// #eevblog — 2019-04-21
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
00:07:13 <T3sl4> for a few years anyway
00:15:20 <rust_collector> plastic can be really random. I have one set of chairs that are... oh, from around 1990. they are changing to white, but are still ok. we have had many other sets come and go.
00:16:07 <rust_collector> It is probably some kind of concentrated cancer plastic, that is banned even in china.
00:16:26 <steve30> I repaired our wooden outside table last year. I was going to 'refurbish' it this year, but I expect the plastic stuff will be replacing it
00:17:28 <rust_collector> I find it is getting harder and harder to find good wood.
00:34:02 <T3sl4> ( ͡° ͜ʖ ͡°)
00:34:14 <T3sl4> I think steve30 can hook you up with that ( ͡° ͜ʖ ͡°)
00:38:56 <steve30> lol
00:45:39 <rust_collector> not going to comment on that
00:49:08 <steve30> I suggest rust_collector makes his garden furniture out of iron so that it can rust :)
00:49:35 <rust_collector> most of my "garden furniture" IS iron, or steel
00:50:07 <rust_collector> I tend to look at it as art. neighbors and wife do not see it that way
01:39:09 Join: Bluenote ([email protected]) to #EEVblog
03:09:59 Quit: Bluenote ([email protected]) EOF from client
03:11:18 Join: Bluenote ([email protected]) to #EEVblog
03:11:20 Quit: Bluenote ([email protected]) EOF from client
03:18:40 <coppice> steve30: plastic furniture is usually OK, as long as it doesn't get too much sun. That might be easier in Yorkshire than in most places :)
03:33:29 <steve30> hehe
03:33:40 <steve30> It feels reasonable quality anyway
04:15:37 Quit: Nothing\ ([email protected]) Quit: Leaving
04:40:18 Quit: ServalSpots ([email protected]) Ping timeout: 121 seconds
04:46:02 <Ampera> good morning
04:46:05 <Ampera> even though it's almost 3
04:46:07 <Ampera> q.q
05:00:48 <rust_collector> hey Ampera
05:03:51 <Ampera> hi
05:04:46 <DHess_US> It's always morning somewhere. Just like I only drink after 18:00 and it is always after 18:00 somewhere.
05:19:46 <rust_collector> I need to put together some kind of load, to test my crappy power supply. Do you people have any pointers for something reasonably simple? I guess 24v lightbulbs would do for now, but I think I would want something adjustable sooner or later
06:10:50 <Ampera> a big ass variable resistor
06:11:12 <DHess_US> It depends on how thorough a test you want to do.
06:11:31 <DHess_US> Worst case tends to be at low currents.
06:12:37 <DHess_US> So run at like 10% load using a resistor to stability tests and of course a 100% load for dissappations tests.
06:14:38 <rust_collector> well, I am not very advanced. This power supply is just the average lm317 with a 2sb1647 transistor, and I guess it is the transistor I am interested in testing, in this case.
06:14:39 <DHess_US> A more thorough test is to sweep a variable load from 0 to 100% while applying a small square wave to the reference and measuring the transient resposne as it changes with load current.
06:14:55 <DHess_US> Yea, those can be tricky sometimes.
06:15:21 <DHess_US> It depends on exactly how you setup the current sharing between the LM317 and transistor. Some ways are more stable than others.
06:16:32 <DHess_US> Oh, a big darlington. That's not a good idea.
06:16:38 <rust_collector> I am not planning on using this power supply as such, it was just the best idea I could come up with to have a way to see if the transistor would work for this kind of use.
06:16:47 <Ampera> https://pbs.twimg.com/media/D4n07uXXoAE_d2m.jpg:large
06:16:48 <Ampera> FEESH
06:16:52 <DHess_US> A 317 will provide more than enough drive for a single transistor.
06:17:22 <DHess_US> It can work with a darlington. It is just more prone to oscillation.
06:17:41 <rust_collector> ok
06:17:50 <Ampera> that's 3.5 pounds of salmon, glazed with a soy sauce maple (garlic & pepper) glaze, about to be smoked over cherry chips
06:18:03 <DHess_US> Give me a minute, I have a link to a book.
