Single ended EL34 amp design

[Stephen1966]

Here's a thread I found interesting as well: https://ampgarage.com/forum/viewtopic.php?f=6&t=30432

[cdemike]

Good article on the SE design based around an EL34 here: http://www.valvewizard.co.uk/se.html

cdemike, have you seen this? https://el34world.com/Forum/index.php?topic=21867.0

I haven't seen that -- thanks for passing it on! I'm consistently surprised by how much gain is possible from triodes. I'm slightly concerned my approach may cause issues with too much output going into the power amp section. I'd hate to derail your thread, but after reading some more, I'm almost wondering if I should go with a different approach to the "big sound, low volume" problem. I'll likely create another thread, but I'm definitely interested to see where your amp goes!

[Stephen1966]

Good article on the SE design based around an EL34 here: http://www.valvewizard.co.uk/se.html

cdemike, have you seen this? https://el34world.com/Forum/index.php?topic=21867.0

I haven't seen that -- thanks for passing it on! I'm consistently surprised by how much gain is possible from triodes. I'm slightly concerned my approach may cause issues with too much output going into the power amp section. I'd hate to derail your thread, but after reading some more, I'm almost wondering if I should go with a different approach to the "big sound, low volume" problem. I'll likely create another thread, but I'm definitely interested to see where your amp goes!

No worries! Of course, if your own discussions are going to branch off in all kinds of ways it's worth posting your own thread - if only for your own sanity It doesn't bother me when people bring alternative ideas to the table, I've been known to derail my own threads from time to time.

By 'big sound/low volume' I am thinking Class A and looking for a higher fidelity tone with more abundant second [even order] harmonics. I've only built push-pull amps before though so this is something of a learning curve... to be continued.

[Stephen1966]

After some thought, and some study, I am thinking of something like the block diagram for the "Champ revisited" for a Class A, SE amp driven by 3 ECC83s and an EL34. Reverb can be incorporated later... I am won over by the idea of the Class A tone, biased just a little on the hot side to give nice, second (and even order harmonics) but with an EL34 to give it in Merlin's words a somewhat "aggressive character" when it starts to bite.

I also want to try a regulated DC supply, either the one detailed in the Champ blog, or one of Merlin's alternatives.

The first step, having decided on the kind of tone I'm chasing, was to think about the transformers and the required power output. I don't mind if it comes out on the lower output side of 8-11W as the idea is for a versatile practice amp which could make itself useful in the studio. The closest thing I've built to this before has been the Tweedle-Dee, and that amp is a window-rattler.

First choices then were to select the PT and OT.

Power transformer.
As I'm intending to run the filaments on their own transformer, the Hammond 290CAX seems a good candidate with the choice of 275-0-275 or 315-0-315 +HT taps. The filament coil would go unused because it wouldn't be enough on its own to supply three preamp tubes and the EL34 at just 2A but it has a 5V/2A coil as well and that will come in handy for driving relays if I go that route.

290CAX.pdf

Output transformer.
For this, I have a choice between the Hammond 125ESE (15W) or the Hammond 125FSE (20W). I don't know which yet but they both offer the option of primary impedances of 2k5, 5K and 10K; also, the usual secondary taps of 4, 8 and 16 Ohm. Presently, I am thinking it would use the Zin of 5k and Zout of 8 Ohm. There is a ton of reading out there about the selection of speaker impedance and its effect on tone but I am going on the fact that 8 Ohm speakers are relatively plentiful and there are some good 10" out there.

125FSE.pdf

The Zin of 5K also seems to support a nice amount of second harmonic distortion with the EL34 looking at the load lines I've drawn.

[Stephen1966]

LOADLINES

The Zaerix tube I am going to use was probably manufactured in Germany (GDPR) by RFT Mullhasuen. That (I suspect) puts it on a par with a Mullard of the earlier era and so for working out the loadlines I have used the Mullard datasheet. What I have been able to glean about this tube so far is that it was commonly used in the early Marshalls and that it has a warm low and mid-low end with clean highs. The tube looks in good condition with no oxidization on the getter. Zaerix, as I understand it, was solely a distributor, based in England, taking in ex-NATO and East European stock and offering them for trade and retail. They seem to be readily available at the moment as NOS tubes and several people I have spoken with have them already, though some US distributors are asking bigger money for them.

That aside, drawing up the load lines for this circuit has been a PIA. Is drawing loadlines becoming a lost art? I begin to wonder because when I came to finding guidance on drawing up the loadlines for power output pentodes it was, shall we say, not easy. I will come clean and confess this is the first time I have attempted this and so it goes without saying, I would warmly welcome any peer review on the ideas I arrived at. For the main part, I followed Merlin's tutorial on designing the SE output stage linked earlier. I didn't find RobRobinette's guide that helpful but that is not to say that wouldn't work as equally well. For my own part, before I started this, I clearly hadn't understood the proper relationship of the plate to the screen grid - hence the need for this process. So please, if you see anything amiss in the following, feel free to chime in.

