• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

12b4a

I have simulated the 12B4A in SRPP and Constant Current draw circuits and I agree it appears to be ideal. I have plans to build a preamp with this in the future. The 417 is a much different tube related to the 6922 - 6DJ8. gain is too high at 42. Try the 12B4A and report back. Microphonic? I have found microphonic tubes of any kind. !2B$A's are cheap enough to sort them out. Check out TubeCad.com
 
Mohan,
could yo please elaborate why?
12B4A and 417A are completely different tubes.

12B4A is the grandchild of the 2A3 single plate, as the granny it is meant to be used for series regultators (and as vertical TV amplifier); it has similar parameters. µ=6.5, G_m=6.3ma/V, R_p=1030Ohm

417A is a supertransconductant tube with high gain. µ=43, G_m=24ma/V, R_p=1800Ohm
AIWI, the 5842 is a WE 417A without the nasty microphonics of the famous tube. Raytheon is to be preferred in this respect.

I plan to build a preamp having a differential pair of 12B4A as output tubes driving a Lundahl LL1660 output transformer. I intend to have the tube running at considerable current: 34mA per tube. I read raves about the 12B4As sonics on the JoeNet; a friend seconded this and so i decided to go with it. As this is a line stage application and meant to be differntial, i am frightened off from using DHTs; the diff.pair's virtual ground would collect too much noise. So i am happy the 12B4A is indirectly heated.
 
dice45 said:
I read raves about the 12B4As sonics on the JoeNet; a friend seconded this and so i decided to go with it.

Dice,

You beat me to that comment. I am going to get some of these and try something out with them too. Here is the recreated data sheet that Bill Perkins did up -- i understand he is sitting on a number of these and selling them cryoed & graded (reason why he went to the effort to create the pdf)

dave
 

Brett

Member
2002-01-07 6:02 pm
dice45 said:
12B4A is the grandchild of the 2A3 single plate, as the granny it is meant to be used for series regultators (and as vertical TV amplifier); it has similar parameters. µ=6.5, G_m=6.3ma/V, R_p=1030Ohm

I plan to build a preamp having a differential pair of 12B4A as output tubes driving a Lundahl LL1660 output transformer. I intend to have the tube running at considerable current: 34mA per tube. I read raves about the 12B4As sonics on the JoeNet; a friend seconded this and so i decided to go with it. As this is a line stage application and meant to be differntial, i am frightened off from using DHTs; the diff.pair's virtual ground would collect too much noise. So i am happy the 12B4A is indirectly heated.
Hello Bernhard,

Please keep us informed with your progress and findings about the 12B4As. I picked up 5 recently, NOS/NIB Sylvanias for $1 ea. Do you know of any preference for a particular brand(s) in terms of performance?

I had also heard good things about them, and wondered how they would fare as a diff stage. I hadn't thought about running them quite as hard as 34mA though. At my first look at the curves, I thought about maybe 25mA with a Vp of 90-100V, loaded by one of Pete Milletts PMCCS, parafeedeing into my S&B TX1022 TVCs.

I love the idea and linearity of DHTs, but using speaks with the sort of efficiencies mine have (104-106dB), I'm concerned about noise and hum. The transfer characteristics are great for an IDT.

Cheers
Brett
 
Dave,

i kiss your feet for posting that 12B4 data sheet link. Thanxalot !!

Brett,

unfortunately my preamp is not to happen soon, 1st i have to get 2 linear tracking tonearms designed ready and built, one open baffle speaker and two power amps built and cooperating.

I do not know any particular preference concerning brand, but i also will go with Sylvanias, i have enuff of them to feed pigs. for my application this is not bad at all, tight matching, yanno.

34mA is recommended data sheet operating point and curves show the tube likes that best. The tube is meant to be the reg tube in a series reg, it has to handle lost of current.

low-level differntial pairs and DHTs: does not work, Thomas Mayer tried that out.
Problem is: the differntial pair has a virtual ground at the connected cathodes and with DHTs, this virtual ground is galvanically connected to all hum and noise the heater supply (trannie) picks up by induction or stray-in in reference to chassis/signal ground.

