layout screw up.

Asymmetric capacitances can change common-mode hum into real hum.

Are you gain switching with toggle switches? That could get tricky. We often cold-switch (namely, only DC on the switches, with the actual gain switching via mux's or relays.)

A gain of 1e6 inside the same box is always interesting.

LOL! 60 dB + 60 dB = 120 dB.

On your preferred 2-sided pcb with no ground plane?

External switch wiring?

Ain't gonna work.

General guidance:

Use ground planes.

Don't switch air wires.

Use low level gain to capture desired signals.

Transfer to separate amplifiers for desired gain.

Use strong decoupling to isolate gain stages.

Shield input and output signals.

Ground everything.

Decouple power supplies.

For your applications, RF starts at DC. Design with this in mind.

The toggles are switching the input R, 10 Meg, ground and open. I'm thinking of using the open to look at high impedance's semi conductors at low K. I was thinking I should stick some bus wire on the toggles and measure the C with an RCL meter at least.

At DC! but there's a LP and offset. I don't think we'll use it all. And I like to leave an easter egg in my circuits.

Compared to the + side which oscillated/ or gain banged into the rails...( it's hard to tell the difference... gain-banging looses any phase coherence after a short while... )

The minus side looked OK, which gives one hope.

George H.

Hey, it's like skiing. If you're not falling down, you're not finding where your limits are.

Most of the time people will want gain here of there. not the full monty, but it's all gotta work.

(I like ground planes, high impedance stuff is weird. A giga ohm and pF, do the numbers.) OK I'm skipping the rest, it's not my first analog circuit with a lot of gain. This one is BW limited to a large degree.. ~a few kHz max, but that could change.

George H. I'm mostly here to talk about circuits and stuff, not to berate people. we all start somewhere.

"Crunch crunch", "Mmmm, good crow"*

Well after some hair pulling, hacking, rebuilding. I got everything 'happy'. I had to add shields to all the input and output lines.

If you want I can post pics of a really noisy gain of 10^6 pulse, with Zin of 10 Meg. I use to have this rule about total gain times bandwidth. I did a box with 10^4 gain and 1 MHz BW. And so I thought

Oh almost forgot the pic of inside.

George H.

The sound of me eating crow, You should picture Laraine Newman from SNL, "Mmmm Good bass"

NICE! Thru-hole and stand-em-up resistors ;-) ...Jim Thompson

-- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

I'm a throw-back to a bygone era. "Surfing on the trailing edge of technology" :^)

GH

Me too ;-) This SMT stuff is too small for me to cope with in hand-made G-jobs... so I do 60mil traces and thru-hole... and don't sweat the Amps ;-) ...Jim Thompson

Very 60s. Wouldn't surprise me if you could solder an SM resistor onto those pads instead.

NT

One thing that's tricky about diffamps is that it's good to have a front-end passive lowpass to keep RF and ESD out and defend against overloads. But if you use RCs or even proper EMI filters, any asymmetry in the frequency rolloffs converts common-mode stuff into a real differential signal. If two RCs mis-match to 10%, you now have about 20 dB CMRR at some frequencies.

Part of the R in those RCs is the customer's.

Hah, I did that on another pcb. It was actually nice having a hole to hold some of the solder, and a place to stick the iron tip.

George H.

Right I've done that myself with 5% caps and 1% R's on a RC Int amp AC input. This has lnd150's (jfets) protecting the front, but that's all.

George h.

Hey, someone must make panel mount bnc connectors with a piece of skinny coax crimped on... or just the connector and we crimp it ourselves. We've already got someone's crimping tool, guess I could start there. OK digikey has a bunch. Do the cheaper ones always have the nut on the outside of the box. I like that better from a 'service' point of view. (I could use the cheapness to sell it to the boss, who is also interested in aesthetics.)

George H.

Den fredag den 3. november 2017 kl. 00.24.24 UTC+1 skrev George Herold:

if you aren't scared of ebay etc. there is tons of stuff like this:

why not make the pcb slightly bigger and use a pcb mount bulkhead bnc?

Something like this?

I like to solder connectors directly onto PC boards and avoid the labor of messing with cables.

Electronics is 70% packaging, and the other 50% is thermal.

External connectors soldered on boards is one of the big failure issues with consumer electronics, the solder cracks. How do you avoid it happening?

NT

Fasten the connector to the board?

Use good connectors and solder. Like, we prefer the big old klunky B size USB connectors, because the Micro AB types tend to rip off a board.

The Vbite BNC is brutal; it hooks onto the board.

And our customers don't abuse stuff like consumers do. Usually.