The Unison Research Simply Two Anniversary is a single ended valve integrated amplifier well known among enthusiasts of the “few watts but good ones” philosophy. It is based on EL34 output tubes and designed to prioritize circuit simplicity, pleasant harmonic structure, and a typically valve oriented sonic character. In amplifiers of this type, components often operate close to demanding real world conditions, especially as valves age and their parameters begin to drift from nominal values. For this reason, beyond the repair itself, it makes sense in the laboratory to carry out a careful verification of critical points, applying small conservative corrections that increase reliability without altering the original design philosophy and without producing audible changes in sound.
In this intervention I describe a typical case: a Simply Two Anniversary that arrived with loud popping noises on one channel, followed by a complete loss of signal. The main cause was identified in the cathode bias network and in the behavior of an EL34 that was clearly at the end of its life. Anyone who works on valve amplifiers knows this well: when an output tube is heavily worn, or simply no longer stable, a bias drift can occur that forces the valve to draw abnormal current, heavily stressing the surrounding components. The goal of this article is to show what I found, why the failure has its own internal logic from a design point of view, and which minimal measures I apply to drastically reduce the likelihood of the problem recurring with aged EL34 tubes.
This Unison Research Simply Two Anniversary was brought to me because, after working normally for some time, it suddenly began producing loud popping noises on one channel. After several episodes, the channel became completely silent. In situations like this it is important to proceed methodically: first a quick check of loads and connections, including connectors, wiring, loudspeakers, and any oxidation, followed by basic measurements on the valves and main supply voltages to determine whether the fault is confined to the output stage or also involves the driver or power supply section.
The first anomaly emerged when measuring the output tubes: one of the two EL34s was completely worn out. “Worn out” does not simply mean reduced emission, but often implies that the tube no longer behaves predictably under voltage and temperature, with a real risk of current and bias instability. At that point, opening the amplifier and performing a visual inspection confirmed that something in the output stage had been operating under conditions well beyond normal. When I opened the unit, I found the cathode resistor in the following condition:
In a single ended amplifier, the cathode resistor plays a fundamental role: it sets the automatic bias of the output tube and dissipates a non negligible amount of power. If, for any reason, the anode current of the EL34 increases beyond its intended value, that resistor is one of the first components to suffer thermal stress, eventually failing. In this case the resistor was clearly damaged, a sign of abnormal operating current sustained over a significant period of time.
As also happens on the S2, from time to time, when valves are old, a bias drift can occur that causes the output stage to draw current in a destructive way. It is important to state this clearly: this is not someone’s fault, nor a reason to point fingers, it is a behavior that can occur in a self bias circuit when the output tube is no longer stable. The fact that the design specifies a 400 volt cathode bypass capacitor is an interesting clue. That component is connected in parallel with the resistor and must survive even under abnormal conditions. If the resistor opens, the cathode can be driven to unexpected potentials, and an underrated capacitor could fail dramatically. In other words, the choice of a high voltage rating is consistent with the idea of making the failure manageable rather than catastrophic for nearby components.
An important detail that emerged from the examination is this: the output tube was probably not already completely dead at the moment of failure, but became so as a consequence of the abnormal condition. When the bias drifts and the valve is forced to deliver extremely high current, internal dissipation rises, temperatures increase, and degradation accelerates. This is the classic scenario in which an EL34 that is already at its limit goes from “still working” to “finished” in a short time, taking the cathode resistor with it.
After replacing the failed resistor, in addition to restoring normal operation I applied a preventive measure based on years of experience with EL34 output stages, especially as the tubes begin to age. In practice, I checked and corrected the value of the control grid leak resistor, limiting it to a maximum of 220kohm. Over time I have observed that keeping this value within 220kohm significantly reduces the tendency for operating point drift and abnormal behavior with worn tubes, effectively preventing the chain of events that leads to uncontrolled current increase and stress on the cathode resistor. For this reason, in this unit I adjusted that value to stay within this limit. It is a minimal and non invasive intervention, it does not alter the philosophy of the circuit and, in practice, it does not produce audible changes in sound, but it does improve operational robustness under real world conditions, meaning also when the EL34s are no longer brand new.
This approach is the same I apply in other similar interventions: small, targeted, and well motivated reliability oriented corrections that do not overturn anything but significantly reduce the likelihood of a recurrence. In short, the amplifier must continue to sound like a Simply Two, not become a different project, but it must also have a safety margin when valves inevitably age.
I replaced the damaged resistor and, as also done on the S2, I modified the value of two other small resistors in the circuit that, based on experience, help prevent the issue from happening again with aged valves. The logic is the same here: small, reversible interventions aimed at improving reliability, not at reshaping the sound.
After completing the repair, I carried out the standard functional checks: verification of main voltages, confirmation of the operating point at temperature, absence of abnormal noise, long term stability, and regular response on a dummy load. I installed two new EL34s and the amplifier returned to correct operation on both channels.
As for output power, the behavior is fully consistent with an EL34 based single ended design. The circuit typically reaches a clean power level of about 6 to 7 watt before waveform clipping becomes evident. If pushed further, output power can rise toward 12 watt and sometimes even a bit more on the meter, but this occurs under saturation conditions, not in linear operation. This is not a defect, it is the nature of this topology and of the single ended compromise: few watts, but delivered with a harmonic structure and texture that many enthusiasts seek precisely for domestic listening with suitable loudspeakers.
Below I report some indicative measurement screenshots, mainly to document that the amplifier has returned to full functionality and to provide an objective reference after the intervention. As always, these graphs must be read in context: they describe behavior on a load under test conditions, they do not by themselves explain listening pleasure, but they are essential to verify that everything is stable and consistent.
THD 1 watt
The spectrum at 1 watt is useful to confirm the absence of spurious components, abnormal hum, or other irregularities that could indicate instability, oscillations, or power supply issues. In particular, after a failure related to bias drift, it is important to confirm that the output stage operates regularly and that no intermittent behavior appears when hot.
Sine wave 1 kHz at 3 watt
The 1 kHz sine wave at 3 watt shows the waveform under a significant operating condition that is still typically within the linear region for this type of circuit. It is a simple but very revealing check: asymmetries, premature distortion, or packet like instabilities often appear clearly here, especially if there are residual issues with the valve or with components related to grid biasing.
Bandwidth at 1 watt
The bandwidth measurement at 1 watt on a resistive load is used to verify that the output transformer and the entire output stage do not show anomalies after the repair. In a single ended amplifier the transformer is a central element, and any significant overcurrent condition can, in the worst cases, thermally or magnetically stress it. The goal here is to confirm that the response remains regular and predictable, without humps, premature roll off, or strange behavior at the frequency extremes.
In conclusion, the intervention on this Simply Two Anniversary was a classic example of repair combined with practical safeguarding: replacement of the failed component, replacement of the EL34s, instrumental checks, and the application of a minimal but effective corrective measure based on experience, to reduce the likelihood of bias drift in the presence of aged valves. If there is a practical takeaway, it is this: in EL34 based single ended amplifiers, thoughtful valve maintenance and small precautions on critical values, such as keeping the control grid leak resistor within 220kohm, can make the difference between an amplifier that occasionally misbehaves and one that remains stable and reliable over time.