Electrical Engineering V3

[F.X Clampazzo]

I got hired as a consultant for a small manufacturing place in town that still did everything hand drawn and beyond old school to bring them into the 21st century, because apparently you can make electronics and still be fucking Neolithic in your process. Hours spent making all their hand drawn trash into clean, beautiful multisim versions, making many back ups and print outs paid off big time.

I was asked today if I can teach their dudes how to use multisim for future designs too. I'm thinking of making them recreate all their co-workers designs in multisim as practice just because I want them to suffer like I did.

I was half thinking to counter offer for a full time job for design and prototype work but tbh the contract money is so good, especially for someone who's still in uni and can't really do full time. Lesson being, even if you're not 100% qualified, if you have the knowledge and confidence put your self out there however you can and work will come in unexpected ways.

[pentium]

I'm designing another thing!

It's a baud rate converter/buffer

[IMG]http://i11.photobucket.com/albums/a166/ballsandy/Computer%20related/IMG_6958.jpg[/IMG]
[IMG]http://i11.photobucket.com/albums/a166/ballsandy/Computer%20related/IMG_6956.jpg[/IMG]

This handy little two-chip devices plugs in between two low-speed devices. In this case a modem running at 300BPS and a Teletype machine running at 110BPS. Data ia received by the modem and transmitted via RS-232. This enters the microcontroller via one of the two channels on the MAX232. From there the microcontroller holds a buffer a few kilobytes in size and retransmits the received data at the slower rate.
Likewise, data transmitted by the teletype enters the microcontroller via the second channel of the MAX232 where it is retransmitted at approximately three times the terminal's speed. The result is that a Bell 101 compliant device is able to operate on the Bell 103 standard.
The whole shebang runs off an external 5V DC source.
I"m still trying to determine though if it is possible to also manipulate other settings. Like how most modern host connections are 8 data bits, no parity and one stop bit but the ASR33 Teletype is naturally 7 data bits, even parity and two stop bits. Basically eliminating the possibility of a remote host not knowing the terminal is not a "standard" terminal.

[No_Excuses]

[QUOTE=No_Excuses;50173242][IMG]http://i.imgur.com/lHb7viu.png[/IMG]

I cannot find this bastard anywhere. It's a DC buck controller from the power board of a dead television. It has the markings 63E06 627 and there's another similar circuit with the same IC labeled 73E14 00A so it seems to be labeled with a serial number/date code and not a part identifier.

I went through every datasheet on digikey for the sot23-6 buck controllers and found nothing that matched the pinout which I am positive is correct. I really truly hate surface mount parts.[/QUOTE]

I went back to this TV and decided to get to the bottom of it. So in the odd event someone ever googles this and comes across this thread....

The parts are On-Bright OB2263 and OB2273 Current Mode PWM Controllers. Can be found on eBay in lots of 10 for $3 at this moment in time. Thank god for Chinese ebay sellers. My hobbies would be suffering without them.

[Lenny.]

So I got this 10pack of mosfets (STP16NF06L) that I'm doing pwm with on a 12v led strip with a nodemcu
Somehow only one of them is capable getting the leds to full brightness

[vid]https://aww.moe/hx0scq.mp4[/vid]

Any idea how that could be possible?

[LoneWolf_Recon]

[QUOTE=LennyPenny;51789953]So I got this 10pack of mosfets (STP16NF06L) that I'm doing pwm with on a 12v led strip with a nodemcu
Somehow only one of them is capable getting the leds to full brightness

[vid]https://aww.moe/hx0scq.mp4[/vid]

Any idea how that could be possible?[/QUOTE]

More than likely the nodemcu (Based on an [URL="http://download.arduino.org/products/UNOWIFI/0A-ESP8266-Datasheet-EN-v4.3.pdf"]ESP8266[/URL]) can't drive the MOSFET's gate voltage high enough (The ESP's I/O can only get to 3v3) to get the RDSon as close to nil ohms as possible ([URL="http://www.st.com/content/ccc/resource/technical/document/datasheet/66/f2/4e/c4/ee/dc/45/e4/CD00002848.pdf/files/CD00002848.pdf/jcr:content/translations/en.CD00002848.pdf"]The MOSFET says its RDSon should drop below 0.2Ohms at 5V VGS[/URL]). Especially since you're running the LEDs off 12V.

Either buy some "logic level" MOSFETs which have a lower gate threshold voltage or you can buy/build a MOSFET driver circuit:

[IMG]https://i.stack.imgur.com/jYZSf.png[/IMG]

Of course with the above circuit, you'll have to invert your PWM duty cycle in software, replace Motor with LEDs & default the pin to high.

