// #eevblog — 2019-01-28
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
01:36:59 Join: Guest|80172 ([email protected]) to #EEVblog
01:37:24 Quit: Guest|80172 ([email protected]) Quit: http://www.kiwiirc.com/ - A hand crafted IRC client
02:45:05 Quit: coppice_ ([email protected]) Connection closed
02:45:25 Join: coppice_ ([email protected]) to #EEVblog
02:45:25 Mode: (by ChanOP) +o coppice_
03:35:16 Join: cappicard ([email protected]) to #EEVblog
04:17:28 <DHess_US> https://i.imgur.com/Gb3LYNP.jpg
04:26:10 <T3sl4> that's a whoops
04:44:57 <steve30> what is it?
04:47:02 <steve30> oh, are they cables?
04:59:53 <rust_collector> very short cables
05:00:25 <coppice_> Some of those cables are set right into the concrete. that's bad original construction
05:04:18 <rust_collector> should be fine, as long as you don't drill holes in it
05:05:02 <coppice_> its not fine at all. The lack of give causes degradation through thermal cycles
05:05:51 <rust_collector> should have plenty of give, in that yellow plastic tube
05:06:17 <coppice_> the power cables are not in a yellow tube
05:06:30 <rust_collector> well, they are not important, lol
05:07:32 <coppice_> the orange cables in the tubes are fine, as long as the strength requirements of the concrete are OK with multiple voids closely spaced like that
05:08:52 <rust_collector> I wonder what that concrete was. but meh. time for coffee
05:10:07 <DHess_US> They make special cable for direct burrial in concrete.
05:11:14 <coppice_> these aren't special. They are basic plastic insulated cables, with no armouring
05:21:25 <rust_collector> well, some times you add something to a building, and there are cables. I guess it made sense at the time
05:21:40 <rust_collector> or maybe it was just cheap
05:40:09 <coppice_> builders can be bodgers might be more on point
05:41:24 <rust_collector> yeah, especially if you want the job done cheap
05:42:17 <coppice_> hey, there's no point in bodging if you are just going to pass the savings on
05:49:37 <rust_collector> if it means you get the job, maybe?
05:51:11 <rust_collector> but, I am looking at pcb material. I wonder what thickness I would like, for general hobby stuff. they have 0.5mm to 2.5mm fr4 I feel 1.5mm would be the one...maybe
05:52:00 <coppice_> 1.6mm is classic PCB material. 1mm is a good choice for smaller boards. 0.5mm can start to be a bit bendy
05:52:57 <rust_collector> yeah, I think 1.5 would be the most universal
05:53:34 <rust_collector> they sell sheets of 510X570mm
05:53:52 <rust_collector> so, have to be something I can use for random things
05:56:52 <DHess_US> If you are doing RF, pick a thickness for convenient microstrip widths. Otherwise 1.6mm or whatever the classic is which can use edge connetors.
05:58:47 Join: DogWomble ([email protected]) to #EEVblog
05:58:57 Mode: (by ChanOP) +o DogWomble
06:02:05 <rust_collector> well, I would not say I "do" anything... I am just tinkering, and the power supply I keep nagging about is the most advanced thing I have done until now
06:06:36 <DHess_US> The standard thickness is good enough for structural elements like panels. Thinner printed circuit board material is too flexible for that.
06:08:52 <rust_collector> sounds good
06:32:58 Join: ninux ([email protected]) to #EEVblog
06:47:45 <ninux> Hi there! Recently I've watched #317 "PCB Tinning Myth Busting". Afterwards I've tried to find a serious perspective on this topic (papers, etc.) but I couldn't find anything narrowed to the use in PCB design. Therefore I'm thinking of doing some research on my own and share the results. So far I could validate Daves' DC experiment results (reducti
06:47:45 <ninux> on of 12% up to 63% of the DC resistance using reasonable techniques for tinning PCB traces). Now I want to analyse the AC behaviour of tinned traces. Currently I'm looking for background information and basic theory on the electrical properties (DC and AC) of layered conductors (e.g. Cu and SnPb, etc.). So far I've been told that there are compani
06:47:46 <ninux> es coating condutors with alloys to improve skin-effect behaviour. Does anyone here have good recommendations on literature for this topic or relevant to the PCB tinning?