Now it is the designer's choice whether to inject that noise voltage into the virtual cathode or into the grid, same amplitude, just phase inverted. On line input level, 2V RMS@0dB, S/N ratio suffers too much.

In a differntial power amplifier with the 1st stage being IDHT, i will try low-µ DHTs for 2nd stage (2x 71A ) and 3rd stage (2x AD1). But even this signal level demands separate and symmetric heater trannies with additional electrostatic shield windings connected to ground.
 
dice45 said:
i kiss your feet for posting that 12B4 data sheet link.

No you don't :^) I think you already owe me one beer if i ever get to Munich, just make it 2 (and get someone to carry me to my room :^)

I did a quick pick of some of the JoeList posts i have archived and attach them here as food for thot:

.............................................................................
Joe Roberts: Not necessarily...how about a SE triode into an output trans with a balanced
output? This is one of the most common topologies in pro gear of yore.

My main line amp is a 12B4 into a 5k:600 ohm MQ output trans. Sounds great. If
I wanted to run balanced, the 125 ohm tap is the effective center tap of the
600 ohm winding.

You can definitely drive the snot out of that long cable with this setup!

Next step is to battery bias that 12B4. Bass is already tight in self-bias,
but I want to hear more of that battery thang.

It was recently mentioned in Valve article by my bud John Day. He took it to a
preamp shootout and it did all the shooting, even though it is one funky
looking piece of gear.

Think I was running the tube at 25 or so mA. low DC B+ (125-135 V or so).
Never wrote it down. Can check if interested. The 12B4 is microphonic so do
something tricky about that.

(Bill Perkins has some of these Cryo treated which is claimed to
reduce the microphonics (as well as other improvements))
-----------------------
Gordon Rankin: This is a great tube for preamps. Make sure you use DC heaters as it is
prone to noise from the filaments.

I used 5K Mills resistor on the plate and 20ma. It worked great.

funny thing I looked up the 12B4, and here on one of those Dog a Day
calandar pages (often littered through my house with schematics on
them...) was my 12B4 schematic. Samurai Spirit (the Dog) Thr, July 23
1998 marked final.

I used the jensen 10K:10K input terminated with a 10K P&G RF11 pot, grid
resistor 100 ohm. Bias on the 12B4 was 90V/-10v/20ma, I used 2 Mills
MRB-12 15K in parallel. Bias resistor 499 mills MRB-5, 33uF/16 BGN cap.
Output 3uf/Hovland, 100K shunt to ground. 6V4 power supply, choke pi,
then split cap.

Proabably would be better with a Jensen P4 input 1:2 using a 20K pot.

I also had some scriblles about battery unit, 8-12V cells, 6V for
heaters, 100 ohm bias (-2V), 100H/20ma choke.

I used regulated dc heaters and placed the 12B4 in Yamamoto sockets and
lifted the socket off the base of the chassis with those rubber standoffs
that Percy sells.

In the end you could bash the damn chassis, nothin...
-----------------------
Kurt Steffensen (the mad Viking & responsible for the Arhus triod Festival); 12B4 is one of my favorite tubes from line signals to driver.
And it makes a magnificent CF.
( Are you guys beginning to explore the merits of this wonderfull little circuit ? ;-)

I use DC heaters , though it is probably not really necassary. Depends upon how much noise you accept.

12B4 is a triode that does what it is told. Nothing less , nothing more.
DEAD neutral !

And it is able to reproduce a clean 24 bit dynamic range.....

Don't know how it will behave as an MC amp.. The electrode structure is to dramatic for such small signals , for my taste.
But I do not doubt , that the white noise could be low. I would worry about microphonics , at such low levels , though.

