EL86 Single Ended PP Power Amplifier

[Irakli]

Hi All,

I am starting the construction of somewhat unusual power amplifier. The ONLY reason I am doing the project is that I have a pair of toroidal power transformers laying around for a long time which I want to use as output ones. The primary windings of these transformers are not center tapped. As a consequence I have very limited options as far as topology choices go.

Again, This is the power amp project for output transformer with no CT and fairly low impedance ratio 1.5K:4 ohm (see below). For any other circumstances I could not recommend this design.

Some additional constrains are:

1. Relatively low (<400V) B+ voltage. This way I can use components rated at 450V (caps, CCS).
2. using triodes.
3. Relatively high power efficiency. Using triodes gives dismal power efficiency regardless what you do but I still wanted the best possible.
4. balanced design (that's just me).
5. No feedback.
6. Simple and cheap.

I considered six topology choices: parafeed, totem pole, SRPP and their bridged versions. Totem pole needs complex driver circuit arrangement so I did not consider it further. Parafeed has prohibitively low power efficiency (something like 5W out with 85W consuming ).
SRPP is the worst choice for power amp because the circuit not only requires constant load but also can be optimized for only one value of load impedance.

After some calculations and simulations using LT Spice I selected the least offending performer as far as power efficiency goes: SRPP bridge with EL86 tubes. The prototype schematics is attached (I have not tested it on bredboard yet).

SRPP in each half of the bridge is optimized for 725 ohm load (giving Rk=333 ohm)

Here are the simulated values (I do not put much faith in simulators but they are adequate for the order of the magnitude estimations):

VB+=400V
P total dissipated =40W
P out = 4W
efficiency 10%
distortion : all harmonics on FFT plot are less then -50db (useless measurement by itself because Spice models are inaccurate. However, good guide to keep input signal at certain level to compare different topologies).

I selected EL86 tubes because they work well with lower VB+.

Comments are appreciated. If you have suggestions please let me know before I start assembling a circuit.

Thanks

[Shoog]

The DC blocking cap needs to be big and none polar. This can be achieved by using back to back electo's with a 1meg resistor at the negative leg cap node, bypass with a film cap.

Shoog

[Irakli]

Thanks Shoog,

I am thinking 15uF pp cap (three 5uF Solen pp caps in parallel because I have plenty of them).

[Zibi]

Ra (R1,R2) ? What for?

[Irakli]

Hi Zibi,

RA is needed to balance currents in the load (it reduces transconductance of upper tube slightly). For high mu tubes it can be omitted. Ideally for unbypassed RK, RA=2*RK.

The Phillips schematic you are referring is different, EL86 operates in the pentode mode.

[Zibi]

Why current balance is needed if blocking capacitor is used?

[hidnplayr]

You could also build a circlotron!
But with EL86, the impedance of your transformer might be a little high...

[Irakli]

Originally Posted by hidnplayr

You could also build a circlotron!
But with EL86, the impedance of your transformer might be a little high...

True, but at the expense of the power supply complexity.

[Irakli]

Here is the power supply for VB+=400V

[tricomp]

A few thoughts ...from a non-pro, which can probably be spotted

How about core-saturation for a toroid in SE? A problem, as no air-gap?

How big's the toroid? Will it handle bigger tubes? Even a 300B runs fine on 400V.

You want more Watt from a SE? Other OP-tube(s) in (P)SE and a CF to allow for low-Z drive/grid-current or is that not kosher?

rgds,

/tri-comp