Electrical Engineering

Shalvenay

12:00 AM

@forest ah, nods I knew that the cardiac effects of a shock were frequency-dependent

but didn't know so much about skeletal muscles

forest

It certainly is. The pain threshold changes too.

Shalvenay

@forest thx 4 the data

forest

I'm reading through the EE meta in general. Is there a problem with a moderator on this site? I see a surprising amount of hate towards a particular mod.

Marla

12:16 AM

@forest you mean Olin?

forest

Nick, according to downvote history and comment upvotes.

It just looks like a moderator who is going against the will of the community and other moderators or something. I mean I'm new to EE so I won't pass judgement.

Just think it's curious how much drama there is about it on meta.

Marla

Actually @NickAlexeev spends a lot of time here in chat.

Being a moderator is tough

forest

Just saw he fights with other moderators, hm. Oh well.

Marla

Dave Tweed generates a bit more than Nick

forest

Does EE have a particularly high amount of drama wrt moderators going against community/meta consensus?

Marla

12:20 AM

Hey, that is axtuallyngood, intra-moderater challenges

forest

Not when it results in migration wars or edit wars. :P

Marla

We are much more vigilant regarding quality of questions

forest

Looks like most of the posts on meta hating on Dave Tweed are downvoted.

It's usually easy to tell when mod hate is the result of a bad moderator vs some disgruntled user who doesn't understand why a community deletes their crap question.

Marla

That is why opinion questions and answers are so vigorously challenged. The desire for high quality results

forest

I see. I asked a new meta question on that topic, asking whether or not it counts as opinion-based if it's well-documented by the community. I can totally understand though. Too many sites just have crappy questions that never get cleaned up...

Marla

12:25 AM

Sometimes a hated or down voted moderator inadvertently sacrifices for the good of the data integrity.

forest

It looks like in some cases, a moderator is hated because the overwhelming majority of meta (and other moderators) consider one thing on-topic, but that one moderator considers it off-topic. At least that's the gist of what I'm getting.

Marla

Debate at the moderator level is good. A site that has no conflict will decay into average.

forest

True. I'll withhold forming an opinion since I'm the new guy here. :P

Marla

@forest . .The hard facts, and not opinion was what kept me coming back here when I first discovered the site. I was SO tired of forums where any idiot could profess knowledge.

forest

Agreed. The community is much smarter here on SE.

Though the fact that downvotes are worth only -2 while upvotes are worth +10 for answers does skew rep in favor of people who post a lot, rather than people who are well-received... That's a problem on some of the smaller sites I'm on.

Nick Alexeev

12:34 AM

@forest W5VO isn't plural

Marla

LoL

Hi Nick

We been debating you

forest

heh, true

Yeah, I just saw some meta posts arguing against your decision to make repair questions off-topic from a year or so ago and was curious about the intensity of the response. Seems like you ignored community consensus or something.

Nick Alexeev

18 mins ago, by forest

It just looks like a moderator who is going against the will of the community and other moderators or something. I mean I'm new to EE so I won't pass judgement.

forest

Yeah I'm not judging, I have no opinion on whose view is valid.

Still reading through meta. :P

Nick Alexeev

I like this back of the hand stuff. First you discuss others, then you say "I don't want to pass judgement."

You want to look... "nice".

You look... "nice".

forest

12:38 AM

:/

I like to get a feel for communities. I asked about some drama I saw on meta. That's not incompatible with not having an opinion yet or passing judgement.

Marla

Hey @forest go do search on Olin Lathrop if you like to see controversy

forest

heh

Nick Alexeev

@forest Communit on EE.SE is about 300 people.

Marla

@NickAlexeev . . we have been heavily discussing opinon stuff

Nick Alexeev

I don't see somebody with 40 per (that's how much one needs to up- and down- vote on meta) as community.

