12B4 Line Stage Amp

[coffeedj]

thanks again coffeedj, let me know if you want some help with your website, i feel like i owe you a favor or two! I noticed quite a few of your links broken and and your linestage design seems to have completley dissapeared. However you may enjoy the diy aspect of html coding.

so, i spent hours trying to fing the CT for the filament heaters, I know its there but i didn't get to it. So i only had a 2watt 100R resistor and i referenced one leg of the ac heater supply to ground. It worked, very minimal hum and i will purchase the 4 neccessary 50R resistors to referece both legs to ground.

OK, so what is roll off? my experience with studio work tells me that it is some kind of curve rolling off the top end/ higher frequencies. I think this could be what i called mushiness, especially seeing as i still have .33uf output cap.

how do i fix roll off? I still have the compnonents you recommended for cathode and anode etc. would this help?

cheers again

stuart

I'd love some help on the website--thanks.

The Center Tap will be found coming out of the transformer at the same place as the 6.3V AC leads come out. Typically it will be the same color, but with a striped band instead of solid color (usually green). 100 ohm is fine--you only need to add two 50 ohm resistors from each lead to the 100 ohm.

Once that is done, play with lead dressing to get hum to zero. Also, these wires now run next to the B+ leads in your cable, so you will need to shield or wrap the AC heater leads in the cable. Tie the shield or wrap to the case at the power supply. Also make sure the B+ final cap is in the amplifier case, not the power supply case. It will eliminate the residual AC hum induced in the B+ run from the Power Supply.

Roll-off refers to the filter response caused by the output Capacitor. In this case there is a one pole filter caused by the combination of the output capacitor with the load resistance. This is called a one-pole butterworth filter, and has the charactaristic of 3 db per octave reduction in the output signal (that is 3db per octave power, voltage would be 6db per octave). Filters can be high pass (this one is), low pass or band pass. We want to pass frequencies over the audio region of 20 - 20KHz without phase distortion, so your 3 db point for a high pass filter needs to be at least two octaves below the lowest audio frequency. At 20Hz that means the 3db point should be below 5Hz. Typically for audio you look for something in the 2-3Hz range, or 3 to 4 octaves down. Each octave is a halving of the frequency. For tube amps, however, I personally feel that 5Hz is sufficient.

To calculate RC filter response at 3db point use the equation: F = 1 / (2 * PI * R * C). Thus for a .22 output cap going to a 10K input load the 3db point is F = 72 Hz. The real value is a little lower, but this is a good enough approximation to work with, and it is terrible for audio. You would loose most of the last octave of a piano. Go to a 2.2uF output and now your 3db point is a 7.2Hz--much, much better. 3uF drops you below 5Hz when other effects are taken into account. 10K should be the maximum value used for output load calculation--regardless of your input impedance at the other end. It is the standard for industry line level input.

 

[barossi]

Hi coffeedj,

finally i found a bunch of 12B4a from Sylvania. They are O.K.

I build your choke loaded line amp. I am using Lundahl plate chokes and i am very happy with the results. I had to match the 12B4a and had to increase the cathode cap from 47uF to ~450uF. It works quit better with my horn system.

With 47uF the bass was boomy and very fat; with 400-500uF it is exactly what i have been searching for.

One last problem remains:

Even with pot turned to zero i hear hissing (braodband noise) coming out of my horns. I changed the Allan Bradley (carbon) grid stopper, 2X120 Ohms connected to 2+7, to 2X1K metal oxid.
This was not the reason for the noise.

😕 😕

KR

Hauke

 

[Zen Mod]

.......
This was not the reason for the noise.

😕 😕

KR

Hauke

what's PSU you are using ?

 

[barossi]

Hi,

i am using a psu with 5R4:



-30H-47uF-500 Ohm-220uF-500OHM-OD3
250-0-250V-4uF-20H-47uF -
-30H-47uF-500 Ohm-220uF-500OHM-OD3


..so i use one 5R4 for both channels, and splitted for right and left after 20H choke and 47uF capacitor. (see above)

KR

Hauke

 

[Zen Mod]

Hi,

i am using .....