I once used it for output tube as well. Cool :)
It makes a good circlotron , Remco ;-)

And by the way , it is an excellent regulator triode :)

Not a single of my 12B4's are microphonic. But I have heard many say this.
( I may have some 200 hundred pcs., though I havent tried them all ;-)

Mines comes from the same series production ( batch ) made by GE.
But I also have a few Toshiba and hmm is it CBS ? ..
None of these suffers from microphony either....
-----------------------
Bill Perkins: The 12B4 is a low mu, single triode of medium transconductance and rp.
It will *operate* at a plate voltage of 500V and is ideally suited to
applications where large output swings are required, particularly with an
IT.
So if you want to drive an 845, for instance, this device will do that
very well.
-----------------------
Dave Slagle: it is really not that hard to just do it from the ground up....
just kiss (keep it simple.)

build a 300V or so supply (the total voltage can be whatever for the firs
go.. 200-500V)

say we have 300V... OK lets use a 10K plate load resistor.... and want to
run 20ma of current ohms law tells us we will drop 200V leaving us with
100v left for the tube

now go to the curves and see that at 100V and 20ma you will be at -11V bias
or so... so you need to chose a cathode resistor that drops 11V at 20ma
which is 550 ohms. toss a 100mfd cap across it add a output cap of 2 mics
or so... a grid resistor and a output resistor and you are done.

it will make music and sound pretty good... now its up to you to work on
the PS and the operating point to make it sound great.
-----------------------
Steven St.Laurent : What I Did This Week (So Far)..

I've been stuck in the hell that is Sun Microsystems Training (free so i
can't complain too much) for the past week plus, so i've had time to plan
things. During my parole time from work I've managed to "finish" breadboarding
my 12B4 one-tube-wonder. There are some cryo'd units I got from our friend
up north, Bill Perkins.

Basically my current setup is a single tube version of the End of Zen by
Tom Stoppleworth (bottlehead forum has big thread). I junked the input
tube and just run the 12B4. Parafeed (Hammond 157G choke + Angela
3.3uf caps) thru some Hammond 125E's. Input isnt optimum, 1k ohm resistors
on the input, need to play with this some. AC filaments, nothing special.
DEAD SILENT, except for the fact this is within 3 feet of 5-6 computers
and several computer monitors plus a halogen lamp with a noizy dimmer!

Sevendust (noize to you older guys) has incredible bass slam, even for a
poor modern recording. Denial (track 2) is quite impressive. Dave Brubeck
Quartet sounds quite good too, cymbals sound very natural. Squirrel Nut
Zippers really rocked, the bass was quite impressive. Played some ZZtop to
good effect also.

I would say its very neutral, but with more bass. Its not going to make
solid state sound any better, if anything it shows its warts. Its worlds
apart from the sound of my 1626's for certain. Its more like a wire with
some gain and a bass boost.


My only complaint is the gain isnt quite what I would like. I need to
play with it some more. I doubt the tube is seeing a pretty load, coming
off the headphone output from a cd-rom. I've not played with output tranny
loads much either yet.

Any suggestions on improving gain and what to put between the cdplayer and
the tube?
-----------------------
Other comments: Julius Futterman used 12B4s as outputs in some of his early, direct-coupled OTL amplifiers.

A 12b4 OTL can be found at <http://www.infomaniak.ch/~bonavolt/12b4.htm>

The GE's I'm using are during the first ten minutes or so of warming
up. Quiet after that. Running them at 90v/28ma. AC filaments,
biased up to 30v.
 
Just as a point of reference, my present preamp is a choke loaded 12B4a (GE, the RCAs are horribly microphonic), using a 60H/30mA Handmade choke (www.hndme.com), with 25mA/155V across the tube. Using Russian 2.2uF PIO + 0.056uF Russian Teflon output caps. It plays good music for me.

I don't think a Lundahl LL1660 with a differential 12B4a is a good idea. The transformer really can't handle the current levels that a 12B4 likes to operate at. You may be better served by other iron, or go with a simple SE common cathode.

David
 
dmcgown,

i have checked that with Manfed Huber and Per Lundahl. 34 mA should be ok. According to the Lundahl catalogue each winding can take 50mA; the trannie is used in balance mode and with matched tubes.