06:19:57 <DHess_US> https://archive.org/details/NationalSemiconductorVoltageRegulatorHandbook1980
06:20:10 <rust_collector> I used this circuit. I figured it would use the transistor more, and the 317 less, so would be better to test the transistor: http://www.bristolwatch.com/ccs/img2/lm317pass1.jpg
06:20:22 <DHess_US> Check out page 7-1 and 7-2.
06:20:29 <rust_collector> thanks
06:20:40 <rust_collector> oh, and salmon looks good
06:21:13 <DHess_US> Yea, that's the general idea.
06:22:26 <DHess_US> With a Darlington a resistor should be in series with the base to limit current. This is not neccessary for a big power transistor because the single base-emitter junction can carry amps without a problem.
06:23:29 <DHess_US> Figure 7.3 shows how to share current between the transistor and regulator so that the regulator's current limiting extends to the transistor protecting it.
06:24:04 <DHess_US> It doesn't really show how to use a darlington this way but it could be done by using two external diodes in series to match the darlington's two base-emitter junctions.
06:32:12 <rust_collector> so, ie you think it is a bad idea to use these. That is ok, I can find something else.
06:34:30 <DHess_US> Oh, they can be used. It just requires more care.
06:35:49 <DHess_US> Take a really good look at what R1 and R2 do in figure 7.3
06:36:10 <rust_collector> I have to look at this a bit, and compare the circuits. This started because I found 6 of these in that old sony amp.
06:36:18 <DHess_US> R1 in series with the emitter of the transistor is called a "degeneration" resistor and it lowers and controls the gain.
06:36:37 <DHess_US> Which becomes doubly important for a darlington because the gain is so high.
06:38:07 <DHess_US> So I'd use that circuit, but with two external diodes in series, 1N5401s are good, and set the ratio of R1 to R2 to control the ratio of current between the darlington and the regulator.
06:38:35 <DHess_US> And pay close attention to how the input and output capacitors are wired.
06:39:10 <DHess_US> Make the output capacitor like 10 to 100 microfarads per amp of output current. Use an aluminum electrolytic.
06:39:18 <rust_collector> so, when I look at the datasheet for the darlington, it has a 70 ohm resistor betwen the emitter and the.... "emitter of the second transistor" built in. that does a similar job as R1 in that circuit?
06:39:51 <DHess_US> C1 can be like 1/10th of the output capacitance. Of course there is an unshown bulk input capacitor which is much much larger.
06:40:05 <DHess_US> No, it is similar to R3.
06:40:21 <DHess_US> You still need R3 as shown though.
06:41:03 <DHess_US> Your circuit should work also. Try it out.
06:41:24 <rust_collector> oh, it does work. I just need to see how well it works
06:42:38 <rust_collector> not tested with any load yet, so it could be useless... but we will see
06:45:06 <DHess_US> The usual problems are overshoot when recovering from a load step.
07:00:06 <rust_collector> yep, it needs some tweaking here. the 317 is getting fairly hot, and the transistor is still cold
07:01:11 <rust_collector> I used a 6ohm resistor
07:07:24 <T3sl4> hi
07:07:52 * T3sl4 oohs, nibbles at Ampera's feesh
07:08:16 <rust_collector> On one side, I could probably just use the 317 as it is, without any transistor, as I don't need much current anyway, it is just a day of messing with parts I have.
07:08:59 <rust_collector> And, I think we need updated pics of that fish
07:17:11 <DHess_US> If you attach the 317 and transistor to the same heat sink and adjust the thermal resistance of the 317, then you can get them to track.
07:18:29 <DHess_US> The idea is that under load, the transistor should heat up just as or almost as quickly as the 317. Then the 317 hits its thermal limit first and drops the current protecting the transistor.
07:23:30 <DHess_US> Adjusting the thermal resistance means trimming the ratio of current between the transistor and regulator and shimming the regulator with extra mica insulators to raise its thermal resistance.
07:24:12 <rust_collector> hmm, interesting
07:24:44 <rust_collector> I will play with that a bit
07:25:01 <rust_collector> thanks
07:55:48 <Ampera> T3sl4, https://pbs.twimg.com/media/D4oE_1kW0AIcSE-.jpg
07:55:59 <T3sl4> :o
07:56:35 <Ampera> My mother and brother had some, and my mom went to culinary school.