I should begin with the loadlines as they looked after my process.

Mullard el34.jpg

Mullard el34 transfer characteristics.jpg

[Stephen1966]

Before I continue, I should link the other pages and blogs I found useful:

My primary source has been Merlin's tutorial: http://www.valvewizard.co.uk/se.html

Along with: https://www.diyaudio.com/community/thre ... st-1119171

This was useful for context as well: https://www.analogethos.com/post/tube-a ... ined-intro

I also consulted the 'Radiotron Designer's Handbook,' 4th Ed. as well as Merlin's books, 'Designing Tube Preamps for Guitar and Bass,' and 'Designing Power Supplies for Tube Amplifiers'.

I will try to keep this as brief as possible.

The load on the plate is reactive, not resistive, but to draw the load line we treat it as resistive - the same as if we were drawing a loadline for any other tube using a fixed resistor. Playing around with the values is instructive.

When the +HT is is 400V the Zin of the available OT exhibits the following behaviour:

• Ia = 400/2500 = 160mA

• Ia = 400/5000 = 80mA

• Ia = 400/10000 = 40mA

When V = 0 (resistive load)

This tells us that the lower primary impedance of the transformer takes the valve closer to Class B and saturation so a 5k primary was chosen. It does not matter which +HT voltage is now chosen, the loadline will remain at the same gradient, i.e. as the voltage increases, the current increases proportionally.

To find the ideal transformer impedance for centre biasing:

Z = Va^2/Pa

When

• Va = quiescent anode voltage = 340V

• Pa = max. power dissipation = 25W

Zout = 340^2/25 = 4624 Ohms

The higher input impedance of the OT, at 5k, biases slightly "hotter" in Class A, providing more second harmonic distortion.

The idle current with this, is:

• 340/5000 = 68mA

This permits us to draw the quiescent loadline (blue line) seen in the graph above.

We transpose this to a point just below the safe operating area. Then, draw a line vertically (in magenta) from max Va to intersect the loadline giving the bias point (blue dot) of 65mA; just below the peak output. Merlin explains why we can do this.

Edit:

SELECTING THE OPERATING POINT

This was a step I should have included earlier but having drawn the first (blue) line on the anode characteristics graph where the +HT (or Va) is 340V and the Ia is 68mA, Merlin assures us we can ignore the DC resistance of the OT and simply draw a line vertically from the Va of - in this case - 340V.

Where this line intersects the maximum dissipation curve (the boundary of the safe operating area) represents a point at maximum plate dissipation or a bias point of 100%. We want to extend the life of the tube so while we can safely bias at 100% and the story is complicated by the fact that the load is actually reactive and non-linear, the received wisdom of these cathode biased circuits is that 90% is a safer choice.

To do this we draw a line vertically up the maximum dissipation curve then draw another line to the left and read off the resulting anode current: in this case when Va = 340V, Ia = 74mA give or take. Simply multiplying 74 x 09 = 66.6 mA. We might NOW transpose the blue line up and across (keeping the same gradient) to intersect the vertical line at 66.7mA and that (magenta) line is now our load line, roughly centre-biased. I did it by eye and arrived at 65 mA for the operating point which comes out around 88% - close enough for jazz!

[Stephen1966]

SCREEN VOLTAGE
This was the most troublesome to calculate as Merlin's logic seems predicated on some lessons I appear to have skipped but after several attempts and interpreting his tutorial as best I could, I found the following...

To give this context we need a schematic of the intended circuit.

IMG074.jpg

*Rg3, the 'screen grid stopper' is included in mine; it isn't designated in Merlin's tutorial but I think it's clear from his context.

We make an assumption here that the screen grid current is "roughly a fixed ratio of anode current" which, from the Mullard datasheet gives us:

• Ia = 100 mA

• Ig2 = 14.9 mA

Merlin appears to have been using the Phillips datasheet, but going with the Mullard figures:

• m = 100/14.9 = 6.7

m, being the ratio of anode current over screen-grid current.

When

• +HT = 340V

• Ia = 65mA

Ig2 = 65/6.7 = 9.7mA

A 1W or 2W rating in the dropping string appears to be more than ample.

Assuming the three preamp tubes draw 5mA, we get the total current running through Rg2:

• 9.7 + 5 = 14.7mA

Which we will call 15mA for argument's sake.

