Electrical Engineering V3 - Developers

[QUOTE=thefreemann;52676670]Uni closed for a week during Harvey and now we're playing catchup :v: Learning somewhat advanced circuit theory and the physics behind material properties at 2x speed is not fun :cry: Also started learning Verilog today! It's actually kinda fun to use[/QUOTE] Hardware design is rosy until you get a test failure, and you trawl through the output of your simulator, only to find that half the time your test case is wrong anyway! Did design verification for a summer, yeah...

is it weird that despite my fundamentals class being incredibly hard, the wealth of knowledge and information we're being taught makes up for it and I actually love the class :v: who knew complex exponentials were so useful

Is there a reason nobody makes pure DC motors/generators with static magnetic fields? I'm fairly convinced it's possible and drew up some designs that [i]should[/i] be really powerful for their size, and I'm pretty sure they'd work, but it seems like someone would already have done it. Right now I'm just waiting for the magnets and copper before I can test it. It's at least not forbidden by physics (see: homopolar motors, though those are pretty useless since they require conductive slip rings), though it can be a bit confusing at first to see why it's allowed. [editline]18th September 2017[/editline] It also doesn't seem like it would've been practical before neodymium magnets appeared, since it requires some pretty strong fields to actually be useful, I guess that could be part of why.

[QUOTE=Tobba;52692646]Is there a reason nobody makes pure DC motors/generators with static magnetic fields? I'm fairly convinced it's possible and drew up some designs that [i]should[/i] be really powerful for their size, and I'm pretty sure they'd work, but it seems like someone would already have done it. Right now I'm just waiting for the magnets and copper before I can test it. It's at least not forbidden by physics (see: homopolar motors, though those are pretty useless since they require conductive slip rings), though it can be a bit confusing at first to see why it's allowed. [editline]18th September 2017[/editline] It also doesn't seem like it would've been practical before neodymium magnets appeared, since it requires some pretty strong fields to actually be useful, I guess that could be part of why.[/QUOTE] IIRC they're not as efficient.

[QUOTE=DrDevil;52692742]IIRC they're not as efficient.[/QUOTE] I can't even find any literature that mentions them, and the only real losses I can see would be copper losses (as long as all the moving parts are non-conductive) and friction. Though right now I'm running the numbers for the linear versions and not getting terribly good results, i.e 60 newtons per ampere @ 1.9 ohms. In theory all you need to do is to add more windings, since the force is linear with the number of turns, but I didn't calculate the copper mass until now... yikes. Need to redo the math for the rotary version but I recall getting pretty good torque on that though, plus, part of the real magic is that you should be able to run these at really high speeds, as long as everything can handle the voltage. Can't wait for room temperature superconductors.

Finally found some mice with the encoder wheels I was looking for. The old Dell mouse I was given seems to fit the bill best, as the wheels are small and have a fine resolution on them. Problem is there's not as much room inside these servos as I thought there would be after adding the wheels. :v: Fitting in the emitter/detector pair will be interesting, as I've no readily available means of producing PCBs. [t]https://i.imgur.com/QmJymS4.jpg[/t] And it turns out the wheels from the "Dexxa" mice don't really fit at all, they rest on a small lip inside the servo case. :/ Dremel time, I guess. *update* Well balls. Turns out this Dell mouse is a Logitech mouse with the Dell name slapped on it. Not a bad thing in and of itself, but apparantley Logitech keeps a tight lid on its devices, which is a tremendous bummer as the detectors used for the encoders are self-contained quadrature receivers, but their output is some sort of serial output that: 1) Isn't documented anywhere that I can find, and 2) Doesn't make much sense on the scope when I probe it and leave the wheel idle/spin it. The Dexxa mice seem to have similar sensors, so even using those means I'll have to figure out what the output is doing in order to do anything with these sensors. :suicide:

[media]https://www.youtube.com/watch?v=i1aT_CqhyQs[/media] This is fucking rad

[QUOTE=DrDevil;52708413][media]https://www.youtube.com/watch?v=i1aT_CqhyQs[/media] This is fucking rad[/QUOTE] I always wondered how they did those impressive graphics back in the 70's/80's when computers were so crappy back then. Now I know it was glorified circuit-bending. :v:

I keep watching those fucking scanimate demo reels: [media]https://www.youtube.com/watch?v=ispW6-7b2sA[/media] And I keep thinking to myself: Couldn't something similar be built with modern electronics relatively easy? I'll keep you guys posted.

[QUOTE=DrDevil;52732346]I keep watching those fucking scanimate demo reels: [media]https://www.youtube.com/watch?v=ispW6-7b2sA[/media] And I keep thinking to myself: Couldn't something similar be built with modern electronics relatively easy? I'll keep you guys posted.[/QUOTE] You'd still need a way to create the "glow" CRT monitors typically had, but the rest is essentially analog circuitbending on a big scale (if I understood it correctly) which on the other hand should be doable with modern parts.

[QUOTE=Van-man;52732599]You'd still need a way to create the "glow" CRT monitors typically had, but the rest is essentially analog circuitbending on a big scale (if I understood it correctly) which on the other hand should be doable with modern parts.[/QUOTE] The idea would be to use my analog scope as a display for now. It has the grid on it, but it doesn't matter for some tests.

Just received my first self designed pcb, which i ordered for a e-bike project i did some time ago. Nothing spectacular nor professional, more like a test how kicad works. i allready made a version 2 which i ordered but havent received yet. v2: [t]https://i.imgur.com/zROrXGh.png[/t] v1: (i forgot to connect the ground pads to the ground net, that was some time intensive debugging before i discovered it...) [t]https://i.imgur.com/MH6utBB.jpg[/t] [U][B]about the project:[/B][/U] [B]Long story short[/B]: i saved myself and a handful of people 550 bucks for a new battery by making a module that tells the ebike its got a genuine battery attached. since about 2 years i own a yamaha e-bike, but the problems with ebikes.. when the battery is low it feels like a steam roller. i looked for a second battery... [B]what the f[/B] 600 bucks for a 400Wh 36V battery, just because it got yamaha written on it. No thanks. so i got myself a logic analyzer, some arduino stuff and found out that its a serial protocol which i can sniff. reproducing this protocol had a few problems, the last byte was a checksum. but thanks to a german member of the pedelec / ebike forums it finally worked. after making a prototype with a arduino and some lose wires, i finally made myself a 600Wh Battery pack for 1/4 the price of the Yamaha one. Some people asked me to help them, since they screwed up the battery. yamaha got a protection built in, much like a fuse but it disables the electronics when current is drawn before the activation command find its way in. A simple LED lamp in parallel to the battery caused it on some of the people. They just replaced the internal PCB with a li-ion BMS and my module. i decided to leave the perfboard and jumperwires method behind and get some professionally made PCBs. osh park seemed to be cheap, but took a while. instead of a 328p chip and the required components, i just could get arduino mini clones.. they got everything on it and are cheaper than the chip itself... atleast now i know how to use KiCad and im prepared for future projects.

Holding my first own PCB in my hands all those years back was one of the key experiences I had with electronics. It was quite surreal for me that something I drew up on my computer became a physical object that did something useful. Good job!

i remember my first pcb i also remember my first pcb that worked :v:

snip

Scored Tektronix 528 Waveform Monitor + 1421 PAL Vectorscope + Sony Trinitron studio monitor The dude who sold it on ebay was the real deal as far as greybeards go. Back in the hayday his company provided most of analog studio equipment for tv stations in the region. He's now retired and rescues old analog equipment that is thrown out. [t]https://i.imgur.com/hO9CJLz.jpg[/t] Don't have any equipment to generate a test image yet, so I can't show it working, but he walked me through all the important points in his lab and showed me all kinds of test images. The trinitron tube has the crispest image I have ever seen, and that shit is analog!

Here's a quickie regarding soft vs hard switching. In this case I'm using 3 half bridge IGBT modules for a sinusoidal inverter/drive. I know soft switching means having a slight ramp up/down in the switching and hard means turning it on and off as quick as possible. What would be better in the case for something that's going to have variable load?

I'm surprised this thing held any amount of charge at all [t]https://i.imgur.com/uXwz71P.jpg[/t] The acid did a number on the copper plating. Good thing all it takes is a little vinegar to neutralize everything.

[QUOTE=papkee;52769950]I'm surprised this thing held any amount of charge at all [t]https://i.imgur.com/uXwz71P.jpg[/t] The acid did a number on the copper plating. Good thing all it takes is a little vinegar to neutralize everything.[/QUOTE] Is that an Anritsu sitemaster by any chance? I have seen quite a few of those with the same battery leakage on the rear cover.

[QUOTE=ddrl46;52770600]Is that an Anritsu sitemaster by any chance? I have seen quite a few of those with the same battery leakage on the rear cover.[/QUOTE] Good eye! Got one for real cheap on fleebay, works great and came with all the calibration standards. I'm planning on building a new battery back for it as soon as I get the new cells in.

Does anyone know what kind of power plug cable I need to buy for this power supply? I can't find which type of cable this is... [IMG]https://imgur.com/a/rjjgp[/IMG]

[QUOTE=quincy18;52779406]Does anyone know what kind of power plug cable I need to buy for this power supply? I can't find which type of cable this is... [url]https://imgur.com/a/rjjgp[/url][/QUOTE] Does it happen to be a high wattage power supply? Because it seems like a variant of the "kettle" socket meant for servers and high current equipment: [url]https://en.wikipedia.org/wiki/IEC_60320#C19.2FC20_coupler[/url]

[QUOTE=Van-man;52779432]Does it happen to be a high wattage power supply? Because it seems like a variant of the "kettle" socket meant for servers and high current equipment: [url]https://en.wikipedia.org/wiki/IEC_60320#C19.2FC20_coupler[/url][/QUOTE] Seems like it, it is actually my friend his power supply so I'll tell him to check it out, thanks :)

Todays project; A precision voltage divider. I wanted to play around with my keithley 155 to see if it's remotely working at the lower ranges, but my crappy analog power supply bottoms out at about 60mv, a far cry from the Keithleys microvolt range. So I made this: [thumb]https://i.imgur.com/RCJvW5J.jpg[/thumb] I dare not show the inside due to my shoddy workmanship, but it's made entirely using dirt cheap 1% resistors off of Ebay, like the ones you can see top right in frame. I simply measured a resistor, and then added in series to make up for the error. For example, in the case of the 100k, I ended up with a 100k in series with a 1k in parallel with a 2.2k, which (according to my uncalibrated hp 3478a) came to 100.002k. The liberal use of crappy resistors in series likely means the tempco is gonna be awful, but so long as I keep it in room temp it should be fine, right? :v: But hey, total parts cost (excluding case and connectors) is like <10 cents. Some anecdotal measurments suggest I managed to end up with about 0.05% accuracy. [thumb]https://i.imgur.com/UCfWTI1.jpg[/thumb] [thumb]https://i.imgur.com/Nntjkck.jpg[/thumb] I also made some measurments over 10, 1 and 0.1 volts and it comes to about the same accuracy. I might make another one later on to get down to those sweet sweet microvolts, for now I'm just impressed with the accuracy that decent multimeter and some patience can achieve.

[QUOTE=chipset;52791954] I simply measured a resistor, and then added in series to make up for the error. For example, in the case of the 100k, I ended up with a 100k in series with a 1k in parallel with a 2.2k, which (according to my uncalibrated hp 3478a) came to 100.002k. [/QUOTE] Couldn't you just get some carbon film resistors and apply a file to them?

[QUOTE=Goz3rr;52797225]Couldn't you just get some carbon film resistors and apply a file to them?[/QUOTE] Probably yes, I might try that. [editline]20th October 2017[/editline] So I decided to get my 2nd HP 8111a out of storage to see if it's working, needs calibration or anything, so I might sell it. 10 minutes into running it, it goes pop. Some troubleshooting later... [thumb]https://i.imgur.com/WLgqYPZ.jpg[/thumb] There's a short, or partial short somewhere because the transformer gets crazy hot and the power supply regulation circuitry shuts down. Thanks to its plug in card design, troubleshooting the modules is a breeze, problem is I've taken out everything that can go out and the transformer still gets hot. Unplugging the transformer itself works so the problem is on the bus board. Let's get desoldering...

I've had an HP signal generator apart up at the radio club for a good couple months now. Somehow even though every component on the 5v PSU circuit tests good individually, when it's all together the 5v rail is at 1.8V and it refuses to work. This is why I hate that all of our test equipment is 20+ years old. At least HP's service manuals are great.

[QUOTE=papkee;52801737]I've had an HP signal generator apart up at the radio club for a good couple months now. Somehow even though every component on the 5v PSU circuit tests good individually, when it's all together the 5v rail is at 1.8V and it refuses to work. This is why I hate that all of our test equipment is 20+ years old. At least HP's service manuals are great.[/QUOTE] If it's pulling the 5v rail low, and it's not a regulation problem, it must be sinking that current into something. Check for hotspots with a thermal camera/tip of your finger. [editline]20th October 2017[/editline] I think I found the issue with my hp 8111a [thumb]https://i.imgur.com/ST6oZp8.jpg[/thumb] Oh... 3300uf 50v cap reading dead short. A quick replacement later, [thumb]https://i.imgur.com/HDxaTyb.jpg[/thumb] All good

i got high today guys [t]https://w3axl.com/images/beamz.png[/t]

I have been designing a 5V-200V@15mA boost converter for a Nixie project. It's designed to be operated in discontinuous mode. SMPS_SW will be driven by an STM32's PWM output, SMPS_PWREN should keep the supply to converter switched off until the input is driven low, this is so I can keep to the USB limit of a 470uF capacitive load. SMPS_SW will be driven at ~50kHz at 3.3V, with a maximum duty cycle of 75%. SMPS_PWREN will be an open collector output, the IO pin used is 5V tolerant. [img]https://i.imgur.com/wXbwBpY.png[/img] Do you see any issues with this as it stands? I have simulated it with LTSpice and it appears to perform as expected, but I don't know if I have missed anything. The FET on the left will probably be a IRLML6402TRPBF (FTDI app note suggested), and the one on the right a BSZ16DN25NS3.