6U8 voltage amplifier and phase splitter in one tube

Mars

AVForums Super Veteran
Joined
Mar 23, 2009
Messages
1,342
Reaction score
137
Location
Paarl, Western Cape
Hi All

I am working on my amp building manual, and it needed a chapter on the 6U8 pre-amplifier and phase splitter. The idea is to explore the details of point to point wiring the 6U8 into a serviceable pre-amp for a push-pull amplifier. The 6U8 (also known as the ECF82)  is similar to the 6AN8, except the pinout on the tube base is different, and readily available from tube suppliers.

The idea to use a single tube as both voltage amplifier and phase splitter is attractive, and the high gain and low distortion of the pentode section makes it more attractive than the dual triode cousins. Dynaco use these types of tube even in their Mark III 60 watt monoblocks.

dsc03906o.jpg


See the minor differences below between a 6U8 and a 6AN8:

6U8
9ae.gif

6u8_057r.jpg


6AN8
9da.gif

6an8.jpg
 

charles

AVForums Super Veteran
Joined
Feb 13, 2011
Messages
1,476
Reaction score
130
Location
Pretoria
Dear Mars,

Both are triode, pentode valves but their characteristics are different. The 6U8 also delivers a higher output.

The 6U8 was used in the Radford STA 15 Series 3 and others.

There is interesting literature regarding the use of the 6U8 valves in the Radford power amplifiers. 

To me the 6U8 is a better valve.

Kind Regards

Charles
 

Ampdog

R.I.P. 23/06/2022
Joined
Aug 27, 2006
Messages
7,611
Reaction score
455
Location
Pretoria, South Africa
Karel,

The low distortion of the pentode is no longer dominant at the high peak output signal required for the typical use of the pentode feeding a cathodyne phase inverter for this application. Also, the distortion is predominantly 3rd harmonic as to 2nd harmonic for a triode input. 

Then something I do not find in data sheets for these valves are anode graphs for the triode part, so that one can determine which has the wider maximum anode swing. (They were intended for r.f. use which did not include such data.) The use as driver for providing up to 160Vpp output makes that important; I am not fond of using that topology for driving output valves directly - still.

From the somewhat higher Gm of the 6U8 triode one would guess that it will have the edge over the 6AN8 output swing-wise - but then it might be slightly less linear. One would expect that in power amplifiers the output stage distortion will be dominant, but that might not be so - EL34s in UL configuration can go as low as 2% distortion at about 50W output.

Still, as you say, a twin-triode topology for this type of work is just about out; also too low gain. There is also the ECF80 (6BL8) for this application, but it is less obtainable though cheaper when I last looked. So perhaps a quick hook-up to determine the maximum output voltage? (Sounds interesting, which I might examine myself over the weekend.) For now, I would prefer the 6U8, but conditionally so.

(This topology is of course ideal for the low grid drive requirement of valves like EL84, and a viable alternative might be using multiple EL84s as in the Manley amplifier, though that does not use a cathodyne phase splitter.)
 

charles

AVForums Super Veteran
Joined
Feb 13, 2011
Messages
1,476
Reaction score
130
Location
Pretoria
Very interesting reading regarding the application of the triode/pentode in Radford's EL34 valve amplifiers

http://www.audiomisc.co.uk/HFN/Radford/revisited.html


Kind Regards


Charles
 

handsome

AVForums Super Veteran
Joined
Jan 23, 2009
Messages
1,358
Reaction score
356
very interesting read indeed. ironically the biggest modification to the Dynaco power amps is of course the removal of the 6AN8 voltage gain/phase splitter. The split load phase splitter is (in my humble opinion of course) the best phase splitter having very low output impedance (when facing a matched load on each phase) but unfortunately needs tons of HT in order to accommodate both phases' voltage swing requirements - sometimes more than the output stage HT in more adventurous designs. Ampdog is correct in pointing out that paralleling many EL84s is an ideal way to get both higher power and sidestep the HT issue and EL84s are rather cheap.....
 

Ampdog

R.I.P. 23/06/2022
Joined
Aug 27, 2006
Messages
7,611
Reaction score
455
Location
Pretoria, South Africa
The relative merits of what is discussed in the article on Radford raises a new discussion outside the scope of this thread - only to say that there is a certain degree of comparing apples to oranges. Just one thing, if I may kind of brush the thread border, one must be careful not to, in comparison, mix types of valves and the circuits they are used in. Those are two separate issues.

Yes, the cathodyne/concertina phase splitter has advantages as stated, but which are often cancelled by the disadvantages of a high output first stage, also bringing into the mix the type of triode or input valve. A high gm moderately high ? triode is preferred, not provided by the likes of the 6SN7 and certainly not by the ECC83. (In my own estimate the ECC83 has little place in any power amplifier design. Whatever the application, it can always be bettered by something else. But that is another story.)

But I am thread-jacking.  I have yet to do my suggested tests on the 6AN8 vs. the 6U8. I have just discovered I am not the proud owner of a 6AN8 any more, only a 6U8 (ECF82) and ECF80 (6BL8). Comparison there is easy as they both have the same pin-out. But the comparison might turn out to be rather academic - I do not expect a meaningful difference in available maximum output amplitude capability. That would be rather more dependant on the h.t. supply as Handsome suggested. At some stage one will further have to consider the relative merits of an l.t.p. plus separate input stage - there is a reason why many top amplifiers prefer this, even at the penalty of an extra valve. But I am digressing again.
____________________________________________________________

PS: I am glad nobody has brought up the 7199 made 'specifically' for audio. An over-expensive valve if I recall correctly, with a rather non-linear triode.
 

