That first achievement is an wide bandwidth amplifier, in VLF-HF range ([50kHz-10MHz]. The initial goal is the HV measurement chain calibration of the modulator of PHIL accelerator (Lal), particularly the capacitive divider. Other Orsay users in Labs are interested, for example for ionic implantation PIII and HiPIMS (LPGP).
In these techniques, fixed ISM frequency 13.56 MHz and its harmonics are attractive, so a narrow band amplifier would be fine, but wide band fits other investigations tests. Power reduction will permit us a first compact equipment, but paradoxically, the VLF Band has brought some difficulties in development.
This wide Band Amplifier can’t operate continuously on the entire Band. You need to switch a set of filters for the output. I selected the elliptic transfer function for these filters.
Technical note (to be completed) : pub2_jlb
Status : the layout have been achieved and the preamplifier and amplifier cards built. Components implantation has been done. Integration in 19″ rack is in course. Some air sink for amplifier is experimented in SME technology. Test of preamplifier : made on 2 versions, the second one optimized on PCB, but no major improvement. nearly 3W on 15 Ohm at 100kHz, decreasing. Frequency raised to 3MHz/0.5W. Conform to Gain/frequency specification. Tests were done in July 2014 on amplifier. After some trivial detected bugs, power has been sourced up to 20 W peak. (20Vp on 20 Ohm) between 100kHz and 1MHz. The limits of the designs start to be well defined. To go beyong 1MHz : 1 improve the preamp, adapt the different ferrites to frequency range. To go beyong 20W : not possible with 1 stage of standard mosfet, achieve a new design (not difficult)
. To improve the quality of the waveform : the large distortions encountered are unavoidable with class AB. It is in state of the art to add a filter bank in output (in course)
. In the end, I improved the design of radiators, so it is anticipated to produce an original solution. That project is in real progress, not about the performances (somewhat mediocre for the moment), but in the knowledge of their limits and the way to improve them…
Technical note on SME
Thanks to Jean Noel Cayla / filip Rudnyckyj