Error Handling in Baby Toys | Hacker News Reader

Error Handling in Baby Toys

brainbaking.com

169 pointsdave-x73 comments

Related but a nod from a toy engineer to parents: my daughter had a toy baby gamepad when she was little. The D-pad and four buttons made different sounds when pressed. I decided to try the konami code (up up down down left right left right b a) and it went through a special sound sequence then said, "you win!"

I have this exact toy for my daughter I think. I would tinker with it a bit, I think just the fact that you hit a certain set of the buttons was enough on its own - I’d have to find the toy to test but I have heard the same voice stab but didn’t do the Konami code.

Sorry to be a spoilsport! :)

Could be that it takes any 10 buttons in a row, but there's a specific sequence of sounds that it does when you enter it, along with "you win!" that I don't recall happening with random keypresses. It's been a while, though.

I had my son's lying around and it definitely happens consistently with the konami code in both the 'learning' and 'playing' modes.

The toy emitting "ut oh" on invalid input reminded me of a phonics toy - I think it was either a schoolbus or caterpillar, where phonemes could be pressed and it would sound them out "g-oo-d" "d-aw-g" - but if as a bored parent you tried "sh-ih-t" or similar it would emit "sh-<heeheehee>" I always wondered if it was localized per-market.

This was likely in response to things like the famous phone incident - https://www.youtube.com/watch?v=ebv51QNm2Bk

If you were the guy coding one of those, and you saw that, I could easily imagine putting in a special case, just for fun. The aha, I could fix that we often feel.

There was an infamous case of some obscure Macintosh educational software that had a feature where it would read back text that you typed in. It even included a swear filter to prevent it from saying bad words.

However, there was also a race condition/out of bounds memory read issue that caused it to stop reading the text you typed in, and instead start reading words directly from the swear list.

https://www.youtube.com/watch?v=sZQmL-PKr8Y

Bonus points to you!

As an opposite experience I bought my daughter a dump truck at Costco and got a good laugh out of it - it has a bunch of phrases, many are common. One very uncommon one is "Wow, that was a big dump!".

reminds me of the better call saul toilet: https://www.youtube.com/watch?v=dR1m29cNVsc

I have one that's a dog with the letters on it and my son was happily pressing away, turns out of you hit "b" then "I" It says "b" "I" "itch".

Bonus humor that it’s a dog. You could just be trying to accurately refer to it!

When I was a kid, I had a toy gun that would start with kind of distorted phrases, then continued with shooting sounds. I think it took me a couple of years before I understand that the voice was yelling something akin to "Baldy! Get the f... over there! <shooting noises>"

This reminds me of the video I think Technology Connections did on old TV censoring equipment back in the 70s. I think it ended up being a simple filter over the subtitle information or similar but thought of this from your comment.

I've had the idea of making clones of the most popular baby toys, but make the volume on them much, much lower. Exact same toy, much less volume. There you go, billion dollar idea. Run with it.

I don't get why all the children's toys need to be so loud. These things are held inches away by a child with presumably good hearing, yet they can be heard rooms away.

If someone misplaced their kid, the could simply start playing with their favourite toy and the kid would invariably come out.

It serves as a baby-retrieval device (BRD in the lingo).

It’s the other way round.

If you hear toy noises from the other room, you know where the baby is and what the baby is doing (and if the toy goes silent you go investigate).

Then again, grown up music is not immune to loudness wars - it shouldn’t be surprising to see the same attention grabbing mechanism in kids toys.

As a parent of young kids, so much this. Many of these toys have a 3-way switch, Off, On-Volume Low, On-Volume High. The high setting is way too loud, and the low setting should be the high setting. And then there'd be a new low setting, lower than the original low.

A piece of clear sticky tape over the speaker grille usually works wonders.

I do this for piezo buzzers in a ported plastic housing. Tape silences them then you poke a hole with a tapered needle to control the volume to the desired level.

Because when you can hear the toy going when you are several rooms away, you know the kid is still there and still playing with it. So you can be in the kitchen doing housework or whatever and hear it over clattering dishes etc.

When the noise stops, you know they're up to something they shouldn't :) Nothing more worrying as a parent of young kids than silence!

