AB1 and Push Pull?

 

[GK]

Yep. A pair of 300B could make a rather nice AB1 P-P amp with a decent power output, provided that the 300B drive requirements were met.

I really wish these tubes went for the same $ as a 6550 or a KT88......

 

[Eli Duttman]

AB1 and Push Pull?

AB2 and push pull. Tubelab's Power Drive takes care of pulling the 300B grids slightly positive. Don't over do it. IRFBC20 FETs will be terrific as the drivers.

A lot of the euphony found in SE 300B amps is going to disappear, due to the inherent cancellation of even order HD products associated with PP "finals". A 6AB4 (1/2 a 'T7) in the voltage gain stage will restore the "moisture". An ECC99 LTP will take good care of phase splitting.

A considerable negative voltage will be needed to bring the 300Bs close to cutoff, for the no signal state. IMO, self biasing will prove very difficult. My hunch is that combination bias is the way to go.

AC heating of the 300Bs should be fine. When the filaments are phased up correctly, hum gets cancelled.

 

[arnoldc]

Gents,

Thank you for the replies. I believe I have to go with fixed bias to attain this goal.

Eli, you think the LL1660S has no place for this?

 

[Eli Duttman]

Gents,

Thank you for the replies. I believe I have to go with fixed bias to attain this goal.

Eli, you think the LL1660S has no place for this?

Arnold,

You did say this amp will be used with commercial speakers. Adequate control of the woofer voice coils is a concern. A few dB. of loop NFB might be the difference between success and failure in the damping factor dept. Interstage trafos inside a NFB loop are a prescription for trouble. A Mullard style circuit, with the voltage gain stage DC coupled to the LTP has but 1 cap. (at the splitter's anodes) in the signal path. Therefore, phase shifts are well controlled. The use of high gm small signal tubes is protection against slew limiting. Roll infrasonic noise off at the amp's I/Ps with high pass filters that "corner" just below 20 Hz. Killing infrasonic crud off protects the O/P trafo cores from FUGLY saturation.

IMO, you should save the Lundahl "iron" for a different project. How about a fully differential circuit using a 12AT7 and triode wired EL34s? No loop NFB, here.

I agree that "fixed" bias is needed, but I would "stand" the 300B cathodes on unbypassed 470 Ohm resistors to provide protection against run away. It's combination bias.

 

[kevinkr]

Hi Arnold,
Long ago (like 10 yrs now) I designed a 30W PP 300B amplifier with regulated supplies which was published in VTV and became part of the SED technical library, (long gone) and while I am not recommending you build this particular design I do recommend you at least read the article which is on my site.

I don't recommend AB2 operation for any 300B, as the grids are rather easily over-dissipated, plus at 400V B+ and 120mA quiescent current you can get 30Wrms out with a primary PP of 3.8K and at less than 2% thd with no feedback. Any transformer in the range of 3K - 5K will provide good results.

I used fixed bias in all my designs and recommend that.

This design depends heavily on the quality of the regulated power supplies and was very carefully designed to provide the required drive on a +400V supply, unfortunately the LTP based driver stage has limited bandwidth when driving the miller capacitance of the 300B.

Lots of people have built these amplifiers, and in addition I sold a fair number of a simpler version of the 300B PP amplifier commercially as well.

Note that I have not found ac heating to be very acceptable long term in my commercial 300B based PP amplifiers, and ended up retrofitting at least 5 of these amplifiers (KTA3030) for dc heating at great expense. All subsequent units had dc heating. All would be well initially and then after a few weeks or months the tubes originally well matched would age and drift out of balance - hum would become intrusive. The procedure of setting the bias and then nulling the hum was tedious for the average user, and not as effective as expected. The JJ 300B is the only tube I have used that did not have this issue over the long term.

Here's the link: http://kta-hifi.net/projects/amp_page/300b_amp/300bpp.html

 

[resident]

Kevin, thanks for the link! Nice schematic.
I'd like one day to build a "universal" P-P amp and try different output tubes. Both triodes and pentodes. And as I read in your article the input stage can provide 170Vpp per plate. I think it's a nice choice. What do you say?

