Hacker News Reader

Comment by javajosh | Hacker News Reader

Am I the only one who looks at these projects with a combination of awe and apathy and jealousy? I do software, not hardware, and I like knowing how things work. When an expert in a different field glances at a PCB and identifies components, their purpose, by sight, and then rummages around a loose parts bin to go deeper, it dawns on me: this is someone's life. The same way I've been swimming in (application) software, and could do similar things with a repo, or even running processes, this person can do in hardware, and it's a set of interlocking skills that took an entire life to grow.

View on HN

> jealousy

For what it’s worth, you could do a similar hack with 5 minutes of work and zero background knowledge if you’re willing to take a leap of faith:

1. Break open a used detergent cartridge and notice that there’s only one chip inside

2. Google the number on the chip and click the first link to figure out that it’s a “I2C Serial EEPROM”. Clearly it’s some sort of memory chip since they advertise its capacity

3. Without knowing exactly what an “I2C eeprom” is, google something like “i2c eeprom rewriter” and buy the first result: a $10 BIOS programmer for desktop PCs (which use the same 24 series eeproms)

4. Clip the programmer onto the memory chip and plug it into your PC’s USB port.

The author went a lot further because that’s what he wanted to do, but you don’t have to. He figured out the USB connector’s pinout, but your programmer comes with a clip that attaches directly onto the chip. He reverse-engineered the data on the chip and built a custom programmer, but you could just click the read button in your programmer’s software to back up the data from a brand new cartridge and then write that data to used cartridges.

For those willing to take leap of faith: be careful about electronic element orientation aka polarity: on each element (where it makes difference) pin number 1 is marked in some way (a dot, cut corner, red wire, coloured band, ...), though rarely there is physical restraint preventing you from putting the clip the other way around. Doing this will fry the eeprom and/or rewriter (on this particular chip it switches power rails and this will cause excessive current flow), but most importantly, it might cause frustration worse than failed compilation for syntax error. Like missing semicolons, this will eventually happen to you. Just be prepared for it.

So, I don't want to sound ungrateful But the what you say is true for ALL the roads someone else has traveled. It's true for electronics and learning the french horn and bread-making. It's true for making jewellery, learning to sew, to cook thai cuisine, and grinding your own optics. That's why the "apathy" in the list of emotions. I want to marvel at others who dwell in interesting spaces, without feeling the need to join them. I know I could, and it's tempting. But I already have a long list of interests, historical and current. Sailing, billiards, photography, piano, writing. I also know how to write in cursive, drive a stick, and change a diaper with one hand. I get that more is available; I want it, and I don't want it.

I was in the same spot when I met this brilliant hardware hacker at this hackerspace who had no formal EE degree but could design circuits in his sleep.

"How do you do it?"

"I read data sheets and a lot of wikipedia".

I already spend tons of time on wiki, so I started reading data sheets. And he was absolutely right, datasheets for components are often chock full of information on how to use them, often with recipe books for circuits that use them.

It's kinda bland and tedious but highly informative.

I think there is a bit more that goes into that (in particular a willingness to fry stuff, that's unavoidable in hardware, and MOUNTAINS of patience), but like software, it's pretty easy in today's day and age to teach yourself hardware stuff.

> datasheets for components are often chock full of information on how to use them, often with recipe books for circuits that use them.

Curious, I looked into the datasheet linked in the post[0]. While I get what you mean by "chock full of information", I can't make the head or tail of it. Where do I get started? (genuinely)

I'm not sure what your question is.

When you don't understand a thing, you can google for it. Like it says it uses a Schmitt trigger for noise protection. How does that work? Wikipedia explains: https://en.wikipedia.org/wiki/Schmitt_trigger

If you have a specific goal for the chip, look up the words that seem to be related to that goal. Once you do that a bunch, you get better at figuring out which words matter.

There is usually an example circuit in the datasheet. I start there. Many ics now also sell that example circuit on a pcb you can buy from the same supplier you buy your ics from. Then to make your own custom thing you just start with that example circuit, change it or add to it as needed for your specific application, and then you can physically make those same modifications or additions to the example circuit board you bought and test your design before ordering custom boards.

It's basically like tracing an outline of a drawing with tracing paper and then filling it in with color. You trace the example design, add your own color.

I (being not a hardware person but able to blink a LED) would start with the diagrams of talking with the device on page 7 and refer to them while reading the text descriptions of same on pages 5–6, maybe refresh my memory on how I2C works[1,2] as well. At the same time, almost any controller you can get is happy[3] to talk I2C for you, so it’s unlikely the details matter beyond “controller writes device address and RW bit then (either!) reads from or writes to the peripheral accordingly until end of transaction” (whence the “dummy write” the datasheet says is needed to set the read address). Many even come with libraries that will talk to I2C EEPROMs specifically (because they are everywhere), so the diagrams might not be so important either.

The rest of the datasheet is (as commonly happens with simple digital components) concerned with “how do I wire this up”, “how good of a digital signal can I expect to get out”, “how bad of a digital signal can I get away with putting in”, a refresher on I2C, and miscellanea like power consumption, operating temperature, and sizes of little plastic boxes it comes in. Necessary if you want to design a board properly, not so much if you’re just going to wire it up for a quick hack (the device seems to have a supply voltage and logic levels of “whatever”).

The only thing you maybe do need these for is to look at the board photo and figure out how they wired it. Dot indentation marks pin 1. Pins 1–4 shorted, picture on page 1 says 4 is ground and 1–3 are address, so per page 4 it ends up with[4] address 1010 000. Power (called Vcc for reasons that are not important here) is on pin 5, wired up to top pin of connector on the close-up photo; the next one down goes to data (SDA as I2C calls it) on pin 8; following one goes under the chip but is almost certainly the bus clock (SCL) on pin 6 because that needs to go somewhere; and the bottom is at first glance only connected to an isolated island of copper but the small holes (“vias”) to the other side of the board probably connect it to the other islands so it has to be ground. That leaves write protect on pin 7, unconnected to anything but Vcc (through a dummy “pull-up” resistor to limit how much power the chip will waste examining it), which, to do a sanity check, page 3 says means writes are permitted (so the name of the pin is a bit unfortunately).

There. Done. You still need to find or make something for the other side of the I2C conversation, but either an off-the-shelf programmer or a tutorial for a cheap board with a controller should be sufficient for that. (Won’t be completely trivial the first time around, but it’s a one-time thing like any other development environment.)

[1] https://learn.sparkfun.com/tutorials/i2c/all

[2] https://learn.adafruit.com/i2c-addresses/overview

[3] https://github.com/dekuNukem/bob_cassette_rewinder/blob/6227...

[4] https://github.com/dekuNukem/bob_cassette_rewinder/blob/6227...

Ah that datasheet might be a bad example because that component is so standardized. In this case you’d be better off searching “i2c eeprom” on YouTube and watching one of many tutorials that cover this chip. Then google “i2c eeprom application note” and skim the manufacturer’s tips on how to use it: http://ww1.microchip.com/downloads/en/Appnotes/01028B.pdf

Same goes for construction materials. So many people fail to read the wood/adhesive/countertop/screw/paint datasheets which sometimes include recommendations against the way the product is often used.

reminds me of medications and antibiotics which include their own datasheets that not all medical providers cross-check against your current medications.

Oh c’mon, if I read those things I wouldn’t ever use a q-tip again in my life.

Just for god’s sake don’t put a Q-Tip in your ear! There’s even a warning on the package

They’re clearly designed and revised over the years for that express purpose. I understand why they have to put that warning on the box and not market it for that use…but they 100% know who their market is.

But it feels soooo good!

The application notes for microcontrollers are usually pretty good. They don't just explain how you do something, but why you do it a particular way.

Consider this a metaphor for any field expert. A chef looks in a fridge and sees a combination of ingredients that will make a tasty dish. A teacher looks at a struggling student and sees what strategies will help them succeed. A lawyer looks at a case and sees specific circumstances that help come to an ideal solution.

There's something very enjoyable about seeing knowledgable people dissecting things. As someone who can do - and has done - both hardware and software work, I also understand why people like to share these experiences. Every step in reverse engineering is a huge success and a nice boost of you ego, confirming that your accumulated skillset is powerful.

I believe you could get to that particular level (identifying and reprogramming an I2C EEPROM) pretty quickly as a software engineer. Just get started!

I’m here to say that you can. I spent many years in pure, high level software, but about five years ago, got an interest in electronics. Started with a learning kit, but just kept at it and now a project like this, while impressive, is 100% within my understanding and, while not underestimating or discrediting the amount of work and ingenuity put in by the author, my abilities. You can do it! Look to places like SparkFun and Adafruit to get started.

I think you have the right mindset, you don't need that much to start hacking digital electronics. Analog is harder for programmers because there is less skill overlap but simple hacks can't be that hard if you are willing to do it.

What always frustrated me with electronics was material considerations. You need all sorts of stuff, like a good soldering iron, oscilloscope, hot air station, magnifying glass/microscope, all sorts of wires, basic components, lab power supply, multimeters, probes, prototyping boards, etc... For programming, you just need a computer, but for electronics, you also need one.

It is not just expensive, in fact it is not that expensive for a full startup kit, thank you China. But what you also need is space, if you don't have a dedicated room for your workshop (i.e. not your bedroom or living room) it is going to be really annoying real quick. You need a bench to put all your stuff, and also enough space to store all the crap you will either repair or cannibalize part from. It also makes moving more difficult.

While it is not an absolute necessity, and there are things like hackerspaces, but there is a barrier that is not present when you are just working with computers.

>It is not just expensive, in fact it is not that expensive for a full startup kit, thank you China.

It is amazing how much more accessible it is these days. An oscilloscope or signal generator (even a used one) in the 80's and 90's, was a lot of money. There's also a lot of multi-purpose test equipment now.

Yeah, I've learned that my new amateur radio hobby is not very compatible with a newly-climbing, curious little human. Thank goodness I have my own little office so I don't have to unplug and carefully pack EVERYTHING every time, but forget anything involving soldering - there's just not enough room for even a small workbench anywhere I can close the door to the little guy who loves anything shiny or with a cable.

You just described my joy of doing hardware better than I would have been able to describe it myself. Thank you.

you should just try it. It’s not as bad as it looks. You’re reading an article in 5 minutes that contains probably a week’s worth of work or something.

trainee/failed grammar nazi here: you mean Envy. Jealousy is what the lord God feels, he's a jealous god: thou shalt have no gods other than him. Envy is "gee, I wish I was as god-mode as that" -which I think is more what you meant.

Nobody else cares, your meaning was clear.

pretty much all of that should be part of any reasonable CS degree

was certainly part of mine, and mine was more theoretical than most

(I certainly remember peering through datasheets of various obscure TI chips during afternoon labs)

Many reasonable CS degrees have no electronics or hardware in the curriculum.

Reply on news.ycombinator.com