This is an interesting problem. I'm not an audiologist, but am sure that any noise above 75dB is harmful to an _adult's_ hearing. If the case can be made that infants and toddlers hearing is more sensitive and should be accounted for, then by leaving the toys so loud we're choosing function over the potential well-being of the toddler's hearing.

Adam Savage of Mythbusters fame has a video on how to quiet those toys. As it turns out, putting a resistor in line with the speaker is all you need.

https://m.youtube.com/watch?v=VvJUWKu6khc

That's definitely the better solution, but I've found it much less work to just pop it apart and shove some kleenex or paper towel in there.

Or just let the battery run down.

Noise-making toys are what you get for parents you hate. Can confirm nearly every one is way too loud.

Unfortunately electronics are so cheap that it can be a challenge to find the one similar toy out of ten on the shelf that isn’t a torture device.

Multiple kids, you give them lightsabers.

If you need to create a rift in the family, a drum set.

My wife gifted her niece a Fisher Price folding wagon --- a couple of years later, when we had our first child, my brother-in-law reciprocated, buying _his_ niece/our daughter the same toy, and it quickly become a favourite of hers --- when they first had occasion to visit, we explained that it was her uncle who had gifted her the toy, so she gladly ran to her room to wheel it in and present it.

When she did so, my brother-in-law exclaimed, "Hey! That's not loud like the one we have!"

It seemed that it was so loud, that it was the subject of complaints, and a modification, so at least for some toys, noise reduction does happen.

I found it fascinating how my baby/toddler used to somewhat often get iPhone UIs into a weird state. Like it was half rotated, or had UI elements weirdly stuck in places they shouldn’t be.

Clearly iOS has bugs where this can happen, but it requires a combination of actions that would be very bizarre for anyone older than a few years to be able to do consciously.

Babies are great at fuzzing UIs. My 6mo daughter managed to call my wife from my locked iPhone, which I didn't even know was possible. The necessary sequence is:

1. Attempt to unlock phone 2. Tap "Emergency" 3. Tap "Medical ID" 4. Tap the emergency contact to call

She's also subscribed us to Amazon channels multiple times playing with the Fire TV remote.

>Babies are great at fuzzing UIs.

So are cats.

My cat hard-rebooted windows 95 pc just by walking on the keyboard.

No dialogs, nothing. Just immediate black screen and clean boot afterwards.

Ctrl+Alt+Del would do it. It was Windows NT which brought the "lock screen" when pressing that combo.

Windows XP was part of the NT line, but still had the Ctrl+Alt+Del combination for 'open Task Manager'. I think it took until Windows Vista for the 'lock screen' to be introduced, but I don't have a copy around to check.

At least the windows 2000 pro and xp machines I used back in the day presented a dialog with both "lock computer", "open task manager", "switch user" or "log out".

It was a setting that could be changed, I think XP is the one that introduced the option to start task manager - because that matches the behaviour of windows 9x line that they were targeting for replacement.

Windows 9x would open a primitive task manager on Ctrl+Alt+Del

That's really interesting. I wonder if the setting was automatically set to 'open Task Manager' when there was only one, unsecured user (and therefore nothing to lock or sign out of). I don't remember changing it, but this was of course well over a decade ago now.

I think XP defaulted to task manager, but I can't be sure, I don't have anything on hand to check.

GPO could set different conditions, including replacing fast user switching (the "new" login screen in XP) with Win2k style including enforcing use of System Attention Key (Ctrl-Alt-Delete on PC) to login.

I'm not sure. It was long ago, but I remember I couldn't repeat the behavior myself and I knew about ctrl+alt+del since before win 95 existed.

Heh, yea Win95 wasn't so stable.

When Windows 2000 came out I started using it at home, and quickly got a lot worse at helping others with their issues, as I simply didn't get much practice at home anymore.

Not really an example of fuzzing the UI, as that's a feature someone at Apple implemented following someone else's design and approval. You were just unaware until then.

Well sure, but she didn’t know that. So she found a feature by fuzzing the UI.

> 1. Attempt to unlock phone 2. Tap "Emergency" 3. Tap "Medical ID" 4. Tap the emergency contact to call

I did not know this existed. That's a great feature, hopefully I'll remember it when it matters.

