• 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.

EL34 schematic confusion

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Originally posted by: Dominique_free
A word about the Marshall/Fender phase inverter in Waine's post.
A such circuit, used among other by Marshall and Fender in guitar amplifier is an heresy for me in a such amplifier.

In order to get a good sound in a guitar output stage when you are playing at high level, you must drive this stage with enough power, 1 or 2 watt per tube for a 6L6.
If you look at the Ig1=f(Ua) characteristic, a 6L6 can take some current on the grid. The driver must be able to drive those grids (at least 2 for a PP class B)

I agree. I posted that link just for some info on LTP's and such. Anyhow I don't think I would use a 6SL7 in a LTP config, there are much better tubes to use in that position. Using a 6SN7 LTP in a Hi-Fi amp would be more than enough drive. I would suggest class AB1 with as much class A operation as possible. Less power but if more power is needed (wanted) biased more towards AB. Class B would be undesirable in a Hi-Fi setup IMHO. :dead:

As promised I'm posting a 6DJ8/6922 input and a 6SN7 LTP. Very similar to the Eico circuit(s). Jax has a 6J5/6SN7 driver-amp posted somewhere here on this forum. Very similar to the one I'm posting here.

I've simmed this and there is plenty of drive here for an PP EL34 or 6L6GC etc.

Cheers
Wayne
 

Attachments

  • 6922wc_6sn7_afamp_ltp_web.png
    6922wc_6sn7_afamp_ltp_web.png
    21.7 KB · Views: 2,568
cogsncogs said:

I like the (Eico) AF Amp/LTP splitter configuration. Since you have an octal there (6SN7) why not use another octal like a 6SL7? shhh ;) , it's closer to a 5751 than a 12AX7. That's what I would do but hey it's your amp!
And yes I have a "bunch" of schematics and circuits I've simmed, I'd be more than happy to post a few or email them to you...

Happy DIY'ing
Wayne :B :D

my private email is navinadv@yahoo.com. pls feel free to send schematics there.

the reason i am thinking of using the ECC88C as the driver is that the gain of the ECC88C allows me to use it as a high pass or low pass filter. this simplfies the circuit. and a simpler circuit that requires fewer parts should sound better.

so in the present design the ECC88C filter/driver will be follwoed by the 6SN7 and then the EL34. I know that the ECC88C is a 9 pin whilst the others are 8pin. I dont know if there is a 8 pin tube that can be used a filter driver in this sort of circuit.
 
A possible arrangement for the input is to use a double triode in cascode as the ECC88 at the place of the EF86. The advantage is at you would get less noise and still have the same high output impedance and high bandwitch as with the pentode.

a simpler circuit that requires fewer parts should sound better.

It is true if it can do the job. And you have to pay attention at the bandwitch. If you take the quad II amplifier and quad 22 preamp, it was a relatively cheap system and it is very good sounding, but in the bass frequencies, it have not much dynamic and a sound that is not clean. The raison is at good condensators are expensives, and in order to get a relatively cheap price, quad was using small condensators and the feedback to get the bandwitch in the bass.

If you have such an amplifier, it is worth to recalculate the bandwitch of every stage, from the input of the preamp to the output.

With good quality condensators, a grid circuit with 1meg between the grid and the ground can have a 470nF bypass condensator coming from the anode of the preceding stage. The decoupling condensator in the cathode circuit must be big enough too.
All that in order to have a good bass respons without feedback. The sound will become more natural, with more dynamic, and it will stil be natural with a lot of dynamic even with the feedback.
 
At duncan amps, you have a free program, the Tone Stack Calculator that can help you design a such filter. It run on windows and on linux with wine.

http://www.duncanamps.com/tsc/index.html

You can adjust the value of the components, Zrc must math the output impedance of the E88CC. I think at the best tone circuit for a hifi amp in this program would be the James. It is a Baxandall without feedback.

For a true baxandall, this page explain the bases: http://www.20sim.com/webhelp4/library/iconic_diagrams/Electric/Examples/Baxandall.htm
Here is a good reading too, with some exemples:
http://www.webace.com.au/~electron/tubes/preamp.html
And here you will found the math:
http://www.headwize.com/projects/showproj.php?file=equal_prj.htm
They are using opamp, but the math are the same, the - input is the grid and the + input is the cathode.
 
