**** BEGIN LOGGING AT Mon May 22 03:00:02 2017
May 22 03:19:50 <thetransformerr[> nerdboy: okk but I havent received BB yet
May 22 03:21:14 <thetransformerr[> jkridner: are you able to provide any tracking details??
May 22 03:33:51 * nerdboy building a 4.4-rt kernel to see what's in there...
May 22 03:44:09 <thetransformerr[> * nerdboy is also doing overtime :)
May 22 04:47:25 <nerdboy> those pwm drivers are at least enabled
May 22 05:49:20 <nerdboy> also this is useful => https://github.com/cdsteinkuehler/beaglebone-universal-io
May 22 05:50:48 <nerdboy> should be able to enable tiecap driver
May 22 07:11:50 <jic23> thetransformerr[:  nerdboy I think Matt Porter was thinking about brushing off the ecap driver he wrote years ago...
May 22 07:12:38 <jic23> M_w have you heard any more on that?
May 22 07:13:24 <jic23> Iirc the old Ti one is more or less raw registers mapped to userspace...
May 22 07:33:56 <nerdboy> the ti stuff is in 4.4
May 22 07:34:48 <nerdboy> if he has a better one he should break it out
May 22 15:12:50 <tlwoerner> m_w: you're looking for grove connectors?
May 22 15:13:13 <tlwoerner> https://www.seeedstudio.com/Grove---Universal-4-pin-connector-p-789.html
May 22 15:13:18 <tlwoerner> https://www.digikey.ca/products/en?keywords=grove%20connector&from=home
May 22 15:14:38 <pmezydlo> tlwoerner: unfortunately we need smd type
May 22 15:15:00 <tlwoerner> ah
May 22 15:27:18 <m_w> yeah
May 22 15:27:49 <m_w> they are made of unobtainium
May 22 15:36:59 <pmezydlo> m_w: I made the necessary corrections on beaglewire, I pushed pull request
May 22 15:42:05 <m_w> pmezydlo: okay
May 22 15:42:26 <m_w> I notice the dip switches might be pulled in the wrong direction too
May 22 15:44:15 <pmezydlo> oh I didn't fix it
May 22 15:44:30 <pmezydlo> I'll do it
May 22 15:58:56 <m_w> wondering if we can do a pull down in the FPGA
May 22 15:59:53 <m_w> spent a little time trying to program the FPGA last night
May 22 16:00:16 <m_w> it appears to be sort of tricky
May 22 16:01:55 <pmezydlo> m_w: What do you mean?
May 22 16:06:53 <pmezydlo> you can use external programmering device
May 22 16:07:06 <pmezydlo> from icestick maybe
May 22 16:07:24 <m_w> well the SPI is connected to both the FPGA and flash
May 22 16:07:54 <m_w> the direction of the lines from the proessor depends on the device you are trying to access
May 22 16:08:50 <m_w> https://hackaday.io/project/7982-cat-board/log/36982-cat-board-all-on-its-own
May 22 16:08:53 <pmezydlo> do you want to take off flash memory
May 22 16:09:41 <m_w> it is possible but I would like to attempt to deal with the flash and fpga direct programming
May 22 16:11:29 <pmezydlo> unfortunately still I have a problem with running ice40-mgr
May 22 16:12:09 <m_w> pmezydlo: I think we should try getting a userspace hack first
May 22 16:12:32 <pmezydlo> ok it's good way
May 22 16:13:25 <pmezydlo> Just spi and muxing ss lines
May 22 16:14:10 <pmezydlo> *multiplexing
May 22 16:14:39 <m_w> pmezydlo: I was going to try programming the SPI flash from Linux to and see if it loads from there
May 22 16:15:19 <m_w> I might have to send you the unit untested because of the timeframe
May 22 16:15:42 <m_w> I have to work today and jkridner made a shipping label for todau
May 22 16:15:49 <m_w> today
May 22 16:16:28 <pmezydlo> It's not problem when I get board after coding start
May 22 16:16:50 <pmezydlo> But I want to hack it
May 22 16:17:19 <m_w> well I will try a few things before shipping it today
May 22 16:17:42 <m_w> I will send you tracking after it ships
May 22 16:18:04 <m_w> hopefully it does not cost too much as this will be coming out of my pocket for the shipping
May 22 16:19:13 <pmezydlo> I'm not sure how much is it
May 22 16:21:04 <pmezydlo> There are no problems with tax
May 22 16:21:37 <pmezydlo> you can use fedex
May 22 16:21:54 <m_w> yeah that is what the shipping label is for
May 22 16:22:12 <nerdboy> m_w: what do you know about small LNAs?
