12B4 Line Stage Amp

[coffeedj]

12B4 GC design

I've also posted a schematic of the psu, although i'm sure it is all kindergarden to you guys..

thanks again, stuart

p.s. i hope i'm not treading on toes posting the linestage diagram, i believe i found it on this site.

One problem with this design--output cap is too small by factor of 10. SHould be 3.3uF , not .33uF.

 

[coffeedj]

Just finished the 12B4 linestage.

I wired up the input harness and cathode measurement jacks and turned it on. Only one slight mistake with my cathode ground-- quickly corrected.

Then plugged in my Sennheiser 650 headphones direct to the output (buffered through a 330 ohm resistor to match the 630 ohm output). Full power with inputs turned down--can't even tell the pre-amp is on. It is that quiet. Then I plugged in my CD and listened to Schlomo Mintz.

Oh my oh my!! The sound is absolutely stunning. Smooth, clear, articulate, extended bass, accurate highs. Wow! This is much better than any 12AU7 or 6CG7 design I've put together.

Had to push the boundary and pulled out my Gina 3D to play back Al Demola recorded from Vinyl at 96KHz 32 bits from my turntable. CD's just don't cut it when you want to test the ability of the equipment. Sound Was Fantastic--absolutely the best I've heard. All that 12AU7 thinness is gone, and the subtle things like hammer ons at pppianisimo are crystal clear. Sounds like a pure passive pre-amp, but with life. And this just with my work bench tubes.

I'll post full design on the website tomorrow. The power supply is definately part of the magic for this one. You may have to scrounge to find equivalent parts--I have a large stock of 50's and '60's parts. (If you want to copy). Don't expect beauty--I build too many to worry beyond functionality.

FYI--the 12B4's drive the Sennheisers way past the point of pain with no loss of clarity or distortion.

One final note: I'm running the tubes a little heater starved (5.9v). That improves the linearity according to my testing. YMMV, but I'm sold on 12B4 as a preamp.

You will notice in the pictures two extra 12AX7 tubes--that is where the phono section will go when I get to it.

 

[barossi]

Re: Just finished the 12B4 linestage.

Then plugged in my Sennheiser 650 headphones direct to the output (buffered through a 330 ohm resistor to match the 630 ohm output). Full power with inputs turned down--can't even tell the pre-amp is on. It is that quiet.

Hi,

i could not understand this. I build mine according to SALAS advices incl.:

-3.3k CC resistors on both pins 2&7
-CCS set to 18mA
-heater 12,2V

i could not stand the microphonic effects. I could even hear my voice in my speakers. During heating up phase (first ~5min) the 12B4A is very,very noisy.

I have now 4 different brands of 12B4A on stock, and i ordered 16 pieces Sylvanias to get rid of the f...cking noise.


KR

Hauke

 

[Salas]

With a CCS almost Mu gain is reached (6.5). THD drops like a stone too. With passive loading, I have measured G=3.56 (Rk bypassed in both cases). A 6k8 10W anode load, and 1k2 capacitor bypassed Rk keeps it exactly at your now bias point for voltage and current but will halve the gain as above. That will halve the manifestation of microphonic effects too. 12B4A is not a tube immune to external vibration. Quiet it is for hiss due to low gain, and no buzz can be achieved with a good PSU. Heroic passive or regulated. Best way to mount against vibration from environment or transformer mechanical hum (if PSU is on same box) is by decoupling elastically some sub plate for them sockets.

 

[mach1]

One problem with this design--output cap is too small by factor of 10. SHould be 3.3uF , not .33uF.

Isn't the optimum cap size a function of the input impedance of the power amp?

 

[coffeedj]

Re: Re: Just finished the 12B4 linestage.

i could not understand this. I build mine according to SALAS advices incl.:

-3.3k CC resistors on both pins 2&7
-CCS set to 18mA
-heater 12,2V

i could not stand the microphonic effects. I could even hear my voice in my speakers. During heating up phase (first ~5min) the 12B4A is very,very noisy.
Hauke

Most likely bad tubes. I once had a series of prized '50s 6SN7s that I sold as part of a pre-amp and had to junk about 40% of them after the buyer experienced the same problems that you described.

Everyone talks about the microphonic 12B4's, but I've only had a very small percentage of my stock exhibit that problem. My tubes are all from the '60s--I don't know if that makes a difference.

