dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
10
share rss forum feed

westom

join:2009-03-15
kudos:1

1 edit
reply to InvalidError

Re: Treadmill start = modem crash (interference)

said by InvalidError:

If you have 20m of #14 wiring from panel to load, you have ~0.33ohms of loop resistance not including receptacle wiring or nuts and a 10V drop if a load draws 30A inrush current for the first cycle or two. This is still perfectly safe and would cause visible dimming for a fraction of a second while the filament re-heats after the momentary dip.

A 20 amp motor on 14 AWG wire is why the bulb dims. Wire is too long and thin for that load. Fix the problem.

Meanwhile, the resulting dimming is near zero (maybe 10%). Ideal voltage for all electronics is when lights dim even to 50%. If voltage is dropping that low, then obviously a wiring problem exists. No responsible homeowner ignores such faults. Fix a defect that, in some cases, might be a major human safety issue.

AFGI obviously does nothing to avert that human safety issue. AFGI is for arcing - due to another and different fault.

UPS has one function. To provide temporary and 'dirty' power during a blackout. A typical UPS does not 'clean' electricity. And does not have to. Electronics are so robust as to make "dirtiest" UPS 'battery backup' power irrelevant. Therefore 'less dirty' AC mains power is also ideal power to all electronics.

What does a typical UPS do when not in battery backup mode? Connects an appliance directly to AC mains - the source of 'cleaner' power. If a UPS is conditioning power, then a manufacturer spec number is posted to define it. No numbers were provided because the typical UPS does not "clean" AC mains power.

If dimming is causing a modem to reset, then irresponsible is to fix it with a UPS. To ignore what might be a major human safety problem. At no time should any motor cause lights to dim to 50%. All electronics are happy and work even when voltage drops cause 50% dimming. But that much dimming implies wiring problems. Never cure symptoms. Fix the problem.

What does a typical UPS do when not in battery backup mode? Connects 'cleanest' power - AC mains - directly to the appliance. No filtering or conditioning exists.


Guspaz
Guspaz
Premium,MVM
join:2001-11-05
Montreal, QC
kudos:23
said by westom:

If a UPS is conditioning power, then a manufacturer spec number is posted to define it. No numbers were provided because the typical UPS does not "clean" AC mains power.

A typical UPS does condition power, although the extent to which they do so varies. Even the cheapest $50 unit from APC claims full time multi-pole noise filtering. They don't get more specific in specs than that. Obviously the higher end stuff does. Smart-UPS reports 5% max distortion, and Smart-UPS online is going to be pretty close to perfectly clean at all times since it's online. Heck, it has a user-configurable output frequency in 0.1 hz increments :P But I won't argue that a $1000 UPS is typical by any means.
--
Developer: Tomato/MLPPP, Linux/MLPPP, etc »fixppp.org

InvalidError

join:2008-02-03
kudos:5
reply to westom
said by westom:

A 20 amp motor on 14 AWG wire is why the bulb dims. Wire is too long and thin for that load. Fix the problem.

You are confusing surge/inrush current with nominal current ratings.

It is not a 20A motor on 15A circuit (this should trip the breaker anyhow), it is a 10A motor with 30A startup amps on 15A circuit, which is perfectly fine, nothing to fix there. Motors having 2-3X their rated amps as startup current is very common and for most applications, wiring only needs to support the steady-state current rating since the startup surge is too short to cause any significant heating.

As I said in my previous post, the majority of PC PSUs have inrush current in excess of 30A but you do not see people rewiring their homes in #8 gauge because of it nor switching their breakers for 30A either because that startup surge lasts only for a few cycles at most, not enough to cause any significant wiring heating since copper wiring has high specific heat and low electrical resistance.

As far as electrical safety is concerned, wiring is sized solely based on continuous RMS current to prevent excessive heat buildup when following installation rules. Thicker gauge for better power efficiency and lower voltage drop is entirely discretionary.

said by westom:

AFGI obviously does nothing to avert that human safety issue. AFGI is for arcing - due to another and different fault.

It does if the excessive voltage drop is caused by a bad connection which would start arcing under heavy loads such as motor startup or a PSU's inrush current.

said by westom:

If dimming is causing a modem to reset, then irresponsible is to fix it with a UPS.

Do keep in mind that I only suggested the "dimming test", the OP has yet to say whether or not he/she has actually done it. In other words, the dimming so far is purely hypothetical.

As for using a UPS as a work-around being "irresponsible", that depends on why the dimming is occurring.

