Its always great to see older electronics repaired and restored , rather than tossed in a landfill somewhere forgotten about. I find these early examples of computer hardware fascinating.
The reason they used a separate 7812 for the head bias was for noise rejection. Linear regulators have over 40db noise rejection, which makes for nice clean reference voltages as long as you don't have anything else on the rails.
I'm surprised at the comments from people suggesting that Marc should have checked the power rails, when in fact, he did. There's sort of an implication that "I would have got that solved straight away," and for that to be meaningful, those folks would have to do that without the benefit of all the legwork Marc put in -- no far watching the video first! You're on your own; you think you could do it? I don't think Marc made any poor decisions. He's got a fairly impressive track record, and shares his thinking processes and experiences, which helps the rest of us. Or... at least... the rest of us who aren't already perfect.
Yeah, I would probably never have figured that out just looking at the schematics. Looking at the simplified diagram he drew up for illustrative purposes certainly makes it a bit easier to understand, but that's definitely not something I would have ever come up with.
That's a use I'm familiar with, but with schottky barrier diodes on the input of rf monitor/generator test equipment. Used to isolate the first rf amp input while transmitting, as a voltage controlled attenuator while receiving and for some amount of lightning protection (sacrificial components when absorbing lightning).
Thank you for your work, Marc. You and other retro tech channels indirectly inspired me to assemble a small DIY amplifier project to make some old Apple Pro speakers work with a 3.5mm input. You could have done it in about 5 minutes with components you already had, but for me this was an exciting project!
Another interesting feature is the bias of the head that does not erase the media. The heads are biased at the exact center of the winding such that there is no net DC magnetization of the head gap. The same technique is used in electron tube audio power amplifiers.
This channel is as much engineering as it is art as it is an epic adventure as it is one of those stories you can listen to make yourself forget all your worries and fall asleep watching someone repair a floppy drive. Thank you for taking us with you and live vicariously through your lab. When I'm finished with my PhD I'll repair a floppy drive as well.
This is one of the best episodes - I came, I saw, I conquered the problems. Thanks, Marc, it was epic. And what's surprising is the complete absence of your crew. You're a wizard, Marc! BTW, I helped a guy at his shop where he repaired the original 5-1/4" full height drives for the IBM-PCs. I asked him if he had ever had to replace the motor's drive belt, and he said he had never seen one go bad. Apparently they used some fibers in the belt that just don't stretch or break - they just last! Amazing for that time in the late 1980s!!
15:30 If you can't fix it with a stick and a rock you're not a real technician, as my father likes to say. Well done M. Marc, I don't think with my limited fundamentals I ever would have figured out how that diode bias switching circuit works. A pleasure to watch as always.
Much as I enjoy the episodes with a bigger team, I also like these with Marc on his own, methodically talking his way through to the solution of a problem. Bonus enjoyment when the solution involves popsicle sticks!
Actually pulled out a 7812 from Motorola today, made in 1979, and still working perfectly. Just modified a board to run off 12V instead of 24V, so the 12V regulator was not needed.
After having seen the AGC reverse engineering video series and recently the laboratory setup of the communication modules of the Apollo missions, watching him repair a floppy drive makes me feel like we are killing a mosquito with a cannon... ;) but it is also interesting and I keep learning from it. Thank you
Interesting: back in the eighties, many professional (broadcast) video monitors used diode switching to select inputs, though usually single-ended circuits rather than differential, with the video signal clamped after the switching.
I'm amazed by your troubleshooting abilities. The only thing more amazing is how much smaller storage devices have become since then. That is one big hunk of metal for such a tiny bit of memory. But it works. Nice job.
Sorry you had to check a buncha things to figure out the problem, but I gotta say it was enjoyable to come along for the troubleshooting thought-process journey!
I SO expected a rickroll with the media file at the end, but you confounded my expectations -which I suppose is the very essence of successful trolling.
My early 1980s HF ham radio transceiver (a TS-130S) uses a very similar diode bias based trick for switching the low-level RF signals at various parts of the signal chain. Let me tell you, as an electrical engineer from a more modern era (and a digital design focused one at that), it took me quite a bit of head-scratching before I figured out what they were doing. A weird but very neat trick!
The utility of a "working example" is hard to overstate. Good one not equal to Bad one, just keep following the schematic until things mismatch, and voila you have at least generalized the area. Then it's a house-to-house search looking at voltages.
Pesky voltage rails! I just fixed a co-worker's boombox and it had a no power issue. I sorted the no power issue, but then it had no sound. Turns out a capacitor had shorted on the power rail for the preamp, but because each chip is fed via resistors, the main rails appeared fine!