 

[coffeedj]

With a CCS almost Mu gain is reached (6.5). THD drops like a stone too. .

I noticed a huge decrease in THD with CCS also, but with a choke load I got better results in the third harmonic--as expected from a non-feedback device. Gain is almost mu as well using choke load.

The two are probably very close in performance, but I prefer chokes for the simple reason that I muck around in the internals all the time--much less likely to damage a choke if I short it. More expensive though if you want to get good quality. Electroprint makes good ones to spec. I get mine from Hong Kong.

 

[Salas]

I proposed an electronic choke too, using cascoded Mosfets. This sounds open, more elastic than a CCS. Next best thing to a real one, and of course costs nothing compared to a quality load choke in the hundreds of Henrie.

 

[Salectric]

I haven't had any noise problems with my two 12B4 linestages over the past 3 years. I have tried a lot of different tubes, mostly RCA and GE purchased on Ebay, and I can only think of one or two that have had any hiss or microphonics.

I wonder if the folks complaining about noise and/or microphonics are using AC on the heaters. I can't recall for sure, but I think I tried AC in the beginning and had some problems with it. The DC supplies have been very quiet.

Incidentally, I take exception to the need for a large gridstopper like 3K. I started with 1K and gradually reduced it to my present 100R which works fine. I did try eliminating it completely and got a loud noise when the volume control was turned down to zero. The gridstoppers prevent that. I am not sure why.

You also do not need to use two resistors, one to each grid connection, as is sometimes stated. I have tried it both ways and a single resistor connected to just one of the grid pins works just fine.

The particular type of resistor used as a gridstopper makes a big difference in sound. I have tried half a dozen types and the best sounding by far (to my ears) is a nude Vishay.

I totally agree that the 12B4 is an ideal tube for a linestage. I experimented with a lot of linestage circuits over the years before settling on the 12B4. As far as I'm concerned, this is my "final" preamp, one that should see me through the rest of my days.

Dave

 

[staggerlee]

I wonder if the folks complaining about noise and/or microphonics are using AC on the heaters.

FWIW, I have no such problems with 12.6v AC heaters. I heard some 'jingling' when i was prototyping, but I have the tubes mounted on a rubber suspension on my real build. I have O ring style dampers on the tubes too, but they didn't improve anything, as there wasn't a problem to begin with.

 

[Salas]

I wonder if the folks complaining about noise and/or microphonics are using AC on the heaters. I can't recall for sure, but I think I tried AC in the beginning and had some problems with it. The DC supplies have been very quiet.

Incidentally, I take exception to the need for a large gridstopper like 3K.
Dave

A pattern emerges from the conflicting experiences so far that confirms the existence of bad and good 12B4As in circulation. 12.2V heaters DC regulated, RCA, Sylvania, many of them. Not one could stop oscillating with less than 3.3k stoppers. Not one would not react to external vibration audibly. Not as much as in Barossi's case though. Could not pick up air born vibration, only impact, short echo. Also no noises during warm up. He is the most unlucky as it seems.

 

[mach1]

My findings on ac heating are the same as staggerlee, and my tubes aren't even rubber mounted. Sure you can hear some microphony when you tap them, but it doesn't appear to be an issue when listening to music.

Modern GE JAN are probably the least microphonic, but they sound like c**p.

 

[coffeedj]

Final results and pictures published for choke fed 12B4 Line Stage

Go to

http://www.galaxysetlabs.com/Designs/Designs_Top.htm

and click on line stages. You will see the 12B4 listed.

Website is only working in a few pages at this point--mainly in the Designs section. Hope to get the rest working this month.

 

[barossi]

My findings on ac heating are the same as staggerlee, and my tubes aren't even rubber mounted. Sure you can hear some microphony when you tap them, but it doesn't appear to be an issue when listening to music.

Modern GE JAN are probably the least microphonic, but they sound like c**p.

Hi,

i have 10 JAN 12B4a produced 6/87 on stock. In terms of microphonic they are horrible :-(

KR

Hauke

 

[mach1]

i have 10 JAN 12B4a produced 6/87 on stock. In terms of microphonic they are horrible :-(

There could be inconsistency between batches - I bought a few sleeves of 80's manufacture (all the same date) and they are definitely quietest on the tap test.

Do you have any others you have compared them with ?