@navin

A dumb question: When you mean:

the reason i am thinking of using the ECC88C as the driver is that the gain of the ECC88C allows me to use it as a high pass or low pass filter. this simplfies the circuit. and a simpler circuit that requires fewer parts should sound better.

Are you thinking of a sort of Integrated Amp with tone controls i.e. bass and treble boost/cut? If so I can help you out there also.
And the simpler the better is very true in that case also. Done passively would be best in 'most' cases. And if you "go" that way, a switch which will take the 'filters' out of the circuit would be desirable IMO. Even better to switch out the lo and hi separately.

What will your input sources be? What will your listening levels be? Speakers/sensitivity if known? This is a Hi-Fi setup, correct? I'd hate to be going down the wrong road, a lot of posters refer to geetar amps here! I don't have a lot of geetar amp experience but I can go down that road also.
Dominique_free's suggestion of cascode operation of an ECC88/6922 is a good idea. Maybe even (waiting for 'gasps' and boos, putting on 'flame suit', :hot: hehe!) a JFET at the bottom position of the cascode stage... :eek:
µ (Mu) stage anyone? There are many possibilities here folks.

I'll work on and look for schematics on my computer for you. I have hundreds. ;)
Also you may want to download Duncan's Power Supply Designer.

www.duncanamps.com/psud2/index.html

Cheers
Wayne :B
 
I don't mix tubes and semiconductors. The raison is at a tube, at least a good one, have less noise as a transistor in practice. Tube are electrons moving in vaccum when solid components are electrons moving in materia. That implie at a tube will introduce less noise.

A big source of noise in a tube is the filament, but it is easy to cancel with the use of a DC alimentation for the filament.
Another big source of noise is the noise in the emission on the cathode and near the cathode. A DC polarisation of the filament in regard to the cathode will remove a very big part of this noise. It is possible to use a baterry, an existing cathode polarisation or a divider to archieve that polarisation.

Old tubes amplifiers don't have print. For a madern amplifier, if it is easier to use other technics for a prototype, a good print will allow a gain of about 10dB on the noise figure of the final version against the prototype.

One component I avoird in tubes design is the zener diode. You want noise, just use it!
It was a few years ago, I was building a stabilised power supply 250-500V. It was very noisy. At the end, I proved to change the zener with a cheap neon bulb as the ones used in many electic cookers as indicator lamps. The result was a 10 time improvment of the noise tension (20dB) at the output.
The only problem with a such bulb is at you get a diferent tension each time you power the system. so I changed it for an OC3. The stability is much better and the noise is unchanged.

Is it thrue for the diode too. Here is a simulation where we can see the combination of the current of each diode in the valve (GZ34) for a double alternance power supply. The result on the condensator is almost a sinus. Very easy to cancel out. The other noises of the valve as the thermal noise are much easier to cancel out as the noise of a solid state diode.
 

Attachments

  • gz34uoio.png
    gz34uoio.png
    83 KB · Views: 1,286
The raison is at a tube, at least a good one, have less noise as a transistor in practice.

That turns out not to be the case.

Tube are electrons moving in vaccum when solid components are electrons moving in materia. That implie at a tube will introduce less noise.

That does not comport with solid state physics. Think about it, what has more noise, a tube or a piece of silver? Yet, the silver is MUCH denser, much more material than even silicon...
 
Hi Navin,

In the beginning you said that too much of a good thing can confuse you. At present you have had such a wealth of good advice that you must really be confused by now!

Seriously I am not altogether convinced that I should add more at this late stage since you are on the right track. But do allow me to perhaps point out from experience what to rather avoid to keep clear of problems.

The ECC83 (12AX7) is a good audio tube, but stay clear of using it for a power amplifier input stage, especially if you want to keep the input independent of pre-amps. It is a hi-mu tube and Miller capacitance can adversely affect feedback stability if not fed from a low impedance source. There are better alternatives aplenty. (In fact, avoid using ECC83 at all in power amplifiers.)

As said by others, stay away from a concertina (split-load) phase splitter for directly driving any but the most sensitive power tubes (e.g. EL84). My preference of everything mentioned would be the old Leak configuration with pentode input, Schmitt phase inverter and feeding an UL output circuit (or triode if you are so inclined, but UL is close to triode in its main characteristics but capable of higher output). Again (regular readers will be getting tired of this) I would NOT use straight pentodes as output tubes in high quality amplifiers. To summarise, their output impedance is too high and to try to restore an acceptable damping factor by a lot of feedback is looking for trouble, although it is done. (Matti Otala wrote a revealing analysis of this decades ago.)