Ampdog

R.I.P. 23/06/2022
Joined
Aug 27, 2006
Messages
7,611
Reaction score
455
Location
Pretoria, South Africa
Although of limited interest, I suppose I should present test findings (TECHNICAL WARNING!).

Firstly, I am going to summarise only because as expected one soon runs into 2nd harmonic distortion, and my distortion analyser is currently on the blink so that I could not test at equal output distortion (2nd harmonics are rather difficult to estimate by sight alone).

Contrary to previous statement the lower ? triode (in the ECF80) fared better. It was stupid of me not to realise that with lower ? triodes anode saturation voltages are mostly lower than with high ? triodes, thus a wider available anode swing.

I tested a 6U8 and ECF80. I do not have a 6AN8 at present, buit it is not very different from a ECF80.  B+ voltages of 300V, 350V and 400V were used and anode loads of 47K and 120K. The test signal was 1 kHz, although reponse is flat over the audio range. (Explanation: testing in normal triode operation was easier, apart from excluding the influence of cathode feedback which makes judgment of distortion (overload) by sight easier. The phase splitter arrangement then simply means dividing the anode load between anode and cathode. Thus the tested load of 47K would mean 22K each in anode and cathode, and the 120K equal to 56K each in anode and cathode.)

For the 6U8 the maximum outputs lay between 180 and 250Vpp with a 47K load (i.e. 22K + 22K)
For the ECF80 these figures lay between 180 and 280Vpp under above load.

For 6U8 the maximum outputs lay between 200 - 260Vpp with a 120K load (i.e. 56K + 56K)
For the ECF80 these figures lay between 230 and 320Vpp under above load.

BUT

The input pentode!

For the above conditions the required triode G1 dc levels lay between 50V to 85V for the split load configuration. For pentodes a low saturation anode voltage of 30V for low distortion is indicated - which then means that for the full above output swings to be realised from the phase splitter, some 120V minimum dc anode voltage is required for the pentode in order to properly feed the phase splitter. Thus, either an RC coupling to the phase splitter becomes mandatory, nullifying the direct coupling convenience, or with dc coupling the phase splitter maximum output will be limited by the pentode, not the phase splitter. No free lunch.

To sum up then: The convenient direct coupled concertina phase splitter is, for high fidelity work, limited to use with EL84s (ideal), 6L6GCs or, pushing it, EL34s, the latter two provided that a B+ of 400V minimum to the phase splitter is used.  For low distortion, use with KT88s or the like is out.

Dc levels in any final design will have to be 'balanced' for a particular pentode-triode with the use of a distortion meter for lowest distortion in this topology (afterwards preferably also including feeding the pentode g2 with dc feedback from the triode cathode instead of B+ to mainatin working points, as used in some designs - same number of components, correctly figured).

[OT: My personal preference: Use of the above only with EL84s. Two could be paralled for some 22W output in UL. Otherwise I lend toward the Schmitt/LTP with pentode input, despite increased cost.]

Karel,

Apology for length and if deviated from your original idea although you raised this configuration. I wish you "voorspoed" with your manual! 

 

Mars

AVForums Super Veteran
Joined
Mar 23, 2009
Messages
1,342
Reaction score
137
Location
Paarl, Western Cape
I searched the 'net for schematics using the 6U8 like I use in my kits, but the results were very limited. The Scott 208 uses a 6U8 as voltage amplifier/phase splitter. And that is about it. So let's look at the Mars Kit 1 schematic:
mars-dynaclone.jpg


Notice a few things:
  • pin 6 and 9 are connected with a wire bridge
  • pin 3 and pin 7 has a capacitor connected between them
  • pin 7 (cathode) is connected to ground via two resistors in series
  • power supply point C has two resistors connected to it going to pins 6 and 3
  • two 0.1uF caps and two 47k resistors are connected to pins 1 and 8

Nothing too mind boggling, but the completed job does look complex. :Ooooooh: So let us draw on paper what we have before us, and plan a layout.  :coffee:



 

Mars

AVForums Super Veteran
Joined
Mar 23, 2009
Messages
1,342
Reaction score
137
Location
Paarl, Western Cape
I've built an amplifier using the above 6U8A. I took a few photos of the progress:

Here I am ready to begin the 6U8 wiring

IMG_7095.JPG


Here you can see the earth bar with the connection to the steel chassis:

IMG_7096.JPG


Here I have mounted the coupling caps as well as the pentode screen grid decoupling capacitor underneath:

IMG_7097.JPG


Next I did the connections to ground, ie. the cathode resistors:
IMG_7098.JPG


Here I did the bridge between the pentode anode and the triode input grid

IMG_7099.JPG

IMG_7100.JPG


Next are the anode load resistors for the pentode:
IMG_7101.JPG


Some steps were done without me taking photos :whistler:

And the job is done

IMG_7104.JPG

IMG_7105.JPG


Amplifier finished
IMG_7110.JPG









 
Top