06:57:04 Quit: DogWomble ([email protected]) Ping timeout: 121 seconds
07:01:14 Part: T3sl4 ([email protected]) from #EEVblog
07:01:18 Join: T3sl4 ([email protected]) to #EEVblog
07:01:18 Mode: (by ChanOP) +o T3sl4
07:03:38 <T3sl4> ninux: AC resistance is where skin effect takes over; in a typical microstrip configuration, most current flow is on the inside surface where you can't get any plating
07:04:06 <T3sl4> and since most of the conductor's cross section goes unused, plating won't help, or will only hurt
07:06:33 <T3sl4> this depends heavily on the frequency and geometry. a typical 2oz copper board will perform better with solder fills up to 10s, maybe 100s of kHz; there will be little difference up to some MHz, then performance will be worse above, idunno, 100s, 1000s MHz
07:13:45 <ninux> T3sl4: I totally agree with you for HF design (as you mentioned microstrips in particular). But how about power traces with harmonics? Those power traces conduct DC up to some hundreds of kHz (typically below the MHz range). Could a tin-coating be benefitial? One question is intriguing me in particular: what happens in the contact zone of the trace
07:13:45 <ninux> (copper) and the tin? All I see in literature is the contact resistance and so on (effects from one medium to another), but nothing about the electrical properties along these contact surfaces.
07:15:45 <T3sl4> so, switching supplies? resistance at DC and fundamental dominates, power above 10th harmonic is <1% of total, tinning will be largely helpful
07:16:06 <T3sl4> the intermetallic layer is very thin, microns; irrelevant for electrical or thermal purposes
07:23:30 <ninux> T3sl4: So you think in SMPS one could consider using traces or plane tinning for improving conductivity? About the intermetallic layer; I was just wondering about the current distribution (looking at the crosssection) for the lower harmonics since the copper can be usually very thin compared to the tin coating. Is there any literature or research b
07:23:31 <ninux> een done on this? I really couldn't find any narrow resources so far.
07:26:47 <T3sl4> probably
07:27:09 <DHess_US> I thought they tinned like that just to raise the cross section enough so they dont' need a wider trace or heavier copper.
07:27:10 <T3sl4> skin effect just goes down by layer, whatever's left from one layer goes on to the next and so on until the current is all gone
07:27:27 <DHess_US> I mean tin like soldered the whole trace.
07:27:53 <T3sl4> DHess_US, exactly
07:44:24 <DHess_US> Switching transformers which do not use copper strips may use Litz wire to some advantage but that is the only place I have seen anybody bother.
07:44:49 <DHess_US> The printed circuits should have short and fat traces where it matters anyway.
07:55:48 <ninux> DHess_US Totally agree. But the ever growing integration of control (small components as QFN and BGA parts) and power circuitry have some opposite design requirements. In some cases tinning traces can be benefitial (as far as I understood). For me the question arose: Can we quantify the effect and how can we make it reliable in production? So far I
07:55:48 <ninux> haven't found any data on this =)
07:58:12 <ninux> Dhess_US It intrigued me since this "technique" is apparently widely used but there seems to be no reliable data available
08:00:45 <T3sl4> quantify? sure, point a thermal cam at it.
08:04:27 <ninux> T3sl4 Sure this is a real proof and I totally agree on that. But what I ment was not a "post-process" but rather a method for prediction before we get the boards i.e. during layout stage =)
08:04:58 <T3sl4> well, so
08:05:26 <T3sl4> your design process goes from a $100 proto run and however many hours spent in layout, plus a one time buy of say $1000 for the thermal cam
08:05:51 <T3sl4> to a multiphysics modeling system plus the training and support to use it, in the $100k range
08:06:06 <T3sl4> you might end up determining that the former process is acceptable after all
08:11:34 <ninux> @T3sl4 I see your point and I again agree with you =) What I had in mind was just a fun research project to work on in my spare time and to maybe gain some new perspectives on some basics in electronics design, not really to use it in a productive enviroment as such
08:12:26 <T3sl4> it wouldn't be terrifically hard to write half of the simulator yourself, assuming DC current flows, and standard formats (gerber?)