Folks make a lot about the extra noise of a pentode over a triode as input stage; I have not found that objectionable over dozens of circuits over the years and it is most input impedance insensitive. But a medium mu triode if you must (e.g. EF86 triode strapped or ECC88).

Enough for now - have fun!
 
Ex-Moderator
Joined 2004
About the 'sound' of the 6SL7, it's parameters are closer to a 5751 than a 12AX7 but it sounds like neither!
It sounds pretty good! I use it as the LTP splitter input stage in my amp and I'm happy with it.

As Johan Potgieter mentions, high-mu triodes can cause problems due to their Miller capacitance. The 6SL7 is certainly a potential culprit in this regard but, if the impedance feeding it is no more than 100k or so, it should be OK. I don't have a problem on the NFB stability issue, because in my case it's outside the NFB loop.
 
cogsncogs said:

Are you thinking of a sort of Integrated Amp with tone controls i.e. bass and treble boost/cut?

What will your input sources be? What will your listening levels be? Speakers/sensitivity if known? This is a Hi-Fi setup, correct?

Dominique_free's suggestion of cascode operation of an ECC88/6922 is a good idea. Maybe even (waiting for 'gasps' and boos, putting on 'flame suit', :hot: hehe!) a JFET at the bottom position of the cascode stage... :eek:
µ (Mu) stage anyone? There are many possibilities here folks.

I'll work on and look for schematics on my computer for you. I have hundreds. ;)
Also you may want to download Duncan's Power Supply Designer.

www.duncanamps.com/psud2/index.html

Cheers
Wayne :B

the speakers are:
tweeter Scan Speak 2905 9900, 90-91db
woofer Scan Speak 18W8546, 87-88db (push push)

input sources are a CD player, Ipod, and CD recorder.

my room is 300 sq. ft. and i normally dont listen too loud. the music is mostly rock from the 60s and 70s and a bit from more recent times and some jazz, classical and blues.

i hoped to have 4 amps.

2 EL34 PP AB1 triodes 15W rms and 2 EL34 PP AB1 Pentodes 30W rms.

The triode amps would have a high pass at 2500hz and the pentodes would have a low pass at 2500hz.

all 4 amps would have a common Power transformer.

the config would be as follows:
5 transformers with the center one being the power tranformer
flanked by the Pentode OPTs and the triode OPTs on the outside.

in front of each OPT would be 2 EL34 (lilke the PP-1C) and in front of the EL34 would be the ECC88C (filter/driver) and 6SN7 (phase splitter).

the chassis would be 22-24" wide to accomodate all this.
 
Johan Potgieter said:


In the beginning you said that too much of a good thing can confuse you. At present you have had such a wealth of good advice that you must really be confused by now!

The ECC83 (12AX7) is a good audio tube, but stay clear of using it for a power amplifier input stage...

...My preference of everything mentioned would be the old Leak configuration with pentode input, Schmitt phase inverter and feeding an UL output circuit (or triode if you are so inclined, but UL is close to triode in its main characteristics but capable of higher output).

Folks make a lot about the extra noise of a pentode over a triode as input stage; I have not found that objectionable over dozens of circuits over the years and it is most input impedance insensitive. But a medium mu triode if you must (e.g. EF86 triode strapped or ECC88).

yup i feel like a character on a old TV show called Vinny Barbarino (from Welcome Back Kotter). In short the more i read the more stupid i feel.

I am thinking of using a ECC88C in the input. I have no idea what a Schimtt phase inverter is so a ckt using it would be welcome.

the leak ckt i got has a EF86 mated to the ECC81 drivng the EL34.
 
It is no need to use a pentode or a cascode for the input stage. Cascode is great when you need a great sensibility, broadband and little noise, but it is not the case here, at least for the sensibility.

And it is another problem, a pentode like a cascode have a too high output impedance to be able to drive the filter. I would prefere a medium mu triode with a so low output impedance as possible. An E88CC would be a good choice.