295

Q: On large communities decaying over time, being nice or mean, and Stack Overflow

Wandering about the Internet, I stumbled across why online communities decay over time. This is a rather good article and should be read in conjunction with a group is its own worst enemy. I strongly advise people to read both of these articles — they will give appropriate background for some of...

discussion

forest

12:41 AM

How many of the votes on meta are from people with a low rep count?

Most of the comments I was reading came from established users.

Nick Alexeev

A lot of votes are from low-ish reputation users. Especially on questions which propose to lower standards.

Marla

One thing of interest is that once you are a moderator, your moderator vote dosnt require other supplemental votes. Which is a double edge sword. You can no longer put your vote to close in among the other 5 required. Its instantaneous closure.

forest

It looks like a lot of the drama is centered around the topicality of repair questions lol

Marla

Thus the MOD seems to be demanding of closure

Because there is a DIY repair site

forest

huh

Nick Alexeev

12:47 AM

@forest If we declare repair question on-topic, we will be swamped.

Marla

Do it yourself repair has separate site

forest

But they are on-topic according to meta. Lemme check again and see if those are just an old consensus before DIY became a thing.

Yeah they are still on-topic if they relate to EE.

Nick Alexeev

Basically, laypeople's electronics and electrical questions can't be on-topic here. Or we'll be swamped. This includes: gadget and appliance repairs, house wiring, car wiring.

Marla

The fundamental here is DESIGN.

forest

Right, it seems the consensus is that repair is on-topic, but it has to meet a minimum standard of quality / good understanding of the device.

Same with identification questions (it looks like).

Nick Alexeev

12:49 AM

@forest Repair questions are off-topic. I'm saying this so you don't get wrong ideas.

Marla

DESIGN

forest

@NickAlexeev No I get it. I'm not under the impression that I can go and ask how to repair a busted device or replace a burst capacitor.

Nick Alexeev

Are you sure you get it? Because you just wrote this:

> ... consensus is that repair is on-topic ...

forest

E.g. electronics.meta.stackexchange.com/a/6095/177824

Marla

Even though questions get closed, we do often enjoy (and allow) questions off topic

forest

12:51 AM

> Questions on the repair of consumer electronics, appliances, or other devices must involve specific troubleshooting steps and demonstrate a good understanding of the underlying design of the device being repaired.

Nick Alexeev

This site is about design. This site is not about repair.

forest

@Marla Ah, so it's a case-by-case basis kind of thing?

Nick Alexeev

This site is about EE design. This site is for designers.

@forest Very much case-by-case. 90% of repair questions get closed. The bar is high.

forest

Yeah that's fair.

Marla

No, not case by case. Just that some questions get lots of response before being closed.

forest

12:53 AM

I think I do understand how repair in EE has a different meaning than repair in the traditional sense that a consumer would understand it as.

Marla

Great point

forest

E.g. repairing a faulty circuit schematic vs repairing a busted RC helicopter.

Nick Alexeev

@forest About your meta question about the undocumented feature. Here's what you could do.

Trial-run it. Write-up the question (if it isn't too long). Post it as an appendix to the meta post.

forest

Well my meta question is more about the policy in general, since undocumented features like opcodes or jtag interfaces would seem to me to be very on-topic. My specific question might not be on-topic for other reasons as well, so I didn't want to poison the question by including variables that are not relevant.

Not to mention, I already got a bit of an answer to the NDS sensor question in here (regarding knowing that the model is ARM946E-S not being enough to answer whether or not the vendor has given it a sensor).

Marla

OK. I am off to watch television Jeopardy

Edited

forest

1:00 AM

Aight, have fun :p

Shalvenay

@Marla DIY.SE deals with Home Improvement, despite the common name -- electronics repair is off topic there

Marla

@Shalvenay oh, yikes.

Shalvenay

@Marla I think the current EE.SE repair policy is adequate

Marla

Do we have a site for electronics repair?

