how-to block ads
|
|
Share Topic
 |
 |
|
|
|
RickStepPremium
join:2002-11-25
Hamilton, ON
kudos:1 | reply to stephiso
Re: Treadmill start = modem crash (interference) said by stephiso:I have an old APC UPS as well with the surge protection, however the battery is dead.
If the battery is dead the device has no use as a UPS.
The treadmill will not generate a surge but rather a voltage drop. There is no device on the market that will correct for a drop in voltage. The UPS works because it is generating the 120 Volts from a battery.
Surge unfortunately is not a good description. A surge (an increase in voltage) could happen if the neutral were improperly connected to the system and if the voltage increased on one half of the 240 volt circuit. Voltage on the other half of the 240 volt circuit would drop by the same amount.
Normal 240 volt circuit: 120-0-120
Unbalanced 240 volt circuit: 115-0-125 or 100-0-140 or any other imbalance.
A defect in the electrical supply system could also cause the voltage to increase.
The typical surge protection device is to remove relatively quick voltage spikes that can show up on any electrical service. The spikes can be several times the supply voltage, and can cause computer type devices to reset, a temporary problem or cause permanent damage requiring the hardware to be repaired. Usually if there is a spike (noise)(voltage increase) there is a corresponding spike in the negative direction. Surge protectors will not remove negative spikes.
said by InvalidError: A PC's PSU does not have enough capacitive filtering to last more than about 20ms either . . .
My choice of 1 to 2 seconds was a bad selection of time. You are correct that most devices won’t stay powered for more than a few milliseconds.
The biggest problem with power adapters is a design where the output voltage is too close to the input voltage required by the regulator circuit to provide a regulated voltage. When the input voltage drops too far, the regulator no longer regulates, and the voltage will rise to maximum voltage.
For a device requiring 5.0 volts, the input voltage needs to be at least 2.0 volts higher at 7.0 volts. If the input drops to 6.9 volts, the regulator doesn’t work and the device is supplied with 6.9 volts, not 5.0. Most TTL & CMOS gates and processors won’t function at 6.9 volts and the device will crash or suffer some permanent damage that could cause recurring problems.
Stabilizing the voltage with a UPS is a good start. Selecting the proper UPS is another issue. If it is too small it won’t help very much. If it is too large you are spending money that you don’t need to spend.
There is an article from APC (American Power Conversion) that explains the difference between watts and volt-amps here:
»www.apcmedia.com/salestools/SADE···1_EN.pdf
said by shepd:Most UPSes won't help unless they're online UPSes.
There are too types of devices on the market. The UPS mentioned earlier and an SPS.
UPS (Uninterruptible Power Supply) generates its own 120 volts from a battery continuously when the device is turned on and is constantly recharging the battery. If the power fails, will run until the battery is exhausted. These are used mostly for computer systems, medical systems and other sensitive or high priority systems. Many UPS devices come with software that can initiate a controlled shut down of the attached equipment.
SPS (Switching Power Supply) which provides 120 volts from the electrical system and at the same time has a battery charging circuit. If the power fails, the device will switch to battery mode and provide power until the battery is exhausted. Used for emergency lighting, emergency power for sump pumps etc. Switching time is usually too slow to prevent computers etc. from resetting.
Rick | |
| said by RickStep:The treadmill will not generate a surge but rather a voltage drop. If the treadmill uses a DC motor fed by a direct thyristor drive, insufficient filtering within the treadmill could also introduce significant distortion on the AC line which would get coupled into AC transformers' outputs.
While a large motor starting may introduce sag, a large motor stopping can also produce a short surge from back-EMF and wiring inductance. With incandescent bulbs, you can often see bulbs dimming while the motor is starting (brownout due to startup amps) and then briefly glow brighter for a split second when the motor switches to steady-state. If in cases of neutral imbalance, the imbalance would persist as long as the motor is running, albeit to a lesser extent after startup. You can also see the startup-dim with the pulse when switching to steady state in homes that have a 240V HVAC compressor... no neutral imbalance there, the steady-state switching surge is simply caused by wiring inductance during a load change.
said by RickStep:For a device requiring 5.0 volts, the input voltage needs to be at least 2.0 volts higher at 7.0 volts. What matters is how much energy above the minimum operating voltage is available in storage caps. It can be achieved through high capacitance, higher voltage or a combination of both.
The 2V higher thing is a relic from LDO/linear regulators. With PWM controllers driving MOSFETs, the only requirement is Vin >= Vout + Imax * Rds(on) or the PWM controller's cut-off voltage, whichever is less. Battery-powered application can even use buck-boost or flyback converters which make it possible for Vout to exceed Vin.
If you open up remotely recent modem or router, you will notice that the first thing they do with their "5V" feed is step it down with buck-PWM regulators to 1.5-3.3V with their input stage often designed to take up to 12V.
In any case, for the devices' internal PWM regulators to do their jobs, their input has to be clean enough to allow normal operation. Here, line noise and surges are likely more problematic than sags: if the line adapter voltage drops, nothing flows through the rectifier diodes and decoupling capacitors might be able to cover the dip. When there is a surge or noise on top of normal current, that junk enters the device's regulators, may cause a surge on the outputs either directly by suddenly altering the dI/dT on the pass inductor or regulator errors from introducing additional noise in the voltage reference and feedback loop.
Neither is more desirable than the other but sags are the lesser of two evils since they are the least likely to destroy equipment. Both can be alleviated by adding an LC filter between the adapter brick's output and modem. | |
shepd
join:2004-01-17
Kitchener, ON
kudos:1 | reply to RickStep
Almost all UPSes (or, as you said, SPSes) sold for PCs are offline. They switch quick enough for the PC to still function. Many of them output square waves too (UGH!)
Put an oscilloscope on them if you don't believe me (follow the necessary precautions so you don't blow it up). When I say almost all, I may as well say all, because unless you are buying for a datecentre, it's going to switch. You can hear the relays if you don't want to connect it to a scope, and then you can usually faintly hear the transformers/caps buzz inside your PSU due to the crappy square or modified sine waves being dumped into the supplies.
Heck, even the several hundred dollar 3000 VA UPS I have is a switcher. PCs ALWAYS have enough power to last for 1/120th of a second power cut, due to the way switching supplies function.
So buying a UPS, unless you are careful, will result in wasted money. | |
|