But as this is design stage and not yet tried out, i may end up with a lower quiescnet current. But not witha nother tube and not with another trannie: i like the trannies sonics and i exactly want the output Z range this combo has: 430 Ohm and cathode resistors can adjust this to exactly 600 Ohm. Which i want to have to drive my constant impedance 600 Ohm gain control.

What i see a bit ticklish: i have the LL1660 PP which does not take much DC imbalance. No risk, no fun. :)

Dave,
thank you for posting all those JoeNet posts. I had lost them.

you now get fed with three Mass of double bock beer and, yes, then i will carry your to your bed. :)
 
12B4-A is a good tube. My comments are based on extended listening to a 300B amp over the Christmas-New Year break. Two versions were made, one using 12B4-A and the other using Raytheon 5842 as input tubes. TJ and WE 300B tubes were used.

Since there were separate power supplies for each tube in the amp, we tried (not exhaustive) several operating (Va, Vg) conditions. Transformer coupling, Capacitor coupling and DC coupling were tried over two full weekends.

In most instances, the 5842 was preferred for its ‘open’ and ‘detailed’ reproduction characteristics. On Carmina Burana – Andre Previn EMI, the 12B4-A clearly had more dynamics to trade.

High transconductance means low noise. 417A/5842 is the clear winner here.

Bernhard: With all due respect to Allen Wright and yourself, I do not prefer differential topologies. The sum and differencing principle always yields less information.

Of the three different couplings tried in the 300B amp;
Transformer coupling enhances the midrange.
Capacitor coupling (Hovland, Ducon Polystyrene, MFD polycarbonate) gives a wider dynamic range but adds a colouration dependant on the capacitor used.
DC coupling had the best dynamic range and transient response

Lowther PM7C in Alfredo Horn was the speaker used in these evaluations.

Brett: What speakers do you use?

Mohan
 
Mohan,

no problem with me, anyone is entilted to his own opinion.
Bernhard: With all due respect to Allen Wright and yourself, I do not prefer differential topologies. The sum and differencing principle always yields less information.
BTW, Allen's and my opinions how a diffential amp should be built differ considerably. And i do not align what i design all too much to is desings. Of course, there are similarities, too.

But i prefer to use the right tube for the right application and i am not afraid like Allen to use signal transformers.

The 5842 is one of my possible choices as phono input tube for exactly the reasons you named. The EC8010 is another one. The 12B4 is my choice as line output tube. I want no DHTs there, so no the 12B4 is just the thing giving me the right impedances and sonics, too.
The 5842 would have way too much gain in this application and ask for another turns ratio. Which would reduce bandwidth. I do not want that. BTW, i fell in love with the LL1660's sonics.

Question: has anyone of you experince with the 12A4?µ=20, G_m=8mA/V, R_p=2k5 ? This tube interests me because it is a it is a one-system envelope suited for matching and has about the parameters i want.
 

Brett

Member
2002-01-07 6:02 pm
Mohan Varkey said:
12B4-A is a good tube. My comments are based on extended listening to a 300B amp over the Christmas-New Year break. Two versions were made, one using 12B4-A and the other using Raytheon 5842 as input tubes. TJ and WE 300B tubes were used.

Since there were separate power supplies for each tube in the amp, we tried (not exhaustive) several operating (Va, Vg) conditions. Transformer coupling, Capacitor coupling and DC coupling were tried over two full weekends.

In most instances, the 5842 was preferred for its ‘open’ and ‘detailed’ reproduction characteristics. On Carmina Burana – Andre Previn EMI, the 12B4-A clearly had more dynamics to trade.

High transconductance means low noise. 417A/5842 is the clear winner here.
Mohan,
With all due repect, I'm not surprised you found there was a difference between the two tubes. As Bernhard has already pointed out (and I'm sure you know) they have very different characteristics. As a typical 300B driver, I doubt the 12B4A would have enough gain to drive it properly from a typical preamp output, and with the difference in gain between the two, it would be impossible to determine definitively the character of each tube, as the circuits for each would also be quite different. Apples and passionfruit.

