• 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 PP schematic

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi All,

I have a 6SN7-12B4A PP amp on the breadboard at the moment and I am thinking about committing it to a more permanent form with a second channel (wow, stereo). Would anyone care to suggest alternatives to the scheme attached before I take that step?

Thanks,
Michael
 

Attachments

  • 12b4a amp rev1_3.jpg
    12b4a amp rev1_3.jpg
    27.7 KB · Views: 2,017
I can't see any glaring mistakes...so

I would go right ahead and build the second channel and listen to it for a while....a real while like a month or two.

You can ofcourse change a thing or two...BUT I like your schematic a lot (and choice of tubes). Simple PP with a transformer doing the splitting early on..Wonderfully simple for a PP.

You can always apply the suggestions that may pop up later! Like a CCS or VR shunt reg PSU, grid stoppers...bla, bla bla. or WHATEVER...but they will remain suggestions..

Right now you have a gem in the making as far as I am concerned...

I am as always curious what people will suggest though.

Cheers,
Bas
 
Drive and grid leaks

Drive is a worthy question. I don't know yet. On the breadboard I am taking my musical input from the headphone jacks of an old Sony onebox amp/cd. May be a problem driving it with a CD player? Suggestions welcomed.

I did omit the gridleaks on the 12B4A, they are 470k.

Thanks,
Michael
 
Driving it...

You'll need something that will be happy with that 500ohm (very low) input impedance. Of course a headphone stage will do, or there are some valve preamps with very low output Z (but don't ask me to name one :rolleyes: )
Another drive thought is a semiconductor one :xeye: , my favourite line driver chip, the DRV134 / SSM2142. If you fed line level in, the differential outputs would drive the input transformer with no problem. These chips sound totally transparent to me, and require no external components apart from 2 decoupling caps.
Of course it'd be much more fun with valves:cool:
 
Thanks Bas

I have had it on the breadboard for about a month now with daily listening and changing. I think, stress think, that this is my favorite operating setup.

The power supply is still to be considered, I am running it from an old Heath tube regulated supply at the moment. I am seriously thinking about VR tubes to regulate the 6SN7 supply and I have been meaning to setup a negative supply so I could test larger Rk for better differential balance and common mode rejection.

Also to be decided is the best volume control scheme. I have tried a 1Mohm variable resistor between the two ends of the input transformer and that seems it may work. Still need to do a little work in that area though. Suggestions welcomed.

Originally this was designed and planned as the frontend for another output tube, but the guy hasn't finalized his order yet. In listening to it I liked it enough that I think I will try it on it's own for myself. As a driver I tested it with a 100 ohm load on the secondary of the opt to reflect back ~50k and the distortion is quite low, I think, with a voltage swing of 400V pk-pk.
 
input impedance

John,

A chip line driver is an idea, though if I had it's differential output I wouldn't need the transformer to split phase, correct?

Another omisson from the schematic, I have a 500 ohm resistor in series with the primary of the input transformer to tame some square wave ringing. Also, the transformer is rated as 500R:50k for single ended use, what does it reflect back when used as PP with a 47k (or larger) load on each side?

I know the breadboard loads a 500 ohm output signal generator but in a pretty frequency independant way. My Pioneer DV-563A didn't seem to have any problem driving a 500 ohm input passive line level filter when I inadvertantly designed the lowpass for the subwoofer filter with that impedance. Do you think it would have a problem with this amp?

Thanks,
Michael
 
Where's the driver's PS bypass cap? :eek: Sure, it's balanced, but gee, it'd be 20 times more effective at keeping signal balance if you put the choke in the 6SN7's cathode circuit. As-is, you'll get nastified distortion for small imbalances!

Speaking of balance, the cathode bypass on the 12B4s is unnecessary if they are class A. (If class AB, you'll need a fixed bias supply instead. And I would recommend AB, given the low power handling of this tube.)

Tim
 
PS cap and cathode choke

Hi Tim,

By PS cap do you mean a cap between the 75H choke and the plate resistors? I tried a 25uF polypro/oil there, both as connected to star ground and to the cathode of the 6SN7 above the Rk. Sort of empirical design method. Anyway, 2nd harmonic voltage increased from 16mV to 56mV and 28mV respectively with 8.2V output into a 16 ohm load. Third harmonic voltage increased from 46mV to 173mV and 116mV. Other orders were less affected.

Another schematic omission- I have a 50R pot between Rk and the cathodes on the 6SN7. Balancing it has a large effect on second harmonic. I tried without the 75H choke and with a 15H choke in the cathode load of the 6SN7. Both led to higher distortion, although I may have made mistakes in the implementation. I thought a choke in the cathode should be very effective? I also tried the 15H choke in the cathode load of the 12B4. Same result.

I thought I was running Class A. I get something around 4W before distortion starts climbing. When I inadvertantly connected the 55uF cap to the star ground while rearranging the circuit I had a 300% increase in distortion levels. It took me awhile to figure out what the heck was wrong with my setup.

Here is a distortion profile of the current setup. I don't really know if this is "good" or not, but it is fun to measure while trying to think of other circuit changes to try.

Michael
 

Attachments

  • 2-8 12b4 pp distortion vs power.jpg
    2-8 12b4 pp distortion vs power.jpg
    26.2 KB · Views: 1,435
Michael,

Even with a line driver chip that gives differential outputs, you'll still need the transformer. It's that that gives you lots of voltage gain.

I'm not clear about the transformer's spec. But from what I gather, the I/P Z won't be much more than 1K, which is still very low.
There's nothing wrong with that IMO, you just need to drive it properly.
 
Sch3mat1c said:
Where's the driver's PS bypass cap? :eek: Sure, it's balanced, but gee, it'd be 20 times more effective at keeping signal balance if you put the choke in the 6SN7's cathode circuit. As-is, you'll get nastified distortion for small imbalances!

