Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail From: Phil Hobbs Newsgroups: sci.electronics.design Subject: Re: Favourite Test Equipment Date: Sat, 6 Apr 2024 11:30:58 -0000 (UTC) Organization: A noiseless patient Spider Lines: 67 Message-ID: References: <9k7j0jlnbhs8qfg5m17pium0835meean83@4ax.com> <1qrjb8o.u2ipkmlx8gzsN%liz@poppyrecords.invalid.invalid> <3d701j9h3n3nc0qhc3mv0r329vfkeu4mes@4ax.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Injection-Date: Sat, 06 Apr 2024 11:30:58 +0200 (CEST) Injection-Info: dont-email.me; posting-host="542ab62d6b685e02ae024874d8db3265"; logging-data="2169877"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1/lxgN2Oa22dacioA/RhIqY" User-Agent: NewsTap/5.5 (iPhone/iPod Touch) Cancel-Lock: sha1:x9kUV9LoGBHOa4UL+Vsm5/D+9dg= sha1:O1a0oL5Q9h9duLv2HLrituVwu9o= Bytes: 3862 Gerhard Hoffmann wrote: > Am 05.04.24 um 17:55 schrieb John Larkin: >> On Fri, 5 Apr 2024 11:38:43 -0400, Phil Hobbs > >>> >>> Try putting BLF03VK600 beads in source and drain. Besides being rated at >>> 3 GHz instead of 100 MHz, it has really nice low Q everywhere. >>> >>> My cascoded lab amp proto is using three SAV-331+'s in parallel like >>> that, running about 2.5 mA each, and shows no sign of trouble. >>> >>> Cheers >>> >>> Phil Hobbs >> >> We tried various fixes, no joy. The circuit is complex, lots of diodes >> and comparators and varicaps and stuff, so there are too many resonant >> stubs. >> >> We prototyped the new sorta-Colpitts circuit, using the BUF602 as the >> gain element, and it's great. Having a closed-loop near-perfect >> follower is nice. >> > > Do you have slowish feedback into the source? From the FETs POV > that makes it look like a capacitively loaded follower. It translates > directly into a negative real part of the input impedance. > > < > https://www.flickr.com/photos/137684711@N07/34701106245/in/datetaken/lightbox/ > > > > It is the input impedance of 2*IF3602 and the negative real part > can get REALLY large, out of bead-land. > In the Smith diagram, S11 is decoded at the marker positions. > Where the trajectory gets out of the circle through 0 and inf, > there comes more energy back from the DUT than the VNA sends to it. > Cannot happen with passive DUTs. > > It is a really hard problem and even in AOE3 is a bad example. > > I got an array of 16* CPH-3910 stable with feedback via a > 3 GHz CFB amplifier. But CFB's 1/f noise easily dwarfed the > noise of the 16 FETs even after 40 dB of gain. > > I tried driving the feedback with 2 * BFQ19S BJTs as a follower. > It seems it kinda works in simulation. Generates a lot of heat. > Not yet built. > The Infineon SiGe parts such as the BFP650 have AC Early voltages around 250V. (The DC curves in the datasheet are corrupted by thermal effect, and so exaggerate VAF, but it’s still very very good.). If you use one or more of those as cascodes, you can probably run a higher first-stage gain, which will help a lot. Those BLF03VK and BLM15BA/BB beads are good for stabilizing them, but that gets harder at high collector current. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics