[
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2008, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]
161.142.97.142
In Reply to: RE: Switching PS for filaments posted by gluca on June 29, 2008 at 12:53:39
Regarding this thread and an earlier question relating to the sourcing of a higher speed smps supply for an amplifier, perhaps I can give some advice on these.
The 83khz is not actually an issue since this technology works at speeds from 66 through to 400 kilohertz.
The speed of operation defines the components, the higher the speed, the smaller the supporting components therefore, the smaller the size. The tradeoff is an increase in radiated and common mode noise as the operating speed increases.
The main issue with smps supporting tube amplifiers is the fact that to make them quiet enough, they require a large number of filter poles, typically 10. Also, a product that attempts to achieve a smooth output using only capacitors will not be quiet enough to be useful.
SMPS can be used to support signal and power tube filaments, 6.3 at 0.8A or 5 volts at 2 amps after that they become excessively large. A 5 volt module sized at 4 by 3 inches and operating at 130 kilohertz for example, supports a single 300B filament thus best suits monoblock SET amps. The cost works out at USD120 or thereabouts.
For those who intend to diy, I would advise not to bother. To get them to work successfully requires an awful lot of work. I shall provide a number of reasons as to why this is so.
The pcb design and layout is very critical for successful operation.
The transformers are extremely difficult to get right and each one will present you with different noise problems until you figure out what to do and get them consistent.
They generate huge amounts of common mode noise to both the output and input. The output side requires up to 10 filter poles for this application and that these have to be mostly magnetic ie common mode chokes.
The common mode chokes each need to cover a different noise band if all of the noise artefacts are to be controlled.
Any capacitor used for smoothing will cause common mode, nanosecond length noise spikes to be generated due to their intrinsic self inductance. These spikes cannot be removed by anything downstream.
The output must float; grounding the output provides a path for the whole of the mosfet noise to pass through the load.
As a note, common mode noise at 5 to 10 megahertz propagates through the filament supply of ALL tube types and will modulate the audio output signal. In certain situations, this may well make an amplifier sound better but it is in fact operating as a short wave transmitter and can disrupt other systems.
In all smps applications, they are designed to be connected to multiple 3 terminal regulators placed downstream and that one side must be grounded to handle the residual noise and to comply with safety regulations. This makes the ground plane very noisy and a tube amplifier cannot support this. A 300B filament supported by a SMPS module must be carefully configured to self cancel the common mode noise artefacts since the filament, acting as the cathode, has to be grounded. In simpler terms, they are quite difficult to set up inside an amplifier.
Despite the negative aspects of the technology, they have a most remarkable effect on the performance of a 300B SET amplifier.
If you are serious about using one of these, feel free to contact octave-electronics.com for more information.
With respect to SMPS derived B+ generation, 450 volts at 100mA is achievable and reliable but, you won’t want to pay for it.
With each one comes a health warning, once you have listened to one, you are addicted.
Regards
John Rankin
Follow Ups:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2008, Audio Asylum, all rights reserved ]