I'm gonna make an addon Thou shalt check voltages in various places I had a boombox that took a while to find the fault, a shorted cap took out the power to one of the chips, but since each chip was fed via resistors the main rails looked fine!
But, but, you need to watch the entire video. I checked the voltages! It's just that there were two different +12V, which was not called out in the schematics. And then there was some power leaking back on the semi-broken +12V rail, making it appear as if the head bias section was working OK. This is a very unusual and rather subtle situation, and made it quite hard to ferret out. But all is good, we prevailed in the end.
@@CuriousMarc hello sir! Naturally I watched the whole thing, anything else is just rude! Thought long and hard about commenting, and came to the conclusion that Dave's first law of troubleshooting still applies in this case. Love your videos and the effort you make to make them!
I've long thought that full height 5.25" floppy drives would be an ideal source for robot parts. They have a large motor for driving around, a big stepper motor, two I/O channels, and an LED. With the parts from a pair of drives a robot could have two motors to drive around, two steppers for a basic arm, and something could be hacked together to connect in place of the four read/write heads to transfer data in and out or activate things or for some type of sensor.
The "solution" for the hinge you came up with is awesome!!!! Here in Germany we call this kind of "solution" a "Sowjet method" or "Russian method". Crude, but it works.
How does aligment work So you are saying all floppy drives in the world have to be perfecly aligned the same way so I can write a disk on one drive to be able to read it in another? Also what does aligment disk contain? How was the first floppy drive in the world aligned then without the disk?
10:39: Wouldn't the head selection work just the opposite? I.e. (-12V)-|>|-(+12V) might be enough of a difference (effectively +24V) to represent a breakdown voltage, so the respective head is enabled, but whatever signal there is on top of (+12V)-|>|-(+12V) isn't anywhere near a breakdown voltage and thus blocked?
Yeah, I saw that too, I guess it's because it has enough bandwidth to pass whatever data is coming from the heads as it passes by, plus a ton of gain. Though I can't imagine it would be anywhere near 120MHz wide. But maybe video amps are also nice and linear within that bandpass and also maybe because they were common chips by then that could do the job. MC1733 is a differential video amplifier. Reading the spec sheet it can be used with tape, drum or disc read heads. So it seems it was designed for this application in mind.
Its always great to see older electronics repaired and restored , rather than tossed in a landfill somewhere forgotten about. I find these early examples of computer hardware fascinating.
Darn multiple +12 volt rails! You win this round...
The reason they used a separate 7812 for the head bias was for noise rejection. Linear regulators have over 40db noise rejection, which makes for nice clean reference voltages as long as you don't have anything else on the rails.
Exactly.
Love the literal "in a pinch" brake repair, amazing work Marc.
I'm surprised at the comments from people suggesting that Marc should have checked the power rails, when in fact, he did. There's sort of an implication that "I would have got that solved straight away," and for that to be meaningful, those folks would have to do that without the benefit of all the legwork Marc put in -- no far watching the video first! You're on your own; you think you could do it? I don't think Marc made any poor decisions. He's got a fairly impressive track record, and shares his thinking processes and experiences, which helps the rest of us. Or... at least... the rest of us who aren't already perfect.
The diode circuit is very clever! Thanks for explaining it.
Yeah, I would probably never have figured that out just looking at the schematics. Looking at the simplified diagram he drew up for illustrative purposes certainly makes it a bit easier to understand, but that's definitely not something I would have ever come up with.
That's a use I'm familiar with, but with schottky barrier diodes on the input of rf monitor/generator test equipment. Used to isolate the first rf amp input while transmitting, as a voltage controlled attenuator while receiving and for some amount of lightning protection (sacrificial components when absorbing lightning).
I LOVE the popsicle stick brake.
I like the mix between electronic and mechanic. Now I am in School but I want to learn something like you did so I can repair such old stuff
Thank you for your work, Marc. You and other retro tech channels indirectly inspired me to assemble a small DIY amplifier project to make some old Apple Pro speakers work with a 3.5mm input. You could have done it in about 5 minutes with components you already had, but for me this was an exciting project!
Way to go!
Another interesting feature is the bias of the head that does not erase the media. The heads are biased at the exact center of the winding such that there is no net DC magnetization of the head gap. The same technique is used in electron tube audio power amplifiers.
This channel is as much engineering as it is art as it is an epic adventure as it is one of those stories you can listen to make yourself forget all your worries and fall asleep watching someone repair a floppy drive. Thank you for taking us with you and live vicariously through your lab. When I'm finished with my PhD I'll repair a floppy drive as well.
We didn't get to see it, but I am pretty sure there was a BIG smile on your face when you found and fixed the culprit ;)
Nice fix of the broken damper on the stepper.
