Hmm, Maybe Evidence for Unseen Intelligence

[Joseph]

Well, I had spent longer than I had expected getting my analog electronic thermostat debugged. I should have wrapped tape around the NPN temperature sensor transistor. It was built in a router motor speed control case. I added the extra components along with a solid state relay to control the on/off state of the compressor. So, the sensor was not isolated.

That stuff is the background information. There is just enough uncertainty about what happened to cause doubt about what really occurred.

OK, so I brought it back into my room and had plugged the cords in. I was moving the temp sensor toward the front of the A/C when it touched a 4 foot long piece of rebar tie wire hanging from a _wooden_ support. It was _not_ connected to anything else except an empty vinyl hose which it was supporting.

Well, a spark flew and I heard a dreadful pop from the modified router control unit. I took it back out to the bench and determined that it was basically BER --beyond economical repair. Mysterious.

[Joseph]

I aways try to find the cause of things and was at a complete loss in this case until now. It could have been that there was a strong electromagnetic field which the unconnected wire was resonant at. When I touched it with the transistor lead, it coupled a lot of RF power between the router control and the wire. I used to almost touch a CB antenna in transmit mode when I was a kid and watch a spark bridge the gap.

If this explanation is right, I wonder if the triac inside the device was producing harmonics which resonated with the wire acting as an antenna. In that case the device was overloaded by transmitting the power instead of receiving it. It seems like way too much power involved, though.

I have heard of surreptitious transmitter devices being focused in certain directions. If that is the correct explanation, instead, then I have more trouble than I am aware of and really should consider getting a metal hat or enclosing my room in a Faraday cage.

[haklesup]

Assuming the wire was not grounded and the sensor that touched it was not biased to a voltage then I would suspect ESD damage.

A vinyl hose being non conductive can acquire a large static electric charge due to its large surface area. Tying that off to a wire provided a path to the charged hose. Touching the wire zapped the component first creating a tiny short then since it was powered, that short turned into a smoke show.

That ESD charge needn't be too powerful, some zaps can be below the level of human preception. I bet the humidity was low that day or the AC dessicated the air in the room. Dry blowing air charges non conductive surfaces through a mechanism known as tribocharging.

Construction methhods of a particular curcuit as well as using sensitive components (like FETs and CMOS devices) can make a project more or less susceptable to this kind of damage. You should use an ESD strap while handling your parts and unhoused projects particularly if the package has that yellow sticker on it.

Also, statistically speaking, ESD is very very common dispite efforts to design devices that are resistant to it.

I doubt it had anything to do with EMI radiation or harmonics. BTW, I see harmonics as more of an adjective rather than a noun as I have seen it used before. Proper wording might go like "Harmonic voltage oscillations were observed on the signal line" and even then without context of what that signal was and what the frequency/amplitude content of the noise signal is, it has little meaning. Once you have made that definition, its safe to proceed just calling that noise "Harmonics" as a shorthand.

Besides, CBs were what, 14W transmitters (someone correct my memory if I'm wrong) you would need a lot bigger EMI source than that nearby so that the wire could passively capture enough energy to cause damage. Not likely with below RF frequencies you are likely using.

Marginally designed SCR circuits usually stick on off or blow up but typically don't oscillate without deliberate effort.

[Joseph]

I think you have a point about harmonics. But I have had a long background period in audio amplification, so the word is rich with meaning to me.

The harmonics I was referring to implicitly are those the fundamental switching frequency of the triac, 60 Hz, produces in parasitic inductances and capacitances in the circuit. It would be the ringing produced by the parasitics. But it is also any frequency above 60Hz which the triac causes in the circuit.


Those parasitic oscillations contain enough power to run the CD74C14 and barely power the solid state relay control. I capture them with a .22uF capacitor and two tiny audio isolation (small signal coupling) transformers. The primaries of these transformers are in series connection while the secondaries are in parallel.

The transistor involved in the incident is a 2N5088. Definitely bipolar.

The spark appeared too big to be from static electricity. It had some body to it and appeared to be lower voltage than that. I was extremely surprised by the event.

The reference to my CB radio days was to exemplify that RF can produce a spark between something acting as an antenna and another conductor grounded or sympathetically resonant to it. It this case, it would have to be grounding, since it was the power line neutral conductor that was involved in the incident.

[Robert Reed]

A harmonic is a sinewave of multiples of the fundamental that produced it, and therefore should be considered as a noun. However it also could be used as a adjective when used to describe a particular situation.
The CB transmitters, at least back when I used one were limited to 5 watts input power. The feedpoint at the antenna is 50 ohms and the voltage is not very high at this point, however the voltage may vary considerably along the antennas length, so it is not surpring that you could draw a minor & harmless arc at some point in its travel (RF wave).
I am not sure I fully understand your situation, but if your triac is switching (such as phase control applications) there will be some ringing around at the triac's switching point (triac turning on and off) but even moreso when power is abrubtly turned off in midstream. I am assuming we are talking the routers inductive load here.
The swithing transients will occur at a 120 Hz rate for maybe 10 usec. to 500 usec. depending on the parasitic properties of the circuit it is attached to. These are usually on the order of 30 KHz to several hundred KHz. I have seen magnitudes as high as a few hundred volts in total power turnoff situations where they were not snubbed. So yes, they could well damage devices under the right conditions.Again I not real clear on your scenario as described,so I am going with this only as I see it in my mind.
But all in all, I think I like Hackle's thought of static electricity being very feasable for doing the damage here.

[Joseph]

The 23 channel CBs put out up to 5w while the later 40 channel ones were restricted to only 4w. I did get a few seconds of skip a few states away once on 4w.

I like how you have described a harmonic. I was thinking of Fourier analysis which if I am not mistaken, is a study of harmonics of varying amplitudes about the fundamental.

I was going to use the triac to control the A/C fan on my window unit while the relay I added was for compressor power.

[Robert Reed]

"I like how you have described a harmonic. I was thinking of Fourier analysis which if I am not mistaken, is a study of harmonics of varying amplitudes about the fundamental. "

I was always taught and firmly beleive that there is only one wave shape in the electrical realm- and that is a sinewave. Any squarewave,triwave or almost any goofy looking wave shape can be broken down by Fourier analysis to a collection of pure sinewaves that vary in amplitude and phase to each other. When we view this on an o-scope trace, we see amplitude versus time of this combined collection of sinewaves that are algebraically added at each instant of the trace and hence the odd wave shapes that might be produced. If we take that pattern and rotate it sideways 90 degrees as with a spectrum analyzer display, we then see all the individual sine wave frequency components that make up that wave form and their relationship to each other. These can be harmonicly related or not, or a combo of both. I have always found this subject fascinating and some of the computer programs out there do a great job of Fourier analysis and lightning quick. Oh oh, I think I am beginning to stray fom the main topic.

[Joseph]

I would add that the sine waves in additional to the fundamental are higher in frequency than it.

I still don't think the static electricity theory fits. The spark had too much current--the main reason why.

[haklesup]

Yeah, we always get off topic but since we are here;

Robert, your generalization of harmonics is as good as any and it belies my point. The meaning of the term Harmonics means different things depending on POV, a radio engineer, a power systrems engineer and a guitarist all have different notions of what harmonics are. Just to toss out the word without nearly a paragraph to qualify it is incomplete and potentially confusing to a reader.

Also Roberts discussion of Transients and Switching noise also eludes to a common mistake of lumping them and other noise sources into the harmonic category. Only certain frequencies qualify. Not saying anyone has done that buit I've recognized it before.

The real issue with harmonic noise as opposed to random noise is that it tends to be additive or resonate with the signal and other noise and perterbs the circuits more so than non harmonic noise.

I was way off on the CB wattage but the point was that it is still more powerful than what 60hZ/120Hz harmonic noise from switching transients could transmit as EMI to be picked up by another object and converted into charge.

ESD failures at system level (assembled circuit powered and operating) are often cascade events. The time between when the initial ESD pulse damages a junction to when that junction becomes molten from execss IR heating can be in the Millisecond range. The spark you saw may have been secondary to the ESD event after the short was established.