For the noise of a triode against a transistor, I persist. It is even for me and in class A, the only one advantage a vaccum tube can have against a FET. The difference is not big, and in both case you must have very good quality component, a very good implementation of the circuit, and a very good print and mechanical design.

If I look only at the moving electrons, electrons in vaccum give less noise as electrons in materia jumping from an atom to the next.

But a tube, as a transistor, are not only moving electrons. You can have impurities in the vaccum. Here you can do nothing, just buy a better valve. You can have a DC power for the filament and polarise the filament against the cathode to concentrate the free elctrons near the cathode. So, you will have much less irregularities in the electon flow.

With the grid, it is a voltage point where the electrons which leave the grid and cross the bypass condenser towards VCC are equal in a number to the electrons returning in the grid from the polarization resistance to take part in the flow of the tube. If you polarise the tube at that point, you get ride of the most of the noise from the grid.

This point is easy to find. You must mount the tube with the same VCC and the same Ra as in the circuit, the cathode at the ground, a high impedance voltmeter (but not too high, between 5 to 10 Mohms) at the place of the grid resistance. The tension on the voltmeter is this point. It can differ with different brands of the same tube.

After, noise can come from the supply. It is even folk doing 2 times each circuit, but in
opposition of phase, in order to have only a constant DC current coming out each node of the supply.
 
If I look only at the moving electrons, electrons in vaccum give less noise as electrons in materia jumping from an atom to the next.

Not only is this not borne out in measurments of voltage noise density, it is a grossly incorrect view of how conduction in solids takes place. Where lowest noise counts, as in scientific instrumentation, solid state is used every time.

There are legitimate reasons for the use of vacuum tubes in a phono stage. This just isn't one of them.
 
navin said:


yup i feel like a character on a old TV show called Vinny Barbarino (from Welcome Back Kotter). In short the more i read the more stupid i feel.

Wrong feeling, Navin. You might feel uninformed (for the moment), not stupid. There is/are probably something(s) you know more about than the rest of us.

A Schmitt phase inverter is just the first user's name for the long-tailed pair. Top triode fed, cathodes connected together and earthed through a large resistor, bottom triode G1 earthed through a capacitor. (Sorry that I am unable to append a diagram now). This is what the 6SN7/other twin-triode is doing in your Leak circuit.
 
Official Court Jester
Joined 2003
Paid Member
VTL ?

in fact-who knows what origin is........

both halves of ECC81 (12AT7) as input
again 12AT7 (or 12BH7) as LTP
and preferred tubes in PP with fixed bias

highish Ua
lowish Ia

tried ,tested

20W in triode
30W in UL

with my OPTs almost same sound in both modes

click on bellow www button,and look for cinema monoblocks to see some pictures (and-no, you can't see that "VTL" schematics there)
 
How about this idea....
take Steve Bench's filter using 6DJ8 (ECC88C) then driving the 6SN7 in the HIF 87 ckt and the 6SN7 feeds the EL34.

this would really involve only 2 ECC88C one for left and one for right channel.

each ECC88C would drive 2 6SN7 and each 6SN7 would drive 2 EL34 PP. There would be no need for the ECC83.

What do you guys think of the EICO HIF 87 ckt and this idea?

Dominique_free said:
It is no need to use a pentode or a cascode for the input stage. Cascode is great when you need a great sensibility, broadband and little noise, but it is not the case here, at least for the sensibility.

And it is another problem, a pentode like a cascode have a too high output impedance to be able to drive the filter. I would prefere a medium mu triode with a so low output impedance as possible. An E88CC would be a good choice.


Johan Potgieter said:

A Schmitt phase inverter is just the first user's name for the long-tailed pair. Top triode fed, cathodes connected together and earthed through a large resistor, bottom triode G1 earthed through a capacitor. (Sorry that I am unable to append a diagram now). This is what the 6SN7/other twin-triode is doing in your Leak circuit.

i think the HIF 87 also has the same schmitt trigger right?
 
Schmitt invertor- a Schmitt trigger is a very different thing. The 6SN7 is indeed in the Schmitt invertor configuration (much more commonly called a long tail pair or a differential amplifier).

The main issue with your idea is gain- or lack thereof. You're replacing a gain block with a gain of roughly 50-60 with one of unity gain. I'd keep the input circuit as is and just add the filter as a separate stage.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.