Bernhard: With all due respect to Allen Wright and yourself, I do not prefer differential topologies. The sum and differencing principle always yields less information.
I strongly disagree. Done correctly, it will yield more information.

Lowther PM7C in Alfredo Horn was the speaker used in these evaluations.
Lowthers have a very distinctive tonal flavour and balance. IMO they would be hard to voice an amp with, without compensating for the speaker.

Brett: What speakers do you use?
Presently my speakers are in flux, as I'm waiting on new elements to be finished. Currently they are a set of modified alnico Klipschorns, with BMS coaxial compression drivers loading a custom flare above about 350Hz. Soon I will have a system very similar to the Edgrhorn Titan , but with LABhorn subs. All front loaded horns (trax flares, hypex bass) 20Hz -22kHz.

Cheers
Brett
 
Mohan,
The sum and differencing principle always yields less information.
forgot to adress this statement.
I am with Brett here.
At Aarhus Triode Audition Festival 2000,
we compared the gorgeous BorderPatrol 300B SET amp (equipped with vintage WE300B) with Allen Wright's preliminary EL34 differential pushpull amp (equipped with noname ECC88 and Svetlana EL34).
Speaker was Peter Bahnsens 3way tractrix horn "Maximator" with passive XO.

As far as sonic detail is concerned, Allen's amp was better, not much but it was better. It was not as beautiful as the BorderPatrol, not as lush, not as seducing.
AFA the low end was concerned, Allen's amp was clearly better and offered more contour, slam, detail, impact. For the record, the Border Patrol had the best low end i heard until then from a low power SET amp.

The most stunning thing with Allen's amp: no weaknesses, equal sonic quality from top to bottom.

Recently i listened to Thomas Mayer's fully differential phono preamp. Thomas managed to get rid of resistors, the thingie was built from many inductors 4 tubes and a few caps... best phono preamp i heard so far. Again, not as seducing as his single ended preamp, but better detail resolution, still very musical, stratoshperic PRaT.

Lowther C series: i better do keep my mouth shut. Our spell checker would paint XXX over the whole statement and i would have to go to the sin bin :) ... for the record, i had a Lowther Acousta with PM6A for ys.
 
Brett & Bernhard,

The preamp used in the listening tests could put out 15V on peaks. I must admit that the listening tests were not made scientifically. This means that not all parameters were the same. They will never be the same if the active devices are different. Whenever we changed components or connections, I used an SPL meter and a 1KHz signal to measure 85dB SPL at a distance of 3.5 metres from the Alfredo Horn.

As to differential topologies, may I hold on to my views until there is an occasion to convince me otherwise. My views are a result of listening to quite a few tube as well as solid state differential pairs over the years. This does not mean that I question your findings. I shall always bear your observations in mind because this forum is for learning from others’ experiences.

I agree that Lowthers have a distinct flavour with most amps. This is because most solid state and tube amps have an upper midrange brightness that complements the Lowther’s upper midrange brightness. I have designed my own dedicated amp for Lowthers.

I have Lowthers at my factory and Stax F83 electrostatics at home. When my amp is driving Lowthers the reproduction is very very similar except for the extended Low Frequency reproduction from the electrostatics.

I envy you guys. You seem to have a lot more gear to listen to. If I had your environment and listening experiences, I probably would repeat your words. Think about it.

Mohan
 
Mohan,

you should not mix up pushpull and differential. A stage with an output transformer can be both, PP and diff. .
The key thing for differntial is that you have two SE devices with their ground or B+ electrodes tied together and connected by a constant current device to either ground or B+. So the total current thru both devices stays rock-solid-constant. You can put what signal you want into the control electrodes of both decives; it is a signal not referenced to ground and the diff. stage only amplifies diffences between its control electrodes. Common mode components are suppressed in the signal. noise, hum, PS ripple, DC from the stage before etc. Common mode signals contribute nothing to the music, they rather spam it, spoil it, deterioate it.

Allens amp was PP and it was differntial even in the power output stage. The cathodes were tied together and connected to ground via a constant current source. TMK, the amp was W/O neg.FB. Huge difference to common PP amps.