It's hard to say, even failure analysts have difficulty pinning the blame on ESD when the damage went all the way to EOS (Electrical OverStress). The evidence of the original ESD failure is obliterated by the collateral damage that follows. Sometimes the diagnosis comes after a string of failures and a detailed investigation.

[Joseph]

The thing is, that with ESD, there still has to be a pathway to complete the circuit. The 4 foot wire is basically isolated by infinite resistance. The only other explanation besides, let's say, extra dimensional beings(s) playing a part somehow is that there was a fairly high power RF signal present from somewhere which the 4 foot wire was nearly in resonance with.

I don't think the failure of the modified router control device produced a big enough EMP (Electromagnetic Pulse) to cause the spark to jump is something that I will add in.

[haklesup]

with ESD, there still has to be a pathway to complete the circuit. The 4 foot wire is basically isolated by infinite resistance.

Not entirely true, ESD pulses typically contain tiny amounts of total charge Q. You do not need a round trip circuit to transfer this charge from the charged object to to ground (or a lower potential of sufficient capacitance). The ground of your circuit was plenty low impedance to absorb a typical ESD spark even if it was just a battery.

The infinate resistance of the hose is where the charge was collected and stored. The wire would have only transported it to the lead of the device where it may have entered the device and damaged a PN junction or MOS gate on its way to ground.

I'm not insisting it was an ESD caused failure but it is plausable, you just as well may have inadvertantly touched that hanging wire to two leads of your sensor and directly caused the short.

[Joseph]

If the circuit in question were a power amp which could be set into, say over-bias by a small spark, like, say, the old Sansui receivers, I could go along with that chain reaction idea. But since the transistor in question as well as the entire temperature-level-set/hysteresis circuit is basically flea power, it hasn't the opportunity to suffer from such a failure mode. It is really a separate circuit all by itself with no interaction with the triac.

For that reason, too, if the base had shorted to the collector, the worse way to short a transistor for damage, the currents are too small to make the transistor fail.

Besides, the incident is emblazoned in my memory. I can see the emitter, the only lead involved, touching the wire in slow motion.

[Robert Reed]

Hi Joseph
I reread your original post and if I understand correctly, the scenario went some thing like this:
You had a router speed control unit only ( gutted from a router or aftermarket?) you had built in an add on temperature sensor. This circuit was to be used to control an air compressor (I am still not sure what it is controlling) in your garage. You were approching the A/C with this circuit in your hand (powered up or down at the time?). You accidentally bumped into a garden hose hanging from a wire from the rafters. The emitter lead of the temperature sensor brushed the wire holding the garden hose and thats when the sparks flew and the whole circuit became history.
Is this correct?
If the wire were to have picked up RF energy from some outlying source, the air path would have offered so much attennuation even at a few feet that its subsequent discharge probably would not have a lot of energy. The hanging garden hose would not have developed much of a static charge on its own unless stroked by an external mechanical means( strong wind blowing on it all day, drag ged across some static producing surface before hanging. All wood has some moisture content and 5% would be considered unusually low. The hanging wire and wooden stick built garage structure would probably be enough to form a drain path of sorts for high static build up.
The question that I have is how did you know you specifically touched the emitter and how was it even exposed? Did you mean a long lead attached to the emitter extending from the enclosed circuit?
Whatever energy was present would be generated from a rather hi-Z source and wherever it came from, I see as being a high reverse bias in your PN juntion as breaking down that device, caused by the negative or positive voltage excusions relative to that particular junction.
Could you maybe post a crude drawing of your device as it would go a long way in trying to solve your mystery?

[Joseph]

.[/img]


Hi Robert,
The vinyl hose goes from a dehumidifier to the sink in the kitchen. The board in question is as dry as can be since I keep the humidity low.

The router speed control has been pressed into service as a window unit A/C fan speed control. I have added a solid state relay to it which thermostatically controls power to the A/C compressor. The 2N5088 transistor is the temperature sensor which I was in the process of placing in the grooves in the front plastic grill of the A/C.

[dyarker]

Exposed 115VAC circuits with no isolation from mains. You're lucky the only thing dead is the circuit! There is a path to ground there somewhere to make the size of spark you described.