May 22 16:26:05 <m_w> nerdboy: low noise amplifiers?
May 22 16:27:16 <m_w> maybe a little
May 22 16:27:25 <m_w> not an expert
May 22 16:32:43 <nerdboy> yeah, that looks like the only part we're really missing from one of those Maxoptix diagrams
May 22 16:33:00 <nerdboy> er, the dev board one i think
May 22 16:33:15 <jkridner> m_w: usually fedex doesn't mind if the date you actually ship is off a day or two.
May 22 16:33:24 <jkridner> they just shift the arrival date.
May 22 16:33:35 <nerdboy> the one with the microcontrtoller that ds2 pointed out
May 22 16:41:38 <m_w> nerdboy: sorry I am a bit out of the loop
May 22 16:41:52 <m_w> nerdboy: too much going to all at once now
May 22 16:44:56 <m_w> nerdboy: any idea of the frequency response range?
May 22 16:49:07 <nerdboy> 40k +/- 10k maybe
May 22 16:49:27 <m_w> so low frequency
May 22 16:49:53 <nerdboy> low compared to what?
May 22 16:50:27 <m_w> other typical uses of LNA
May 22 16:50:43 <m_w> usually in the GHz range
May 22 16:55:02 <m_w> jkridner: okay I might ship tomorrow then
May 22 17:00:32 <pmezydlo> m_w: if shipment will be expensive I can pay part of the cost
May 22 17:02:52 <nerdboy> should be a simple transistor circuit maybe?
May 22 17:02:53 <ds2> LNA?????
May 22 17:03:06 <nerdboy> i don't want to use tubes...
May 22 17:03:18 <nerdboy> although it would sound better
May 22 17:03:47 <ds2> for 40KHz, a nice quiet op-amp should do nicely
May 22 17:15:50 <m_w> http://www.ti.com/lit/ds/symlink/lmh6629.pdf
May 22 17:18:26 <nerdboy> yeah, i'm seeing these http://www.mouser.com/ProductDetail/Broadcom-Avago/MGA-631P8-TR1G/?qs=sGAEpiMZZMvplms98TlKY9GybS%252bm2o7GGd5BnfqHGPk%3d
May 22 17:18:47 <nerdboy> they start at 400 MHz
May 22 17:18:59 <m_w> those are for RF stuff
May 22 17:19:14 <ds2> m_w: or one of the ADI choppers
May 22 17:19:24 <m_w> yeah
May 22 17:19:29 <ds2> nerdboy: those won't work well done at 40KHz
May 22 17:19:34 <nerdboy> yeah, and that's all the comes up on mouser searching for LNA
May 22 17:19:56 <ds2> only gotcha with the LMH6629 I see is the compentation
May 22 17:22:12 <ds2> LNA is not the term you want.... low noise op-amp
May 22 17:22:21 <ds2> I suspect you can probally even use a LMV324
May 22 17:22:43 <ds2> since it is at 40KHz, it can be AC coupled
May 22 17:24:46 <m_w> https://www.sparkfun.com/products/9816
May 22 17:24:51 <nerdboy> well, if we can knock that out he might have a chance at building this thing
May 22 17:25:35 <nerdboy> pru code will need to be pretty tight
May 22 17:26:13 <nerdboy> but i *think* it's doable with the amplifier/prudaq/muratas/prus/pwm
May 22 17:26:44 <ds2> why not drop the DAQ
May 22 17:27:07 <ds2> run the output to a limiter and just look for edges from the PRU
May 22 17:27:29 <ds2> should reduce timing to +/-10nS or so
May 22 17:27:30 <nerdboy> can you make a diagram?
May 22 17:27:39 <ds2> one sec
May 22 17:29:37 <nerdboy> simpler is better as long as the data looks good...
May 22 17:29:39 <ds2> https://pastebin.com/Uq76Ba3s
May 22 17:29:50 <ds2> all you care is when the signal arrives
May 22 17:31:40 <m_w> this seems like the typical ping module
May 22 17:32:22 <m_w> HC - SR04
May 22 17:33:29 <ds2> m_w; what seems like the SR04?
May 22 17:35:15 <m_w> the time of flight measurement thing
May 22 17:35:35 * nerdboy still thinks phase difference would be higher "resolution"
May 22 17:36:27 <ds2> is that the one that boils down to an arctan function?
May 22 17:36:34 <nerdboy> but if tof has enough, with maybe some signal averaging, it should work more or less
May 22 17:36:57 <ds2> m_w: pretty much... the SR04 assumes fix speed of sound; in this case we have a constant distance but different speeds of sound
May 22 17:37:42 <ds2> nerdboy: doesn't changing the air speed change the frequency too?