08:12:57 <T3sl4> to add thermal to that, you could add some dissipation rules, but getting actual convection in a system, who knows
08:13:12 <T3sl4> i don't know of such a tool offhand though
08:13:33 <T3sl4> i think there are tools offering current distribution plots, which would get most of what you want
08:14:08 <T3sl4> i forget if AD18/19 was supposed to introduce something like that, or if one of the higher tier EDA tools does that already
08:15:16 <DHess_US> The only place I have dealt with this is RF applications like helical resonators where standard procedure would be to use silver plating. But the benefit is limited and in most cases, the loss in Q can be made up for by making the resonator 10% larger.
08:15:58 <DHess_US> At least if you are using copper. Cavity reconators use something like invar so silver plating would be more important.
08:16:35 <T3sl4> heavy plating at that, to ensure all the skin effect is in the plating
08:17:06 <DHess_US> Hmm, I've seen heavy plating used on old oscilloscope circuits. I wonder if that was why.
08:17:28 <DHess_US> Not consistently so I guess not for that reason.
08:17:48 <T3sl4> or because they wanted the bling
08:18:09 <DHess_US> No, and heavy gold plating was a reliability problem later.
08:18:33 <DHess_US> Learing to fix those types of circuits was fun.
08:21:14 <DHess_US> What is the best way to remove heavy gold plating without overheating and damaging the printed circuit board substrate?
08:21:27 <T3sl4> cyanide
08:21:44 <T3sl4> ...for some values of "best"
08:22:13 <DHess_US> hehe. My solution was *low* temperature tin/lead solder.
08:22:45 <T3sl4> yes, dissolving it off with metals would be the next best
08:23:12 <DHess_US> The trick was figuring out that lower temperatures dissolved the gold slower but damaged the substrate even slower yet.
08:23:17 <ninux> T3sl4 I was thinking of using Empire XPU as it allows to import/export gerber designs and one can simulate very easily typical materials used in electronics designs (since the tool is dedicated to electronics design simulation).
08:23:20 <T3sl4> a standard aerospace operation, soaking gold (and unleaded tin, when lead is desired/required) in solder
08:23:33 <coppice_> dipping in several tin/lead solder baths is a common gold removal technique. as the first bath gets seriously polluted you send it for gold recovery (they pay you for it), and put a claen bath at the end of the row
08:24:01 <DHess_US> A NASA article was the inspiration but I had to figure out that lower temperatures worked best myself.
08:24:33 <T3sl4> a curious kind of "pollution"
08:24:39 <coppice_> its done a lot with RF power transistors to get the bloody gold off the tabs, so it doesn't embrittle the joints and cause transmitters to catch fire
08:25:11 <DHess_US> That was the problem these old heavy gold plated boards had.
08:25:36 <T3sl4> ninux: that looks alright i guess, don't know that it will be as well suited to cross sectional current density or setting up platings or analysis at SMPS frequencies, but that would be something to look into
08:26:26 <DHess_US> I know I've read articles which discussed silver plating and what surface finish worked best for skin effect applications but it was a long time ago.
08:27:32 Quit: DL3CE ([email protected]) Quit: Nettalk6 - www.ntalk.de
08:28:01 <coppice_> silver plated RF parts all have a matt finish, but I don't know the details of those coatings
08:28:33 <T3sl4> yeah, those RF guys aren't very brite
08:29:19 <coppice_> try operating a silver plated RF connector on a transmitter in HK or Singapore, and in less than a week its pitch black. Such is the effect of a warm high humidity atmosphere
08:30:02 <T3sl4> warm high humidity atmosphere with a million cars below, even
08:30:30 <DHess_US> The tarnish doesn't have much effect.
08:30:57 <coppice_> The tarnish makes the resistance soar, and things gert very hot and fail
08:30:57 <T3sl4> at least it's easily wiped off, or usually is
08:31:13 <DHess_US> haha. I've never seen it easily wiped off.
08:31:45 <coppice_> what kind of oxide ever wipes easily off?
08:32:06 <T3sl4> dihydrogen monoxide certainly does
08:32:09 <DHess_US> I have a 500MHz silver tuned filter which the stuff is apparently indellible.
08:32:26 <coppice_> and where are these millions of cars in cities with low car ownership?
08:32:39 <DHess_US> That just means there are even more people without cars.
08:33:57 <T3sl4> coppice_: or more to the point, whatever nearby coal power plants and shipping ports there are
08:34:50 <coppice_> T3sl4: most cities have worse atmospheres, but you can get away with using silver plated connectors long term if the humidity is low
08:36:44 <DHess_US> There are easy chemical ways to remove silver sulfide but I didn't want to damage the surface of the plating so I just left it.
08:37:14 <DHess_US> Where you live makes a huge difference for silver tarnishing.
08:37:49 <coppice_> in HK it takes very little time for the entire silver layer to turn into a salt
08:37:52 <T3sl4> yes, humidity accelerates it
08:38:18 Quit: ninux ([email protected]) Quit: AustNet Chat Network - WebIRC
08:38:19 <coppice_> if you remove the tarnish after a couple of week you will find no metallic silver left
08:38:50 <DHess_US> It's actually not difficult to turn it back into silver on the surface.
08:39:13 <T3sl4> add a reducer
08:39:18 <T3sl4> downside is you get a porous deposit
08:40:00 <coppice_> you can turn the salt back to metallic silver, but it no longer forms a consistent coating
08:40:55 <DHess_US> The silver sulfide protects the silver under it from further tarnishing so I just left it.
08:41:08 <DHess_US> coppice_, exatly
08:42:16 <coppice_> I've never seen the initial tarnish protecting anything. connectors just continue to corrode, and the tarnish starts to fester
08:43:20 <coppice_> this seems to be like aluminium in busy streets, where the oxide layer provides little protections, and aluminium objects decay into a festering mass of corrosion
08:44:16 <DHess_US> connectors have surface rubbing. Silver plated RF surfaces are just exposed.
08:45:56 <coppice_> my experience is mostly with connectors which are rarely touched, so the rubbing is not a big deal
08:46:33 <DHess_US> I have a lot of old silver plated connectors and they have a similar dark hard tarnish except where the sliding surfaces are.
08:46:50 <DHess_US> I suspect there is an issue with exactly how the silver is applied.
08:48:07 <coppice_> silver eventually darken in most places, unless you use dessication. That's why there are dessicators all over the place in lots of RF equipment. The speed in place like HK is just so fast. A connector won't last a month
08:49:34 <T3sl4> there's also "silver preserver" sheets, which i don't think are quite the same as silica gel packets, but apparently something to help
08:51:52 <DHess_US> I wonder what would be a good practical test for RF surface conductivity. Would a VNA and microstrip be enough to measure the differences?
08:52:19 <DHess_US> Hmm, I guess the problem there is that the surface that matters is on the bottom.
08:52:24 <T3sl4> right
08:52:33 <DHess_US> So a VNA and some type of plated transmission line.
08:52:47 <DHess_US> Would a VNA be good enough to measure the differences?
08:52:59 <T3sl4> that works well enough, for sufficiently accurate VNA, and length of TL, and ability to separate loss of the substrate from loss of the trace
08:53:13 <DHess_US> I know a cavity resonator would be good enough to measure the differences but that is more complicated to assemble.
08:53:36 <T3sl4> why not both?
08:53:37 <DHess_US> I was thinking like a plated air line.