Shalvenay

@Marla we do not

Marla

1:07 AM

Probably just as well. If you don't know enough about repair to make it a valid here on EESE then you shouldn't be doing it

Shalvenay

@Marla yeah, "how fix X?" questions from newbs don't fit into the Stack model anyway

forest

Except Stack Overflow, which probably has some of the lowest quality questions. :P

Shalvenay

@forest :P some of us have been in SO review queues before shudders although EE.SE queues can get pretty gnarly as well

the queues here make me wonder "is something wrong with how we teach EM in high school physics?!"

forest

My home site is Information Security, so we get a bunch of "help I just got a virus!" and "how do I write a virus also pls hack my gf's facebook for me plz", and sometimes even "i know the cia is after me they i know they hacked my computer and my ps3 and my tv remote control and it persists even after i buy a new computer halp".

Shalvenay

@forest yeah, I sympathize with the reviewers there

my problem half the time when I ask questions is that they just get...no traction

I'll write up a detailed if dense question, and it'll barely see any play in terms of views, comments, voting, etc

forest

1:12 AM

Same. Some of my questions go above the heads of most people (often because it's on some aspect of information security that is not popular, i.e. anything that isn't web app exploitation and sqli, not because it's particularly complex).

Or 1k+ views, 5 votes, one comment asking what an acronym means, no answers...

Shalvenay

@forest heheh. I'll toss a few examples of mine from EE.SE here

Nick Alexeev

@Marla ifixit.com helps with electronics repairs sometimes

Shalvenay

@NickAlexeev they have some interesting teardowns too

(although my best example got zapped by Community for having such low traffic!)

electronics.stackexchange.com/questions/227869/… (10k+ link)

forest

I asked a quesetion about measured boot which I think is a good question. It's not even that complex, yet it has 1 vote and a little over 50 views, probably because it was about hardware and not web exploitation. I also asked a historical question about a crappy design decision in JavaScript, and got 41k+ views and 206 votes.

Or the one I mentioned before with 1k+ views and 5 votes, but no answers...

Nick Alexeev

@Shalvenay DIY.SE stack is as badly named for Home Improvement. Judging just from the URL, someone might think that it should cove DIY appendectomies. Similar thing with Electronics.SE for Electrical Engineering stack. In the minds of most people iPhone and remote control are filed under "electronics".

forest

1:18 AM

@Shalvenay I don't have 10k rep on EE. :(

Shalvenay

as to stuff that you can see:

electronics.stackexchange.com/questions/275277/… (wound up self-answering this one after a comment convo with SamGibson helped me troubleshoot it)

forest

Yup, it's always the questions that require someone to think and not regurgitate common knowledge that trip people up.

Shalvenay

and this one electronics.stackexchange.com/questions/219458/…

(related -- the runt pulse theory seems plausible but not something I can really confirm at this point)

Nick Alexeev

SPI is a tradition

Shalvenay

@NickAlexeev hahahaha :D yes!

0

Q: How does the CMRR of a fully differential amplifier change with input impedance imbalance?

Most pro-audio microphones provide a balanced output -- this is intended to provide improved noise rejection, and provided the output and line impedances are appropriately balanced, it does. However, with conventional differential-to-single-ended conversion stages, source and line impedance imba...

impedance differential common-mode balanced

1

Q: What's with the minimum voltage rating on *some* IEC 60947-2 MCBs?

It seems that some, albeit not all, miniature circuit breakers have a minimum voltage rating under the IEC 60947-2 scheme (often 12VAC/DC). For instance, see the Sprecher/Schuh L8 series (PDF page 8) as compared to, say, a Schneider Electric Multi9 C60SP which has no documented minimum voltage r...

circuit-breaker

and have two other stumpers I still have no satisfactory answer to to this day

forest

1:22 AM

Well at least you have some answers. :P

1

Q: Verifying that the CRTM is read-only for the purpose of trusted computing

When using a TPM to provide integrity measurements for a system via SRTM, the trusted computing base is reduced to only the TPM and CRTM. The CRTM is a component of the BIOS which executes first and allows the TPM to hash the remaining contents of the BIOS, in effect making it so that even compro...

hardware tpm bios trusted-computing boot

I got nothing ^

Shalvenay

@forest that's me with the CMRR question :P

forest

Damn, two years ago too.