 

[Tubelab_com]

Is there a way to do some 20W or more if I use push-pull?

I built a P-P 300B amp several years ago. It started out as a variation on Kevins design but morphed into something different to work with the parts that I had. I have used this amp on and off for 6 or 7 years. I usually grow tired of my amps after a few years and sell them, trade them away, or rob them for parts. I still have this one. There is something about the sound that I have not yet been able to duplicate. Last year the enevitable happened, I flipped the power switch, and zap... poof....twang...it was dead. I am sure that one of the electrolytics shorted leading to some toasted diodes.

I never drew a proper schematic, but am now attempting to reverse engineer one of my own amps to figure out what I did. I know this:

The input stage is an LTP phase splitter using a 5751. It feeds two seperate SRPP driver stages using 6CG7's. The SRPP stages are capacitor coupled into the 300B grids. Fixed bias is used. AC heating is used on all tubes. The power supply is a SS bridge on a 350 VAC large toroid feeding a choke input filter. The B+ is about 375 volts. - 200 volts is generated using some more diodes and capacitors somehow.

I measured the amp when I built it and it made about 25 watts. The output tubes are biased so that they operate in class A for power levels up to about 10 watts. I am using a 6600 ohm OPT.

I made this amp with some cheap parts with the idea of upgrading it later. I used some guitar amp OPT's that I got for $15 each. The power supply components came from computer power supplies, the coupling caps came from ceiling fan controls. There is some synergy happening between all of the junk because every time I have attempted to upgrade it the amp sounds worse. I learned to leave it alone.

Now that it is dead, what do I do? My current plan is to change the dead parts and do nothing else. I want to keep it as a standard by which to measure my next attempt. I plan to do another 300B P-P amp, but with more updated circuitry. PowerDrive, Augmented Cathode Follower, I don't know yet, but it better sound as good as this one, or it gets dissected and I start over! I want the driver board to have "universal" drive capability, but if it can drive 300B's it can drive most anything else except for maybe 845's.

 

[arnoldc]

Gents,

thanks so much for all the help... I'll try to address them one by one...

Eli, I came up with this schematic-

[IMGHTTPDEAD]http://img233.imageshack.us/img233/7536/dsc00523ts7.jpg[/IMGHTTPDEAD]

Did I interpret what you said properly?

edit:

There's a flaw in the schematic, as the grid leak resistors appeared after the grid stopper.

Hi Kevin,

You know, I realized I have that file, and I was perusing it last night, when it struck me, it sounded like you. There was no Kevin Kennedy in the document, until the last diagram I will be reading it thoroughly, and you can expect me to ask questions.

update:

Hey Kevin, I just got this SMS from a good friend of mine whom I asked if I should go for PP or PSE. He said-

"i would go for a push pull 300b i got a PP 300b by Kevin Kennedy before one great sounding amp. look for the amity 300b by Lyn Olsen for inspiration. PSE 300b tends to sound like a good SS amp for me."

Hi George,

I was actually reading the 300Beast yesterday, and I got scared with the pic But knowing it's your prototype, I'm not suprised. Thank you for describing the circuit, I'll try to make a diagram off it.

 

[arnoldc]

Kevin, he further said-

"magaling yan. sayang hindi nag take off yung company nya na KTA before. but patok yang PP 300B."

Translated to English-

"He's skillful. It's a pity his company KTA did not take off. The PP 300B is a winner."

 

[Tubelab_com]

I'll try to make a diagram off it.

I found the Eagle schematic on an old computer and saved it on a CD. It is for the driver PC board only. If I can get the new Eagle version to read the file created in an old DOS version of Eagle, I will post the schematic. Understand that I did this amp several years ago. I have learned much since then, but it sounds real good.