08:53:47 <T3sl4> how about, plated tube slid inside a coaxial resonator structure
08:53:49 <DHess_US> A cavity reconator is just more complicated.
08:55:11 <DHess_US> The simplicity of an air line removes a lot of other variables.
08:55:34 <DHess_US> Might not be able to make it long enough though for good enough sensitivity.
08:58:57 <DHess_US> Yea, some kind of resonator is the way to go. Forget using a transmission line.
08:59:49 <DHess_US> Or maybe a transmission line resonator would be acceptable.
09:05:03 <EEVBlog> New EEVBlog Tweet: I'm now LIVE! with a Q&A https://t.co/0W6KPUwFoc
09:23:44 <Ampera> people are showing Dave with superchats
09:23:59 <bigbadwolv> lol
09:27:35 <bigbadwolv> poor Dave... He can't keep up with the vanishing donation comments haha
09:27:55 <Ampera> lol
09:32:42 <coppice_> is be putting out the begging bowl again?
09:33:10 <Ampera> he doesn't have to put out anything, he's being browned with cash
09:33:22 <Ampera> he's probably got over a hundred bucks atm
09:33:36 <bigbadwolv> I reckon, He's smashing it
10:48:31 <damian> is that a 4cpu motherboard behind him
10:48:34 <damian> dang
11:13:16 Join: Guest|54578 ([email protected]) to #EEVblog
11:13:55 Quit: Guest|54578 ([email protected]) Quit: http://www.kiwiirc.com/ - A hand crafted IRC client
11:24:12 Join: ThunderSqueak ([email protected]) to #EEVblog
11:24:12 Mode: (by ChanOP) +o ThunderSqueak
12:27:19 <DHess_US> https://tinyurl.com/y8d7f2z2
12:51:29 <ThunderSqueak> DHess_US, I saw that the other day :)
13:05:44 <DHess_US> I bought some to keep next to my microscopic space fleet. https://i.warosu.org/data/tg/img/0340/02/1407583961500.jpg
13:19:39 <mrpackethead> righht.
13:49:32 <DHess_US> Dude, you never know when you might need a microscopic space fleet.
13:50:30 <mrpackethead> how much did dave make for sitting talkign for 5 hours on youtube
13:53:49 <DHess_US> He now owns Australia.
13:54:22 <Ampera> so 5 USD
13:56:34 <DHess_US> Yup
14:20:07 Join: Halcyon ([email protected]) to #EEVblog
14:20:21 Quit: Halcyon ([email protected]) Changing hosts
14:20:21 Join: Halcyon ([email protected]) to #EEVblog
14:20:30 <Halcyon> Meow
14:23:38 <damian> hey man
14:24:29 <Ampera> Eyy, Halcyon
14:24:38 <damian> ey Ampera
14:25:18 <Ampera> Did you ever bring up the concept of an IRC Q&A to Dave, Halcyon?
14:35:08 <EEVBlog> New EEVBlog Tweet: Elon musks- Flying Car.. BUSTED!! https://t.co/gSSjybfvCD
14:42:50 <mrpackethead> no body but a few old farts uses IRC any more
14:42:59 <Ampera> I'm 17
14:43:12 <mrpackethead> your an old fart thats 17 then
14:43:20 <Ampera> probably
14:43:33 <damian> old fart in a young farts body
14:43:46 <Ampera> T3sl4, you're missing all the farting
14:55:30 <Halcyon> Heya
14:55:40 <Halcyon> I did briefly Amps
14:55:52 <Halcyon> Havent really discussed it properly
15:04:18 <EEVBlog> New EEVBlog Tweet: Ask Dave LIVE: https://t.co/I4htgsmedk via @YouTube
15:05:06 <Halcyon> There you go, looks like he has a Q&A on YouTubes Ampera
15:05:20 <Halcyon> Well did
15:05:36 <Ampera> I know he had a YouTube one, I'm just asking because I wanted to know if he still wanted, or even knew about doing it via IRC
15:05:43 <Ampera> as a way to bring attention to the channel
15:05:49 <Halcyon> Yeh I think he would be keen
15:06:34 <Ampera> Well since you're in primary contact with him, and now we have the bot going, maybe it would be a good time to do that
15:06:46 <Halcyon> Yeh I'll hit him up
15:07:14 <Ampera> alrhough, I wonder how eager he will be do it because of simply how much money he got from streaming
15:07:22 <Ampera> he got probably a couple hundred bucks just from superchats
15:07:45 <Halcyon> He could always stream the IRC window
15:07:49 <Halcyon> On YouTubw
15:07:50 <Halcyon> YouTubw
15:07:54 <Halcyon> YouTube
15:07:54 <Halcyon> Fuck
15:08:08 <Ampera> That is true.
