Thanks Giame, both for your consideration and interest 
So, I still haven't given up on the no NFB goal. I drew a loadline on the 6N6P curves, as shown below. I was hoping to use Ia=7.5 mA to minimize the drain on the 150mA PS from the input stage and leave as much for the output stage as possible. Is this an acceptable region of linearity? And is 15mA for the CCS OK?
So, I still haven't given up on the no NFB goal. I drew a loadline on the 6N6P curves, as shown below. I was hoping to use Ia=7.5 mA to minimize the drain on the 150mA PS from the input stage and leave as much for the output stage as possible. Is this an acceptable region of linearity? And is 15mA for the CCS OK?
Attachments
My EL84 Experience
I've rebuilt my EL84 Ultralinear Push Pull about 5 times to various schematics. It started life as a "Bervois Valley" Ultralinear.
It is now at the best I've ever heard it. Circuit is now a minor variant of the the Yves Monmagon circuit in Post #4.
In place of the ECL86 I use JJ Electronics EL84 pentodes and the new JJ Electronics ECC803S for the front end. I also cathode bias the EL84s as per schematic in Post #1. The ECC803S (premium 12Ax7) is entirely adequate to drive EL84s.
Component changes to suit Post #4 schematic:
I use 15K for R12 for a little more shunt feedback.
Global feedback is 13K from the 4 Ohm tap with 1K from -ve input grid to 0V (10K in place of the 13K lost a little imaging, did'nt try 12K). High value resistors here react with grid current to unbalance the diff amp and in particular limit high frequency response - so get that 100K on the -ve input down to something like the 1K I use.
Current source as per schematic in post #1 but set for 1mA total (change the 68R to 1K) and the 330R and 680R can be 3K3 and 6K8 instead.
Definitely ditch those 100pF caps across the output tube grid resistors.
The final tweak to give it REAL bass authority:
ditch the 270R cathode bias resistors and replace with BC547B + MJE340 "ring of 2 transistors" current sources. I used 470uF/50V Blackgate caps for the bypass caps initially AND then finally changed to 4700uF/ 63V Nichicon caps (I left the 470uF BG and just added the 4700uF across them) to get real bass authority. The current sources for cathode bias neatly steps around the usual problem of blocking distortion with high value caps across cathode bias resistors.
Hope some of this is of interest.
Cheers,
Ian
I've rebuilt my EL84 Ultralinear Push Pull about 5 times to various schematics. It started life as a "Bervois Valley" Ultralinear.
It is now at the best I've ever heard it. Circuit is now a minor variant of the the Yves Monmagon circuit in Post #4.
In place of the ECL86 I use JJ Electronics EL84 pentodes and the new JJ Electronics ECC803S for the front end. I also cathode bias the EL84s as per schematic in Post #1. The ECC803S (premium 12Ax7) is entirely adequate to drive EL84s.
Component changes to suit Post #4 schematic:
I use 15K for R12 for a little more shunt feedback.
Global feedback is 13K from the 4 Ohm tap with 1K from -ve input grid to 0V (10K in place of the 13K lost a little imaging, did'nt try 12K). High value resistors here react with grid current to unbalance the diff amp and in particular limit high frequency response - so get that 100K on the -ve input down to something like the 1K I use.
Current source as per schematic in post #1 but set for 1mA total (change the 68R to 1K) and the 330R and 680R can be 3K3 and 6K8 instead.
Definitely ditch those 100pF caps across the output tube grid resistors.
The final tweak to give it REAL bass authority:
ditch the 270R cathode bias resistors and replace with BC547B + MJE340 "ring of 2 transistors" current sources. I used 470uF/50V Blackgate caps for the bypass caps initially AND then finally changed to 4700uF/ 63V Nichicon caps (I left the 470uF BG and just added the 4700uF across them) to get real bass authority. The current sources for cathode bias neatly steps around the usual problem of blocking distortion with high value caps across cathode bias resistors.
Hope some of this is of interest.
Cheers,
Ian
ECC803S vs 12AX7
I did the rebuild and set to work using EI 12AX7. I then substituted some NOS Aussie production 12AX7 (from the Philips Hendon factory in Adelaide - it closed 20 years ago) which are one of our Aussie secrets, fantastic tubes (usually branded AWA). They were better than the EIs. The JJ ECC803S were at least as good as the NOS Philips and are what I'm running today.
My unit uses some good quality Output Trannies which helps but the shunt feedback in conjunction with the Ultralinear connection allows very good drive (low output tube rp) and the best bass authority I've ever heard in a 10W per channel amp plus crisp, tingly highs. It leaves my 20W per channel 845SET for dead in ALL areas. I'm still not sure that Yves appreciates just how clever his little circuit is!!!!
I also built an ECL86 version for a lady friend who is a classical music and opera fan and she absolutely loves the sound. Her amp was called the "Baby". The EL84/ECC803S version I've called "Baby Huey" although "Jack" (the giant killer) might have been an appropriate name.
I have a third version running at the moment but with the shunt feedback replaced by cathode feedback from the output transformer secondary and using 6V6 and Hammond 1608 trannies. Its just not in the same league - probably due to the Hammonds not being quite good enough but also the cathode feedback (determined by the 4 Ohm taps) not providing the same level of local feedback as the shunt feedback scheme.
As for the tube newbies "What Amp should I build" question which turns up time and again this circuit is ideal. Its the best EL84 amp I've ever heard and I've heard quite a few. In fact it comes close to being the best amp I've ever heard. Perhaps I should post the schematic in a new thread. I'll draw it up and get it scanned.
Cheers,
Ian
I did the rebuild and set to work using EI 12AX7. I then substituted some NOS Aussie production 12AX7 (from the Philips Hendon factory in Adelaide - it closed 20 years ago) which are one of our Aussie secrets, fantastic tubes (usually branded AWA). They were better than the EIs. The JJ ECC803S were at least as good as the NOS Philips and are what I'm running today.
My unit uses some good quality Output Trannies which helps but the shunt feedback in conjunction with the Ultralinear connection allows very good drive (low output tube rp) and the best bass authority I've ever heard in a 10W per channel amp plus crisp, tingly highs. It leaves my 20W per channel 845SET for dead in ALL areas. I'm still not sure that Yves appreciates just how clever his little circuit is!!!!
I also built an ECL86 version for a lady friend who is a classical music and opera fan and she absolutely loves the sound. Her amp was called the "Baby". The EL84/ECC803S version I've called "Baby Huey" although "Jack" (the giant killer) might have been an appropriate name.
I have a third version running at the moment but with the shunt feedback replaced by cathode feedback from the output transformer secondary and using 6V6 and Hammond 1608 trannies. Its just not in the same league - probably due to the Hammonds not being quite good enough but also the cathode feedback (determined by the 4 Ohm taps) not providing the same level of local feedback as the shunt feedback scheme.
As for the tube newbies "What Amp should I build" question which turns up time and again this circuit is ideal. Its the best EL84 amp I've ever heard and I've heard quite a few. In fact it comes close to being the best amp I've ever heard. Perhaps I should post the schematic in a new thread. I'll draw it up and get it scanned.
Cheers,
Ian
This is what I want!! I plan to build this EL84 amp to drive my subwoofer, a W frame design with 2x 98dB speakers, so the "low" power and the somewhat "high" output impedance won't be a problem (the high output will even be good, as it increases the Qts of the speaker). I already had plans to use a CCS at the cathodes (to get a balanced DC current through the OPT, for better bass), but was going for a 317 or 337, because of its simplicity. Soon I will have more choices!!
Gingertube, many thanks for your efforts!
Erik
Re: ECC803S vs 12AX7
I would like to see that very much. Failing that, I would really be interested to find out what current you set your output stage CCS to. John Swenson set his to 70 mA but his B+ was very low.
Not sure that's a good idea. CCS's on the output stage are all the rage and everybody raves about them, but the weakness is that clipping is very harsh when the stage does get overloaded. And a 20W sub amp will likely be overloaded on occasion.
gingertube said:
I would like to see that very much. Failing that, I would really be interested to find out what current you set your output stage CCS to. John Swenson set his to 70 mA but his B+ was very low.
ErikdeBest said:
Not sure that's a good idea. CCS's on the output stage are all the rage and everybody raves about them, but the weakness is that clipping is very harsh when the stage does get overloaded. And a 20W sub amp will likely be overloaded on occasion.
Hi Leadbelly
Thanks for your reply on the CCS. The subs will be used just for audio at moderate sound levels, and each will contain two 98dB speakers, that's why I tought I could go for a low power amp to drive them. Well, I will build the amp anyway, but take this in account. If it gets overloaded, I will put a beefier amp on them (PP EL519)
Erik
Thanks for your reply on the CCS. The subs will be used just for audio at moderate sound levels, and each will contain two 98dB speakers, that's why I tought I could go for a low power amp to drive them. Well, I will build the amp anyway, but take this in account. If it gets overloaded, I will put a beefier amp on them (PP EL519)
Erik
el84PP
I am also planning a el84PP.
My idea was to use the front of the "cold war amp" that can be found at: http://www.ultranalog.com/
Quite similar of what we see here.
What do you think ? It uses 2c51/5670 or 6n3p tubes.
El84 with Gary pimm CCS under them. Thought about 300 volt/40 mA per tube.
I am also planning a el84PP.
My idea was to use the front of the "cold war amp" that can be found at: http://www.ultranalog.com/
Quite similar of what we see here.
What do you think ? It uses 2c51/5670 or 6n3p tubes.
El84 with Gary pimm CCS under them. Thought about 300 volt/40 mA per tube.
Progress
Here's a pic of progress so far. Only thing wired up so far is the heaters. As you can see, I took many comments to heart, and driver stage is now 6SN7, rectifier is 5AR4, and an extra transformer in the back corner to provide 2 more 6.3VAC windings, 1 used for the 5AR4.
I tried making a DC heater section for the 6SN7/EL84 tubes, but was discouraged by how much heat I had to contend with, so I now have wired it as AC. I am musing over whether to pursue my original intent of using the 2nd 6.3V winding for a CCS, or to maybe use that for DC heater power for just the 6SN7's, and deal with the CCS bias by other means, say an input decoupling cap? Opinions welcome.
Here's a pic of progress so far. Only thing wired up so far is the heaters. As you can see, I took many comments to heart, and driver stage is now 6SN7, rectifier is 5AR4, and an extra transformer in the back corner to provide 2 more 6.3VAC windings, 1 used for the 5AR4.
I tried making a DC heater section for the 6SN7/EL84 tubes, but was discouraged by how much heat I had to contend with, so I now have wired it as AC. I am musing over whether to pursue my original intent of using the 2nd 6.3V winding for a CCS, or to maybe use that for DC heater power for just the 6SN7's, and deal with the CCS bias by other means, say an input decoupling cap? Opinions welcome.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.