If the dimming is occurring due to a motor's startup current or a PSU's inrush current, it is simply unavoidable.

If the dimming is occurring due to high inductance in the local power utility's distribution grid, it is also unavoidable and will cause visible dimming even on other circuits. The HVAC compressor at my mother's house is on a dedicated 240V/20A circuit and lights dim all over the house when it starts.

In my previous apartment, the neighbor's AC did cause my lights to dim and my modem to lose sync/reset every now and then, which is when I decided to mod my modem's power brick. Since the two apartment units have separate 100A circuits all the way to HQ's power meters, wiring from the breaker box to my lights/modem clearly had nothing to do with the problem.

Much like the OP, simply connecting my modem to my UPS (APC RS-1000) prior to modding was insufficient to prevent disconnect/reboot.

westom

join:2009-03-15
kudos:1

1 edit
reply to Guspaz
said by Guspaz:

A typical UPS does condition power, although the extent to which they do so varies. Even the cheapest $50 unit from APC claims full time multi-pole noise filtering. They don't get more specific in specs than that.

First, only UPS that does that line conditioning costs about $1000 or higher. A typical UPS connects an appliance directly to AC mains. When switching to battery power, a UPS output looks more like what this utility demonstrates in a tech-tip:
»www.duke-energy.com/indiana-busi···p-03.asp

My 120 volt UPS in battery backup mode outputs 200 volt square waves with a spike of up to 270 volts. The manufacturer calls this a sine wave output. And he is not lying. Those square waves and a 270 volt spike are nothing more than a sum of pure sine waves. Since he is marketing 'conditioned' power in advertising, then he only need make subjective claims. A sine wave output from the UPS is what the utility’s tech-tip shows. My sine wave output is square waves.

UPS provides 'dirtiest' power during a blackout. That is also sufficient to claim 'conditioned' power.

View numbers for that filter. Near zero filtering. View numbers for its protection. Near zero surge protection. But just enough above zero so that advertising and a majority will call it surge protection and 'conditioned' power. The claim is subjective. So lying is legal.

The UPS has one function. To provide temporary and 'dirtiest' power during a blackout.

Second, what happens if a worst case load (or a ten amp motor) causes lights to dim 20%? Nothing. All electronics must work even if the lights dim and remain at 50% intensity. If dimming causes a modem to trip out. then a building wiring problem may exist. In most cases, is easily solved by fixing defective connections.

Third, will a motor drawing 30 amps or more trip a circuit breaker or AFGI? Of course not. A 20 amp breaker can conduct 30 amps even for minutes without tripping. The AFGI breaker would do nothing for wiring defects that might cause dimming. But that dimming, in some cases, identifies a major human safety threat. Fix the problem. Never cure symptoms.

Obviously a UPS to solve dimming is only curing symptoms. A UPS is for temporary and 'dirty' power during a blackout. To do more (ie conditioning) means spending about $1000 or more. Does not matter how clean that output power. The first thing done inside electronics? It makes that power even dirtier. Then uses superior filters and other circuits to clean that power all over again. Why is a 'dirtiest' UPS also ideal power for electronics? Because best "power conditioning" is already inside electronic appliances. Electronics will convert even 'dirtiest' UPS power into ideal perfect power. Electronics works just fine even when incandescent bulbs dim to 50% intensity.

In one case, homeowners ignored dimming light. Fortunately nobody was home when the fault because worse. Causing other failures and an exploding house. Only warning of a serious problem was dimming lights. They ignored it.


Guspaz
Guspaz
Premium,MVM
join:2001-11-05
Montreal, QC
kudos:23
said by westom:

First, only UPS that does that line conditioning costs about $1000 or higher. A typical UPS connects an appliance directly to AC mains.

Again, the cheapest UPS that APC makes, for $50, claims noise filtering.

said by westom:

When switching to battery power, a UPS output looks more like what this utility demonstrates in a tech-tip:
»www.duke-energy.com/indiana-busi···p-03.asp

Yes, but line-interactive UPS don't run through the inverter for normal usage, so your showing us pictures of an approximated sine wave on battery power tell us nothing about noise filtering on normal operation.

said by westom:

My 120 volt UPS in battery backup mode outputs 200 volt square waves with a spike of up to 270 volts. The manufacturer calls this a sine wave output. And he is not lying. Those square waves and a 270 volt spike are nothing more than a sum of pure sine waves. Since he is marketing 'conditioned' power in advertising, then he only need make subjective claims. A sine wave output from the UPS is what the utility’s tech-tip shows. My sine wave output is square waves.