This is one of the best episodes - I came, I saw, I conquered the problems. Thanks, Marc, it was epic. And what's surprising is the complete absence of your crew. You're a wizard, Marc!
BTW, I helped a guy at his shop where he repaired the original 5-1/4" full height drives for the IBM-PCs. I asked him if he had ever had to replace the motor's drive belt, and he said he had never seen one go bad. Apparently they used some fibers in the belt that just don't stretch or break - they just last! Amazing for that time in the late 1980s!!
Randomly stumbling upon this channel is one of my greatest YT discoveries of ~18 years. You have officially eclipsed NumaNuma.
You know a floppy drive is happy when it can read its own victory jingle.
Ce type est quand même étonnant. Quelle patience, compétence et intelligence.
15:30 If you can't fix it with a stick and a rock you're not a real technician, as my father likes to say. Well done M. Marc, I don't think with my limited fundamentals I ever would have figured out how that diode bias switching circuit works. A pleasure to watch as always.
la réparation de la tête à base de touyettes en bois est mythique ^^ Bravo Marc, tu nous étonneras toujours !!
It’s always a good day when you see a notification for curiousmarc!
It's so satisfying to see all this old tech repaired in such a methodical fashion. Well done!
I'm in the process of working through some IBM PC 5.25 360k floppy drives. Very similar design. Thanks for posting. The video is helpful.
quite a welcome sight to see the old system coming back to life again under the hand of a real master ....
Much as I enjoy the episodes with a bigger team, I also like these with Marc on his own, methodically talking his way through to the solution of a problem. Bonus enjoyment when the solution involves popsicle sticks!
swapping a 7812 and jamming some lolly sticks in there, now that's my kind of repair!
Actually pulled out a 7812 from Motorola today, made in 1979, and still working perfectly. Just modified a board to run off 12V instead of 24V, so the 12V regulator was not needed.
Head scratching of the correct kind in floppy diagnosis - just an unbiased observation…
That was a sneaky bastard indeed! Good thing you found it. Still makes me wonder why a voltage regulator would fail just like that.
After having seen the AGC reverse engineering video series and recently the laboratory setup of the communication modules of the Apollo missions, watching him repair a floppy drive makes me feel like we are killing a mosquito with a cannon... ;) but it is also interesting and I keep learning from it. Thank you
All cheer for the popsicle stick which made it as a permanent element of the drive! Yay!
Interesting: back in the eighties, many professional (broadcast) video monitors used diode switching to select inputs, though usually single-ended circuits rather than differential, with the video signal clamped after the switching.
I'm amazed by your troubleshooting abilities. The only thing more amazing is how much smaller storage devices have become since then. That is one big hunk of metal for such a tiny bit of memory. But it works. Nice job.
Sorry you had to check a buncha things to figure out the problem, but I gotta say it was enjoyable to come along for the troubleshooting thought-process journey!
Ces épisodes où tu répares quelque chose sont toujours aussi fascinant. Merci Marc!
I SO expected a rickroll with the media file at the end, but you confounded my expectations -which I suppose is the very essence of successful trolling.
Marc! Excellent fault-finding and troubleshooting as always!
Confidence and competence. These are good things. 👍
The Dutch expression 'houtje touwtje' (literally wood and ropes) solution comes to mind😂
Love the diagnostics process. Thanks!
Any day you pull out the diff probe is a good day! Imagine having one and never getting to use it... sad.
My early 1980s HF ham radio transceiver (a TS-130S) uses a very similar diode bias based trick for switching the low-level RF signals at various parts of the signal chain. Let me tell you, as an electrical engineer from a more modern era (and a digital design focused one at that), it took me quite a bit of head-scratching before I figured out what they were doing. A weird but very neat trick!
Nice to have a working drive to compare, prob saved you a bit of time there. Enjoyed the video Thanks
Popsicle sticks!🤣 Well done, Marc!
MarcGyver at work 😎
@@RicoD5 Haha, that works! I was thinking Fred Flintstone.
The utility of a "working example" is hard to overstate. Good one not equal to Bad one, just keep following the schematic until things mismatch, and voila you have at least generalized the area. Then it's a house-to-house search looking at voltages.
Door to door combat...
Quite genius break repair =D
With that pile of drives, I'm getting HP clock board deja vu. ;)
Tnx for the insight!
Always interesting!
Pesky voltage rails! I just fixed a co-worker's boombox and it had a no power issue. I sorted the no power issue, but then it had no sound. Turns out a capacitor had shorted on the power rail for the preamp, but because each chip is fed via resistors, the main rails appeared fine!
Marc - Well done! Now you'll have to score yourself an 8" alignment disk.