 

[arnoldc]

Hi George, looking forward to it. GIF/JPEG please

 

[kevinkr]

Hi Arnold,
I made a lot of mistakes... It sure is nice to know that my work is appreciated, although at this point I would consider that design a bit dated.. It was a set of well balanced compromises in order to make something that others could actually copy. Performance with UL connected 6550 is pretty stellar and it does a pretty good job with the 300B too. I think it was probably the best sounding PP amplifier I designed at that power level, and I still have them - not something I can say for most of the other amplifiers I've built.

These are the ones I kept or built since the last big clean out:
300B PP
300B SE
2A3 SE
71A SE
45 SE
EL84 PP Integrated (2)

The only one currently in use is the 300B SE amplifier which is one of my best ever.

 

[Tubelab_com]

GIF/JPEG please

Jpeg is included, but it looks kind of ugly because I had to reduce the resolution to make it fit the attachment restrictions. PDF in the next post. I can email you the original eagle file or the gif or jpeg if you want. The forum email has no attachment feature, email me and I can reply with the file you want.

The older DOS version of Eagle used to make this schematic did not allow component values, so they are not on the schematic. I can add them the next time I get the amp out of the closet.

The amp design is fully differential. Just ground one of the inputs if you have a SE source.

 

[Tubelab_com]

PDF

 

[arnoldc]

Hi George, it's tiny! I'm at 400% zoom to read the values. But thanks much, I can figure this drawing out. Maybe make a Visio version

 

[Eli Duttman]

Gents,

thanks so much for all the help... I'll try to address them one by one...

Eli, I came up with this schematic-

[IMGHTTPDEAD]http://img233.imageshack.us/img233/7536/dsc00523ts7.jpg[/IMGHTTPDEAD]

Did I interpret what you said properly?

edit:

There's a flaw in the schematic, as the grid leak resistors appeared after the grid stopper.

Hi Kevin,

You know, I realized I have that file, and I was perusing it last night, when it struck me, it sounded like you. There was no Kevin Kennedy in the document, until the last diagram I will be reading it thoroughly, and you can expect me to ask questions.

update:

Hey Kevin, I just got this SMS from a good friend of mine whom I asked if I should go for PP or PSE. He said-

"i would go for a push pull 300b i got a PP 300b by Kevin Kennedy before one great sounding amp. look for the amity 300b by Lyn Olsen for inspiration. PSE 300b tends to sound like a good SS amp for me."

Hi George,

I was actually reading the 300Beast yesterday, and I got scared with the pic But knowing it's your prototype, I'm not suprised. Thank you for describing the circuit, I'll try to make a diagram off it.

Arnold,

Even with the stoppers correctly positioned, the circuit you drew will not work. You can't get the voltage swing needed to drive 300Bs from a 12AT7 LTP. What you drew strongly resembles both Poindexter's "Musical Machine" and "El Cheapo". You don't need a RC Network at the non-inverting triode's grid. Simply connect the non-inverting triode's grid to ground. That circuit has just enough gain to drive triode wired EL34s.

Get a hold of the Mullard 5-20 schematic. That topology, with a 6AB4 in the voltage gain position and an ECC99 as the LTP, HAS the voltage swing needed to drive Class "AB" PP 300Bs. Remember, at most, only a FEW dB. of loop NFB will be needed, mostly for damping factor.

Unbypassed 470 Ohm 300B cathode resistors are fine, but you also need a negative bias supply, with a trim pot. for each grid.

KK's remark about the grids in 300Bs not being especially tough should be heeded. Power Drive is fine, but you want to go down to zero and no more than a tiny amount positive. Doing it that way maximizes power O/P, without overly stressing the costly 300Bs. Leave the stoppers off the 300B grids, if you use Power Drive. Given the type's low gm, stoppers are probably unnecessary, almost all the time.

 

[arnoldc]

Hi Eli,

I think I have that white Mullard project book that has the 5-20 schema. I'll dig it up and read from it first.

Gotcha, on the biasing, and other tips. Thanks much...

I hope I can pull this project off... This is going to be BIG (at least for me)