The triode section of the 6GW8 (ECL86) is identical to each triode section of a 12AX7. The actual anode voltage is not super critical but the "usual" requirement/recommendation for a triode stage applies. That it, it should be between 1/2 and 2/3 of the rail voltage.
I hope the following makes sense to you:
The advantage (for a front end) of running toward the upper limit (higher voltage end) of this range is that the stage then needs a little more bias voltage for the same operating current. The effective grid to cathode voltage is the sum of the bias voltage and the applied signal. As this composite "effective" bias is reduced on positive input signal swings you can run into grid current. For any particular input signal level voltage, grid current is less likely if the bias voltage is higher. The effective signal + bias voltage from grid to cathode stays further away from 0V.
The distortions resulting from a triode section running into grid current on positive input signal peaks are fairly nasty sounding.
For the 6GW8 variant of the Baby Huey with a 300V rail that means the anode voltages should be between 150 and 200V. The rail to the triode stage of the modified/experimental version we are discussing should be much higher than the +180 you sketched. When using a resistor + capacitor to generate another rail for a preceding stage, the amount of voltage you drop across the resistor is not the overiding concern. The time constant is what is important. For say 1Hz = 0.16 milli seconds time constant and using a 22uF cap the resistor needs to be >7.2 kOhms, so I used 10K. with 1.2mA in the diff amp that means the rail to the diff amp will be about 12V less than the main B+.
Also with the CCS in the diffamp tail to enforce diff amp balance there will be practically 0 net AC current draw from this power supply node anyway so you don't need to go overboard with the decoupling for this supply.
A bit of a rave but I hope there is some stuff of value in it.
Cheers,
Ian
I hope the following makes sense to you:
The advantage (for a front end) of running toward the upper limit (higher voltage end) of this range is that the stage then needs a little more bias voltage for the same operating current. The effective grid to cathode voltage is the sum of the bias voltage and the applied signal. As this composite "effective" bias is reduced on positive input signal swings you can run into grid current. For any particular input signal level voltage, grid current is less likely if the bias voltage is higher. The effective signal + bias voltage from grid to cathode stays further away from 0V.
The distortions resulting from a triode section running into grid current on positive input signal peaks are fairly nasty sounding.
For the 6GW8 variant of the Baby Huey with a 300V rail that means the anode voltages should be between 150 and 200V. The rail to the triode stage of the modified/experimental version we are discussing should be much higher than the +180 you sketched. When using a resistor + capacitor to generate another rail for a preceding stage, the amount of voltage you drop across the resistor is not the overiding concern. The time constant is what is important. For say 1Hz = 0.16 milli seconds time constant and using a 22uF cap the resistor needs to be >7.2 kOhms, so I used 10K. with 1.2mA in the diff amp that means the rail to the diff amp will be about 12V less than the main B+.
Also with the CCS in the diffamp tail to enforce diff amp balance there will be practically 0 net AC current draw from this power supply node anyway so you don't need to go overboard with the decoupling for this supply.
A bit of a rave but I hope there is some stuff of value in it.
Cheers,
Ian
I experimented further with my ecl82 set up.
- I kept the garter (blumlein) bias of the pentode output stage (á la Shoog).
- I changed the front end into the original baby huey LTP with the CCS (2 x BC547B + 1 red led), set for about 1 mA
- I added on top of the triode a single LND150 mosfet and used the MU-output. I forgot to draw it but I bypassed the top ccs with a 2 meg resistor to prevent fighting between the two CCS's.
- I used the cross coupled feedback from plate of pentode to grid of the opposite triode.
I like the idea of using the original Kimmel mu-stage in a differential topology and with pentode on top, but for now it was less work to put a mosfet in.
I only had a 10K pot lying around so I used it but I change that into a 100K pot as soon as I find one.
I had only a short time to listen and had to use a 88 dB speaker which is not sensitive enough. The layout is a mess or better a disaster because it is an experimental setup that had a lot of changes. I used cheap components and I only have 1 mono version.
But it still sounds very good.
I hope the well known experts will try this set up to compare it with the original baby. I suspect it will sound gorgious with a el84.
May be the LN150 should be cascoded.
the best thing would be if some egghead transformes this into a direct coupled version, the we have only one cap left in the feedback string.
See my original Rembrandt drawing.
- I kept the garter (blumlein) bias of the pentode output stage (á la Shoog).
