Power outage while the Nord keyboard is on

[maxpiano]

“… Power dips can be more damaging than power spikes. ….

Why?

Because the various components are expecting certain voltages. Those voltages get generated by transformers who basically work on X voltage comes in and it is stepped down to Y voltage that goes out. When the power dips the ratio of power in and power out changes and the components start working in unknown ways that can potentially cause damage throughout the whole device because it keep trying to work. In a massive power surge typically the first component or two get blown up and the device shuts down.

Are you sure? Afaik a "transformer" always keeps the same ratio regardless the voltage, because the ratio is determined by a physical factor (the number of turns on the primary winding divided by the number of turns on the secondary coil), otoh a switching step down converter (or a voltage regulator) would shut down if input voltage falls below the minimum required to keep the output voltage correct. (PS: "voltage" and "power" are 2 different things, let's not confuse them).

Do you have any link/reference with examples or technical explanation on how a voltage drop may affect a PSU output and/or damage a circuit powered by it (Seriously, I am curious)?

[23skidoo]

Are you sure? Afaik a "transformer" always keeps the same ratio regardless the voltage, because the ratio is determined by a physical factor (the number of turns on the primary winding divided by the number of turns on the secondary coil), otoh a switching step down converter (or a voltage regulator) would shut down if input voltage falls below the minimum required to keep the output voltage correct. (PS: "voltage" and "power" are 2 different things, let's not confuse them).

Do you have any link/reference with examples or technical explanation on how a voltage drop may affect a PSU output and/or damage a circuit powered by it (Seriously, I am curious)?

Yes, the voltage ratio remains constant: so if the voltage IN drops (as during a brownout), so does the voltage OUT. This has a variety of ramifications in a brownout context, not limited to the simple ones I am about to explain.

One of the direct consequences is an increase in the current sunk by linear regulation (which I believe most if not all NORDs are, for reasons of reduced RF EMI) and in extreme cases reduces the regulation headroom to a negative value, causing the output rails to actually drop. Running a 3.3V microprocessor on lower voltages, if the chip is not designed to work at lower voltages, can cause transient data communications and internal processing inconsistencies. It can also introduce persistent data corruption in flash write operations, among other issues. The increased current sunk by amplifier and other circuits biased against fixed voltage standards (such as ADC input amps and power supplies) also leads to early failure of the component, sometimes exponentially so depending on the topology of the circuit and the additional heating brought on by the increased current.

In addition, a glitching microprocessor might "command" peripherals, including power supplies, incoherently or randomly, which might put THEM in a state that they or something downstream from them gets damaged.

In most cases, brownouts just make the unit glitch and it's fine after a reset. Sustained operation in moderate brownout conditions won't usually hurt anything beyond the power supply lifespan. But depending on the design, the system, the severity and length and pattern of the brownout, etc., there are absolutely ways in which a brownout can cause data loss, accelerated component failure, or outright damage. Operation of any system outside design specs, whether over or under, can have (sometimes serious) consequences. For this reason, a lot of electronics (but by no means most) have brownout-detection-and-reset circuits to simply shut the entire system down when a brownout exceeds the internal "safe operating" boundaries of the system design parameters.

[maxpiano]

@23skidoo: all understandable, I just wonder what is the likelihood of a brownout being wide and sustained enough to actually create such a dangerous condition; if I look and the input/output curve of a common 3.3v regulator, for example, the range of input voltage that would cause the output to be lower that the nominal, but still not 0, is really narrow. Of course, all can happen, just maybe not so likely (at least statistically)

[23skidoo]

A linear 3v3 regulator cannot boost the voltage, so if the input voltage falls below 3.3V plus the silicon ("diode") drop internally through the regulating element, the output will be below 3v3. The Nord power supplies are linear, to the best of my knowledge, and are NOT buck-boost, so if the input to each regulation stage falls below that minimum headroom, the rail will necessarily droop. Usually regulators are chained to minimize total voltage drop at each stage, so you'd have, say, AC voltage in -> +/- 12V (or 15V) for the audio rails, off of which you'd get +/- 5V for 5V logic, and then off that +/-3V3 for the mcu logic. Sometimes on the mainboard there's even lower voltages (down to a common 1v8 for ultra low power stuff these days). If you drop the 12V out of regulation, the 5V will usually stay in regulation until the 12V rail drops below 5V.

But if they use multi-tap transformers (I don't know, just hypothesizing), then sometimes you'd have a tap that provides 12V (plus headroom), 5V (plus headroom), and sometimes 3v3 (plus headroom) directly, each feeding their own regulator directly. In that case, a brownout browns out all of the supplies at once due to the linear droop of the input. The advantage is that the later stages don't place higher requirements on the upper stages (e.g. the power for the 3v3 rail doesn't need to be provided by both the 5V and 12V regulation), which makes the supply more efficient (each stage has losses more or less linear with the power provided and the voltage drop). The downside is more rails are vulnerable to brownout simultaneously.

Just one example of a topology choice where one design is both optimal (efficiency-wise) and more vulnerable to total brownout compared to another that's less efficient (and more demanding on the components) but less vulnerable overall.

But we're veering way off-topic here.

[baekgaard]

Maybe a final comment on the brown-out causing damage: Most electronics have some sort of start-up, under-voltage lockout and power-on reset circuitry that ensures the system start in a valid state, gets shut off when voltage drops and/or resets on power-on.

These can be tricky to design, and are perhaps not going through very extensive testing on system level (even if they do on IC level) as validating and testing across all imaginable "unexpected situations" can be difficult. One easily identifiable such "thing" is what I guess is a relay in Nord boards that seem to only connect the outputs to the internal amplified when the board has initialized fully, to prevent clicks in your speakers.

In case of a brown-out, a board may end up with some ICs/ASICs/circuits that manage to shutdown and others that don't. When power comes back some mS later, some circuits may then properly reset whereas others don't and thus are in an unexpected and internally inconsistent state.

Further, a brownout could possibly be repeated drops in input voltage followed by almost normal levels -- i.e. your lamps may blink a couple of times. That could likely make things worse as you now partly shut down some parts, leave others in undefined state and then power up again just to repeat the process. All this can cause increase wear on the components.

In practice, it happens rarely -- but most of us have maybe experienced "glitches" in our boards or computers in case mains get unstable (brownout) for a short period. I usually need to reboot at least some of my computers and switches at home when I see a brownout.

[WannitBBBad]

Hey, I was wondering to anyone this has happened to, what happens when the power goes out or you accidentally unplug the Nord keyboard while it's on? I don't actually own a Nord keyboard yet, but I'm planning on getting one soon and I was wondering if that's something that can damage a Nord keyboard.

The most likely scenario for you is an outage on the electric distribution system serving your home/venue that could last just seconds (momentary outage) or remain out for minutes/hours (sustained outage). Though most of the time losing power is a non-event, no one can guarantee that it will always have the same effect as simply turning off a switch. A larger danger with any electronic device is the voltage spike that can occur when power is restored. Part of the design of the electric distribution system when there's an outage is to have reclosers designed to immediately try to complete the circuit again - you see the effect when electric is dropped and comes right back. As it can occur quicker than one can reach back and turn off their keyboard or any other components, folks often use a UPS with surge protection to deal with that type of outage and ride through long enough to power down, or keep playing if the power is restored quickly enough. Other equipment like power conditioners discussed above are also useful for other power circumstances, but the UPS with surge protection is a good investment no matter of the keyboard you are using. Good luck!