15:08:47 <Ampera> Could also do it to bring attention to the forums, have the questions asked on the forums over a span of days, that way we can get more interesting and serious questions that he has good time to answer
15:21:24 <EEVBlog> New EEVBlog Tweet: YouTube To Curb Hoax And Junk Science Videos https://t.co/GWWxIMHV7U
15:21:59 <damian> yeah, but you tend to have more conversation if you span questions over time, it makes for more work for the host and not as good feedback
15:25:08 <Ampera> I'm confused as to what you mean by that, damian.
15:25:23 <Ampera> The idea would be to have conversation, that's what people like to watch.
15:25:33 <Ampera> That's effectively a podcast at that point.
15:34:15 Join: Guest|60376 ([email protected]) to #EEVblog
15:34:51 Quit: Guest|60376 ([email protected]) Quit: http://www.kiwiirc.com/ - A hand crafted IRC client
15:39:32 <damian> if you post questions for a few days before the sit down, then there's conversations based on those questions on the forums already, which the host would need to read through in order to answer and respond to the entire thing, unless they literally just take out posts that have questions and don't bother reading the rest
15:39:43 <damian> it can be very time consuming
15:41:25 <damian> https://www.reddit.com/r/techsupportgore/comments/akbk2p/customer_wants_to_repair_the_charging_by_himself/
15:41:25 <damian> haha
15:46:12 <Ampera> That's exactly what we'd do.
15:46:29 <Ampera> The thing is, there would be questions for Dave, to be answered by Dave, not other people.
15:46:50 <Ampera> They will look for more personal answers than general one.
17:08:33 Join: DL3CE ([email protected]) to #EEVblog
17:51:16 Quit: Halcyon ([email protected]) Ping timeout: 121 seconds
17:52:10 Join: Halcyon ([email protected]) to #EEVblog
17:54:22 Nick: Halcyon is now known as Halcyon1
17:57:34 Quit: DL3CE ([email protected]) Ping timeout: 121 seconds
18:02:24 <EEVBlog> New EEVBlog Tweet: Ask Dave - Live - 28th Jan 2019: https://t.co/nsYkpRHMON via @YouTube
18:08:25 Nick: Halcyon1 is now known as Halcyon
18:08:28 Quit: Halcyon ([email protected]) Changing hosts
18:08:28 Join: Halcyon ([email protected]) to #EEVblog
18:08:33 <Halcyon> Shitty NBN
18:15:28 <EEVBlog> New EEVBlog Tweet: I added a video to a @YouTube playlist https://t.co/I4htgs4DlM Ask Dave LIVE - Jan 28th 2019
18:30:33 Join: DogWomble ([email protected]) to #EEVblog
18:30:43 Mode: (by ChanOP) +o DogWomble
19:59:31 Quit: ThunderSqueak ([email protected]) Quit: Leaving
20:10:31 Quit: Halcyon ([email protected]) Ping timeout: 121 seconds
20:11:43 Join: Halcyon ([email protected]) to #EEVblog
20:13:48 Nick: Halcyon is now known as Halcyon1
21:31:31 * Ampera throws Halcyon1 one end of an ethernet cable plugged into his fibre
21:37:49 Quit: DogWomble ([email protected]) Ping timeout: 121 seconds
Search this day's buffer: