I built a power amplifier with the 807 some time ago:
807_amplifier
Now I decided to do some modifications on it:
- B+ reduced to 350V
- First tube will be EF86 or something similar
- Cathode follower after phase splitter cancelled
- Output transformer turns ratio 20:1 (3.2k plate-plate load)
- UL operation with 40% g2 tap
- 12 dB negative feedback
- Combined bias (fixed/cathode bias)
807_amplifier
Now I decided to do some modifications on it:
- B+ reduced to 350V
- First tube will be EF86 or something similar
- Cathode follower after phase splitter cancelled
- Output transformer turns ratio 20:1 (3.2k plate-plate load)
- UL operation with 40% g2 tap
- 12 dB negative feedback
- Combined bias (fixed/cathode bias)
Interesting, I will take a look at that thanks. Ironically one of the reasons I thought the 2-20 design may have been suitable was because people are saying it had too much gain from the EF86 stage and was very sensitive! There are even instructions out there on reconfiguring the EF86 as a triode to reduce its gain.
Anyway, done some more digging and found this schematic, am interested to get some feedback on it.
It appears to be based on the 5-20 design.
The EF86+6SN7+6L6 push-pull circuit - Amplifier_Circuit - Circuit Diagram - SeekIC.com
I'm pretty sure there would be no problem with gain with the original 5-20 circuit. The EL34 cathode is at 32 volts in the 5-20 vs 36 volts for the 6L6 in the circuit linked in post#22, a negligible difference. Sensistivity for the original 5-20 was 250mV at full output. That said, I prefer the circuit in post #22 due to the 6SN7 phase splitter (you could also use 6cg7). Like you said for the 5-20, you could also configure the EF86 as a triode to reduce gain / feedback.
That was a lovely looking build, with great details! Did you buy a bulk lot of old can caps?
I was looking at the way the NFB was configured and its quite different to what ive seen before, it looks as if its taking the NFB from across most of the secondary, perhaps its done this way to balance it?
Interestingly enough, ive also noticed the Mullard 5-20 design is also recieving NFB from the primary winding on the OT coupled to one of the UL taps by a capacitor.
Most designs ive worked with simply take the NFB from the top of the secondary on the OT.
Ive often found that this is part of a circuit that requires tweaking on past builds ive done.
The original mullard 5-20 design uses a ECC83/12AX7 as a phase splitter which was a rather bad choice.
I had been looking at this improved design which uses a 6CG7 which is essentially similar to a ECC82 or 6SN7
Mullard EL34 Push-Pull Tube Amp Schematic (Dynaco A420 Transformer)
I might end up going with the design in post #22, although it looks like its not that hard to adapt the 5-20 to use 6L6 / 807 if I want to.
It could very well be OK, im just going by what im reading in various threads regarding the 5-20 design.
Seems a few think a 12AX7/ECC83 was a lazy choice from what Im reading, but not sure on exact reason, but was wondering if it was too high gain?
I didnt think there was anything wrong with a 12AX7 personally, but I know a 12AT7 can drive more current than the likes of a 12AX7 which would be beneficial for 807 tubes. Im not too sure where the 12AU7 sits, but this "improved" schematic has swapped out the phase inverter for a 6CG7 which is comparable to a 12AU7 or 6SN7 in regards to gain.
Mullard EL34 Push-Pull Tube Amp Schematic (Dynaco A420 Transformer)
nzoomed, you may want to look at some loadline tutorials and apply them to your target schematic - the likely important aspect for you is the voltage swing capability of the PI tube in the intended circuit - the 12AU7 or 12AT7 allow a larger voltage swing (as saturation voltage is lower), and that is where the most benefit may arise.
Note that if you need to drive 807 grids more (ie. a larger voltage swing - not current capability) then your options are to increase the amp's input voltage (sensitivity) and/or lower the feedback level. The caveat is that none of the stages in your amp reach clipping level (which is where the 12AU7 may or may not be needed depending on your voltage rails and what PI tail voltage you want to operate at).
Note that if you need to drive 807 grids more (ie. a larger voltage swing - not current capability) then your options are to increase the amp's input voltage (sensitivity) and/or lower the feedback level. The caveat is that none of the stages in your amp reach clipping level (which is where the 12AU7 may or may not be needed depending on your voltage rails and what PI tail voltage you want to operate at).
Wouldn't it be nice if tubes had datasheets? Where you could actually look-up and ponder tube performance? From white-shirt junior engineers sweating in pre-airconditioning labs? Instead of ill-formed opinions from "the web"?
Attachments
Yes ive got a couple of books on amplifier design that talk about this.
Been going through some of it. Im at the point that Im confident I will be able to get something working fairly well with little if any modification required.
My other main concern is the OT, i have bought a bunch of winding gear and wire from a person who used to wind transformers and has since deceased.
Ive been going through the basics on transformer theory, as i may as well experiment making my first OT's at the same time!
Its not like i have not read the datasheet for the 807, as you can see, it confirms that the 807 tube requires a higher grid voltage which is what I already know, its actually higher than what I thought, I had 40V in mind from what i remember, but its about 10 years ago I did anything with 807's.
For obvious reasons 6L6 and 807 don't work in UL, which is why nobody bothers.
On my rebuilt and fully restored 807 amps the bias is -30V with HT of 600V and screen at 325, all fully stabilised.
It's running AB2 CT choke fed via CF 6SN7, and knocks out 75W clean no sweat.
I have a pair of 125W units using a pair of 807s, which I am upgrading with a quad of STC 5B25xm to make 175W RMS total.
Last edited:
PRR's point was that for the conditions he highlighted, the 807 actually required slightly lower drive voltage compared to the EL34, and that one shouldn't believe everything one reads on the internet
Im not too sure how he is getting those figures. TBH those philips datasheets are pretty hard to understand, that figure of 22.7V doesnt even have a label saying its the grid1 voltage, nor does it say anywhere that its RMS volts.
Im still not sure how 60V peak would work out to only 21.21V RMS either, my calculations give me 42V RMS from a peak voltage of 60V which is the figure ive always had in mind for a while.
Its not just on the internet either, an old ham guy helped me with my last build about 10 years ago and was telling me the same stuff about the 807 drive voltage, so is something ive always accepted.
Last edited:
Hi,
the data sheet from the Philps EL34 is written more in EU standards.
Then take a look at the Mullard Datasheet.
https://frank.pocnet.net/sheets/129/e/EL34.pdf
there ist Vin 55V g1-g1.
Phillips give the Vgin in Vrms against ground per tube. Veff means Vrms and is dutch/german abreviaton for effektiv. Veff x root2 = Vpeak . 22,7V x1,414 = 32,09V peak . 2 Tube with 180 degrees means twice the Volts grid-grid 64,2V peak-peak
It is not a big difference between the grid ac Voltage of an EL34 or an 6L6GC/ 807.
An 807 can easily driven with an ECC83/12AX7 without clipping. Its got more swing than the 807 are need for full blast.
73
Wolfgang
the data sheet from the Philps EL34 is written more in EU standards.
Then take a look at the Mullard Datasheet.
https://frank.pocnet.net/sheets/129/e/EL34.pdf
there ist Vin 55V g1-g1.
Phillips give the Vgin in Vrms against ground per tube. Veff means Vrms and is dutch/german abreviaton for effektiv. Veff x root2 = Vpeak . 22,7V x1,414 = 32,09V peak . 2 Tube with 180 degrees means twice the Volts grid-grid 64,2V peak-peak
It is not a big difference between the grid ac Voltage of an EL34 or an 6L6GC/ 807.
An 807 can easily driven with an ECC83/12AX7 without clipping. Its got more swing than the 807 are need for full blast.
73
Wolfgang
The data sheet clearly said peak to PEAK grid voltage. Not just one peak, but twin peaks.
Think about it. The peak grid voltage is what is required to overcome the BIAS and take g1 to zero. You may actually clip at a bit less depending on your load line, which for a real speaker is elliptical and variable. So you do want a bit of overhead. Whatever vg1 you set the bias to is what you need to drive it with, give or take a volt or two. Then you don’t even really need a data sheet to determine your output stage gain or drive requirement. Set it deeper into AB and it takes less drive for full output, with the penalty being heat.
Ham radio operators often run true class B or C, possibly very low or even zero idle bias to get efficiency and the most possible output power. The g1 bias is more negative. And they often run into positive g1 territory to get even more power. That not only requires the extra voltage (maybe another 10 volts), but the driver stage must also provide g1 current. The usual capacitor coupled stages cannot do this - direct or transformer coupling is required. Most audio amps with pentodes run AB1 (or B1) because there are other ways of getting the extra current (ie, increase Vg2) and do not require a special driver stage to do this. The 807 (and 6L6 variants) were designed to take vg1 current with no damage but others audio pentodes may not be. Most audio amps I’ve seen here that run AB2 use triode outputs, where the positive g1 drive gets you similar output voltage swing to pentodes - eliminating the usual efficiency penalty associated with them.
I'm curious...
Since the 807 (and older pre-"GC" 6L6) have screen grid limited to 270V, wouldn't one be able to use these with a UL OPT if the plate supply is kept down at 270V to maybe 300V? While you won't get anywhere near max power that way, you could get perhaps 15W, with good damping (by running the pentodes at low voltage on the plate/screen grid but relatively high current, perhaps Vp,sg = 300V and Ip+sg = 75mA). A lower impedance primary could also be used for the OPT, which potentially makes design and manufacture easier/cheaper.
Is there a reason other than being limited to low output power that this will not work?
Also, how different is the screen grid construction of an 807 from the 7027A? The 7027A was pitched for use with UL OPTs, and it's basically similar to a 6L6GC (I'm not sure there is any difference, actually).
Honestly curious, which is why I ask. Thanks.
Last edited:
I can say I have run the 1625 which if the 12 volt version of the 807 UL for a few years without problem.
Operating with the Dynaco A470 transformer with about 410 volts. Likely about 400 volts on the plate and screens. Main use was as a car amplifier above 100HZ.
You have to remember that we are mostly worried about total screen dissipation when considering the voltage limit. Screen grid current increases as the plate voltage swings lower than the screen and towards 0 volts. With UL, the screen voltage drops with the plate so the worst case screen dissipation will not be as bad.
Operating with the Dynaco A470 transformer with about 410 volts. Likely about 400 volts on the plate and screens. Main use was as a car amplifier above 100HZ.
You have to remember that we are mostly worried about total screen dissipation when considering the voltage limit. Screen grid current increases as the plate voltage swings lower than the screen and towards 0 volts. With UL, the screen voltage drops with the plate so the worst case screen dissipation will not be as bad.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.