Can the DSP Audio Signal Processing be programmed for Speaker Tuning and Speaker Protection?
Yes, the DSP Audio Signal Flow and the Signal Processing parameters can be accessed using the Audio Canvas III Programming GUI, and change or optimized for a given application. Care must be taken not to disrupt the existing Audio Signal Flow to assure that the EVAL KIT functions are not affected.
Full-bridge Class D (V+ to GND): Other than increased size, are there drawbacks with introducing large decoupling caps to remove DC bias?
You can certainly add the capacitors to the output, but it’s not obvious that it would accomplish the goal of removing the bias without introducing a wide range of additional issues that you would have to deal with, some of which would be dependent on the capacitor size and characteristics. In short, it would not be recommended.
You talked a fair bit about the tradeoff between efficiency and performance, but how do you measure or quantify performance?
As discussed during the Webinar, THD+N and Noise Performance are tightly coupled to Efficiency. Efficiency is also considered a performance specification, and the trade-off between Audio Performance and Thermal Efficiency is critical. Both Switching losses and Conduction losses contribute to Thermal Efficiency, and Deadband selection, switching rates, switching frequencies, output inductor selection and PCB Layout all contribute to the trade-off.
How is average power determined in a class D amplifier since audio is very dynamic?
Average Power for ANY Amplifier is generally calculated based on the expected Peak vs. Continuous Power expectation for a given application. The Continuous (or Average) Power can range from 1/8 to 1/3 Peak Power, depending on the expected audio content and application. The ratio if also dependent on the expected Frequency application for the Amplifier – Subwoofer, Full-Range Mid-Range or Tweeter. Toward the lower end of the spectrum (Woofers/Subwoofers) Continuous Power and Peak Power can become very nearly the same.
What’s the Audio Precision bandwidth in the presentation?
For the captured Audio Precision plots, the Audio Precision 2700 had the AES17 Filter enabled, which limits the bandwidth to 22kHz.
What is the GaN Systems reference design performance at 50 kHz?
The initial data shows +0.08/-1.0 dB. This can be optimized for a particular load. We typically design the Output Filter for 6 ohms and can optimize the response for 4 ohms, and smooth it out even further.
GaN has a very short (sharp) ON/OFF edge, which transfer to more and stronger harmonics. Does the sharp edge help audio performance?
As with other differences between the actual switching waveform and the ‘ideal’ waveform, any ‘slant’ or ’tilt’ in the turn-on or turn-off will contribute to THD+N measurements. This ‘slant’ or ’tilt’ also represents time in the active region of the GaN device, which contributes to thermal inefficiency.
Is everything available with the Reference Design for me to translate this into my own GaN-based Class-D Amplifier or SMPS system?
Yes, the Reference Design includes Schematics, BOMs, PCB Layout files, Gerbers, Audio Canvas III Project Files and Firmware.
Are Class-D for audio always hard-switched, or is there room for zero-voltage-switching techniques?
As a rule, Class-D Amplifiers are ‘hard-switching’. Most designs cannot tolerate the increase or change in Deadband that might come as a by-product of Zero-Voltage-Switching.
Is deadband caused by the dead time between top and low side FET?
Yes, a small amount of deadband is desired to prevent ‘cross-conduction’ between the Top-side and Bottom-side devices. However, this deadband also contributes to loss of audio performance, and in addition reduces the Modulation Index, which ultimately reduces the amount of power output from a given power supply voltage.
Is there a large turn off delay or is this just the graphics?
In the capture for the presentation, some if this is actually just the graphics, but there is a slight delay in ‘turn-off’, due to total Gate charge that must be depleted.
Is there no Qrr with GaN because in GaN FETs there is no body diode?
Yes, the Qrr is associated with the ‘body-drain’ diode.
Is the GaN Class-D Amplifier implementation more expensive than a comparable Class-AB or Silicon-based Class-D Amplifier design?
As noted in the presentation, it is best to consider cost implications at the ‘system’ level. While an individual component might be slightly more expensive, that cost difference is usually more than offset by the reduction in cost of other components or mechanical construction.
Where is the industry at with respect to typical GaN switching frequency AND achievable modulation bandwidth while retaining acceptable THD? (Interested for RF Envelope tracking)
GaN Systems devices perform well at very high frequencies used in 6.78 MHz and 13.56 MHz wireless power transfer applications up to 80MHz fast switching industrial applications.
Have SiC JFET’s made any inroads in the audio amplifier space? Any shortcomings with that approach?
Most SiC devices are not appropriate for Class-D Audio Amplifier designs. Present SiC devices are more targeted at SMPS applications, and do not have the characteristics of a high-quality, lower-voltage GaN device.
Have you encountered issues with Dynamic Rdson in GaN?
Dynamic Rdson was a topic of discussion several years ago. GaN Systems and other GaN companies have designed their products so this is no longer a concern. Refer to this webinar review on the topic from GaN Systems >
Speaker: Skip Taylor
Skip Taylor’s 40+ years of audio experience has all culminated in the creation of Elegant Audio Solutions, Inc., where he serves as President and leads a group of elite technologists and engineers. Numerous clients use Elegant Audio’s expertise to move their product and technology developments to revolutionary levels of performance. Prior to establishing Elegant Audio Solutions, Skip founded D2Audio Corporation, where he served as Chief Technology Officer. He was also an Executive at Cirrus Logic, where he established the Audio Digital Signal Processing Group, was Director of Engineering at Peavey Electronics, Inc. and led the development of state-of-the-art techniques for target recognition, digital image processing and digital signal processing at Harris Corporation. Skip Taylor has received BSEE (1970), MSEE (1972) and PhD (1977) Degrees in Electrical Engineering at Mississippi State University.
Moderator: Paul Wiener
VP Strategic Marketing at GaN Systems
Paul Wiener is GaN Systems’ Vice President of Strategic Marketing. Prior to joining GaN Systems, Paul led the power magnetics business unit at Eaton. Paul brings more than 25 years’ experience in operations, sales and marketing, and business development. His experience includes vice president of sales at Fultec Semiconductor Inc. and several management roles at Genoa, BroadLogic, and Raychem.