Not sure if this should be a new thread, but while I'm asking questions, can anyone explain the driver stage of this amp, the audio research D40? I assume its running AB2, because the quoted maximum clean output is 42W, which I don't think you can get from 6L6's without going into AB2.
The thing I don't get is that it looks like the driver stage is common cathode with NFB from the OPT. How do they get low enough Ra out of that setup to drive the 6L6's in AB2? Traditionally I thought you would use cathode followers there for their low output impedance; thats the way I think I'm going to go.
Interesting schem anyway, looks like there's feedback applied in about a thousand different ways.
The thing I don't get is that it looks like the driver stage is common cathode with NFB from the OPT. How do they get low enough Ra out of that setup to drive the 6L6's in AB2? Traditionally I thought you would use cathode followers there for their low output impedance; thats the way I think I'm going to go.
Interesting schem anyway, looks like there's feedback applied in about a thousand different ways.
Attachments
Last edited:
on my SE amps, I used no capacitor bypass. This was a large EL156 running at 450VDC and a 50V zener added for grins. I sold this amp to a friend, but eventually one of the amplifiers started drawing a touch more current than the other (they had built in current meters) ... it took a bit of troubleshooting to figure out that one of the zeners shorted out and let the full B+ pass to the screen - weird as I expected it to go open. Obviously they went over the spec at some point! 
Removing the zeners did not seem to have any outward change on the sound either and now the amps continue to plug along sans silicon - with lower B+ and olde Tungsol 6550s
Removing the zeners did not seem to have any outward change on the sound either and now the amps continue to plug along sans silicon - with lower B+ and olde Tungsol 6550s
Last edited:
No, not even any glow. Good tubes like old Sylvania 6BQ5's and new JJ EL84's seem to have no problem at this level. Most of the Russian tubes, and Chinese tubes do NOT like it.
Even I wouldn't put 700 volts to anything in a 9 pin bottle. I get 30 to 32 watts with 430 volts on the plates and 325 volts (regulated) on the screens in a Simple P-P board. This is using cathode bias at about 13 volts so the tube sees about 415 volts. If the 700 volt number is real, then I guess that I need to turn the knob up some more!
That amplifier claims a damping factor of 12.5 and a very low static dissipation. This can only mean a low idle current (near class B) and a lot of feedback.
The capacitor coupling between the driver and output stages will not allow going very far into class AB2 due to bias shift and recovery time issues. I don't think 42 watts is out of the question for 6L6GC's in AB1 with 400 volts across them. I have cranked far more than that. I have seen 100 watts from a pair in AB2 with only a little glow! The driver stage is a pair of cascaded LTP's. I have a similar circuit driving the same 6L6GC's that made 100 watts. Audio Research is famous for their symmetrical feedback loops and cathode feedback to the output stage from the OPT secondary. Trust me, you need a good OPT to do this. Try is with a Walmart grade transformer and the distortion will get worse, not better. The Plitron toroid OPT is the only one that I have that works in this manner.
Diodes, especially zeners, almost always fail to a short. It is not uncommon for a rectifier to short, and then blow open because of the excessive current. I may be known for melting tubes, but I have blown up far more silicon than glass!
This was one of my main concerns with using a Zener: whether I would even notice if it blew out. I guess if I put an LED in series, then the LED would also burn out (or at least get a lot brighter), and I could perhaps tell that something bad had happened.
OK, this leaves me only one choice....turn the knob to the right! My big power supply only goes to 650 volts, but puts out 1.7 Amps, so if the tube blows up it's going to look really cool. I got to set up video for this one!
Seriously 700 volts at 15 mA of idle current (per tube) is a rather unusual operating point, but obviously it works for him. If the idle voltage on the plate is 700 volts the plate voltage can swing from near zero to almost 1400 volts. Cheap plastic tube sockets need not show up for work here!
I agree with him there. This is especially true for the EL84's but valid for many output tubes.
Regular, or Extra Crispy!
I won't have time to blast any EL84's for a while. I am stuck at work (yeah its Sunday afternoon). If I escape soon enough, I have the red board on the bench, maybe I'll turn that up some more. I have been hesitant to go past 550 volts since some parts exploded last year. 1.7 amps made a BIG BANG, and several parts dissapeared. I never found the remains of all of them.
Haha this is all making me want to build a big test bench PSU and driver... At the moment I'm just using the PSU from an old SE 6v6 radio with one 6X5 that really isn't even up to driving a pair of EL84's.
Anyway, do you guys think its worth using a tertiary UL winding with the 6P3S's? In sims I'm getting fairly low distortion just using NFB.
Also, I'm planning on using silicon rectification, think I should chuck a thermistor in there for a slow start on B+ and/or heaters? I've just read the thread about applying B+ on cold tubes and it seems like a standby switch is a waste of time, but slow start might not be a bad idea.
Thanks for all the info. And the brutal talk about abuse of EL84's.
Anyway, do you guys think its worth using a tertiary UL winding with the 6P3S's? In sims I'm getting fairly low distortion just using NFB.
Also, I'm planning on using silicon rectification, think I should chuck a thermistor in there for a slow start on B+ and/or heaters? I've just read the thread about applying B+ on cold tubes and it seems like a standby switch is a waste of time, but slow start might not be a bad idea.
Thanks for all the info. And the brutal talk about abuse of EL84's.
mnturner, getting back to 6L6G in UL, I've recently been playing with an amp with 6P3S output tubes. The B+ is 370V and it is self-biased so there is about 340V across the tubes, and Pd of about 15W/tube. At first I tried UL with 51V zeners from the UL taps to the screens. Then I took out the zeners and the sound improved, which put around 340V on the screens. I ran it like that for quite a while, couldn't see anything glowing. Then I ditched UL mode and ran pentode mode with screen regulators set to 280V, this was better again. schematic All this was without global feedback. Very listenable now but still a work in progress. I wouldn't be too worried about thermistors but they sure wouldn't hurt either, up to you. Yeah, been an interesting thread, 700V on EL84's, been meaning to draw up a loadline to see what it looks like! Except, who's got a graph for 350V on the screens?
Ian.
Ian.
My 6P3S order never made it to my door, so it looks like I'll be haggeling wiht the seller to either get another shipment or my money back and place a new order with someone else.
Bottom line is I won't be working with 6P3S for at least another month so I'll still be playing with 6P1Ps. I'll be forced to abuse them to take out my frustration.
Bottom line is I won't be working with 6P3S for at least another month so I'll still be playing with 6P1Ps. I'll be forced to abuse them to take out my frustration.
I never got the second shipment, and he has offered to send a refund but has not as yet. May be $110 down the tube.
It looks like 6P1P abuse will commence once I get the workbench cleared off, motorcycle back together and take the significant other on a vacation to Charleston SC.
I'm still looking at the 6P3S and would like to do an AB1 design. I've got the cash to order OPTs but am a bit confused as to impedance. Based on KEN RAD CEB36 it looks like 8500ohms is optimal (page 17 top graph) and yet on the same page they claim 6600 is optimal (bottom Graph).
It looks like 6P1P abuse will commence once I get the workbench cleared off, motorcycle back together and take the significant other on a vacation to Charleston SC.
I'm still looking at the 6P3S and would like to do an AB1 design. I've got the cash to order OPTs but am a bit confused as to impedance. Based on KEN RAD CEB36 it looks like 8500ohms is optimal (page 17 top graph) and yet on the same page they claim 6600 is optimal (bottom Graph).
Its just a sim, and a rather strange one at that....
But it does hint as much as 60W clean sinewave
power might be possible in AB1 without violating
conservative spec sheet limits of a 6L6.
Its a faked McIntosh topology with twin OPTs.
I'm not sure where the extra power comes from
exactly... Could just be a mathematical error?
You need LTSpice and Duncan Amps 6L6 model.
I'm not sure it proves anything...
But it does hint as much as 60W clean sinewave
power might be possible in AB1 without violating
conservative spec sheet limits of a 6L6.
Its a faked McIntosh topology with twin OPTs.
I'm not sure where the extra power comes from
exactly... Could just be a mathematical error?
You need LTSpice and Duncan Amps 6L6 model.
I'm not sure it proves anything...
Attachments
Are the cross coupled plate to screen connections a form of negative feedback? It looks to me like it is, but does it certainly does not follow UL, nor Schade.
In triode mode, the plate and screen are locked together.
In Pentode(tetrode) mode, the screen is fixed and the plate can go above or below the screen, which has a strict limit on it's voltage.
In UL, the ratio of plate to screen voltage is fixed by the transformer tap , but the plate will not go below the screen unless the screen is also going below it's normal bias position so this is ok as well.
In the case above it is possible for the screen to go greatly positive wrt the plate (the full peak to peak plate swing value). Is this acceptable?
In triode mode, the plate and screen are locked together.
In Pentode(tetrode) mode, the screen is fixed and the plate can go above or below the screen, which has a strict limit on it's voltage.
In UL, the ratio of plate to screen voltage is fixed by the transformer tap , but the plate will not go below the screen unless the screen is also going below it's normal bias position so this is ok as well.
In the case above it is possible for the screen to go greatly positive wrt the plate (the full peak to peak plate swing value). Is this acceptable?
I suggest to fugedallaboud UL for this project. Since UL is nothing more than local NFB, just use parallel local NFB. O. Schade has already done the work for us, and recommends that 10% of the plate peak voltage be fed back to the control grids. That's the figure I used for an 807-based project. This figure was spot on, and the inclusion of local NFB really helped in eliminating that pentode nastiness the 6L6-oids like to produce.
Given that these types are designed expressly for getting that screen voltage requirement down (to improve efficiency) they really aren't amenable to UL anyway. Parallel local NFB and regulated screen supplies is the best way to go.
Another good thing about parallel NFB is that it's tweakable. With UL, once the primary of the OPT is wound, the die is cast. You just have to hope they got it right.
Another 6.0dbv of gNFB to clean up the remaining nonlinearities, and it was good to go.
Vixen Main Schemo
Given that these types are designed expressly for getting that screen voltage requirement down (to improve efficiency) they really aren't amenable to UL anyway. Parallel local NFB and regulated screen supplies is the best way to go.
Another good thing about parallel NFB is that it's tweakable. With UL, once the primary of the OPT is wound, the die is cast. You just have to hope they got it right.
Another 6.0dbv of gNFB to clean up the remaining nonlinearities, and it was good to go.
Vixen Main Schemo
The cross coupled Plate to Screen connection is a pure Pentode mode.
There is however, 50% cathode nfeedback. Which is a form of Schade.
Screen strapped to own plate = Triode
Screen strapped to own 43% tap = 71.5% UL
Screen strapped to B+ = 50% UL
Screen strapped to opposite 43% tap = 28.5% UL
Screen strapped to opposite plate = Pentode
Remember what counts for "Screen" is really Screen minus the Cathode.
In both the McIntosh, the cross coupled, and the previously mentioned
D40 to some extent, the Cathodes of these designs are not sitting still.
So the Screen feedback maths may not be as you are accustomed.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.