Nick Alexeev

Abuse of SPI tradition is a sport. I've worked with a fellow who had a silver medal in abuse of SPI tradition. Analog Devices has a gold medal in abuse of SPI tradition. This sport can be played by a team too.

2

Shalvenay

@forest although, I do have a theory on how they (or at least I would) implement what you're talking about in your IS.SE question if that helps you?

Marla

It is normally at this time that I post the link to the Monty Python abuse clinic.

forest

1:24 AM

@Shalvenay Oh?

Marla

I won't do it now :)

Shalvenay

@forest can you entertain the assumption that the BIOS Flash chip uses a SPI interface (not some legacy interface such as LPC or parallel)?

forest

Yeah

Shalvenay

lemme grab a representative D/S for a SPI flash chip then

forest

I would have assumed that the flash chip just has an address region which is locked and refuses write commands. At least since the CRTM (bootblock) needs to be in the flash chip anyway and can't be stored anywhere else.

Shalvenay

1:26 AM

ww1.microchip.com/downloads/en/DeviceDoc/… if you wish to follow along, btw

forest

The idea is that the CRTM is read-only but all it can do is verify the hash of the rest of the BIOS (well, send the hash to a TPM PCR which verifies it) and then jump to it.

loading pdf now

Shalvenay

although this particular chip seems to be a more interesting beast than usual

forest

One thing that frustrates me is that terminology often changes depending on what point of view someone is coming from. Like a security person will say CRTM but a BIOS developer will say BIOS bootblock. Who knows what a flash chip designer would say.

Shalvenay

older Flash chips had fixed-location bootblocks that were used for this, yes -- the SST part I linked does something a bit different though

it has an OTP force locking register -- burning a bit in that register locks the corresponding Flash block permanently

forest

So sort of like 8051 write protect lock bit?

Shalvenay

1:30 AM

@forest right, although truly OTP instead of being erased by a chiperase command

forest

ah

Where does it say that in the datasheet? Still reading it.

Oh, section 4.2?

Shalvenay

4.1.3

forest

Ah

> The non-Volatile Write-Lock Lock-Down register is an alternate register that permanently prevents changes to the block-protect bits. The non-Volatile Write-Lock Lock-Down register (nVWLDR) is 136 bits wide per device: one bit each for the eight 8-KByte parameter blocks, and one bit each for the remaining 32 KByte and 64 KByte overlay blocks.

@Shalvenay It could also be the volatile hardware protection, since the threat model assumes an attacker that can send arbitrary writes over SPI (anything they can do with ring 0 on the main x86 CPU), not remove or tamper with the actual chip.

Shalvenay

I think that style of mechanism is the common mechanism, yes

and since it's OTP floating gate cells (old-style EPROM cells minus the UV window), the only thing an attacker who isnt' into decapping could do is make it so your next BIOS update bricks the computer

Marla

@Shalvenay . .In case you missed the earlier discussion on water cooled cables and bussbars, review the earlier chat. Might come in handy some day.

Shalvenay

1:39 AM

(by causing a partial update due to unexpectedly locked blocks)

forest

Well an attacker with physical access could do much scarier things, like add a BIOS to LPC or even over legacy PCI! Or just swap the BIOS chip. :P

Shalvenay

@forest yeah XD they have way better options than trying to tamper with the existing BIOS flash at that point

forest

@Shalvenay A good BIOS vendor will have recovery code in the bootblock.

Shalvenay

@forest true, true

forest

Is there any way to detect the state of these lock bits over SPI?

Shalvenay

1:41 AM

@Marla I saw! kinda makes me wonder at what point you'd consider that for building wiring

Marla

@Shalvenay . . you don't willingly do water cooling when air cooling will work

And, I suppose, that is the point where you consider it

Shalvenay

@forest it's possible, although the normal tools for manipulating BIOS Flash chips don't seem to support it -- you'd write all zeroes to the block protection register, then readback the block protection register and see what bits were still set. thing is, flashrom can't read or write protection regs at all. you'd need to poke the SPI with a custom tool.

at least, for the chip in the linked datasheet -- parts probably vary somewhat in their semantics

forest

Yeah that's fine. Even writing a custom kernel driver would be OK. :P

Shalvenay

and then you'd compare the protected blocks to a memory-map of the Flash chip

and see if the CRTM block is perma-write-locked or not

forest

So it really depends on the specific flash chip then, huh.

Shalvenay

1:47 AM

if it is? then well, attacker is going to need to crack the case and swap chips. if not? then well...it can get scribbled on by say a malicious BIOS update :P

@forest yup. flashrom should at least be able to tell you that much

forest

Sadly flashrom often has no idea what a given chip is. :(

Especially if it's in a laptop or obscure server.

Shalvenay

@forest yeah, that can be annoying. sometimes there's no substitute for getting eyeballs on the chip

forest

yeah

It's a shame that isn't integrated into the chipset.

Shalvenay

@forest actually, the problem is chips that don't play nice with JEDEC's Flash interface standards :P

because every Flash chip on this planet is supposed to have a unique JEDEC ID code you can read back

forest

I mean the x86 chipset, since it is the best place to hold that kind of thing. It knows where to find the BIOS, so it should be able to hash the BIOS into the TPM's PCR0.

Shalvenay

1:50 AM

using a standard command

but nooooo....some vendors are too daft for that :P

forest

Yeah, hardware vendors don't always get many things right...

Shalvenay

(never mind ECs that react badly when the BIOS Flash gets poked at)

forest

Especially RAM.

Shalvenay

@forest which is embarrassing, because that's about as standardized as it gets :P

forest

Heh, or even arbitrary IO ports. I love that scary warning on lm_sensors saying that probing for a temperature sensor might result in the system crashing. :D

@Shalvenay APCI is even worse. 1000+ page specs yet well standardized, and even still ACPI tables for many OEM laptops are just badly broken or corrupt (esp DMAR).

Shalvenay

1:55 AM

@forest even Intel's mobos sometimes have ACPI issues :P

forest

Yup

Which is a problem for me, because every system I work on needs proper DMAR.

ACPI pisses me off. I hate having to consult ACPI tables just to power off a machine. I much prefer being able to write to one single IO port to shut down, not go through a complex hardware-specific dance to find the right magic value that lets me poweroff.

@Shalvenay Do you know if there are one-time programmable fuses for modern x86 Intel chips which determine whether or not JTAG (ITP-XDP) is supported?

Just question I had on my mind from reading more about these fuses.

Shalvenay

2:15 AM

@forest I do not -- you'd have to check Intel's datasheets for the details

forest

Requires signing an NDA to get the JTAG datasheets sadly. :(

Oh well.

JRE

10:34 AM

Waiting for the police to show. "Microsoft" called and told me my (Linux) computers are being used by hackers. If I don't let "Microsoft" access my computer to remove the hacker software, then they will have to call the police and have my computer confiscated.

I told them to go ahead and call the police. They'd be very interested in the people commiting fraud. The German police would be really thrilled to get such a call from India (in English.)

Marla

1:26 PM

@JRE . . Sounds like a very good framework for the scam.

barlop

4:26 PM

is a latching switch and a toggle switch the same thing?

Shalvenay

@barlop no -- toggle generally is a physical form-factor/styling, while latching is a behavior

one can have momentary toggle switches and latching slide or rocker switches

barlop

so can a toggle switch be non-latching?

Shalvenay

@barlop yes

barlop

so latching and momentary are behaviours and opposites?

Shalvenay

@barlop correct

barlop

4:36 PM

thanks

Nick Alexeev

5:18 PM

@Shalvenay (cc: @barlop) I don't readily agree. But you can convince me with an example of a toggle switch that's momentary.

Rocker switch and lever switch are form factors. Momentary switch and toggle switch are behaviors (and opposites). Afaiu.

Shalvenay

@NickAlexeev ah, what you're calling "lever switch" is what I (and it seems much of the industry) calls "toggle switch" no?

Nick Alexeev

@Shalvenay Not quite. Lever switch is usually a toggle switch. Momentary lever switches exist too. [Similar thing with rocker switches.]

Shalvenay

@NickAlexeev hrm. terminology differences are annoying? :)

Nick Alexeev

@Shalvenay Especially for fringe cases like momentary lever switches. (Now imagine a 3-position lever switch where middle is neutral, one side is toggle, the other side is momentary. Makes me go numb on terminology and just use part numbers.)

Shalvenay

5:36 PM

@NickAlexeev yeah

Nick Alexeev

6:13 PM

2

A: Is there such thing as supreme board for international arbitrators?

They can't even write English on their certificate. ... training of good stand, and is in accord with American and European standers... That's an obvious typo. Of course it should say ... training of good sit, and is in accord with American and European sitters... Is it worth pay...

Marla

6:34 PM

@NickAlexeev. . I often wonder if they really have people fall for it.

Nick Alexeev

7:01 PM

@Marla That O.P.'s reltive came fairly close to falling for it.

telegraph.co.uk/technology/microsoft/9346371/… includes links to a Microsoft research paper with exactly this conclusion - "by sending an email that repels all but the most gullible the scammer gets the most promising marks to self-select" — AakashM Sep 18 at 7:43

barlop

8:53 PM

I guess there's no doubt though that momentary switch and latching switch are opposites and behaviours. If I go to google images then search for 'toggle switch' seems to be a form factor as Shalvenay said, and
looking up "toggle switch", I see this dictionary, https://www.merriam-webster.com/dictionary/toggle%20switch which says "an electric switch operated by pushing a projecting lever through a small arc" Would you agree though that "lever switch" is a form factor, while "momentary" and "latched", are behaviours?

Nick Alexeev

9:12 PM

@barlop In my view, toggle is equivalent to latched.

barlop

@NickAlexeev well I get that but it disagrees with others and the dictionary

Nick Alexeev

With all my respect to Webster dictionary, it was written and edited mainly by linguists, rather than practicing engineers.

A rocker switch is a toggle switch, even though it has got no "projecting lever". [Except for the momentary rocker switches, of course.]

@barlop Anyway. ~~What do you need these definitions for?~~ What are you writing?

barlop

how do you define 'need' ?

Nick Alexeev

@barlop I'm not going to go all Maslow here. ;)

barlop

I'm digesting information. At some point I might have to ask a question regarding it, and it helps to have the right definitions and definitons that are widely understood and agreed upon , in my mind, so that if they need to come out of my mind in a question on a forum or elsewhere, then i'll be understood, and if I want to read other things i'll understand what they are saying

Nick Alexeev

9:25 PM

Over-communicate, and you're likely to be understood. One can always say "toggle [latching, bistable]". Add a part number of the switch for a good measure. That multiplies your chances of being understood.

barlop

Seems latching, momentary, and lever, are agreed upon. The first two being opposite behaviours, the third being a form factor. I don't need to refer to a 'toggle switch'.

Nick Alexeev

@barlop That does it.

barlop

would you say that "push button switch" is a form factor?

Nick Alexeev

@barlop Yes, that's a form factor. You'd also have to specify the behavior. As in "momentary push button switch".

barlop

thanks

Transcript for

Electrical Engineering