07:56:58 <Ampera> she said it was restaurant quality :3
09:46:05 Topic Is: Official IRC channel of the EEVBlog - www.eevblog.com
09:46:05 Set By: (lego.jp.austnet.org) on: Sun Mar 17 01:18:37 2019
10:07:53 Quit: arekm ([email protected]) Ping timeout: 121 seconds
10:09:30 Join: arekm ([email protected]) to #EEVblog
12:28:07 <DHess_US> https://i.imgur.com/EY1kkIU.jpg
15:37:06 <steve30> morning
17:14:23 Join: hsn ([email protected]) to #EEVblog
17:14:37 <hsn> hey guys, i've stupid question
17:15:14 <hsn> RS485 has differential pair but the current is returning through ground (3rd wire) ?
17:17:17 <hsn> i've PLC rs485 module that has (5v gnd A B). sensor that needs (24v gnd A B). the (5v and GND) of plc module is isolated. that means they will not share same GND if i provide 24v and gnd to sensor.
17:17:56 <hsn> unless i take the gnd of plc to the gnd of 24v but then i broke the isolation
17:18:03 <hsn> tie them ^
17:19:06 <Ampera> you gonna ask a question?
17:19:45 <Ampera> Do you mean to ask does current return in RS485 via ground?
17:20:22 <hsn> there is (A) and (B) in rs485 which are differential signals. is the current moving from A to B
17:20:30 <hsn> or from A to GND and from B to GND
17:21:32 <Ampera> If you have a differential pair, then that means both members of the pair have the same signal moving in the same direction
17:22:41 <hsn> hmm, so they are both above ground.
17:22:42 <Ampera> As for where it goes, it all ultimately goes to an effective ground, but if you're talking about the signaling for RS485, I'd imagine it goes through some circuitry first.
17:23:11 <hsn> their potential is above ground.
17:23:23 <Ampera> It's confusing to know exactly what you're asking. You have an RS485 module, and wish to connect it to a sensor that has data lines, but a different power supply
17:24:26 <hsn> hmm its like this: its siemens PLC and module is connected to CPU module. i connect the CPU to 24v
17:24:36 <Ampera> Then I would presume that the current moving from AB to AB would go to each of the devices respective grounds (presuming the standard is two way communication, I'm not familiar with RS485)
17:24:58 <hsn> and then i measure the voltage between the 24v and (GND of rs485 module). it was 1v
17:25:43 <Ampera> So you are measuring between a 24V supply and an isolated ground?
17:26:01 <hsn> i assumed its isolated after i saw it 1v :D
17:26:24 <Ampera> So what is the question?
17:27:16 <hsn> i wanted to know if rs485 current goes to gnd or only between (A) and (B)
17:28:11 <hsn> if it goes to GND then i've to connect the isolated GND of PLC module to GND of supply ? doesnt make a lot of sense. :X
17:28:46 <Ampera> ¯\_(ツ)_/¯
17:29:01 <Ampera> Still not entirely sure what you mean
17:30:00 <hsn> you said that : If you have a differential pair, then that means both members of the pair have the same signal moving in the same direction
17:30:15 <Ampera> Yes, that's how differential pairs work.
17:30:24 <hsn> are you refering to current here?
17:30:49 <hsn> the current same direction ^
17:31:37 <Ampera> Ah, the signals are complimentary, my mistake.
17:31:52 <hsn> yes.. this make sense now.
17:32:01 <Ampera> but as for direction of current, should still be moving the same direction.
17:32:29 <Ampera> The idea is that with a differential pair, you use the difference of two conductors to get the final signal, thereby negating (or at least reducing) any signal interference.
17:32:34 <hsn> if you say that direction of current should be same. then both signals' potential are above gnd
17:33:02 <Ampera> Well yes, I'd presume that the potential of the signals are above ground/0v
17:33:20 <hsn> in this case, both sensor and module should share same gnd
17:34:14 <hsn> the module has isolated gnd :X and this doesnt make sense right?
17:34:17 <Ampera> Honestly I'm not sure either. In my very simplistic mind it shouldn't matter what ground the signals end up at, so long as they are both grounds, but I could be missing quite a few important details.
17:35:11 <Ampera> T3sl4 should probably be able to tell you better than I
17:35:16 <Ampera> this is above my minimal levels of understanding
17:35:38 <hsn> you are also an expert. thank you :)
17:36:19 <Ampera> just know that I have no clue what I'm talking about
17:36:24 <Ampera> so if you break shit, not my fault
17:36:37 <hsn> :D no worries
17:36:48 <Ampera> As always, the forums are a good place to get help.
17:37:21 <hsn> yeah although i saw this question about rs485 asked many times but when i read i dont reach to conclusion.
18:29:13 Join: Bluenote ([email protected]) to #EEVblog
18:35:10 Quit: Bluenote ([email protected]) Quit: Leaving
18:53:02 <EEVBlog> New EEVBlog Tweet: RT @brettroberts: I see what you did there... https://t.co/0BVXOnv0TA
19:02:06 <EEVBlog> New EEVBlog Tweet: RT @spaceabba: let’s talk about the fact that rey has to be mentally running the numbers on death star scrap here https://t.co/3ONsMIzmQZ
19:44:46 <DHess_US> <hsn> RS485 has differential pair but the current is returning through ground (3rd wire) ?
19:45:03 <DHess_US> No, the current is only between the signals. It is truely differential.
19:45:35 <hsn> through termination resistors right?
19:45:48 <hsn> so the current is opposite in the wires.
19:45:54 <DHess_US> yes
19:46:46 <DHess_US> Now you *can* return the currents through ground deliberately but if you then disconnect ground, the currents return through the signals. That simply means you can use single ended terminations if you want.
19:46:46 <hsn> so I can leave the ground of sensor isolated from the ground of module? :X
19:47:23 Join: ThunderSqueak ([email protected]) to #EEVblog
19:47:27 <DHess_US> Sort of ... the receiver has a limited common mode input range. The bias currents still return through ground.
19:47:33 Mode: (by ChanOP) +o ThunderSqueak
19:47:54 <DHess_US> Usually the ground *are* connected but the signal currents don't return through it.
19:48:28 <DHess_US> If you want true galvanic isolation, then something else is required.
19:50:12 <Ampera> hi ThunderSqueak, how do
19:50:37 <DHess_US> for instance a transformer can be used with RS-485 signalling for true galvanic isolation. Or a floating power supply can be used on one or both sides.
19:50:57 <DHess_US> The later is common with galvanically isolated RS-232.
19:51:09 <DHess_US> The former was used with Appletalk.
19:51:35 <ThunderSqueak> Hi there
19:52:08 <Ampera> hullo
19:54:24 <hsn> sounds like the signal isolation are usually within rs485 driver chips
19:54:24 <ThunderSqueak> https://twitter.com/spaceabba/status/1118970642852909057
19:54:59 <ThunderSqueak> discovered today that the noise floor in my apartment is really high because of the APC battery backups that I run
19:55:02 <DHess_US> It's actually within the receiver chip. The receivers have a common mode input range of more than +/-5 volts. Some are as high as +/-15 volts.
19:55:06 <ThunderSqueak> they are terrible
19:55:24 <ThunderSqueak> err cyberpower, not apc
19:55:55 <DHess_US> If ground was not connected at all between them, then one would drift compared to the other until an input protection diode was forward biased to limit the difference.
19:56:35 <DHess_US> So it isn't what would be considered galvanic isolation. You *could* increase the electrical isolation by using high impedance buffers at the receiver.
19:56:56 <DHess_US> You could even increase the common mode input range to hundreds of volts.
19:58:12 <DHess_US> But if you want true galvanic isolation, then a transformer or other means is still required. Just differential signaling by itself is not enough.
20:02:04 <hsn> The receivers have a common mode input range of more than +/-5 volts.
20:02:31 <hsn> but the standard is -7 to +12 right? so those chips are not meeting the standard?
20:03:44 <hsn> i mean (1 = +200mV , 0=-200mV) but allowable -7v to +12v
20:07:28 <hsn> oh that's the differential.
20:11:13 <DHess_US> They are intended for systems which are connected together in the same building which has only one power service entrance.
20:12:20 <DHess_US> Even then, common mode transients can be large enough to disrupt it.
20:36:06 <ThunderSqueak> https://pbs.twimg.com/media/D4pkhDvXsAYTkHQ.jpg
20:52:41 Quit: hsn ([email protected]) Ping timeout: 121 seconds
21:14:49 Join: hsn ([email protected]) to #EEVblog
21:50:14 Quit: ThunderSqueak ([email protected]) Quit: Leaving
23:18:28 Quit: hsn ([email protected]) Ping timeout: 121 seconds
Search this day's buffer: