04:51 <azonenberg> AKL-AD1 prototype assembled

04:51 <azonenberg> About 17 mV of DC offset. I'm sure i could improve this with some sort of calibration

04:52 <azonenberg> There's a weird artifact on the falling edges i need to investigate

05:08 <azonenberg> Ok so, back to basics... When powered up, it pulls about 82 mA from both the +2.5V and -2.5V rails

05:09 <azonenberg> With both inputs terminated with 50 ohms to ground, and the output terminated with 50 ohms to ground, I measure 15.26 mV offset at the output

05:09 <azonenberg> With no load on the output (high Z) I measure 19.2 mV offset

05:15 <azonenberg> aaand it looks like i need to put shielding on the probe lol

05:16 <azonenberg> i did a noise spectrum

05:16 <azonenberg> and saw a nice big 2.4 GHz peak

07:31 <azonenberg> https://www.antikernel.net/temp/ad1-s21-dcblock.png

07:31 <azonenberg> lain: S21 of AKL-AD1 from P/N inputs to output with SMA DC block on output, and other output terminated into ground with a capacitor + 50 ohms

07:31 <azonenberg> apparently this amp doesn't like the output being DC coupled to ground, despite the docs saying it's fine as long as the common mode is compatible

07:32 <azonenberg> i think what's happening is that the natural common mode offset of the amp is causing a DC current flow, so i could probably get away with DC coupling the output if I carefully tuned Vcm with a DAC to null out the offset of the internal Vcm buffer

07:32 <azonenberg> But realistically for high speed work, i think an AC coupled amplifier is fine. What I'm less thrilled at is the mismatching between P and N

07:33 <azonenberg> From 350 MHz to 1.2 GHz they're very well matched

07:35 <lain> :o

07:36 <lain> yeah o.o

08:17 <azonenberg> I'm also trying to figure out how to properly merge a bunch of 2-port s-parameter measurements into a s3p/s4p

09:24 <azonenberg> https://www.antikernel.net/temp/ad1-eye.png

09:24 <azonenberg> lain: also ^

09:25 <azonenberg> interesting rising vs falling edge asymmetry

11:17 <lain> oo, interesting indeed