Im wanting to build a class AB amp in UL configuration with cathode bias for driving a pair of 807 tubes on each channel and thought the mullard 5-20 design would be a perfect fit.
Obviously there will need to be some modifications to the phase inverter to suit the load lines of the 807, something im not 100% familiar with.
What I do know is that an 807 needs far more grid current to drive the tubes compared to an EL34 which is much more sensitive.
One bonus though is the EF86 voltage amplifier will help somewhat, as its regarded as a fairly sensitive amplifier.
Im considering running the tubes at 400-450V since ive found thats quite a sweet spot for the 807 in my guitar amp.
IIRC, 4K impedance is what I need on the OT, since my guitar amp runs with 2K on the primary with a quad.
Anything else I should know?
Obviously there will need to be some modifications to the phase inverter to suit the load lines of the 807, something im not 100% familiar with.
What I do know is that an 807 needs far more grid current to drive the tubes compared to an EL34 which is much more sensitive.
One bonus though is the EF86 voltage amplifier will help somewhat, as its regarded as a fairly sensitive amplifier.
Im considering running the tubes at 400-450V since ive found thats quite a sweet spot for the 807 in my guitar amp.
IIRC, 4K impedance is what I need on the OT, since my guitar amp runs with 2K on the primary with a quad.
Anything else I should know?
The 807 only needs grid current if you drive it into AB2, i.e. grid current. In AB1 by definition it doesn't take ANY grid current. Decide what class you want to operate in and then you can figure out what you need from the driver.
It would be AB1 then, from what I understand AB2, is essentially class B, but seems to be loosley defined what class B really is.
All I know is the mullard 5-20 is class AB, but not 100% sure if its AB1 or AB2.
Last edited:
No, the amplifier classes are quite well defined. Read up on it so you can make an informed decision.
The original Mullard as most copies are AB1, i.e. no grid current. Regardless of what output tubes are used. You can design yours for whatever class you want.
If you decide on AB1, then your OP reference to grid current is beside the point.
The original Mullard as most copies are AB1, i.e. no grid current. Regardless of what output tubes are used. You can design yours for whatever class you want.
If you decide on AB1, then your OP reference to grid current is beside the point.
Yes thats all I need to know, ill stick with AB1. Still not sure what component values will need changing on the phase inverter, but will talk to a couple of friends who have more experience in this.
I was confused regarding the different classes, since I was reading a thread here that basically said that class AB1 and 2 are essentially class B and that there isint really any such thing as class AB.
Im pretty sure an 807 does require alot more than an EL34.
I built a huge guitar amp about 10 years ago, and was the last thing ive done with 807's but its coming back to me. At the time I was told to use 47K on each anode of the long tail pair with a 12AT7, it seems to drive a quad of the things rather well.
This is the 5-20 schematic here for reference, i may be better off changing the inverter tube out for something similar gain to a 12AT7.
Mullard EL34 Push-Pull Tube Amp Schematic (Dynaco A420 Transformer)
807 IS a 6L6 with a top cap and UX5 base.
As a result it like the KT66 and KT88 makes huge amounts of IMD and odd harmonic distortion.
The only octal version is the TV scan version made during the 1960s and dirt cheap in Russia.
The only difference with 807s really is the improved anode material which means you can run them red plating for hours without them melting, and up to 850-900V on the anode.
If you check it out, the harmonic 5 is the worst and the screen grid is aligned, so it take small amounts of screen current, compared with the much more modern EL34, but has much lower gain, so you have to multiply the drive voltage by 2-3x.
807s have huge production spread so you will need at 10 to get anything approaching a matched pair or 2.
Like the 6L6 they only start to sound nice and linear when they are close to blowing up or melting.
It's a 1930s design what do you expect?
Last edited:
Ive got a box of 807s here and most of them actually read in a very similar range, i definitely have a few matched quads here, anyway with cathode bias you dont need to worry about matched tubes.
The OP obviously erred with the reference to grid current in the first post, and really meant grid voltage swing.
nzoomed, it is worthwhile checking the OPT you have for its UL tap %, and cross check that with what a 6L6 likes best. You will also be making compromises throughout the circuit as you are not directly copying the original 5-20 part-for-part. As every change you make will have some impact on operation, and that can get very technical to both explain and also make informed decisions on whether the change is really a concern, it is probably best to get ready for a learning curve and start trying things out yourself and working out how to make measurements that indicate where you stand. Trying to pre-design your variation too much is likely somewhat futile as there would be so many unknown's or what-ifs, as you would need to provide a starting schematic with all part details and some base level design info.
The most obvious aspects are going to be drive voltage swing available to the 807 grids, and the need to dramatically lower or disconnect negative feedback until you have advanced the nominal operation of the rest of the circuit, and have a plan for how you can measure stability performance at both low and high frequency ends.
nzoomed, it is worthwhile checking the OPT you have for its UL tap %, and cross check that with what a 6L6 likes best. You will also be making compromises throughout the circuit as you are not directly copying the original 5-20 part-for-part. As every change you make will have some impact on operation, and that can get very technical to both explain and also make informed decisions on whether the change is really a concern, it is probably best to get ready for a learning curve and start trying things out yourself and working out how to make measurements that indicate where you stand. Trying to pre-design your variation too much is likely somewhat futile as there would be so many unknown's or what-ifs, as you would need to provide a starting schematic with all part details and some base level design info.
The most obvious aspects are going to be drive voltage swing available to the 807 grids, and the need to dramatically lower or disconnect negative feedback until you have advanced the nominal operation of the rest of the circuit, and have a plan for how you can measure stability performance at both low and high frequency ends.
Last edited:
I dont know alot about UL taps on the transformer as far as they go, but was reading that its usually just 43% of the primary winding, but it doesnt surprise me if its different depending on the tube type.
I didnt think there was much that had to be changed other than the phase inverter and grid stopper resistors, etc.
Ill need to look into this, but am trying to find other schematics for a similar AB1 class amp with 807 / 6L6 tubes.
My marshall 2203 clone I buult was exactly the same, except i changed the phase inverter for a 12AT7 and changed the resistor values on it and it seems to work fine.
Thats a guitar amp, so is certainly not hifi quality but works perfect as far as guitar sounds.
Thanks for that.
Ive been looking at this schematic and it appears its based on the mullard 5-20 design, its not UL however, but it is set up with KT66 tubes which 807/6L6 would basically drop right into.
Its the one at the top called "PARMEKO"
100 Amplifiers – Part 3 , 1955 – 59 | Lilienthal Engineering
Perhaps it may be easier for you to just get an amp working - that could be reasonable if you can measure signal levels through the amp and then look at loadlines to check you are seeing acceptable max voltage swings and clipping occurring to know where each stage is at. Just be aware that if you don't get the power or performance you think you should get, or aren't sure if the stages are working appropriately, then it is up to you to make measurements and work through why. One hassle with many schematics is that there is no detailed design/measurement doc to go with it - some magazine or kit or clone style projects go some way by giving at least DCV levels throughout the circuit, and some go to AC signal levels expected within the circuit, and scope waveforms that show the signal levels at onset of clipping at the output of certain stages.
Thats what I think too.
Its not too difficult to modify later anyway with the same chassis, the tube sockets will still be OK to use if i need to swap out tubes.
I spoke to an old mate who reckons that significant modding wont be necessary and will be mostly around the phase inverter, particularly around the anode load resistors.
I think I may as well give it a shot and see how it goes, only other things that will need changing are bias resistors on the cathode on the output tubes.
Most critical thing is basically the right matched OT at this point.
I can always tweak the design easily enough if needed.
Just a thought, the 5-20 design may not have enough gain, EL34 has almost 2x the Gm of a 807/6L6.
Were it my design, I'd use something like a Williamson front end with the 807s in UL. If your speakers are efficient enough, maybe run the 807s triode strapped in AB2 with a MOSFET follower on the the grids.
Works here. If you have access to it, read Morgan Jones - Valve Amplifiers, he does a very good job of analysing the Williamson topology vs the Mullard 5-20.
Were it my design, I'd use something like a Williamson front end with the 807s in UL. If your speakers are efficient enough, maybe run the 807s triode strapped in AB2 with a MOSFET follower on the the grids.
Works here. If you have access to it, read Morgan Jones - Valve Amplifiers, he does a very good job of analysing the Williamson topology vs the Mullard 5-20.
Pentodes & Beam tubes are critical for load impedance, dependent both on plate & screen supply volts. If the conditions in your guitar amp are correct then with the 4K be sure to use the same supply conditions. Some guitar amps (Peavy) are real hifi, the FX does the distortion.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.