You keep calling UPS output "square wave". It's not. Square waves only have two levels (low/high), approximated sine waves have three (low/med/high). What is shown in the pictures you linked to is *NOT* a square wave.

said by westom:

UPS provides 'dirtiest' power during a blackout. That is also sufficient to claim 'conditioned' power.

An approximated sine wave is fine for most devices. If it is a problem, the APC Smart-UPS line, which uses a pure sine-wave on battery mode, starts at $260.

said by westom:

View numbers for that filter. Near zero filtering. View numbers for its protection. Near zero surge protection. But just enough above zero so that advertising and a majority will call it surge protection and 'conditioned' power. The claim is subjective. So lying is legal.

The UPS has one function. To provide temporary and 'dirtiest' power during a blackout.

No, a UPS is intended to provide power protection. The three axis of that are surge protection, power conditioning, and power backup. Which of those are performed and how effectively they're performed vary by pricepoint.

said by westom:

Second, what happens if a worst case load (or a ten amp motor) causes lights to dim 20%? Nothing. All electronics must work even if the lights dim and remain at 50% intensity. If dimming causes a modem to trip out. then a building wiring problem may exist. In most cases, is easily solved by fixing defective connections.

Some modems (or their power supplies) are incredibly sensitive to voltage fluctuations. A midrange UPS can typically guarantee voltage swings of no more than roughly +/- 5-10%, depending on the model (my older BR1500 had a much tighter range than my BR1500G). They may have to switch to battery to do that, since they don't have terribly effective autotransformers.

I'd keep replying but you just keep spouting more and more of the same FUD. Nobody is saying wiring faults don't need to get looked at.
--
Developer: Tomato/MLPPP, Linux/MLPPP, etc »fixppp.org

InvalidError

join:2008-02-03
kudos:5
said by Guspaz:

Square waves only have two levels (low/high), approximated sine waves have three (low/med/high).

The three levels on stepped approximation are +/0/-, alternating positive and negative square pulses separated by some amount of dead-time to allow time for diode recovery in SMPS PSUs and greatly reduce harmonics.

Stepped sine is a nearly cost-free 'trick' that significantly improves reliability, efficiency and compatibility.


Guspaz
Guspaz
Premium,MVM
join:2001-11-05
Montreal, QC
kudos:23
said by InvalidError:

said by Guspaz:

Square waves only have two levels (low/high), approximated sine waves have three (low/med/high).

The three levels on stepped approximation are +/0/-, alternating positive and negative square pulses separated by some amount of dead-time to allow time for diode recovery in SMPS PSUs and greatly reduce harmonics.

Stepped sine is a nearly cost-free 'trick' that significantly improves reliability, efficiency and compatibility.

I'm not disagreeing with that, just pointing out that it's not a square wave, since square wave jumps directly from one extreme to the other.
--
Developer: Tomato/MLPPP, Linux/MLPPP, etc »fixppp.org

InvalidError

join:2008-02-03
kudos:5
said by Guspaz:

I'm not disagreeing with that, just pointing out that it's not a square wave

I was not disagreeing with what you said either, just pointing out that there is no "low/medium/high", it is alternating rectangular pulses, all or nothing in either direction.

low/medium/high (multiple output levels) sounds more like what a triangular approximation UPS would be doing. These used to be somewhat common about 10 years ago but are largely forgotten today.

westom

join:2009-03-15
kudos:1
reply to Guspaz
said by Guspaz:

I'm not disagreeing with that, just pointing out that it's not a square wave, since square wave jumps directly from one extreme to the other.

While arguing sementics, you ignore the bottom line. A UPS output is so 'dirty' as to be potentially harmful to electric motors and power strip protectors. That same 'dirty' electricity is also ideal perfect for all electronics. Stick with the facts. Electrics are so robust as to make even 'dirtiest' power from a UPS irrelevant. Also so robust that voltage is ideal when incandescent bulbs dim to 50% intensity.

Why would anyone buy a UPS to fix a trivial and slightly lower voltage? Or to clean mains power when mains power is typically the 'cleanest power'.

Again, you say, "the cheapest UPS that APC makes, for $50, claims noise filtering." And again you are making subjective claims. Subjective claims can prove anything. Where is the number? APC has near zero noise filtering. Just enough above zero so that you will promoting it as 100% filtering.