Hauke

same noise in both channels ?

did you tried changing OD3s ?

or to put 1n~2n2 across each OD3 ?

 

[coffeedj]

Hi coffeedj,

finally i found a bunch of 12B4a from Sylvania. They are O.K.

I build your choke loaded line amp. I am using Lundahl plate chokes and i am very happy with the results. I had to match the 12B4a and had to increase the cathode cap from 47uF to ~450uF. It works quit better with my horn system.

With 47uF the bass was boomy and very fat; with 400-500uF it is exactly what i have been searching for.

One last problem remains:

Even with pot turned to zero i hear hissing (braodband noise) coming out of my horns. I changed the Allan Bradley (carbon) grid stopper, 2X120 Ohms connected to 2+7, to 2X1K metal oxid.
This was not the reason for the noise.

😕 😕

KR

Hauke

You are right--the input buffer resistors don't affect the noise. You proved that by turning gain to zero--which takes the input resistors out of the equation. No need to swap.



Your noise would normally be coming from one of two places given the information--noisy tube or noisy PSU (including DC heaters if you are using them). If you are using DC heaters, disconnect and go back to AC as a first check. The OD3's should give very clean B+, so I doubt the noise is coming through the B+.

A noisy tube can do it for sure--do you have extra tubes to try?

I'm also a little surprised by the 47uF vs 470uF effect. I suspect it may be more a function of a non-linear bypass cap rather than the actual value. If your mids & highs are attenuated, that is the equivalent of giving boomy bass. This is why I like to add a non-polar bypass to the bypass at about 1 / 100 of the value. One way to check this is to lift the bypass cap completely. That will reduce gain, but it will tell you if your system response is linear. If you hear boomy bass with the bypass lifted, something else is going on.

What is the DC resistance of the Lundahl chokes?

 

[coffeedj]

Hi,

i am using a psu with 5R4:



-30H-47uF-500 Ohm-220uF-500OHM-OD3
250-0-250V-4uF-20H-47uF -
-30H-47uF-500 Ohm-220uF-500OHM-OD3


..so i use one 5R4 for both channels, and splitted for right and left after 20H choke and 47uF capacitor. (see above)

KR

Hauke

Another possibility for boominess is power supply oscillation. Reverse the 5000 ohm and 500 ohm resistors and see if that changes the equation. You really don't want a high impedance from your last ps stage feeding the OD3's. 100 to 500 ohms is about right. The OD3's are your voltage control 'reserve'. They need quick response from the power supply to function at top form. THe 220uF cap is the first line of resupply, so you don't want that pathway constrictied. As it is now your system is unbalanced. The RC for 5K/220uF is much larger than 500/47uF. Reversing these two resistors places these two dowstream RC's within ~8 db of each other and should give better overall performance. I haven't modeled it to prove this assertion, but I think it will be correct.

You may be using Duncan to model, but I find that Duncan is only a starting point. The impulse response often shows a much different picture on a scope than the Duncan program would lead you to believe. If you can look at the scope response of your power supply to a square wave input--that will tell you what is really happening under load. For example, if you remove the 4uF input capacitor, Duncan shows minor instability. But in point of fact, the system will probably function better without that capacitor--you need to test.

 

[barossi]

Hi coffeedj,

thanks for your reply. The noise is on both channels, even with shorted inputs!

It is audible with headphones and of course with my horn system with ~115dB efficiency. (Bi-amping)

Yes, i swapped 12B4a tubes but the noise is still there.
I am using DC heaters; generated by LM317. Each 12B4a has its own DC regulator.
I am using Elna Cerafine caps in the power supply, and Panasonic FC caps (low ESR) as cathode cap, Caddock MP930 resistors in the PSU.