May 22 17:37:59 <ds2> i.e. it "stretches" or compresses the medium
May 22 17:38:09 <ds2> if so, can't we measure the wavelength instead?
May 22 17:44:35 <nerdboy> one coffee is not enough for the physics hat
May 22 17:45:06 * m_w should probably understand the aperatus ... googling
May 22 17:46:04 <nerdboy> i have the acoustic remote sensing book in the google-y folder
May 22 17:46:22 <m_w> https://www.decagon.com/en/canopy/canopy-environment/ds-2-sonic-anemometer/
May 22 17:46:35 <m_w> ds2: they have one named after you :)
May 22 17:46:36 <nerdboy> didn't you get a share for that?
May 22 17:46:44 <m_w> I got it
May 22 17:49:13 <nerdboy> the anemometers have gone from a housing with big spikes/ceramic transducers to a smaller plastic (6" diameter maybe) housing to a micro-sized cylinder with everything internal
May 22 17:49:25 <nerdboy> i have no idea how that last one works
May 22 17:49:51 <nerdboy> saw one at a vendor booth at AMS
May 22 17:50:04 <m_w> sounds expensive
May 22 17:50:24 <ds2> hahaahahah
May 22 17:50:39 <nerdboy> the measurement distance for that one has to be pretty small
May 22 17:51:08 <ds2> why is that?
May 22 17:51:31 <ds2> isn't bigger easier (i.e. so you have more time to turn aroun from switching the signal to receiving it)
May 22 17:53:30 <ds2> <-- looking at things from first principles
May 22 17:56:30 <jic23> Hmm how high a frequency can ecap run at?
May 22 17:56:48 <jic23> Would be interesting to try if using a chopper.
May 22 17:57:36 <ds2> Oooooooh I like eCAP
May 22 17:57:39 <jic23> Hmm CPU frequency?
May 22 17:57:45 <jic23> Sysclk...
May 22 17:58:50 <jic23> Should do for accuracy :). Certainly might be worth a detour..
May 22 17:59:57 <jic23> I can dig out Matt's original driver though for prototype doing just this would only be a few lines of code.
May 22 18:00:56 <jic23> This thing doing a sine wave?
May 22 18:01:12 <jic23> Could use eqep instead :)
May 22 18:01:36 <jic23> https://e2e.ti.com/support/microcontrollers
May 22 18:02:23 <m_w> that might work for phase detection
May 22 18:03:11 <jic23> Yup. That's what I thought.
May 22 18:03:19 <jic23> Nasty :)
May 22 18:03:32 <ds2> I was thinking of period measurement
May 22 18:03:52 <m_w> eCAP is better for that
May 22 18:09:21 <m_w> are there IIO drivers for these peripherals?
May 22 18:12:24 <jic23> Not really. Matt Porter did ecap when back at linaro. Not updated since....
May 22 18:12:46 <m_w> still out on the list somewhere?
May 22 18:13:00 <jic23> Someone mentioned he was planning to dust it off and have another go a mainlining..
May 22 18:13:13 <m_w> he mentioned that at scale
May 22 18:13:18 <nerdboy> these are in the current 4.4 kernel
May 22 18:13:19 <jic23> I'd like him directly.
May 22 18:13:21 <nerdboy> http://processors.wiki.ti.com/index.php/AM335x_PWM_Driver's_Guide
May 22 18:13:40 <nerdboy> eHRPWM and aCAP
May 22 18:13:47 <nerdboy> *eCAP even
May 22 18:13:53 <jic23> Yeah read down...
May 22 18:14:08 <jic23> Ecap currently only supports pwm
May 22 18:14:11 <jic23> Mode
May 22 18:14:15 <nerdboy> yup, it's the register/sysfs interface
May 22 18:14:38 <m_w> mdp: are you hearing all of this? you are in for some work!
May 22 18:14:42 <nerdboy> eqqp is there too
May 22 18:15:22 <jic23> Not looked at that one..
May 22 18:15:30 <nerdboy> who named those things?  someone from latvia...
May 22 18:15:55 <jic23> :)
May 22 18:16:52 <jic23> The real question is what cute fluffy pronounciation the hardware engineers use for them
May 22 18:17:24 <nerdboy> <M> EQEP Hardware quadrature encoder controller
May 22 18:17:33 <nerdboy> under misc devices
May 22 18:17:48 <jic23> For initial prototype I'd just bodge a single purpose driver..
May 22 18:18:03 <nerdboy> help doesn't say anything helpful...
May 22 18:18:14 <jic23> Yeah, I vaguely recall looking at that in discussions about blue...