Also verify that the MOSFET can handle the current load of the LED strip at full brightness.

[Lenny.]

I didn't hope that was the case as I saw them mentioned in some random guide, guess I'm back to waiting on aliexpress orders

Still weird that one of them is able to fully open
Any ideas for an even cheaper alternative to this? :v: [url]https://www.aliexpress.com/item/10PCS-IRLB8721-TO220-IRLB8721PBF-TO-220-free-shipping/32714364118.html?ws_ab_test=searchweb0_0,searchweb201602_1_10065_10068_10000074_10000032_119_10000025_10000029_430_10000028_10060_10000067_10062_10056_10055_10000062_10054_301_10059_10099_10000022_10000013_10103_10102_10000016_10101_10096_10000018_10000019_10000056_10000059_10052_10053_10107_10050_10106_10051_10000053_10000007_10000050_10084_10083_10119_10000047_10080_10082_10081_10110_10111_10112_10113_10114_10115_10116_10037_10000041_10000044_10078_10079_10077_10000038_429_10073_10000035_10120_10121-10051_10116_10120,searchweb201603_10,afswitch_1_afChannel,single_sort_3_price_asc&btsid=3d4b5980-c5a4-43f7-870f-2973a22a08cd[/url]

[LoneWolf_Recon]

[QUOTE=LennyPenny;51790439]I didn't hope that was the case as I saw them mentioned in some random guide, guess I'm back to waiting on aliexpress orders

Still weird that one of them is able to fully open[/QUOTE]

Note on the datasheet that the threshold voltage varies quite a bit (1V to 2.5V), you just may have gotten on of the easier to turn on ones from the lot.

However the 2N3904 (The EU equivalent is the BC548, or really any general purpose NPN bipolar transistor) is super common so you might have luck picking it up from some RadioShack equivalent store or even salvaging it from old electronics (Pre-2000s electronics as they have alot more through hole components than SMD).

[Lenny.]

I doubt that 2N3904 can handle 12v5a

[editline]7th February 2017[/editline]

Oh yea wait I can just use those to drive the mosfets I got as seen on the pic sorry

[editline]7th February 2017[/editline]

Will that add any significant latency? I'm making a realtime music visualization and can't really have any delays

[editline]7th February 2017[/editline]

Wow datasheets are cool
I can deal with 70 ns delay and rise time

[LoneWolf_Recon]

[QUOTE=LennyPenny;51790499]I doubt that 2N3904 can handle 12v5a

[editline]7th February 2017[/editline]

Oh yea wait I can just use those to drive the mosfets I got as seen on the pic sorry

[editline]7th February 2017[/editline]

Will that add any significant latency? I'm making a realtime music visualization and can't really have any delays

[editline]7th February 2017[/editline]

Wow datasheets are cool
I can deal with 70 ns delay and rise time[/QUOTE]

Ooh realtime music visualization. Me and my roommate are working on something similiar with our RGB strips. What algorithm are you using to transcribe the audio data into colors/intensities? We've yet to figure that out.

(Also I recommend putting a heat sink on the MOSFET, at 12V 5A, it'll be producing about 1.8W of waste heat.)

[Lenny.]

Right now I'm just getting fft data from the windows audio stream in form of an array of floats.
Then I split that into lows mids and highs (by averaging), after that I map those floats from 0.0 to 1.0 to make them more usable.
Then it's just about finding a function to scale those values
Right now it's just something I came up with after fiddling around a lot so its pretty retarded ([URL="https://aww.moe/mll086.png"]looking at in geogebra[/URL] I just wanted an x^2 that rises slow at first but then rises very fast)

bass = (pow(bass, 17.6) +  1.25 * pow(bass, 3)) * 1500;

(don't worry about the * 1500 that's just because my pwm range is from 0 to 512 on each led color and with maxmimum BASS I want all the leds on)

After that it's just spamming the nodemcu with udp packets in the form of "r g b" each being a value of 0-512

Today in maths class we talked about logarithms again and I remembered that we also hear logarithmically, so I should probably find a scaling function based on logarithms instead

Here's a video where I (ab-)used the mosfets on the actual controller (which apparently are of logic level type) of the leftmost led strip to drive both the led strips

[hd]https://www.youtube.com/watch?v=vf7kiASVwSw[/hd]

In the future I want to control both stripts individually so I can play around with more effects containing not just the bass.

Ideally one would not simply scale the fft ranges anyway but rather find patters and/or changes in patters or out-liars to drive even more awesome effects but I guess I'll just start with some simple beat detection that's already in the library I'm using to get the fft data

[editline]7th February 2017[/editline]

I really want one of these apa102c or ws2812b led stripts in decent lentgh and led count, but those are hella expensive
I'm always making up cool ways to make these go to music in my head

[pentium]

Anyone here have experience with the LM301 OP-AMP?

[Chryseus]

[QUOTE=pentium;51814259]Anyone here have experience with the LM301 OP-AMP?[/QUOTE]

op-amps pretty much all work the same, they're designed to be general purpose amplifiers.

[LoneWolf_Recon]

[QUOTE=Chryseus;51816490]op-amps pretty much all work the same, they're designed to be general purpose amplifiers.[/QUOTE]

Some of them vary in terms of voltage supply though, so if he's trying to operate it it single supply mode its not gonna work (Alteast for the LM301 which requires a bipolar supply).

[sp]I've made that mistake one too many times..[/sp]

[pentium]

Nah this was a question about a pair in a power supply I was repairing. One of them was bad and I was wondering if they had to be replaced as a matched pair or not in this particular circuit.
You don't in this instance, but it wouldn't hurt to swap both anyways.

Here, instead have a pic of a cable I built to match an official Playstation 2 VGA adapter:

[IMG]http://i11.photobucket.com/albums/a166/ballsandy/New%20Bucket/IMG_7035.jpg[/IMG]

[img]http://img.techpowerup.org/070126/vga.jpg[/img]

[Lenny.]

It baffles my mind that there is any kind of common relation between these different connectors.

[F.X Clampazzo]

[QUOTE=LennyPenny;51818528]It baffles my mind that there is any kind of common relation between these different connectors.[/QUOTE]

Half of it comes down to electronics companies had this bright add idea that if they didn't standardise their connectors, you would pay big bucks for replacement/spare cables. They're partially right tbh and it's sorta the single most annoying thing in the world. Though for power cables it does keep morons from plugging in power supplies that their device isn't rated for and destroying the device.

[LoneWolf_Recon]

[QUOTE=F.X Clampazzo;51820048]Half of it comes down to electronics companies had this bright add idea that if they didn't standardise their connectors, you would pay big bucks for replacement/spare cables. They're partially right tbh and it's sorta the single most annoying thing in the world. Though for power cables it does keep morons from plugging in power supplies that their device isn't rated for and destroying the device.[/QUOTE]

Thankfully alot of engineers avoid that bullshit and use common stuff like IEC C13 (wall to PSU connector), USB, D-Sub (VGA, RS-232), RCA, etc.

No one likes it when you have to use a dedicated connector hence why Apple's "courage" is bitting them in the ass and why 99/100 phones today connect via microUSB.

[pentium]

What's weird about the Sony A/V connector is that there's physically a difference between the standard composite cable and the higher end RGB cables. They have more than one plastic injection mold. One has slots milled for all 12 contact pins. The other omits slots for unused pins so it's not possible of being converted and composite cables are WAY more common than the component/RGB cables.

To make that cable I had to use the connector from a knockoff cable with all 12 pins, the plastic shell and ferrite core from an official composite cable and the video and RCA plugs from an older ATI All in Wonder card.

In a way though, it is kinda nice that Sony stuck to one standard A/V connector for so long instead of consoles needing ports for composite, S-video, component, VGA and L/R RCA.

[LoneWolf_Recon]

Storytime: We received a high intensity UV panel back from one of our customers in full working order (We were tasked with upgrading it to a new LED type with better thermal/performance characteristics). We double checked that everything was working and then proceeded to swap in the new panel. I had to turn up the current drive for each LED string on the panel (Two driver boards with 40 drivers each) so that we can get to spec. First driver board (Half of the panel) turned up just fine to current and we were getting great thermals (it ran cooler than previously) with higher intensity. Then I came to the second driver board....3/4 through tuning the entire board I realized that one of the drivers refused to run. I verified that the LEDs were still fine and then me and my supervisor diagnosed the chip needing a replacement. (End of Tues)

Yesterday, I swapped the old chip in with a new replacement, double checked everything, powered up the board and *POP POP POP*. Three other chips released the blue smoke before I scrambled to get the supply off. I replaced them as well. We couldn't find the problem, the voltage rails were stable, the thermals were in spec, nothing was shorting/beyond the current limit of the LEDs/chips, and the fucking chips say they have thermal shutdown. :quotes:

Today I replaced the last couple bad chips, verified the fuck out of continuity and thermal connection (There's a thermal pad underneath each chip). And just when we thought it was over and I was finalizing one of the strips (Only 14 LEDs on the edge of the board), said 14 LEDs popped. We started our reflow as usual using our heat gun and lo and behold the fucking heat gun went up in smoke (It was 7 years old). Thankfully the damage is contained to only 14 out of 1120 LEDs (The rest of the panel and other driver board are fine), the voodoo driver board has been quarantined and a replacement board & heatgun will be in tomorrow. I'm not above user error, but I was anal/cautious/triple checking everything. Thankfully my supervisor agreed.

[B]TL;DR When Karma hates your fucking guts that it fucks you over twice by killing the tools needed to repair the board.[/B] :suicide:

[pentium]

How much rocket science would need to go into a 555 circuit that runs off 5v and outputs a 60hz square wave at 3.3-5Vpp?
I've found a diagram for doing 60hz as a sine wave but 555 timers are still an unknown for me.

[editline]asda[/editline]
Adding to that, it needs to be a STABLE 60hz source.

[Van-man]

[QUOTE=pentium;51845922]How much rocket science would need to go into a 555 circuit that runs off 5v and outputs a 60hz square wave at 3.3-5Vpp?
I've found a diagram for doing 60hz as a sine wave but 555 timers are still an unknown for me.

[editline]asda[/editline]
Adding to that, it needs to be a STABLE 60hz source.[/QUOTE]

[URL="http://www.falstad.com/circuit/circuitjs.html?cct=$+1+0.000005+18.278915558614752+64+7+50%0Aw+112+192+112+224+0%0Aw+112+224+176+224+0%0Ac+112+224+112+288+0+1.0000000000000001e-7+2.816889247053185%0Ag+112+288+112+304+0%0Aw+112+64+240+64+0%0AR+112+64+80+64+0+0+40+5+0+0+0.5%0AO+352+160+416+160+0%0A165+176+96+192+96+2+5.000000000000001%0Aw+240+64+304+64+0%0Aw+304+64+304+128+0%0Aw+352+160+304+160+0%0Aw+112+192+160+192+0%0Aw+160+192+176+192+0%0Ar+112+192+112+128+0+105000%0Ar+112+128+112+64+0+105000%0Aw+176+128+112+128+0%0Ac+240+256+240+288+0+1.0000000000000001e-7+3.333333333333242%0Aw+240+288+112+288+0%0Ad+176+128+176+192+1+0.805904783%0Aw+352+288+240+288+0%0A162+352+160+352+208+1+2.1024259+1+0+0+0.01%0A162+352+208+352+288+1+2.1024259+1+0+0+0.01%0Ao+2+32+0+4099+5+0.0001953125+0+2+2+3%0Ao+6+16+0+4106+10+0.00009765625+1+1%0A"]Something like this?[/URL]

But be warned you need to doublecheck your circuit, and probably add a precision trimpot in series with each of the two resistors to fine-tune it in order to compensate for other components.

[LoneWolf_Recon]

[QUOTE=Van-man;51846162][URL="http://www.falstad.com/circuit/circuitjs.html?cct=$+1+0.000005+18.278915558614752+64+7+50%0Aw+112+192+112+224+0%0Aw+112+224+176+224+0%0Ac+112+224+112+288+0+1.0000000000000001e-7+2.816889247053185%0Ag+112+288+112+304+0%0Aw+112+64+240+64+0%0AR+112+64+80+64+0+0+40+5+0+0+0.5%0AO+352+160+416+160+0%0A165+176+96+192+96+2+5.000000000000001%0Aw+240+64+304+64+0%0Aw+304+64+304+128+0%0Aw+352+160+304+160+0%0Aw+112+192+160+192+0%0Aw+160+192+176+192+0%0Ar+112+192+112+128+0+105000%0Ar+112+128+112+64+0+105000%0Aw+176+128+112+128+0%0Ac+240+256+240+288+0+1.0000000000000001e-7+3.333333333333242%0Aw+240+288+112+288+0%0Ad+176+128+176+192+1+0.805904783%0Aw+352+288+240+288+0%0A162+352+160+352+208+1+2.1024259+1+0+0+0.01%0A162+352+208+352+288+1+2.1024259+1+0+0+0.01%0Ao+2+32+0+4099+5+0.0001953125+0+2+2+3%0Ao+6+16+0+4106+10+0.00009765625+1+1%0A"]Something like this?[/URL]

But be warned you need to doublecheck your circuit, and probably add a precision trimpot in series with each of the two resistors to fine-tune it in order to compensate for other components.[/QUOTE]

To add to that, I recommend using a CMOS version instead like the TLC555CP. High impedance on the inputs and all that jazz, regulated supply and a tight tolerance capacitor.

Bonus points if you want to make it temperature independent by adding PTC resistors as the timing elements (Aim for a coeff of ~+1500ppm/K to counter the [URL="http://myukk.org/sm_pdf/SM0661.pdf"]NTC of the internal divider[/URL]). Use Tantalum instead of Aluminum Electrolytics for this as well.

[pentium]

This should work more reliably.

[img]http://www.eleccircuit.com/wp-content/uploads/2007/07/circuit-60hz-clock-pulse-generator.jpg[/img]
[quote]IC1_____MM5369AA
X-Tal____3.579 MHz
R1______20M____Resistor 1/4W
R2______1K_____ Resistor 1/4W
C1______47pF 50V___ceramic
C2______Trimmer 5-15pF[/quote]

The circuit has to be as small as I can get it because it needs to piggyback on a cable to convert the wire harness for a PDP-11 to operate with a modern ATX power supply. I already have several people sending me offers if I'm planning to build more.

[t]http://i11.photobucket.com/albums/a166/ballsandy/Computer%20related/IMG_7062.jpg[/t][t]http://i11.photobucket.com/albums/a166/ballsandy/Computer%20related/ATXtoBA23.jpg[/t]

The 5369 can be run at 5v without modifying the schematic so I can wire it to run off the +5SB.

[DrLuke]

[QUOTE=pentium;51845922]How much rocket science would need to go into a 555 circuit that runs off 5v and outputs a 60hz square wave at 3.3-5Vpp?
I've found a diagram for doing 60hz as a sine wave but 555 timers are still an unknown for me.

[editline]asda[/editline]
Adding to that, it needs to be a STABLE 60hz source.[/QUOTE]

Define stable? A 555 timer will not be stable, as all the oscillator components are highly temperature-dependent.

[nikomo]

You live in Canuckistan, you have a "stable" (over time) time source coming out the wall.
Drop the voltage with a transformer and then build some sort of detector circuit - probably half-rectify the AC and then use a comparator.

[Van-man]

[QUOTE=nikomo;51849365]You live in Canuckistan, you have a "stable" (over time) time source coming out the wall.
Drop the voltage with a transformer and then build some sort of detector circuit - probably half-rectify the AC and then use a comparator.[/QUOTE]

I'm getting the feeling that's what he wants to avoid.

[fpro]

Anyone here ever build a radio telescope? I've been playing with the idea of re-purposing an old dish that is pretty large, maybe 3 meters in diameter.

[LoneWolf_Recon]

[QUOTE=false prophet;51850878]Anyone here ever build a radio telescope? I've been playing with the idea of re-purposing an old dish that is pretty large, maybe 3 meters in diameter.[/QUOTE]

I haven't built anything larger than a repurposed DISH network antenna (Reused as a WiFi Cantenna). However if you're serious about this, I'd recommend saving up atleast $300 to $500 on electronics alone, those $20 RTL-SDRs from Amazon won't cut it for anything too useful IMO.

I'd recommend looking for an SDR with the following specs:
-Atleast 10-bits in ADC resolution (Really helps with SNR)
-Something that can receive in the 70cm (~440MHz) to S-Band (~2.4GHz), especially if you want to do Hydrogen-Line measurements/monitoring.
-Minimum bandwidth of 10MHz (20MHz preferable)
-Bonus if you can mount it in an metal case to cut down on EMI/RFI.

[URL="http://www.rtl-sdr.com/roundup-software-defined-radios/"]Here's a good list[/URL], I'd recommend the SDRPlay RSP1 or 2, MyriadRF or the BladeRF as all are proven and have GNURadio support.

In terms of LNA, go for one with a low noise figure (<2dB preferable) and low VSWR (<1.4 preferable) and then choose whatever frequency range you wanna listen too.

Good sources for LNAs within budget are [URL="http://rfbayinc.com/subcategory_product.php?cPath=83"]RFBay[/URL], [URL="http://lna4all.blogspot.com.au/"]LNA-For-All[/URL] is great (They target SDR specifically) and then various ebay sellers ([URL="http://www.ebay.com/itm/Ultra-Low-Noise-Amplifier-10-6000-MHz-RF-LNA-40-dB-Gain-0-8-dB-Noise-Figure-/282319840770"]this is a good one[/URL]).

[pentium]

[QUOTE=Van-man;51849565]I'm getting the feeling that's what he wants to avoid.[/QUOTE]

It's Plan C. The original power supply did exactly that but the retrofit needs to be as simple as possible so I can manufacture more and it doesn't require people to install additional retrofits. Simply install the PSU, install cable and you're running.

[fpro]

[QUOTE=LoneWolf_Recon;51850985]*snip*[/QUOTE]
Wow this is great stuff. I'll definitely save some money because I was looking at some cheaper stuff.