Virtual Earth Mixing and Alternative Output Transformers

I mentioned in the previous post that the gain setting resistors of the channel amplifiers have been zero referenced making virtual earth mixing a possibility. I have now received prototype PCBs of the V2 Twin Line Amplifier (TLA) which incorporates the relevant improvements added to the channel amplifiers:

• Provision for Molex KK connectors for internal wiring
• Amplifier gain preset pots moved to be accessible from the front panel
• Tubes moved so as not to foul the enclosure
• Ground reference the gain setting resistors so each amplifier can be used as a virtual earth mixer

In addition, a couple of other improvements have been made:

• Inclusion of a series 1N4007 in the HT+ line to protect against accidental reversal of the HT supply
• Addition of a 100nF capacitor across the HT supply where it enters the PCB to prevent the instability that can occur when several modules are connected.
• Moved one output capacitor so as not to be directly above a 6922 which makes it too hot.

The new TLA V2 is shown below:

Apart from the modifications listed above, the new TLA is functionally identical to the original TLA and can be used wherever the original was used. The one thing this TLA can do that the original cannot is virtual earth mixing and this is what I wanted to test out. I used my new test rack (shown below) to feed the output of the new channel amplifier to one of the buses via a 47K resistor and connected one of the new TLA virtual earth inputs to the same bus. I measured the output of the the channel amplifier and also the output of the TLA virtual earth amplifier. They were within 0.2dB of each other. Allowing for tolerances in the two output transformers included in the test circuit, this is a good result, implying the virtual earth mixer gain is very close to unity as expected, However, when I measured the noise, the result was rather disappointing at only -63dBu; clearly not good enough for any serious mixer. I checked the output of the channel amp in case this was the source of the noise but this was much lower so the noise clearly was introduced by the TLA. I then realised I had forgotten to connect the +ve input of the TLA (the one used for passive mixing) to 0V. I added the necessary link and the noise dropped below -80dBu.

This means virtual earth (VE) mixing is now a potential alternative to the normal passive mixing I use in my mixers. The reason I use the word 'potential' is that VE mixing has it own set of problems, the most serious of which is instability caused by bus capacitance. This is particularly problematic where op amps are used. They have a fixed open loop gain but the closed loop noise gain depends on the number of sources feeding the VE. This means the amount of negative feedback (NFB) varies which affects stability. Steve Dove, in his excellent series on mixing console design goes into more detail about this problem. To paraphrase Steve:

"Bus capacitances, from all the cabling and PCB tracks between the channel amplifiers and the virtual earth amplifier, appears across the input of the virtual earth amplifier. This has the effect of eroding the phase margin and can lead to instability and even oscillation. This capacitance reacts against the feedback impedance to to increase the closed loop gain at high frequencies. Even a few pF is enough to tilt up the closed loop response well within the open loop parameters, threatening instability. In a real mixer with cables from many channels, hundreds of pF may be present. This makes ensuring the required phase and response characteristics very difficult. Sometimes a small series limiting resistor can be added to to define just how much this unwanted gain can rise, but this is at the expense of the ‘virtual earth’ now being determined by this resistor (which rather defeats the object)."

There is another solution to this problem and that is to vary the open loop gain as the number of channels varies. Regular op amps do not do this but many years ago a discrete op amp was designed that does exactly this. It is called a Trans Amp (note that the term Trans Amp is also used to refer to other topologies so this can be a bit confusing when conducting an on-line search). The basic idea is that the resistor that sets the closed loop gain also sets the open loop gain. If this is done correctly then the difference between the open loop gain and the closed loop gain (which is the amount of NFB applied) is practically constant which means the stability criteria are the same no matter what the gain.

This is exactly what the EZ Tube mic pre does and for similar reasons. In the mic pre, we want to be able to vary the gain over a wide range ( in this case from 6dB to 40dB) whilst maintaining stability. The EZ Tube mic pre achieves this by gradually shorting out the cathode resistor of the first tube stage of the mic pre. This gradually increases the gain of the first stage as the resistor is reduced and also gradually increases the closed loop gain. The net effect is the NFB is nearly constant and stability is assured. The TLA uses exactly the same amplifier so it also automatically adjusts its open loop gain as its closed loop gain is varied so it should make a stable VE amplifier that is insensitive to the number of channels feeding it. Note that varying the open loop and closed loop gain by gradually shorting out the first stage cathode resistor is not a new idea. The first instance I know of its use is in the V76 amplifier.

Why are we even thinking of using VE mixing when passive mixing already works very well. There are several reasons:

• Improved crosstalk.
• Reduced requirement for low source impedance bus drive.
• No need to ensure unselected bus signals are connected to 0V.
• No need to adjust the mix amp gain depending on the number of channels

The only disadvantage is that you cannot easily include a mix group fader directly across the bus before the mix amp as you can with passive mixing. This means the mix group fader has to come after the the VE amp. As this cannot drive an output directly you need another amplifier to buffer the fader and drive an output transformer. The only alternative would be to have a transformer directly after the VE amp and use a 600 ohm balanced attenuator at the output as the mix group fader.

On balance, VE mixing has a lot to recommend it so I plan to incorporate it into the 8 tracker build,

Which brings us rather neatly onto the topic of output transformers and their ability to drive 600 ohm loads. One of the most expensive items in the EZTubeMixer design is the input and output transformers so I have been looking for lower cost alternatives. Edcor is well known for its low cost transformers so I got a couple of their XSM 2.4K/600 transformers (listed at $12.97 compared to the £25 for the Carnhill VTB2291). The XSM 2.4K/600 is physically smaller than the VTB2291 normally use in the EZTube Mixer but it is rated at 2.5 watts which is well over what we need. The primary inductance measures 28H at 100Hz and the secondary is 9.8H at 100Hz, both satisfactory values. I have given this transformer a thorough test and I am pleased to report it performs almost identically to the VTB2291. In both cases, the EZTube output amplifier runs out of steam before the transformers do so I am happy to recommend this as a cheaper alternative. At +26dBu into 600 ohms the distortion with the Edcor was 0,49% and at +29dBu it became 1.1%. These results are identical to the performance of the VTB2291. Frequency response was -1.8dB at 20Hz but this is entirely due to the 4.7uF output capacitor which is 3dB down at 14Hz with the reflected secondary load.

To facilitate testing of new 6U and 3U modules and PCBs I have built a sub-rack based test rig.

On the left is space for two 6U modules and the top right row can accommodate four 3U modules. Underneath the top 3U section I have tacked a couple of input/output panels. One panel connects directly to bottom connector of  the left most 6U module to which its on board output transformer is wired. At the moment the right hand panel connects to the right most 3U module on the top row. As 3U moduels have no room for an output transformer I have mounted an Edcor one to the sub-rack behind the panel.