I use the Lundahl LL 1667/20mA anode choke:

http://lundahl.se/pdfs/datash/1667_68.pdf

Inductance ~200H; DC resistance ~2.4K Ohm

Coffeedj, i am using 2X500 Ohm Caddock MP930 resistors (not 5000 Ohm!) in the PSU; which gives a very good slope and no oscillation can be seen in Duncan PSU.

I do not have other OD3 to test. For the moment i am using the Chatham military version OD3W CHATHAM VR150/30.

I will test the following:

I will remove the first cap in the PSU ( only choke loaded!) which will drop the voltage to 150V. I will then remove the regulators (OD3) and see if the hiss/noise remains.

KR

Hauke

 

[coffeedj]

Hi coffeedj,

thanks for your reply. The noise is on both channels, even with shorted inputs!

Yes, i swapped 12B4a tubes but the noise is still there.
I am using DC heaters; generated by LM317. Each 12B4a has its own DC regulator.

Hauke

Both channels leads me to suspect DC heaters again. Solid State regulators are notorious for mucking up sound.

You can just pull the OD3's to see if they are contributing hiss--won't affect anything except to add a little hum (maybe).

 

[bequerel]

same noise in both channels ?

did you tried changing OD3s ?

or to put 1n~2n2 across each OD3 ?

Barossi,
The broadband noise is typical for non bypassed OD3s and similar.
Try a bypass as Zen Mod suggests.🙂

 

[Zen Mod]

Barossi,
The broadband noise is typical for non bypassed OD3s and similar.
Try a bypass as Zen Mod suggests.🙂

when properly executed , I found un-bypassed gass stabs good for even first stage in phono preamp .

anyway - test without them , or bypassing them will reveal where is the culprit .

another nice touch can be re-arranging LM317 as CCS for heating .... and elevating heaters to some pos. potential

 

[barossi]

Hi,

thanks; yes i will try to bypass the OD3's with a 1-2.2nF cap.

Just for my understanding.
How to connect the OD3😕

Sylvania:
http://frank.pocnet.net/sheets/137/0/0B3.pdf
page 2: Pin5 to unreg B+ and Pin2 to ground

Tung-Sol:
http://frank.yueksel.org/sheets/127/0/0D3W.pdf

page 5😛in5 to unreg B+, Pin2 to ground, Pin3+7 bypass in-out

IMHO it should not make any different 🙂

KR

Hauke

 

[Zen Mod]

Hi,

thanks; yes i will try to bypass the OD3's with a 1-2.2nF cap.

Just for my understanding.
How to connect the OD3😕

Sylvania:
http://frank.pocnet.net/sheets/137/0/0B3.pdf
page 2: Pin5 to unreg B+ and Pin2 to ground

Tung-Sol:
http://frank.yueksel.org/sheets/127/0/0D3W.pdf

page 5😛in5 to unreg B+, Pin2 to ground, Pin3+7 bypass in-out

IMHO it should not make any different 🙂

KR

Hauke

bypass pins are there - to implement them as path from reg point to load ...... in case that some smarta$$ pull reg tube out , with powered gadget ........ so - no fireworks if you're smart to use 3&7

ask me how I discovered that trick ?



(tip - ZM smarta$$)

 

[barossi]

Hi,

so, after a couple of testing szenarious; no news; the 12B4A Line Amp the noise (hissing like a waterfall) is still present and audible with my horn systems.
I did the following:

-bypass OD3 with 1,8nF cap's --no change
-change heater from DC --> AC heating with 6,3V --- no change
- remove first cap in PSU to lower B+ and remove OD3's --- no change
-swap some caps in PSU (bigger, smaller) --- no change
-change tubes; 12B4A; 5R4G, OD3's -- no change
-finally i use my high voltage power supply (Hewlett Packard) and plugged it to first cap --- no change

Any ideas guys!?