 

[blaaberg]

Hey coffeedj

Thanks for the big work that you have been done javascript:smilie('')

Is this the final schematics?:
http://www.galaxysetlabs.com/Designs/12B4ChokeFed.pdf

I have a great trafo from Amplimo with sec. 300V, 6.2V, 10V and 20V, can I use this in your schematics, and can you deliver the chokes for the PSU?

cheers, Jan Jensen

 

[coffeedj]

Hey coffeedj

Thanks for the big work that you have been done javascript:smilie('')

Is this the final schematics?:
http://www.galaxysetlabs.com/Designs/12B4ChokeFed.pdf

I have a great trafo from Amplimo with sec. 300V, 6.2V, 10V and 20V, can I use this in your schematics, and can you deliver the chokes for the PSU?

cheers, Jan Jensen

Yes -- that is the final design. Your Tranny sounds great--you can use the 10V tap for the DC heaters and the 6.2V tap for a 6AX5GT rectifier. Voltage drop is a little more, but you have plenty of headroom and the current is well within 6AX5 limits. Your dropper to the OD3 tubes will need to be sized appropriately. If you want to be really clever you could use the 20V to make a B- supply for the grid, but I have no idea how that would sound and you have to add an input cap.

Don't have spare PSU chokes--my chokes were salvaged from old 40's and '50's B&W TV's. I use flyback transformers and clip the secondaries. These things are built like tanks and have great filter performance. http://www.tubesandmore.com/ has a Hammond choke part #P-T193C that should do the trick in pairs, but the dropping resistance is less than my design. Alternatively, I'm putting together a design using .440H chokes in a triple and quad PI filter design that should be awesome. I use these for guitar amps because they have incredible speed; DCR is 16 ohms. I'm going to get some boards made to standardize the placement and methodology.

I do have plate chokes and tubes. I'll activate my blog, so feel free to use.

 

[Salectric]

Salas said: <<A pattern emerges from the conflicting experiences so far that confirms the existence of bad and good 12B4As in circulation.>>

I'm not sure what pattern you're referring to. I just see a lot of inconsistent results. FWIW, I question your repeated statement about a 3.3K gridstopper being necessary to prevent oscillation. While I don't doubt that you are reporting your experience correctly, I think your problem must be due to a layout issue or some other anomaly for the simple reason that others including myself have not had this problem.

Between my son and myself, we have built 4 linestages using a simple 12B4 circuit----one with a plate resistor and 3 with a CCS for plate loading. We have tried regulated B+ (both ss and VR tube) and unregulated. All with unregulated DC heater supplies. Between us we have probably tried 40 to 50 different 12B4 tubes none of which have been tested or matched in any way. We have tried gridstoppers ranging from 100 to 1K. NONE of these preamps in any configuration oscillated at any time. The only problem whatsoever was with my first preamp when I tried eliminating the gridstopper completely. However, even that may be a layout issue because my son's preamp doesn't have any gridstopper and his is dead quiet.

Food for thought.

Dave

Dave

 

[Salas]

Yes -- that is the final design.

That is a ''heroic'' 12B4 build and it must be sounding it! Congratulations.

 

[Salas]

I'm not sure what pattern you're referring to. I just see a lot of inconsistent results. FWIW, I question your repeated statement about a 3.3K gridstopper being necessary to prevent oscillation. While I don't doubt that you are reporting your experience correctly, I think your problem must be due to a layout issue or some other anomaly for the simple reason that others including myself have not had this problem.


Dave

Its 99% the tubes. Way too much head scratching and way too many iterations have gone in to it. 2 at work as we speak. Must have bought from a bad batch. For the simple reason that Barossi has the same issue too (but stronger).

EDIT: On some Tungsol ones in the stash there is microphony as much as Barossi gets. He has built it with different ready made depletion Mosfet cascode anode CCS, using tight, tried and tested on many applications layout on PCB, mine was enhancement mode on perfboard. No single 12B4 among 30, including RCA, Sylvania, Tungsol pairs, and some Philips singles, has zero reaction into knocking the box its housed in. That stays with resistor anode load too, only lesser as the gain goes down vs active loading. As for how large a grid stopper I have a suspicion that it may be related to the regulated PSU HF coupling level between channels. What I report is with regulated common PSU, valve or Maida. But then again if all those tubes in the particular stash gotten from ebay are original rejects, there will be some trouble anyway.