Lowther:
if you say your Lowther is sonically very close to the Stax ELS F83, then your particular sample could qualify to judge the amp. Both Stax, the F81 and the F83, belong the the most neutral and to the most musical speakers ever made on this planet and i am not the only one saying so. Provide the amp can make them move their butt, they have incredibel low efficiency.

One of the Lowther "fancies" is the huge manufacturing quality variation. Which not always leads to defective units but provides not a pair of lowthers sounds the same, TME. In general, I found the C series to be brittle, technical, HIFIish, not musical. PM2A and PM6A are great, are musical, but on the long run i experienced permanent listening fatigue with the PM6A. So did my buddy with his PM2A. the PM4A i cannot stand except in a frontloaded horn, then it's gorgeous.

I left the Lowther family, bought a pair of Fertin FLB20EX. At the moment they are mounted in a 70x100cm piece of corrugated cardboard as open baffle. no inner urge to build the wooden open baffle i have drafted in the CAD. Big smile ear2ear evertime music plays. Long term listening fatigue gone completely. Okok, slender lowend if present. :)

More listening experience: well, i am into audio since 18 ys and my ear/brain is trained well and has good sonic memory. Never skipped an occasion to listen to interesting gear, never skipped occasions to listen to unplugged life music.
I worked in hifi shops and Germany was a wealthy country and so most expensive gear came across. No big merit, just came with the years. BTW, most expensive gear i despised.

Maybe you all make it to Munich next autumn? The Triode Audition Festival will take place there and we could meet and have lots of listening experience? :). Good occasion to meet so many audio nuts and master minds in one place.
 
Bernhard,

Thanks for the update on Differential and PP topologies. I have been designing tube amps for over 25 years.

Like Lowthers, the Staxes are single driver (single panel) electrostatics. Both are very musical and I enjoy both. Single Panel electrostatics are the least efficient and that is the penalty we pay for being purists.

PM7 and PM5 are my favourite Lowther drivers. To date I have not been successful at taming the PM2A.

Perhaps, I may have a chance to come over for the Triode Festival. It will be a good opportunity to put a faces to names. Look what the internet is doing to us!

Mohan
 

Brett

Member
2002-01-07 6:02 pm
dice45 said:
you should not mix up pushpull and differential. A stage with an output transformer can be both, PP and diff. .
The key thing for differntial is that you have two SE devices with their ground or B+ electrodes tied together and connected by a constant current device to either ground or B+. So the total current thru both devices stays rock-solid-constant. You can put what signal you want into the control electrodes of both decives; it is a signal not referenced to ground and the diff. stage only amplifies diffences between its control electrodes. Common mode components are suppressed in the signal. noise, hum, PS ripple, DC from the stage before etc. Common mode signals contribute nothing to the music, they rather spam it, spoil it, deterioate it.
Well said. In a well designed and implemented diff amp, there is an amazing reduction of low level noise and grain, even if it doesn't measure much different. There is a 'blackness' that the sound emanates from.

Allens amp was PP and it was differntial even in the power output stage. The cathodes were tied together and connected to ground via a constant current source. TMK, the amp was W/O neg.FB. Huge difference to common PP amps.
Bernhard you said you heard Allens amp over a year ago. Since then he has added Lundahl output transformers (LL1663) and PSU chokes, replacing the Chinese(?) ones he had.

Lowther: ...etc
The only comment I want to make on full range drivers, is they have at least as many design compromises and limitations as multi-ways, and they leave out too much info for my tastes. They are no more 'pure'.
However, I don't mean to be offensive to you guys, I'm glad you enjoy what you have.

Cheers
 
Brett said:
The only comment I want to make on full range drivers, is they have at least as many design compromises and limitations as multi-ways

So far in my experience a good FR ends up being a 3-way (at least a good FR that doesn't cost an arm & a leg -- i understand Dice is keeping quiet about his prosthetic leg :^) and he is still planning on adding bottom.

100-10k from the FR, some help at the top and active help at the bottom...

dave