- I changed the front end into the original baby huey LTP with the CCS (2 x BC547B + 1 red led), set for about 1 mA
- I added on top of the triode a single LND150 mosfet and used the MU-output. I forgot to draw it but I bypassed the top ccs with a 2 meg resistor to prevent fighting between the two CCS's.
- I used the cross coupled feedback from plate of pentode to grid of the opposite triode.
I like the idea of using the original Kimmel mu-stage in a differential topology and with pentode on top, but for now it was less work to put a mosfet in.
I only had a 10K pot lying around so I used it but I change that into a 100K pot as soon as I find one.
I had only a short time to listen and had to use a 88 dB speaker which is not sensitive enough. The layout is a mess or better a disaster because it is an experimental setup that had a lot of changes. I used cheap components and I only have 1 mono version.
But it still sounds very good.
I hope the well known experts will try this set up to compare it with the original baby. I suspect it will sound gorgious with a el84.
May be the LN150 should be cascoded.
the best thing would be if some egghead transformes this into a direct coupled version, the we have only one cap left in the feedback string.
See my original Rembrandt drawing.
Attachments
Sorry,
Have not progressed this and haven't even been listening to it.
Got sidetracked by 2 other projects.
I've wanted a decent guitar amp for a long time to replace converted PA amp I've been using so have started to build one, all octal tubes, 6SL7s for preamp and phase splitter and 6V6s for output.
Also (blush with embarrassment) I've been off playing in the sand pit, doing some upgrades on a SS Amp - you know one of those things with 3 legged fuses in it.
Will try to get back to it sometime soon - meanwhile it seems plenty of others have things to contribute.
Cheers,
Ian
Have not progressed this and haven't even been listening to it.
Got sidetracked by 2 other projects.
I've wanted a decent guitar amp for a long time to replace converted PA amp I've been using so have started to build one, all octal tubes, 6SL7s for preamp and phase splitter and 6V6s for output.
Also (blush with embarrassment) I've been off playing in the sand pit, doing some upgrades on a SS Amp - you know one of those things with 3 legged fuses in it.
Will try to get back to it sometime soon - meanwhile it seems plenty of others have things to contribute.
Cheers,
Ian
John,
No updates on this idea, but I have got two bits of good news.
Finally got the Tabor up and running and stable - sounding goooood !!!
I also have just got my latest Super Linear Cathode Follower preamp working. This version is using ECF80's. Sounds fantastic - a big improvement over my last version. Still got a bit of case work to do, and the stepped attenuator will have to wait a bit.
I feel a trip over in Frans direction coming on - what do you think ?
Shoog
No updates on this idea, but I have got two bits of good news.
Finally got the Tabor up and running and stable - sounding goooood !!!
I also have just got my latest Super Linear Cathode Follower preamp working. This version is using ECF80's. Sounds fantastic - a big improvement over my last version. Still got a bit of case work to do, and the stepped attenuator will have to wait a bit.
I feel a trip over in Frans direction coming on - what do you think ?
Shoog
ECL82 symmetric PP
Hi!
I would like also to build a simple but good amp with only two ECL82 tubes and input transformer. Something like the attached schematic.
Could somebody help me to calculate the missed values or correcting the errors?
The 250Vdc PSU and the PP output stage are given already.
Thanks!
Tyimo
Hi!
I would like also to build a simple but good amp with only two ECL82 tubes and input transformer. Something like the attached schematic.
Could somebody help me to calculate the missed values or correcting the errors?
The 250Vdc PSU and the PP output stage are given already.
Thanks!
Tyimo
Attachments
There is partial feed back in my design and that is what makes it possible to run the outputs as pentodes (though I have them triode strapped). The design will work better with the outputs as pentodes.
Look at the 100K resistor on the original schematic.
Also it will have excessive gain and the triode grids will overload easily. You are missing the oportunity to run the front end as a LTP which is easy with an input transformer. The Garter bias on the outputs keeps perfect DC balance in the output stage which allows for high quality toroidal transformers.
All in all you have missed all the design features which make my design sound good
Shoog
Look at the 100K resistor on the original schematic.
Also it will have excessive gain and the triode grids will overload easily. You are missing the oportunity to run the front end as a LTP which is easy with an input transformer. The Garter bias on the outputs keeps perfect DC balance in the output stage which allows for high quality toroidal transformers.
All in all you have missed all the design features which make my design sound good
Shoog
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.