Q-tron Audio bias calibrator technical background
Technical background about adjustment of bias, (operating point)
Aug 22, 2021 · 50 minuter läsning

Q-tron Audio Automatic Bias Calibrator

Automatic control of bias and offset in our OTL amplifiers and automatic

tuning when changing power tubes

• Automatically adjust bias and output offset voltage to optimal values

• Controls the real quiescent current, (other "Autobias" circuit instead controls the average current which introduce cross-over distortion, see below for an in-depth analysis).

• Automatic tuning of bias when replacing power tubes

Technical background about adjustment of bias, (operating point)

For optimum operating conditions of a tube amplifier, we want the operating point to be stable and therefore the quiescent, (idle) current in a power amplifier should be kept constant and at the designed value.

If the quiescent current is lower than the design value distortion will increase, and power will be lower than intended if the quiescent current is too high tube life will be reduced.

In push-pull amplifiers, it is also important that both tubes are set to the same quiescent current as otherwise, even-order distortion will increase.

The quiescent current or idle current in a tube is most often controlled by the grid voltage setting which then can be adjusted to a given operating point of anode current, this adjustment is often referred to as "setting bias".

Drift in tubes operating parameters

Electronic tubes can drift, meaning that anode current and operating point may vary over time, this variation can be due to:

  • ageing, (emission decreases when tubes age)
  • heater voltage variation
  • anode voltage variation
  • due to internal heating, (internal heating can cause expansion of anode sleeve or changing the distance between cathode, grid and anode thereby changing parameters).

There is a common misconception that tubes themselves drift all the time, i.e., that parameters are not stable after reaching equilibrium conditions, this is not at all correct! Drift due to ageing is very slow, a typical power tube has a life expectancy of 1000's hours so the drift day by day is very small.

Methods for reducing drift in tube amplifiers

Reducing drift due to variation of anode and heater voltage variation 

In an amplifier connected to mains voltage, heater, grid, and anode voltage vary if the mains voltage varies.

Heater power, anode and grid voltages have a large effect on anode current and a common problem is that anode current therefore varies with mains voltage.

When mains voltage increases both grid and anode voltage increase and results in decreased anode current, and increased anode current when mains voltage decreases. This effect is somewhat counteracted by the fact that a change in mains voltage will also change heater voltage, but this doesn't eliminate the effect on anode current due to a change in mains voltage.

However, it is possible to eliminate anode current variation due to mains voltage variation by simple means and we believe it should be done in every tube amplifier. This doesn't need to include stabilisation of any voltage but rather modifying the effect the different voltages have on each other so that the net effect is stable anode current even if mains voltage varies.

However, it is very rare to find any compensation circuits in most commercial tube amplifiers which we find surprising.

Description of circuits that (intend) to automatically adjust bias in tube power amplifiers and why they don't work as intended

Several so-called auto-bias circuits have been invented and described earlier; they normally consist of a DC loop that aims to stabilise the measured average anode current by comparing it to a fixed value. The DC loop must be filtered at a very low frequency but many of these loops seem to work well if the amplifier is biased in class A where the average anode current theoretically is constant.

However! even in class A amplifiers these loop circuits don't work as intended!

Average current is not the same as quiescent current, average anode current in a class A amplifier is constant in theory but not in real amplifiers!

An amplifier in Class A will still have shifts in average current caused by second-order distortion in the output stage, normally the average current will increase as current from 2nd order distortion is added to the quiescent current.

This will affect a loop bias circuit as it will try to keep the total average current constant and therefore the quiescent current will be reduced, and the result is cross over distortion on sustained loud passages and after sustained loud passages as the loop is slow to react.

The problem arises from the fact that a loop cannot separate quiescent current and the additional current from 2nd order distortion!

Cross over distortion is often said to be the main reason why transistor-based amplifiers have harsher sound than tube amplifiers so naturally, we want to avoid this as much as possible. Tubes compared to transistors have a much smaller tendency of cross over distortion, (tubes don't switch off abruptly at low currents which transistors do) so the idea to deliberately introduce a cross over distortion mechanism in a tube amplifier we think is a really bad idea.

For class AB or B amplifiers, the problems get worse as anode current varies also with signal level, the common method trying to solve this problem with a loop bias circuit is to clamp the measured anode current so that the circuit does not react when current increases above the set quiescent current value. Many traditional tube amplifiers biased in class AB have a small difference in current at no or full signal so then the clamping method give a fairly well-working circuit, however, cross over distortion is still a problem and often gets worse in a class AB amplifier. 

In amplifiers biased in AB with low quiescent current or amplifiers biased in class B the difference in current with no or full signal is higher and the problem with induced cross-over distortion gets even worse, (as the additional current from 2nd order distortion can be of similar value as the quiescent current, the circuit can adjust the bias to a point where quiescent current is almost zero which then introduce heavy cross-over distortion.

A new look at the bias stabilisation problem

Analysing the problem of bias stabilisation thoroughly gives the following observations:

  • There are 2 main problems:
  1. Drift due to changing supply voltages

  2. Drift due to ageing

Summary

  • Drift due to changes in power supply voltages can be controlled by simple means as we have done in all OTL amplifiers.
  • Drift due to ageing is very slow, a typical power tube has a life expectancy of 1000's hours so the drift day by day is very small.
  • Common described auto-bias circuits don't work as they can't stabilise the quiescent current and therefore creates bad sounding cross-over distortion.
  • For all amplifiers, including class A it is not possible to measure and control the quiescent current while playing music, control of quiescent current must be done while there is no signal.
  • All tube amplifiers need heating time before they reach optimum operating conditions, often up to 30 minutes or more. This means that if the quiescent current should be adjusted before playing music the listener would need to wait this time, obviously, this is not acceptable.

From these observations we draw the following conclusions:

  • There is no need to constantly control the quiescent current, drift due to ageing is very slow and other drift can be controlled by other means.
  • Control of the quiescent current must be made when music is not playing to make sure that it is the real quiescent current that is measured and adjusted.

Q-tron Audio has developed a unique circuit for controlling the real quiescent current and output DC offset in our OTL amplifiers.

The circuit which is based on a microprocessor automatically measure and if necessary adjust quiescent current and output DC offset just before the amplifier is being switched off. The circuit works in the following manner:

When the power switch is pressed to switch off the amplifier the microprocessor circuit will start measuring the quiescent current and output DC offset and adjust these if necessary, when these values are correct the amplifier will switch off completely. The whole procedure normally takes less than 1 minute even if the bias values need to be adjusted and the values for bias are then stored in a non-volatile memory. Normally the bias values don't need to be adjusted at all and then the amplifier will switch off immediately after measuring current and offset.

To ensure that the amplifier has been thoroughly warmed up and have reached its stable condition when the adjustment is performed, the circuit will only perform calibration if the amplifier has been switched on and operating for at least 30 minutes before it is switched off.

When the amplifier next time is switched on the stored values for bias will be applied.

The circuit will also give an alarm if it is not possible to adjust bias correctly, if so an orange "service" LED is lit, the amplifier can still be used but will not be optimally adjusted.

If any value is completely out of specification even after adjustments, (as in the case that a tube catastrophically has failed) there will be a red "alarm" LED lit and the amplifier can then not be used for playing music, (the inputs and outputs will be shorted) until the fault has been corrected.

Automatic bias tuning of new power tubes

The bias auto calibrator also automatically set correct operating parameters for new power tubes.

When old power tubes are exchanged for new the owner should set the amplifier in test mode and switch the amplifier on. The amplifier will then heat the amplifier to stable conditions and automatically adjust the bias and offset to correct values, when the amplifier next time is switched on the stored values for bias are applied.