Mosfet Devices Versus Bipolar Junction Transistors
I will try not to be biased in this article. Pun intended. Again, no re-invention of the wheel so we will guide the newbie in the right direction.
There’s different trains of thought when it comes to the different transistor types out there and which is suitable for what.
The first Mosfet amplifier I built was the Elektor 100W amplifier using Hitachi’s 2SJ49 and 2SK134. Bomb-proof it was supposedly and in reality, looking at the thousands that were built and repaired they were surprisingly good. In fact they still are but for the price of the output devices they would still be popular. Lateral Mosfets are just so insanely expensive.
Vertical and HEXFET
Switching Mosfets are of the V (vertical) or Hexfet type and have very fast switching speeds and are of course used in switching supplies and inverters. They are also very much cheaper than the lateral variety. Hexfets are by their very nature, as will be understood later through links, not very good at linearity which is why audio designers take a flying duck when coming down to design aspects.
Bipolar Junction – NPN / PNP
Then we have the tried and trusted BJT, making amp design cheaper and better understood. BJTs are comparatively linear when biased properly. If one had to take every design technique of an audio amplifier known to man we would be hard pressed to discover a new improved design which would make the world a better place. Not so!
It’s all in the design
There are many design engineers plying their wares on the web and if one had to position who had the best design, which one would sound better, which one had the best spec in all honesty you would end up with a migraine. A well designed amp sounds better than one which was badly designed. A properly designed amplifier is going to be more reliable. It would run cooler (if not class A), it would not distort, have a great power bandwidth and take a knock or two.. Where to start though?
Why Class D
If one had to look at the various class of amplification I would start at Class D. Period. This one is not going to go away, has improved in leaps and bounds over the last thirty years and pretty much inexpensive, Watts/ZAR. Audio engineers are taking this class very seriously. Unfortunately they are rated as amongst the worst by the purists. It’s a tough world out there!
For class A and AB designs you will find various trains of thought, whether Mosfets are better than BJTs or vice versa. Don’t sweat about it, again proper design is key. Mosfet designs seem to be the order of the day when looking at mono-blocks of hundreds of Watts. And these quite commonly use HEXFET devices, so don’t be put off by the Laterals and BJTs.
Nelson Pass / Rod Elliot / Douglas Self / Robert Cordell
We as a rule can dig out the necessary information from the pages of either Rod Elliot or Nelson Pass, both mentors key to experimenters and professionals alike. Although you may see a leaning towards either transistor type, I do believe we should have enough understanding from either of these two sources before making idle boasts at a nerd party.
Manufacturers do not recommend using HEXFETs in audio. Period. So before looking around and noticing that laterals are exorbitant in price, catch a glimpse of Rod Elliot’s Using HEXFETs in High Fidelity Audio and then look at the TDA7293 and TDA7294, both DMOS devices made by ST Microelectronics, inexpensive and brilliant. We ran an article on these two chips found here. DMOS, either Lateral or Vertical – these were borne from the technology used in Double-diffused Metal Oxide Semiconductor or more specific, LDMOS or VDMOS. (LDMOS is used preferably in RF work above 1GHz).
HEXFETs and DMOS in Class D and switching supplies
I emphasise the importance of vertical or HEXFETs because of two very important reasons (i) Efficiency (ii) TCO or total cost of ownership. Texas Instruments is forward thinking, promoting the use of their Class D chip range. Savings should be passed on to the commercial and pro audio sectors. More HT systems are coming out with switching supplies and amplifiers. Costs are still high but R&D around this topology cannot be cheap. Professional audio manufacturers all seem to be flexing their muscle by producing insanely powerful switching amplifiers for outdoor venues – at very good price points. And yes, reliable as well.
Most Class D amplifiers use DMOS output stages. LDMOS is preferred in very fast switching applications. RDS or the resistance between Drain and Source when saturated should be as low as possible. 0.27 Ohms for the IRF520 is a good start. At a source current of 9A dissipation is about 2.5W. Charge pumps are used to control the gate switching. Transistors must be switched efficiently and cleanly, not an easy task. Not doing so results in inefficiency, overlap gate drive and almost always (for me at least), calls on the smoke wizard. Controlling deadtime can be a bastard.
So there you have it, different strokes for different folks. My opinion. To throw a spanner in the works we have IGBT as well (insulated Gate Bipolar Transistor). Better RDS and the best of both worlds: Bipolar and MOSFET. More about this later. There was a circuit in Elektor years back which I need to pull out of the archives.
Design Considerations for Class D Audio Power Amplifiers – Texas Instuments
How to choose MOSFETs for Class D amplifiers – EE Times
Using HEXFETs in High Fidelity Audio – Mitch Hodges and Rod Elliot
The Sony VFET Amplifier 40 Year Commemorative – Nelson Pass (not be confused with VMOS or HEXFET
Pass Linsley-Hood Amp – IRF244 – Nelson Pass
Audio Power Amplifier Design – Douglas Self (advertisement)
Why I prefer Power Mosfets – Robert Cordell
A MOSFET Power Amplifier with Error Correction – Robert Cordell
Designing Practical High Performance Class D Amplifier – International Rectifier
200W Audio Amplifier with Mosfet BUZ905P-BUZ900P – Hobby Electronics
Sony VFET Amp parts – DIY Audio / Nelson Pass
Sony VFET Information and schematic – First Watt / Nelson Pass