Now, the voltage drop across Rg2, when Rg2 = 560 Ohm

• Ig2 x Rg2

• 0.015 x 560 = 8.4V

The power dissipation of this resistor is calculated as

(0.015^2) x 560 = 126 mW

A 1W or 2W resistor in the B+ dropping string would appear more than adequate.

[Stephen1966]

SCREEN VOLTAGE: Rg3 - the screen grid stopper

This one gave me the most trouble because Merlin's last schematic doesn't tally up with his calculations and I had trouble interpreting his results. However, following Merlin's process for transposing grid curves down, to find the screen grid stopper value, I calculated what follows. The purpose of the screen grid stopper is to "limit dissipation," and to induce sag in what is described as a "sliding screen operation." Basically to prevent the screen from going nuclear under sustained peak conditions.

For this to make sense, referring to the grid curves in the plate characteristics chart above, it appears to require that we take our orientation from the -8V curve which corresponds to the voltage drop across Rg2 (8.4V), it further requires us to treat this as the new Vg1 = 0V curve. This requires that we transpose the -8V curve down (2 steps) tp the -16V curve which we now treat as the new -8V curve - are you with me so far? We now take our measurement from the intersection of the loadline with the -16V curve (standing in as the -8V curve). This curve corresponds approximately, to the voltage drop across the screen grid dropper (Rg2). The intersection, marked with a green dot, now gives us the figure:

• -8V, 115mA

Plotting this on the Transfer Characteristics chart (green dot) we see an indicated screen voltage, rounded up to 240V. Indicating, under peak conditions, the screen voltage drops:

• 340V - 240V = 100V

To get this value of resistor using the voltage drop and peak anode current (130mA), we have:

• R = 100/0.130 = 769 Ohms

And so using a bit of judicious rounding up for the available resistor values we arrive at 820 Ohms

Power dissipation again...

(0.015^2) x 820 = 184.5 mW

Granted, this is mostly an intuitive understanding at best, so I would welcome any clarification if you see I've miscalculated.

[bepone]

After some thought, and some study, I am thinking of something like the block diagram for the "Champ revisited" for a Class A, SE amp driven by 3 ECC83s and an EL34.

go for it

[bepone]

from operating point above, 350V, 0.065A,

Rk needed for bias =470R 5W-10W, including g2 current in calc.

Ck-from 47uF till 470uF / 63V depending of desired bass content

Rg1stopper- 1-2k carbon comp
Rg1 to gnd, max 100k

[bepone]

Rg2 and Rg3 is a mistake, this is 1 resistor only

[Stephen1966]

Rg2 and Rg3 is a mistake, this is 1 resistor only

I am quoting Merlin here:

Screen Voltage: The screen voltage is usually set by a dropping resistor in the HT supply (Rg2) (see the section on smoothing and filtering), or a choke, plus a small screen-grid stopper.

The italics are my emphasis. Schematics are logical but not spatial and let's say: that Rg2 and Rg3 are not a mistake. It might be better to say that Rg2 is normally part of the B+ dropping string on the power board... if it isn't incorporated into the design as a choke. I think we often see the choke only in its primary role as a smoothing choke, but it is also a resistive load and induces a voltage drop; virtually the same as a fixed resistor would. Rg3, the screen-grid stoppers are typically (in Dumble's layouts at least) installed right on the tube sockets connected to pin 4 for the screen grid, (g2) across to pin 6 (n.c.).

For sure, we wight have come up with a better designation - Rg2 as Rht and Rg3 as Rg2, or something like that for example - but I was just wanting to get around the fact that Merlin doesn't include the second resistor in his first schematic and this is one of the difficulties of his tutorial - it isn't always clear what is going where...

[bepone]

anyway you will mount this resistor on the pin 4 directly and it will have function of dropping and grid stopper in the same time.. but you dont need to worry too much, put there 1kohm and finish the story, i dont believe that you need some dropping resistor there for SE amp at that "low" anode voltage (350V)

[bepone]

also there is detailed vintage telefunken datasheet of EL34 with many output stages defined, just pick up one which you like, also schematics are provided, so no need to think too much for that simple thing

[Stephen1966]

anyway you will mount this resistor on the pin 4 directly and it will have function of dropping and grid stopper in the same time.. but you dont need to worry too much, put there 1kohm and finish the story, i dont beleive that you need some dropping resistor there for SE amp at that "low" anode voltage (350V)

Agreed! The general wisdom seems to be that the values are not too important. And yes, we could treat the dropper and the stopper as one and the same. It's 340V in my reckoning. I am going to use both though because I like the idea that there is a little overhead - a few volts difference provided by the dropper for the stopper to sag and provide the sliding screen operation. The screens are the most delicate part of the valve and it offers greater protection this way.