OK. Just a note - USB mini cables are quite rare, the micro->B combination is several times cheaper Micro USB 5 Pin Male to USB 2.0 B Female Panel Mount Type Cable with Screws | eBay and nanos come with micro too, e.g. Nano Micro USB With the bootloader compatible Nano V3 Red controller for arduino CH340 USB driver 16Mhz Nano v3.0 ATMEGA168P|buck converter|dc-dc step downstep down dc-dc - AliExpress . Many options 
Minor update. I recalculated the values for the resistors in the divider. I made a bonehead mistake the first time around, and each step was about 9.5 dB of attenuation rather than the 10 dB I intended. The updated values are 680, 1.47k, 4.64k, 14.7k and 46.4k.
Unfortunately, that 46.4K should really be at least a 1/2W part, and there aren't a lot of 1/2W thin film 1206 or 1210 parts, so I may end up modifying things a bit since I suspect that using a thick film part for the 46.4K resistor will be a serious problem. Actually, as I understand it, using thick films will be problematic wherever linearity is a major concern.
Unfortunately, that 46.4K should really be at least a 1/2W part, and there aren't a lot of 1/2W thin film 1206 or 1210 parts, so I may end up modifying things a bit since I suspect that using a thick film part for the 46.4K resistor will be a serious problem. Actually, as I understand it, using thick films will be problematic wherever linearity is a major concern.
Last edited:
Looking at the layout my only real concern is the spacing if this will be used for anything over 50V. It should be enough to meet the safety requirements in the standards. A secondary consideration is getting enough separation from the controller and its noise. That can cross in since the internal impedances can be high and even if low the HF currents can generate fields which can couple.
Another thought (borrowed from the Boonton) would be input fuses. SMT fuses are quite small and way preferable to replace than other circuitry.
To reduce the distortion from the varying voltages across the protection diodes and Zeners maybe bias the Zeners on with a small current? The 1n914's will be reverse biased at all times then.
Another thought (borrowed from the Boonton) would be input fuses. SMT fuses are quite small and way preferable to replace than other circuitry.
To reduce the distortion from the varying voltages across the protection diodes and Zeners maybe bias the Zeners on with a small current? The 1n914's will be reverse biased at all times then.
Thanks, do you think that there needs to be more distance between the microcontroller and the divider network? I was thinking that it's already pretty far away.
I'm actually going to redo a bit of layout to change some resistors around and adjust clearances. I should have done a bit more research on available parts before I selected footprints. For some reason I was under the impression that prices on metal film 1206 parts (usually called thin film) was lower, but it seems that 30 cents each is about as good as it gets without buying 5000 of them. Since there are some areas where the use of SMD really isn't needed, I'll probably change them out for through-hole. Is the extra cost of using thin-film SMD resistors an issue? Probably not, but if I can reduce the BOM cost without sacrificing significant performance, I will do so.
For those reading who aren't aware...
Distortion caused by resistors is a very real and measurable thing, caused by something called the Voltage Coefficient of Resistance. That is, the resistor value actually changes due to the voltage across it. Unlike changes in resistance introduced by heating and cooling with different parts of the AC cycle, this distortion mechanism is not limited to low frequencies.
Metal film and thin film (essentially the same thing) resistors perform much, much better than carbon film, metal oxide, thick film or (especially) carbon-comp resistors. The vast majority of SMD chip resistors that you see are actually of the thick-film variety, which is unfortunately not ideal for achieving low distortion.
Thick film resistors will be used in parts of this project. Nobody cares about the temperature coefficient or the voltage coefficient of the resistors in series with the LEDs, for example.
The idea of biasing the zeners is a good idea. Thanks for that suggestion, a couple resistors should do the trick and will probably save me many hours of troubleshooting.
Progress has been a bit slow lately. I've been pretty busy with work and school, and I carefully timed the purchase of a BSS EPC780 project as a distraction from unpleasant events. Just as planned, it has taken up much of my free time.
Here's a schematic and board layout for the first in what will be a series of expansion cards. This one is for a fixed frequency twin-T notch filter, though it does have trim pots on it.
Here's a schematic and board layout for the first in what will be a series of expansion cards. This one is for a fixed frequency twin-T notch filter, though it does have trim pots on it.