Noise filtering alread inside every modem is so good as to make irrelevant increased harmonics, noise, and other anomalies.

Meanwhile, with all this discussion of expenisve miracle boxes has made impossible a discussion of why a modem would lose synch. A suspect list includes more than just noise on AC mains.

InvalidError

join:2008-02-03
kudos:5
said by westom:

Noise filtering alread inside every modem is so good as to make irrelevant increased harmonics, noise, and other anomalies.

The modems I have taken apart have little to no noise filtering on their power supply input. They have a tiny HF choke, some decoupling capacitors and those directly feed DC-DC converters within the modem. Little more than the bare minimum necessary to avoid radiating EMI through their DC power cable, definitely not enough to filter junk power from the AC adapter. The AC adapters are either "iron lump" types with no filtering whatsoever or self-oscillating inverters with little more than EMI filtering, nothing that would protect the load or the adapter itself from any junk on the line.

said by westom:

120 volt electronics could withstand 600 volts. Today's electronics are even more robust.

I would not bet the farm on that.

Most of today's consumer electronics are designed with cost-cutting and disposability in mind which means much thinner tolerances if not engineered for timely failure (just after warranty expiration) outright. If you take apart consumer electronics, most stuff hard-wired for 120V only has line-facing components rated to 200V which are likely to get damaged from any sort of exposure to 600V if not immediately destroyed. If you step up to universal PSUs with APFC, most of those have components rated for 400-450V so 600V peaks would still be quite likely to cause some permanent damage even if MOVs eat the brunt of it.

westom

join:2009-03-15
kudos:1
said by InvalidError:

Most of today's consumer electronics are designed with cost-cutting and disposability in mind ...


A product designed with such obsolete technologies would, so obviously, cost too much. Nobody uses those obsolete designs. Therefore costs are lower AND electronics are more robust.

Obviously, denials are from one who thinks that tiny device is a choke. That choke would be much too tiny to accomplish any useful filtering. Why would they spend so much for a choke that does nothing? They don't. The mythical choke is something completely different. A device that means electronics easily exceed 600 volt spec. As defined by the APC salestool.

Is APC lying when citing design standards from 1970? Of course not. Only problem is one who has no idea how easily power is conditioned inside electronics. And therefore cost less to build.

All these denials do nothing helpful for the OP or the topic. Does not address other more likely reasons for that modem cutout. Numbers, even from an APC salestool, say why modems can operate just fine even with 'dirtiest' power - from a UPS.

InvalidError

join:2008-02-03
kudos:5
said by westom:

A product designed with such obsolete technologies would, so obviously, cost too much. Nobody uses those obsolete designs. Therefore costs are lower AND electronics are more robust.

Which 'obsolete' designs would those be? They are all still in use today. For devices like routers and modems, brick-style iron-core transformers are still very common but plug-style oscillator or SMPS designs are becoming more popular. All three options provide some degree surge protection simply from line-load isolation but are not surge-protected themselves. A 600V spike of any significant duration (say one cycle) would likely destroy most semiconductor-based AC adapters I have dissected since the only "protection" they have is a fuse between line power, their rectifier bridge and 200-250V cap. They also have an EMI choke but those don't do squat against surges since EMI is common-mode noise while surge is differential.

said by westom:

That choke would be much too tiny to accomplish any useful filtering. Why would they spend so much for a choke that does nothing?

You would already have the answer to your question if you actually read the message: EMI filtering to prevent the modem's switching regulator noise from going out on on the adapter cable. These things only have a few uH of inductance. Enough to block most of the switching regulators' EMI and ineffective against surges from the AC adapter.

said by westom:

Only problem is one who has no idea how easily power is conditioned inside electronics.

Proper filtering of 60Hz line power uses up lots of space, there is no miracle work-around to that.

HeadSpinning
MNSi Internet

join:2005-05-29
Windsor, ON
kudos:5
said by InvalidError:

For devices like routers and modems, brick-style iron-core transformers are still very common but plug-style oscillator or SMPS designs are becoming more popular.

Unfortunately, the new energy efficiency regulations in many areas have forced manufacturers to stop using the iron-core transformer style power supplies and start using SMPS designs.

We have several thousand modems in the field, and note that occasionally power disruptions will confuse the switch mode devices, requiring them to be unplugged for a few minutes to allow the capacitors to drain and the circuit to re-set in order to function again.

Very annoying.
--
MNSi Internet - »www.mnsi.net