There was that kid who found a way to bypass the Ubuntu lock screen IIRC.

They should bring back the Monkey.

https://www.folklore.org/Monkey_Lives.html

It's fun being an engineer and a parent. You find yourself just naturally exploring these things. I came to the same conclusions about the toys with the binary switches, and also tried to do the unused combos.

But my favorite were the toys that require the kids to twist or push buttons in a sequence. It had no batteries so it was all mechanical, and it was fun trying to reverse engineer the mechanics so that it only opened the door when the right combo was used, but also didn't seize up when the wrong combo was used, and you didn't have to "reset" it to start over.

I remember encountering an arithmetic toy which had buttons for 1-9, as well as +-*/. If you pressed 6/2 it’d go “three”, but 7/2 would be “Haven’t learned that yet”. Though that was pretty fun.

Kind of reminds me of Gray code, who's whole shtick is being a sequence which only changes one bit at a time. If you see two bits change you know it's an error condition and can handle that explicitly.

https://en.m.wikipedia.org/wiki/Gray_code

Or just a very quick change?

With a Gray code you typically try to oversample to where it's either not an option to change that quickly, or represents an error in its own right if it does.

One of the vocabulary words in the TI-99/4A's speech synthesizer accessory was "UHOH". If you used CALL SAY from Speech Editor or Extended BASIC, it would spell out words that weren't in its vocabulary and emit an "UHOH" for any symbol it didn't understand (and it understood very few, mainly + to reduce the delay between words like "SIX+TEEN" and # for multiword vocabulary items like "#TEXAS INSTRUMENTS#" and "#HANDHELD UNIT#" (early TI-ism for joysticks)). If you made the mistake of leaving alpha lock off and typing your speech string in lowercase, you got a long stream of UHOHs.

Not exactly a toy, but there was a talking wristwatch for the blind that would sometimes say something like "the time is: three, equals, memory clear, PM". Apparently the company, who also made other hardware for the blind, decided that it was less costly to manufacture one kind of chip for both the calculator and the watch, and then have some way of telling it which device it is, probably a pin that was either always 0 or always 1. When the battery was running low, something weird would happen and the wrong sample would sometimes play.

Once you recognize this you see it all over the place in kids toys. The current "toy record player" for example uses this - the "grooves" in the disk are just to press the switches in the "arm".

This one: https://www.walmart.com/ip/Fisher-Price-Classics-Record-Play... (the old 1978 one was just an actual "record" player).

I have one of the old ones. The head is actually a music-box style mechanical movement, with each "track" of the record corresponding to a particular tone. A standard music box has the notes of each tone embossed on a cylinder, or has pins installed at the location of the notes. As the cylinder rotates, the lands (that is, the raise part) of the embossing hits or plucks the strings/wires of the musical part. With the fisher-price toy, instead of a cylinder, its a disk.

I've been meaning to figure out how to spec out an appropriate torsion spring to replace the one in mine, as all of the songs sound a bit slow these days. I'd also love to write a program to convert some songs from midi into a 3d-printable file to make my own records for it.

I had no idea its from 1978, although that tracks with how old my siblings would've been when they got it.

It was made for quite awhile; they show up on eBay and thrift stores quite often.

I actually have a half-dead one, but I'm probably never going to actually fix it.

Such embossed discs were quite common in the late 19th century for various music boxes and orchestrions.

It seems like they then went out of fashion in favour of paper rolls, presumably due to the increased playing time that could fit on a roll. Then of course the gramophone record was invented, so back to discs... Enter audio cassette tapes: miniature rolls. CDs. Are we in fact stuck in an eternal cycle? Where are my optical tapes?!

They offered both the music-box model and the actual record player model back in the '70s. Source: I was there, and had both of them. I still have the actual record player, and a VERY beat up copy of "Mickey Mouse Disco"

That is amazing, and also kind of a durability nightmare. I wonder how they designed the needle assembly to withstand the rigors of 6-year-old operators.

I must've been 12 or so before my parents would let me even touch the record player on their stereo. Not without reason, to be sure: the stereo's 8-track player had been irretrievably broken before I was born when one of my siblings decided that it looked hungry and fed it a peanut butter sandwich.