Speaking of balance, the cathode bypass on the 12B4s is unnecessary if they are class A. (If class AB, you'll need a fixed bias supply instead. And I would recommend AB, given the low power handling of this tube.)

Tim
PS cap? Maybe. I wouldn't call 2H & 4H "nastified distortion"...

Cathode decoupling cap? Well, on amps with lots of NFB this is almost true. On amps with none, and an imperfect balance, the presence or absense of that cap will make a big change to the distortion spectra. It'll affect the "sound", which, I believe is what we're talking about ;)
 
You observed an increase in distortion because 2nd H. is an unbalanced type of distortion. By allowing the PS rail (that the 6SN7 draws from) to supply AC currents (by adding the cap), you allowed those unbalanced currents to proceed. Thus, without it would seem to be preferable, but if a large unbalanced current gets in the system, the plate voltage will utterly bottom out and make a big PHHUT sound (electrically speaking... that was a joke). This is utterly bad design practice; the simple correction (while maintaining the exact same circuit function) is to place the choke in the cathode circuit instead. Since impedances here are much lower, the choke can be as small as <SWAG> 3H and have *better* balance.

The goal of PP is to be balanced, and this means eliminating even harmonics (and class B considered...it does a damned fine job I might add!), so if you need it, go build a PSE amp. Anyway, back to designing... a bypassed output cathode has the same effect as bypassing any other balance point - it'll increase 2nd H by reducing balance. It's better to be unbypassed here.

Tim
 
Run that by me again?

Could you run that by me again, I don't quite understand what you are saying about the plate voltage and large unbalanced currents on the 6SN7. Also, when I removed the bypass or ultrapath cap on the 12B4 the distortion levels increased ~300% according to my notes.

Michael
 
"Crappy filtering" aside......

Sch3mat1c said:
You got crappy filtering or something? Any cathode bypass will increase distortion and gain as compared to none.

Tim
In terms of THD this is often, though not always, the case.
You said yourself that generally, the odd H get cancelled out.
Apart from the obvious effect on degeneration, fitting a bypass cap also reduces Rp, thus reducing distortion.
Additionally, this is not just any cathode bypass, but is coupled to the ideal point, thus bypassing a lot of the PS noise.
 
Slightly off topic, but I've been wondering about trying 12B4s in a common PP EL84 amp - in my case a Leak Stereo 20. I assume that with seperate cathode resistors on the 12B4s I'd use 1.4K each valve and bypass them with 100uF polypropylenes (currently in place). My present topology is a 6EW7 - second beefy section (mu=6) into the first section as concertina (22k+22k), zero NFB. Any thoughts? I assume this gives about 4 watts instead of 10w with the EL84 in UL PP. How it would compare with trioded EL84s I don't know.
 
choke in cathode circuit

I decided to retry a 15H choke in the cathode circuit of the 6SN7 this morning. To reminimize distortion I had to raise the total Rk to 2118 from 740 (both plus the 50 ohm balance pot). This changed the bias voltage to 8.2 from 5.6 volts. There was a significant increase in 2nd at all power levels, a decrease in 3rd at higher power levels and a decrease in 5th until clipping. Any comments on the reasons for these changes?

Haven't listened to it yet because it's too early in the morning.

Michael
 

Attachments

  • 15h choke distortion graph.jpg
    15h choke distortion graph.jpg
    26.4 KB · Views: 672
Re: "Crappy filtering" aside......

dhaen said:

In terms of THD this is often, though not always, the case.

I would be interested in knowing what condition would produce other results.


You said yourself that generally, the odd H get cancelled out.
Apart from the obvious effect on degeneration, fitting a bypass cap also reduces Rp, thus reducing distortion.
Additionally, this is not just any cathode bypass, but is coupled to the ideal point, thus bypassing a lot of the PS noise.

You have the order of function reversed. Distortion will ALWAYS be higher because it applies NFB. You can run the numbers yourself. You can also run the numbers and see that Zo increases, but this will only affect bandwidth.

I'm intrigued by Audiobot's results, can you post a complete updated schematic?

Tim
 
new schematic

Tim,

Here's the schematic as built. May still contain omissions or errors, but I think it's right. I have to say that there is no doubt, distortion increased greatly when I removed the cap on the 12B4 cathode. Why I am unsure.

Adding the choke to the cathodes of the 6SN7 required a major revamp in operating point, as seen from the voltages and current on the new schematic, to reminimize distortion. Any thoughts as to why?

I have listened (quietly as I live in an apt) to a couple of CDs now. Not sure if there is any noticable change, but seldom play at the volume levels where the major changes took place, except the general increase in 2nd that is.

Michael
 

Attachments

  • breadboard as built rev1_4.jpg
    breadboard as built rev1_4.jpg
    33.5 KB · Views: 821
For a fair comparison the operating points should be the exact same. BTW I don't know anyone who designs experimentally based on measured distortion - it's not like you're trying to get an exact figure. Just because distortion is low doesn't mean it's working right. (But that statement just doesn't sound right :apathic: )

The only modification necessary should be to subtract the choke's DCR from the 6SN7 bias. The original shows 802R, while your choke measured 703, so a 100 ohm resistor in series should bring it up good.

Uhm...hmm your 6SN7s are biased so far off, plate is only 211V? Yikes! You get far less distortion when it's down around 1/3 to 1/2 times supply voltage (which in this case is around 250V). No wonder it's so distorted, you *barely* have enough headroom to drive the 12B4s.

Aside from that, I wouldn't doubt it if the plate choke is causing the trouble - I'd replace it with a 1 to 4.7k resistor of adequate wattage and bypass it to ground.

Tim
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.