Just did! A Patreon sent one in! What a supportive bunch of people they are, the Patreons.
You forgot the first rule: "thou shalt check voltages"
I'm gonna make an addon
Thou shalt check voltages in various places
I had a boombox that took a while to find the fault, a shorted cap took out the power to one of the chips, but since each chip was fed via resistors the main rails looked fine!
Also, thou shalt check the ripple. Especially in newer stuff
But, but, you need to watch the entire video. I checked the voltages! It's just that there were two different +12V, which was not called out in the schematics. And then there was some power leaking back on the semi-broken +12V rail, making it appear as if the head bias section was working OK. This is a very unusual and rather subtle situation, and made it quite hard to ferret out. But all is good, we prevailed in the end.
@@CuriousMarc Humans win again!
@@CuriousMarc hello sir! Naturally I watched the whole thing, anything else is just rude!
Thought long and hard about commenting, and came to the conclusion that Dave's first law of troubleshooting still applies in this case.
Love your videos and the effort you make to make them!
I've long thought that full height 5.25" floppy drives would be an ideal source for robot parts. They have a large motor for driving around, a big stepper motor, two I/O channels, and an LED.
With the parts from a pair of drives a robot could have two motors to drive around, two steppers for a basic arm, and something could be hacked together to connect in place of the four read/write heads to transfer data in and out or activate things or for some type of sensor.
Да, были времена... Когда девайсы делали почти на одних проводах:-)
Moral of this repair is: Thou shall check voltages first - ALWAYS. Never assume 78xx regulators are good. They can fail in weird ways.
the old rule applies. If you don't know where to start repairing, first check that all power is on.
The look on the next guys face when he sees the popsicle sticks...
What a boss
6:40 2035:looks vintage to me
Watching Mentour Pilot followed by Curious Marc... half asleep... Marc says "seeking to 20"... I hear "climbing to 20..."
CDC was big on the ANSI logic symbology. No more "shovels and spades". :-)
I kept wondering what all those silly little boxes were... :)
3D printing some of the Plastic/composite parts would be good idea.
I'm just here for the theme music.. 😉
The "solution" for the hinge you came up with is awesome!!!!
Here in Germany we call this kind of "solution" a "Sowjet method" or "Russian method". Crude, but it works.
if it looks stupid but it works, IT AIN'T stupid
0:50 Holy heck, does the Dolch have GPIB (IEEE 488)?
You bet it does!
Love it ;)
damn, why didn't I salvage several 8" drives when I had the chance!
Good candidate for the 3D printer.
Missed a trick at the end there.
Thought we were going to get rick rolled
Thanks for sharing your videos. The amazon-links page is not working for me.
Thats' probably because you have an add blocker. Try disabling it on that page.
@@CuriousMarc No ad blocker here. No go on firefox and edge.
👏👏👏👏👏👏
kool
I have stacks and stacks of 8” drives at my old work… ya want some?
Sweet ice cream 😍😛
Not so fast with the ‘no stickers’. What about the temporary tongue depressor brake?
How does aligment work
So you are saying all floppy drives in the world have to be perfecly aligned the same way so I can write a disk on one drive to be able to read it in another?
Also what does aligment disk contain?
How was the first floppy drive in the world aligned then without the disk?
one "step" at the time
10:39: Wouldn't the head selection work just the opposite? I.e. (-12V)-|>|-(+12V) might be enough of a difference (effectively +24V) to represent a breakdown voltage, so the respective head is enabled, but whatever signal there is on top of (+12V)-|>|-(+12V) isn't anywhere near a breakdown voltage and thus blocked?
How much is one of these later Dolches nowadays? Insane money still, I guess, so I'll have to make do with a keycap set lol?
You need cat eye disk to align head. 😉
Back later…
Know anyone who repairs vintage computer equipment? I need to find someone qualified. You are very good...how much do you want?
yikes too early for HD resolution I'll be back later :D
The compression crust is too much for you? :3
15:18 норм заколхозил 👍
A brake? Stepper motors don't generally have brakes.
21:41 did I hear video amplifier?
Yeah, I saw that too, I guess it's because it has enough bandwidth to pass whatever data is coming from the heads as it passes by, plus a ton of gain. Though I can't imagine it would be anywhere near 120MHz wide. But maybe video amps are also nice and linear within that bandpass and also maybe because they were common chips by then that could do the job. MC1733 is a differential video amplifier. Reading the spec sheet it can be used with tape, drum or disc read heads. So it seems it was designed for this application in mind.
i got a CDC unit - but not this design - it's packed in a curved topset.
Goude djob.
It’s 2023, what’s a floppy drive?