KR Hauke

 

[Zen Mod]

@barossi - tried cap from Ua to cathode ?

you didn't tell what's sensitivity of your amps ?
or I didn't saw that ...

what's sensitivity of your horns ?

 

[barossi]

@barossi - tried cap from Ua to cathode ?

you didn't tell what's sensitivity of your amps ?
or I didn't saw that ...

what's sensitivity of your horns ?

Hi,

what do you mean cap from Ua to cathode?

I will use the line amp as a smart driver for my bi-amping Horn system. Solid state amp for the 15" woofers; Single Ended 45 (Bugle) for the 2" Radian drivers and JMLC Horns from Musique Concrete.

..on my test bench i am using headphones; and of course it is very good audible with them ...

KR

Hauke

 

[Zen Mod]

Hi,

what do you mean cap from Ua to cathode?

......

to inject some PSU noise from PSU , hopefully in antiphase with present noise at anode;

at least you'll see is noise going to decrease .

start with 1uF upwards and inform us what's happening

 

[Salas]

Any ideas guys!?

KR Hauke

You have a very sensitive for SPL/W system and 12B4A is not the best for hiss. Has big cathodes and heaters. Your best bet is to keep the amplification factor low IMHO. Dynamic loading (CCS, Gyrator) or choke pushes the mu. Use resistor loading for the less mu and hiss possible, but expect more second harmonic, i.e. classic softer sound with less fine analysis.

 

[coffeedj]

Hiss with 12B4

You have a very sensitive for SPL/W system and 12B4A is not the best for hiss. Has big cathodes and heaters. Your best bet is to keep the amplification factor low IMHO. Dynamic loading (CCS, Gyrator) or choke pushes the mu. Use resistor loading for the less mu and hiss possible, but expect more second harmonic, i.e. classic softer sound with less fine analysis.

I have the Terasonic speakers which are also extremely sensitive--104 dB. At full volume, with my ear 12" from the speaker, I cannot hear any hiss or hum--absolutely dead quiet. Same is true if I plug high impedance headphones directly into my 12B4 pre-amp. Amplification with 12B4 is already low--even with choke load. Using a resistor load won't change hiss by much, if the hiss is tube noise. You will drop only a few dB.

I have tried several 12B4's and I do not have hiss problems with any of them either. I did, however, have a bad 0D3 that caused hissing. However, I'm not sure I can put my finger on the problem. There can be a huge difference between tubes. In addition, heater voltage can really affect hiss as well. I'm running my 12B4's slightly starved--about 5.5V DC. I've noticed with 6CG7's that increasing heater voltage from 6.0 to 7.0 can multiply hiss by 4X or even 8X. I've also noticed that by selecting 6CG7's I can cut noise by a factor of 4. Based on the posts from some of the other members, it appears that there is a lot variability with 12B4's. I may just be lucky with a good batch. Then main thing I have noted is variability with microphonics.

But, with the huge hiss problem noted--one has to suspect something else is going on. The debugging done seems to eliminate the B+ and LT as potential sources of hiss. All that is left is the 12B4 circuit. Grids shorted directly to ground at the tube still hisses so that takes out the grid circuit. Choke fed eliminates the possibility of noisy load resistor, so unless the Lundahl chokes sing, that is not the problem. Output caps can leak--causing hiss, but I assume this has already been checked. Cathode circuit shouldn't audibly affect broadband hiss. All that is left is the tube itself and the input connection to the next stage amp.

Have I missed anything? The only other thing I can think of is external noise. I once spent hours trying to find a hiss source on my bench, and then realized it was coming from my lab light shining into the unshielded component area.

I confess I'm stumped at the moment.

 

[Salas]

I use about 6V with 12B4 and the hiss lowers. Starved heater works. I agree.

P.S. If your speaker uses a Lowther DX3, 104dBSPL/W is highly optimistic especially after 15kHz. DX3 is quoted 99.5dB alone.
Horns have completely different projection and if the tubes aren't the best or something more has to be done for hiss still, they can be very well accentuating.