May 22 18:18:53 <jic23> Registers shoved to sysfs iirc. Fine for this use but not great for general purpose userspace :)
May 22 18:18:58 <jic23> :(
May 22 18:19:20 <m_w> causing rapid mood swings
May 22 18:20:04 <nerdboy> mdp: <cough> ecap driver? <cough>
May 22 18:21:12 <jic23> https://github.com/beagleboard/linux/blob/4.4/drivers/misc/tieqep.c
May 22 18:21:26 <nerdboy> https://nathanielrlewis.com/archives/87  <= let's see how helpful this is...
May 22 18:21:33 <jic23> Not that bad at a quick glance...
May 22 18:22:12 <jic23> Ah same guy :)
May 22 18:22:53 <jic23> Simple interface, should work fine for basic encoder use.
May 22 18:23:50 <mdp> nerdboy: sounds awfully unrelated to my qml/javascript/c++/UI work these days ;)
May 22 18:24:20 <nerdboy> so maybe you need a diversion...
May 22 18:24:50 <nerdboy> waddaya got to look at?
May 22 18:24:58 <nerdboy> code-wise...
May 22 18:25:36 <jic23> *laughs*anyhow got to run. Ideally time for an IIO pull request tonight... Hohum.
May 22 18:26:31 <mdp> nerdboy: it's on the list as jic23 pointed out. the driver is fine for the feature support. the issue was getting an api to handle every type of waveform since that will be ABI
May 22 18:27:09 <mdp> drivers == easy, kernel ABIs to live with forever == harder and less fun
May 22 18:29:27 <m_w> with the new counter support things might be a little easier these days
May 22 18:29:54 <mdp> one other issue is that the pwmss shim that was dumped in drivers/pwm needs to be moved to (likely) drivers/soc/ti/
May 22 18:30:13 <mdp> doesn't mean anything to me
May 22 18:30:54 <nerdboy> nope, that blog was not helpful either...
May 22 18:31:05 <mdp> if you have a link that would help..i don't really follow upstream kernel stuff any longer
May 22 18:31:52 <mdp> at the time, jic23 wanted a complete waveform api defined to encompass a wider array of devices that need to describe attributes of a waveform.
May 22 18:32:37 <nerdboy> sounds useful, i can try patching it in, maybe move that other one
May 22 18:32:41 <mdp> iirc, things like mark/space, etc. to that level.
May 22 18:32:59 <m_w> that is a different creature altogether
May 22 18:33:38 <ds2> hmmmm
May 22 18:34:20 <nerdboy> mdp: patchwork account?  or just list limbo...
May 22 18:36:50 <m_w> https://lkml.org/lkml/2014/1/29/418
May 22 18:37:33 <m_w> https://patchwork.kernel.org/patch/3554651/
May 22 18:40:12 <m_w> v3 seems to be the latest patch revision
May 22 18:40:15 <m_w> https://lkml.org/lkml/2014/2/5/373
May 22 18:43:18 <m_w> counter support in iio came around 4.10 I think
May 22 18:43:34 <m_w> http://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1241195.html
May 22 18:43:44 <nerdboy> do we need 4.10 bleeding?
May 22 18:44:55 <m_w> at least
May 22 18:46:22 <nerdboy> where is the canonical iio.git repo?
May 22 18:46:41 <nerdboy> or a current github mirror maybe...
May 22 18:47:03 <m_w> https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git/
May 22 18:48:14 <mdp> m_w/nerdboy, https://patchwork.kernel.org/bundle/ohporter/ecap-v3/
May 22 18:50:10 <mdp> do you guys actually have a use case or just making up projects to do? ;)
May 22 18:50:11 <nerdboy> hmm, i did something similar last year with some iio patches...
May 22 18:50:40 <nerdboy> mdp: sonic anemometer project
May 22 18:50:55 <mdp> ok, I've heard you mention it before
May 22 18:51:20 <mdp> fwiw, the above driver is used in an ultrasonic salt level monitoring system.
May 22 18:52:15 <nerdboy> somewhat similar reqs
May 22 18:52:25 <mdp> so it's providing that pulse measurement on request for the distance calculation in userspace so I know when my water softener salt is low.
May 22 18:52:30 <mdp> great
May 22 18:53:14 <nerdboy> is the code for that on github/somewhere?
May 22 18:53:19 <mdp> let me know if you are going to tackle updating the api part and I'll take it off my list
May 22 18:54:10 * nerdboy tends to do (almost) everything in public...
May 22 18:54:23 <mdp> nerdboy: it appears I never pushed that out ;) I'll take a look for my tree and shove it out there
May 22 18:55:24 <mdp> nerdboy: it's not like there's any intentionally non-public..it sorta like taking photos..I choose to live life instead of pausing to take a photo every 5 seconds like some people. ;)
May 22 18:55:56 <mdp> I suspect I was spending most of my time resubmitting patches until linaro took over all my time ;)
May 22 18:56:04 * nerdboy sucks at taking enough pictures
May 22 18:56:22 <mdp> I know I'm the only person here that has unfinished projects
May 22 18:56:36 <nerdboy> riiiigght
May 22 18:57:45 <nerdboy> so the kernel i just built yesterday isn't what we want today
May 22 19:23:57 <jic23> Yeah new classes always run into that issue. Disadvantage of trail blazing. A lot of stuff needs figuring out to ensure the ABI will extend to cover similar parts.
May 22 19:25:50 <jic23> Unfinished projects? *Laughs manically*
May 22 19:27:13 <jic23>  I love the sadly rare occasions where someone else finishes my half done work.  Seeing as I rarely finish anything these days, sometimes it is the only way I get anything new done.
May 22 19:27:17 <nerdboy> so if we go this route we'll need at least one more "part" meaning the actual encoder
May 22 19:27:53 <jic23> Confused.
May 22 19:28:14 <nerdboy> <shrug> i do a ton of stuff that goes up github mainly for other people to use
May 22 19:29:15 <nerdboy> https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg1242883.html <= i mean thsese patches
May 22 19:30:06 <jic23> Using RCN kernel?
May 22 19:30:07 <nerdboy> Add IIO support for the ACCES 104-QUAD-8 => means we'd have to buy 104-QUAD-8 or 104-QUAD-4
May 22 19:30:51 <jic23> No. Good luck finding one of you did!
May 22 19:30:57 <nerdboy> or something similar enough to hack that driver
May 22 19:31:13 <jic23> It's a pc104 board from quite a while back.
May 22 19:31:38 <nerdboy> then what good is iio support for a part we don't have?
May 22 19:31:45 <jic23> Now m_w was suggesting that as a reference for the ABI for an encoder driver.
May 22 19:32:25 <jic23> Actually has expanded more recently for some st parts but that doesn't matter
May 22 19:33:12 <nerdboy> ah, so not directly useful unless we fill a hole?
May 22 19:34:01 <jic23> Discussion drifted into mainlining ecap eqep etc drivers. MDP tried a while ago and I did typical maintainer and asked for an abi generalization. He got busy so never did a final version (was very close iirc)
May 22 19:34:05 <nerdboy> okay, i'll probably go with 4.10 boneblack kernel and add stuff
May 22 19:34:18 <jic23> RCN backport IIO.
May 22 19:34:39 <nerdboy> which kernel?
May 22 19:34:50 <jic23> His 4.4 tree has stuff I only sent upstream a few hours ago
May 22 19:35:05 <jic23> https://github.com/beagleboard/linux/commits/4.4/drivers/iio
May 22 19:35:09 <nerdboy> meaning ti-dev kernel
May 22 19:35:50 <jic23> RCN bases on ti Dev I think then grabs other stuff as useful.
May 22 19:37:29 <jic23> Actually closer inspection puts him on 4.11 IIO I think..
May 22 19:38:28 <nerdboy> are you looking at beagleboard repo or rcn github repo?
May 22 19:39:05 <nerdboy> bb-kernel is up to 4.12-rc1-bone0
May 22 19:39:34 <nerdboy> or 4.4.68-bone17
May 22 19:39:40 <jic23> Beagleboard repo
May 22 19:40:08 <nerdboy> doesn't that stuff go from bb-kernel to beagleboard?
May 22 19:40:31 <jic23> Not sure.
May 22 19:40:43 <jic23> I tend to build mainline
May 22 19:41:02 <nerdboy> it pulls from mainline and adds patches
May 22 19:41:25 <jic23> Where is 'official' tree?
May 22 19:41:36 <nerdboy> official what?
May 22 19:42:11 <nerdboy> rcn (mostly mainline) is usually official enought for me
May 22 19:42:35 <nerdboy> or i make a gentoo-sources-ish ebuild and add patches as needed
May 22 19:43:05 <jic23> Ah. Bb-kernel Vs the one on github beagleboard account.
May 22 19:43:38 <nerdboy> pretty sure first one pushed to the second one
May 22 19:43:51 <nerdboy> *pushes even
May 22 19:44:35 <jic23> Could be. Backports are in second. First has too many branches so I can't figure out which to look at :)
May 22 19:44:44 <nerdboy> we need to get everybody in front of a few pitchers of beer and get it straight...
May 22 19:45:09 <nerdboy> i usually go with the latest branch
May 22 19:45:23 <jic23> Beer always good. Hmm back later maybe.
May 22 20:07:31 <nerdboy> hmm, iio has a "A kthread based hammering loop trigger" but no counters in this kernel yet
May 22 20:30:34 * nerdboy slamming his laptop with multiple kernel builds each with -j4
May 22 21:40:32 <nerdboy> well, ti-dev kernel is there if anyone needs it => http://www.gentoogeek.org/files/arm-bb_kernel/ti-dev/
May 22 21:40:53 <nerdboy> just a smidgin newer than packages and slightly changed
May 22 21:41:40 <ds2> nerdboy: is the maxim dev boards approved and ordered?
May 22 21:43:42 <nerdboy> not yet, jkridner just mentioned we need to add any "deltas" to the spreadsheet
May 22 21:44:01 <nerdboy> that didn't look like the best answer
May 22 21:44:07 <nerdboy> *to me
May 22 21:44:31 <nerdboy> but if you have a particular hack in mind i need to understand it
May 22 21:45:00 <ds2> nope.. just looking at time
May 22 21:45:06 <nerdboy> right now the opamp idea plus pwm/ecap looks better
May 22 21:45:41 <nerdboy> so we need to test that
May 22 21:45:49 <ds2> but coding is suppose to start next week
May 22 21:46:19 <nerdboy> the answer from beahlelogic/prudaq said the round-robin example sounded like what we need
May 22 21:46:25 <nerdboy> *something like it
May 22 21:46:49 <ds2> opamp with what sensor?
May 22 21:47:05 <ds2> the whole point of the maxim board is to mininize HW work
May 22 21:47:07 <nerdboy> well, i think he should probably start with pru code for that ^^
May 22 21:47:40 <nerdboy> we have muaratas from last year and that's what is currently in the spreadsheet
May 22 21:47:49 <nerdboy> *and on-the-way i assume
May 22 21:48:05 <ds2> how do you plan to drive them?
May 22 21:48:18 <nerdboy> that's what the opamp is for
May 22 21:48:57 <ds2> what op-amp?
May 22 21:49:04 <ds2> the one we talked about earlier!
May 22 21:49:05 <nerdboy> the other alternative would be to maybe hack an SRF04
May 22 21:49:06 <ds2> ?
May 22 21:49:13 <nerdboy> yup
May 22 21:49:19 <ds2> uh.... that won't work well
May 22 21:49:30 <nerdboy> the LNA then?
May 22 21:49:32 <ds2> these devices want high voltage (relatively) and low current
May 22 21:49:41 <ds2> no... the voltage is too low
May 22 21:49:47 <ds2> the are ultimate piezo electric devices
May 22 21:50:09 <ds2> depending on the exact P/N, we need to drive them with a high enough voltage or you get little or no output
May 22 21:50:22 <ds2> that's the point of the maxim board...it generates the higher voltage
May 22 21:50:27 <nerdboy> so what kind of hardware can we buy/make easily?
May 22 21:50:38 <ds2> hacking the SRF04 may work but I don't see there being enough time
May 22 21:50:44 <ds2> the maxim board
May 22 21:50:57 <nerdboy> maxim board can drive the muratas?
May 22 21:51:11 <ds2> no, the maxim board comes with matching piezo devices
May 22 21:51:32 <nerdboy> yeah but we need one to rx and one to tx
May 22 21:52:07 <ds2> okay, let me start again ---- what is the overall HW combo you are proposing (for both RX and TX)?
May 22 21:52:12 <nerdboy> do we unmount one and use low-impedance cable?
May 22 21:52:34 <ds2> No, get 2 boards. No HW mods, no soldering
May 22 21:52:39 <nerdboy> the muratas are already separate parts
May 22 21:53:16 <ds2> step 1 - use the demo sw to generate the 40KHz output on one board; step 2 - connect the second board to the Beagle using jumpers
May 22 21:53:26 <ds2> step 3 - try to get output
May 22 21:53:36 <ds2> yes, the murata requires a lot of "HW" work
May 22 21:53:49 <nerdboy> it didn't look like enough control on those boards
May 22 21:53:57 <nerdboy> one can just listen only?
May 22 21:54:20 <nerdboy> plus that software is windoze-only afaict
May 22 21:54:32 <ds2> yes, look at the schematic
May 22 21:54:37 <ds2> yes. that is step 1
May 22 21:54:45 <nerdboy> didn't look like a very friendly dev board
May 22 21:55:16 <ds2> once that is working, then we can high-z the pins (i.e. put the local mcu in reset) and drive it from the Beagle to eliminate the Windows stuff
May 22 21:55:31 <ds2> it is friendly the building HW for the time frame we have
May 22 21:57:37 <nerdboy> it says "one tranducer for send/rec or two transducers, one to send and one to rec"
May 22 21:57:50 <nerdboy> on the same baord
May 22 21:58:08 <nerdboy> we need to work one on each
May 22 21:58:47 <nerdboy> single transducer mode each board will think it needs to send *and* rec
May 22 21:59:19 <nerdboy> unless you see in the schematic we can make it work that way
May 22 22:00:34 <nerdboy> so, back to the high-voltage output, can we step up the 5v from gpio or do we need separate psu?
May 22 22:02:17 <nerdboy> but i need to print that schematic to really read it so...
May 22 22:03:06 <nerdboy> can you tell me the boards can do what we need?  were you thinking two transducers on each board?
May 22 22:03:36 <ds2> From the pix, it comes in 2 transducer mode
May 22 22:04:01 <ds2> you can step up anything to anything... just a matter of some power engineering
May 22 22:04:14 <nerdboy> yeah, so how do we get one to only listen and one to only send
May 22 22:04:31 <nerdboy> *one board
May 22 22:04:33 <ds2> you connect the wires to the right place?
May 22 22:04:37 <ds2> :D
May 22 22:05:01 <ds2> xmit is easy
May 22 22:05:10 <ds2> rcvd - grab it at the jumper
May 22 22:05:16 <nerdboy> plus the same clock
May 22 22:05:17 <ds2> AFAICT, it is a stock 0.1 jumper
May 22 22:06:11 <nerdboy> if you get them first, can you actually test it?
May 22 22:06:39 <nerdboy> just the part about setting it up in that mode
May 22 22:06:51 <nerdboy> s/it/them/
May 22 22:07:50 <nerdboy> so assume that part works, we can get thetransformerr[ started on pru/other interfaces to go with it
May 22 22:18:12 <ds2> not before GSoC starts
May 22 22:25:31 <nerdboy> i'd ask stephanie except she has security homework right now
May 22 22:26:00 <nerdboy> er, wormo even
May 22 22:27:11 <ds2> probally be next fri before I can get to it
May 22 22:36:40 <nerdboy> there might be some beer (or something) in it if you get to it faster...
May 22 22:40:23 <nerdboy> the manual *does* say "... and GPIOs, UART, SPI port, timers, and a watchdog."
May 22 22:41:04 <nerdboy> so the main thing with two boards would be the bb clock source driving both of them
May 22 22:41:35 <nerdboy> if they expose the right stuff...
May 22 22:44:43 <ds2> unless you can get me the boards like tomorrow morning...
May 22 22:44:58 <ds2> the rcvd side is not an issue
May 22 22:45:09 <nerdboy> asking jkridner ^^^
May 22 22:45:11 <ds2> the xmit side might be trickier but I recall a few tap points
May 22 22:45:43 <nerdboy> i have no idea what the spi device might expose
May 22 22:45:52 <nerdboy> or what gpios
May 22 22:47:46 <ds2> got other commitment for the long wknd
May 22 22:48:53 <nerdboy> says timer0 on a couple of pins
May 22 22:49:49 <thetransformerr[> ds2: as you said above about resetting local mcu on maxim board,i understood it but how do you plan to recieve with jumper thing???
May 22 22:51:21 <thetransformerr[> nerdboy: so for now we have three ways for our project-1) use PRUDAQ 2)use Phase difference Method 3) use eCAP module
May 22 22:52:09 <nerdboy> might be both prudaq/pwm/ecap
May 22 22:52:51 <nerdboy> depends on what you end up driving partly, also what you can and can't do in the prus
May 22 22:53:03 <nerdboy> space/time constraints
May 22 22:54:29 <thetransformerr[> PRU are very fast, we have to care for timing constraints for sensors and their drivng circuits only
May 22 22:56:03 <ds2> thetransformerr[: connect the jumper to the prudaq
May 22 22:56:52 <thetransformerr[> I wanted to know is it much difficult or much error prone circuit, if we build our own circuit with some rs232 and opamp like some projects for anemometer have used
May 22 22:56:53 <ds2> can someone tell me if phase method == something that results in an arctan to get timing?
May 22 22:58:07 <thetransformerr[> ds2: it seems nice the jumper thing,
May 22 22:58:16 <thetransformerr[> let me check for phase...
May 22 22:58:24 <ds2> trying to get a better understand of the alg side so I can suggest things or at least give meaningful answers
May 22 23:10:35 <thetransformerr[> ds2: the gist of phase shift method is like this, at 40 kHz , we have period of 25 microseconds, therfore                   25µs/cycle * 340.29µm/µs = 8507.25µm or 8.5 millimeters (1/4 inch) per cycle
May 22 23:11:06 <thetransformerr[> now if we assume wind speed changing the speed of sound
May 22 23:11:53 <ds2> that sounds very touchy
May 22 23:12:09 <ds2> if the input clips, you are screwed... if it is too low, you are screwed
May 22 23:12:38 <ds2> resolution pretty much is related to how much of the ADC range you use
May 22 23:13:28 <thetransformerr[> here are some discussions from forums
May 22 23:13:30 <thetransformerr[> http://www.electro-tech-online.com/threads/ultrasonic-distance-measurement-using-phase-shift-detection-help-with-atmega32.96700/
May 22 23:13:46 <thetransformerr[> https://forums.ni.com/t5/LabVIEW/Distance-measurement-using-Ultrasonic-sensors-and-phase-shift/td-p/3131083
May 22 23:14:32 <thetransformerr[> http://community.atmel.com/projects/ultrasonic-distance-measurement-using-phase-shift-detection
May 22 23:15:23 <ds2> hmmmm this is a different usage of phase
May 22 23:16:23 <thetransformerr[> and actually I had read about phasing extensively during proposal time, so I cant recall much about phasing
May 22 23:16:47 <thetransformerr[> but I would get back to you by today
May 22 23:19:32 <ds2> just brain storming here... I wonder if you can just do something like this: sample both the xmit and rcv sides with the 2 channels of the ADC
May 22 23:19:48 <thetransformerr[> the good thing about phase method would be it looks like continous method, we wont have to trigger anything
May 22 23:19:49 <ds2> multiple the 2 channels then run a high order LPF through it
May 22 23:19:59 <ds2> and detect the timing difference from the residual values
May 22 23:20:46 <thetransformerr[> ds2: exactly same thing I was working on but I was not able to connect the points
May 22 23:21:07 <ds2> what do you mean connect the point?
May 22 23:21:21 <ds2> what I am describing is from RF stuff
May 22 23:21:40 <thetransformerr[> means I couldnt understand this method
May 22 23:21:44 <ds2> Oh
May 22 23:21:55 <thetransformerr[> :)
May 22 23:22:05 <ds2> think of it this way (let me know if I am going over your head) -
May 22 23:22:27 <ds2> cos(wt) and cos(wt+p) are your 2 signals, rcv and xmt respectively
May 22 23:22:40 <thetransformerr[> okk
May 22 23:22:55 <ds2> cos(wt) * cos(wt+p) will roughly give you somethling like f(p) + cos(2tw); essential DC + 2xfreq
May 22 23:23:01 <ds2> LPF it to get rid of the 2xfreq
May 22 23:23:11 <ds2> the remains should be the phase difference
May 22 23:23:23 <ds2> you can look at phase detectors in a PLL setup if you want another example
May 22 23:24:40 <ds2> for LPF, you might be able to use very simple filter (not an alg person so someone else should confirm or refute this) - run a long rolling average
May 22 23:24:53 <ds2> maybe 128 or more samples
May 22 23:25:14 <ds2> or if you have enough bits, just a rolling sum (same as an average scaled by num of pts)
May 22 23:25:28 <thetransformerr[> and I assume it works on continous method i.e. it continously transmits and receive
May 22 23:25:38 <ds2> SDR folks have a term fo this kind of filter but i can't recall the name at the moment
May 22 23:25:44 <ds2> thethransformerr[: yes
May 22 23:26:28 <ds2> the forums you point me to looks more like comparing edge dtect times... crude but not what I would call phasing but that is a term. issue
May 22 23:28:07 <thetransformerr[> and last question is can we sample the 2 channels for each dimension any time
May 22 23:28:19 <thetransformerr[> in round robin fashion??
May 22 23:28:34 <thetransformerr[> and get phase diff
May 22 23:29:57 <ds2> yes
May 22 23:30:13 <ds2> donno if you have to offset it or something
May 22 23:31:37 <thetransformerr[> yet they all require the driving circuit
May 22 23:37:38 <ds2> different circuits
May 23 00:04:19 <ds2> hmmm
May 23 02:11:58 * nerdboy testing patches on top of 4.12
May 23 02:12:45 <thetransformerr[> would you please share some link to these patches
May 23 02:13:14 <thetransformerr[> or any task that I should handle
May 23 02:14:11 <nerdboy> they'll be on github shortly
May 23 02:17:04 <nerdboy> scroll back for the lkml link jic23 gave
**** ENDING LOGGING AT Tue May 23 03:00:01 2017