Record player needles are surprisingly durable if you don't really care terribly much about fidelity.

And the most common issue can be solved by bending it back down so it rides the groove again.

I also recently analyzed the binary encoding of a baby toy.

https://aaron.axvigs.com/content/leapfrog-fridge-phonics-dec...

Ha I did the same thing after that apple was on our fridge a few days. I wondered "how does it know which letter tile is inserted?" Looked at the back of a couple of the tiles, sure enough, binary. Good catch on the ASCII values!

We have this remote control spider thats fun but a pretty low quality Shenzhen special. One of the silly things about it is you can “crash” the code by managing to press forwards and backwards at the same time. The result is that it permanently tries to go forwards until you remove the batteries.

Toy software and error handling is fascinating and often fun to find the bugs.

Now if only we could get error handling in products for adults! It is astonishing to me and my friends how utterly useless so much modern tech is when things go wrong. It seems the default now is just an "oopsy doopsie, something went wrong! sowwy" message usually accompanied with some attempt at charming art or just a sad face emoji. Half the time we don't even get an error code, let alone a course forward to debug it.

When I was a kid, I saw a toy phone that would normally play DTMF tones when you pressed a key, but also had hidden sounds for different key combinations. This was at a family gathering, so I was naturally very bored and brute-forced them all, at least the 2-key ones. I don't remember what all the hidden sounds were, just different ringtones I think.

Now the engineer within me wonders if the DTMF tones it played were actually accurate.

I am pretty sure I had a variant of that chip. DTMF but a police car siren, a ring tone, and "merrily we stroll along" as a tune.

I wonder if one of the hidden sounds could've been 2600Hz.

I couldn't really understand this paragraph. Maybe I need a worked example?

> Bit rotation can be implemented using a combination of bit shifting and or operations, and on a hardware level, with different locical XOR ports. That’s of course just one way, and you still need a way to identify the number without confusing it with another figure. Although I don’t think that last was implemented: if you select your starting bits in a smart way, there won’t be any overlap. The identification system here uses 14 bits, and all of the figures have either 6 or 7 1s, making it even possible to completely ignore bit rotation and just rely on the relative position of the enabled bits to do the identification. I’m of course merely guessing at this point.

Not sure what it means to "completely ignore bit rotation" and "just rely on the relative position of the enabled bits", those sound like the same thing to me.

To throw my speculation out there, I wonder if this is just done with a lookup table. If we use a byte (generous) to identify each figure, then a table indexed by all 2^14 bit patterns will take less than 17kb. No need to do anything more clever.

I think the codes follow the pattern 1111110xxxxxxx. 1111110 is the header, and xxxxxxx is the data. The program rotates the bit string until the first seven bits match the header. This always produces the same data because there isn't a 14-bit string that can be rotated in multiple ways to have 1111110 as the header but different data.

> That said, this article proves we have way too many battery-powered toys that make noise…

I was never bothered by that. The real trial by fire for a parent are wooden toys - really loud when they hit the floor, can actually hurt and don't have any batteries to pull out.

Anyway my daughter has the most recent generation of Furby given by a friend of mine working for Hasbro, whose annual bonus is apparently paid in kind.

The toy is pretty sophisticated with actual, though rudimentary, voice recognition and the ability to detect and record speech, which it repeats to you with different effects, but outside from that follows a pretty strict script and I was unable to produce any errors.

Had two with tiny chiptunes ICs. 1990s vintage. A keyboard had a "turn me off" beep sequence and then did a 2 tone "battery flat" and stopped. The other was a flip phone which played "merrily we drive along" if you pressed a magic key sequence we never memorised, and faked out a police car siren on another, which was moderately distressing.

The keyboard had 2 key rollover and was a bit bistable which one persisted. I think the keypress reader algorithm was possibly on a slow timer?

I’ve spent a bunch of time disassembling a baby toy recently to do some circuit bending / adding to my synth setup. I’d be curious if anyone here has done any of that and has any fun results to share. I got stuck because I need a more precise soldiering iron, but hope to post some results at some point.

'Toet toet autos' and 'Zoef zoef dieren' seem the same thing, electronically. I think they even share track width. I forgot what happens if they meet each other's codes.

Reply on news.ycombinator.com