Most of it was allready sayed by TG (post #17).
-Vg of less then 1V starts grid current, better to be avoided, so I changed the first stage.
The grid leak resistors of the EL84 are a big load for the concertina, made them a bit larger.
And bigger cathode decoupling for a more stable bias (Ia peaks in class B)
Mona
Ok, thanks. I'm gonna start ordering parts I guess.
Just out of curiosity:
It's said a few times that a 12AU7 would be better for the concertina. This is I presume Because of the larger Ia? This would probably also mean that the grid leak resistors of the EL84 are less of a load to the concertina.
I was trying to use the 2 triodes of the same tube in the same channel. But I could use the 2 half's of an 12AX7 as input stages for 2 channels and the same with the 12AU7 as concertina for the 2 channels. However would this not also mean there's a high risk of x-coupling between channels? Maybe it's better to just add 2 extra tubes then and just use one half of each tube. so every channel has it's dedicated tubes. But that's for when I'm experimenting.
First build this, listen and learn
Just out of curiosity:
It's said a few times that a 12AU7 would be better for the concertina. This is I presume Because of the larger Ia? This would probably also mean that the grid leak resistors of the EL84 are less of a load to the concertina.
I was trying to use the 2 triodes of the same tube in the same channel. But I could use the 2 half's of an 12AX7 as input stages for 2 channels and the same with the 12AU7 as concertina for the 2 channels. However would this not also mean there's a high risk of x-coupling between channels? Maybe it's better to just add 2 extra tubes then and just use one half of each tube. so every channel has it's dedicated tubes. But that's for when I'm experimenting.
First build this, listen and learn
Do you have a chassis? Once again, BTB do some Hammond chassis if you're interested, though there are plenty of places to buy chassis and you may want to be creative!!
Just make sure the chassis is big enough! Usually bigger than you think.
https://btb-elektronik.de/suche/?searchterm=Gehäuse+hammond
I don't think you need more than 2 input sockets however you wire them up. However, if you had 4 sockets you could experiment with two EF86 in triode as first stage which has always been popular. And then two single 9 pin tubes as the concertina. Plenty of choice to optimise that.
Just make sure the chassis is big enough! Usually bigger than you think.
https://btb-elektronik.de/suche/?searchterm=Gehäuse+hammond
I don't think you need more than 2 input sockets however you wire them up. However, if you had 4 sockets you could experiment with two EF86 in triode as first stage which has always been popular. And then two single 9 pin tubes as the concertina. Plenty of choice to optimise that.
Last edited:
Nope a chassis, I will need to order as well. I already have a clear Idea how it will look. but first it will need to work:
Step 1: make it work
Step 2: make it prety
Step 3: experiment and blow everything up
Already found some distributors of parts in europe.
Tonefactory, Tube-Town and now BTB. The only thing I cannot seem to find with these distributors is 'resistors' in different power ratings and values. They all have some, but not always the ones I need.
I might go to Digikey, Farnell or Conrad for those maybe, albeit the latter being the most expensive probably
Step 1: make it work
Step 2: make it prety
Step 3: experiment and blow everything up
Already found some distributors of parts in europe.
Tonefactory, Tube-Town and now BTB. The only thing I cannot seem to find with these distributors is 'resistors' in different power ratings and values. They all have some, but not always the ones I need.
I might go to Digikey, Farnell or Conrad for those maybe, albeit the latter being the most expensive probably
Its quite silly to use this one of the best pentodes in history for a pseudo triode.
This is a low noise, high amplification tube made for small signals to boost with several hundred mu. And you will make it a low gain, high level triode. Btw, the real NOS EF86 are very rare and so they are expensive. Sorry, have to laugh about those ideas. There are plenty of low gain triodes out there, from cheap as dirt to expensive prices. To each his own.
Have a look at this Tube-Town page. It covers both the pentode and triode uses of the EF86. As you say, it's often - but not always - a small signal tube.
Tube-Town GmbH – How To Use a EF86
"Figure 3 shows the EF86 in triode connection. When this is done, the resulting triode is among the lowest-distortion electronic amplifying devices ever made. Typical second-harmonic distortion with 10v RMS output is on the order of 0.05%. This result is achieved without any negative feedback, so this circuit is an excellent choice for high-end audio design. The voltage gain will be about 25 in this connection, making it suitable for line stages."
Tube-Town GmbH – How To Use a EF86
"Figure 3 shows the EF86 in triode connection. When this is done, the resulting triode is among the lowest-distortion electronic amplifying devices ever made. Typical second-harmonic distortion with 10v RMS output is on the order of 0.05%. This result is achieved without any negative feedback, so this circuit is an excellent choice for high-end audio design. The voltage gain will be about 25 in this connection, making it suitable for line stages."
Try Gotron | Home page | ElektronicaspecialistAlready found some distributors of parts in europe.
Tonefactory, Tube-Town and now BTB. The only thing I cannot seem to find with these distributors is 'resistors' in different power ratings and values. They all have some, but not always the ones I need.
I might go to Digikey, Farnell or Conrad for those maybe, albeit the latter being the most expensive probably
or Reichelt: Security Check
Mona
Last edited:
A possibility to consider is a 12AX7/ECC83 and a 12AU7/ECC82, in each channel. However, only 1 section of the tubes is used. Wire the in service sections asymmetrically, between the 2 channels. When you see signs of wear, flip/flop tubes between channels and get doubled service life. A small benefit of the method is somewhat improved channel separation.
Triad Magnetics shows 5 distributors in Belgium. Look here.
Triad Magnetics shows 5 distributors in Belgium. Look here.
This might end in death cathodes in tubes that are run withoud cathode current for a long time. It would be better to wire both triodes in parallel and adapt the resistor values. Anyway, I don't mind to use both triodes of the same tube in both channels. Optionally you could opt for a 12DW7/7247 or their J/J equivalent, either ECC832 or ECC823 (I don't know exactly how it is spelled...)
Best regards!
Best regards!
Only heat one cathode. It means you’re stuck with 6 volts, rather than the option of running off 12.
Jones880,
If you want to explore this direction I recommend a review of the diytube design by Parks Audio that uses one 12ax7 section for voltage amplification and both 12au7 sections for phase splitting. I built it and it is a very fine EL84 amplifier. Your concern about cross-coupling between channels in one double triode is valid, but in practice in this application that proved to be a non-issue.
Unfortunately the PCBs are no longer available, but it would not be difficult to hardwire build it.
See: http://www.diytube.com/st35/ST35REVE.pdf
The EL84 is a very easy to drive tube.No problem for the ECC83.
Philips EL84 datasheet gives max Rg1 (the load for the concertina) 1M with auto(cathode) bias ande 300k with fix bias.
Most amps use 470k with cathode bias but there are stories about runaway tubes.
So, to be on the save side, I proposed 330k no need to go lower.
330k//100k= 77k the real AC-load.10Vtop (max drive) is a current change of 130µA on 740µA bias, no problem.
Mona
Philips EL84 datasheet gives max Rg1 (the load for the concertina) 1M with auto(cathode) bias ande 300k with fix bias.
Most amps use 470k with cathode bias but there are stories about runaway tubes.
So, to be on the save side, I proposed 330k no need to go lower.
330k//100k= 77k the real AC-load.10Vtop (max drive) is a current change of 130µA on 740µA bias, no problem.
Mona
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.