Old trick - you don't need a buck convert to get 7V, just connect the fan between 12V and 5V lines! Back in the days you could even get brand-name Molex to 3-pin fan adapters wired this way in retail packaging. Having easy access to 7V is neat since many 12V fans won't start on 5V but will on 7V (they will often RUN on 5V but not enough torque to reliably start).
@@Epictronics1 Cool old hack, but dangerous for a power supply. While most PSUs have some way of short protection on each rail, next to none have protections for shorts between the rails. The way you did it is much better.
Do not do that. Power supplies are not meant to SINK current on positive rails. It will "work fine" IF.... the lower rail has more load than the higher rail (to absorb the current you're sinking into it), and _nothing goes wrong._ But that does not make it safe for the fan or the PSU. It just means you're "probably" going to get away with it. Just use a buck converter, or run it at 5v/12v.
@@nickwallette6201 5V runs the PC mainboard and cards it has massive load compared to 12v which mostly just runs some motors. It will always have greater load. spec show 5V load is 10x 12V load
@@nickwallette6201 I can't imagine a typical group-regulated power supply getting damaged by this. If you do the hack without any 5v load it will just raise the voltage of the 5v rail. This will tell the control chip to reduce the duty cycle (likely to zero) and the 12v rail will reduce and likely cycle or lock out. The exact same happens if you run the supply with only 12v loads, they need 5v load to run stable. If your supply has independent rails, then your chance of damage is much higher.
Nostalgia overload for me. I remember testing and repairing these machines back in the early '90s. MCA architectecture was quite state of the art. Using edge triggering to speed up the I/O request process and so on. IBM still thought they could lead the market with their own designs but this is when the competition really started to hurt them. ESDI and MCA ideas were never fully adopted by the market. Instead the ideas were taken and redeveloped by the competition. The '90s saw a massive increase in technological development. Ideas came and went so quickly it was hard to know what to invest in.
I think you're best doing things the way you did, removing the card and getting the 165 code, then installing the new card. Skipping a step or even sometimes completing 2 or more changes at the same time would cause terror in these machines back in the day. Much better off going slow and steady, being sure to take every step in order. Plus it's super fun working on these old IBM machines. What a great system! I am so very jealous. Thank you as always for sharing this with us all!
The case looks OUTSTANDING. Great work on the crafting of the front panel cover. It looks amazing. Seriously. Nice job! Your work brings joy to my heart, and keeping these old machines going thrills me beyond imagination. I miss the 90s.
Yeah, the 90s were super fun. (Actually 1987 to 2001-ish) Computers were changing and evolving incredibly fast compared to the 80s and the 2000s. Love this era of computers too
19:00 Two batteries is only to extend the life of it. You don't have to install two batteries since now it can be changed easily. Just install one and change it next time it is flat. You can also install a pin header and plug in a small battery holder that contains 2x CR2032 in parallel (+ to + and - to -, same voltage, double the capacity).
To avoid those "dents" on a PCB while desoldering, I usually use couple mm wide capton tape taped down very close to and along the row of pins. Both sides of the row on the PCB's side I work on. So hot nozzle does not directly touch PCB surface, but only the solder joints.
That's pretty clever. I wish I knew about this trick earlier :) My new nozzles have arrived now. They are small enough to only cover the solder pad, so they don't scratch the PCB
I wish caddy loading CD drives had become the standard. It would have been so nice and professional looking to keep a collection of discs in caddies you could just pop in at any time instead of bare discs… maybe in another timeline it caught on…
The only problem with this is that we would be missing a lot of manuals. jewel CD case have a insert for a manual in them and if every big box had a caddy instead of a jewel cd case we would be missing a lot of Manuals and CD Keys. I suppose manufactures could print the CD Key onto the caddy. But other than that I 2 would of loved for the Caddy Loading CD Drives to have taken off :)
Yeah, they're neat, and I enjoy using them in old drives, but their days were numbered from the start. For one, they're very expensive. This wasn't SUCH a concern when CD-ROM discs were new and very expensive. But, in the days where mass-producing discs was cheap, it wouldn't have made sense to keep spending so much money on complex shells and elaborate loading mechanisms in the drive. For two, they wouldn't have worked at higher speeds. They were always going to be limited to the early, slow drives. But how cool is it that all those manufacturers agreed on a common caddy design? It blows my mind that there weren't six different variations that were totally incompatible with each other.
It is amazing that it took so long for someone to try depotting the dallas modules with heat only for it to turn out to be super effective and easier than trying to drill or dremel your way in.
I got from a liquidation at a old factory a box with a lot of those long pins sockets. Apparently they were used for wire-wrap circuits. I used to cut theirs legs when I needed to use them. I also used to make small circuits (spectrum and MSX modules) with sandwitched pcbs and those sockets came in handy by passing from a pcb to another, given both structural strength and electrical connections.
Considering that for a long time, computers weren't really color coordinated pieces of artwork, and were more variations of beige, I think your cosmetic panel turned out well. Perhaps, the fact that it's a different shade even makes the overall presentation a little less boring. Personally, I might have tried to find some sort of functional accessory to fill the bay, whether it's a slot fan, a 5 1/4 / 3 1/2 combo drive, or a 6x 5 disk CD changer. My machine has a memory card that's similar to that, except it uses 20 pin SIMMS and it's EISA based. It has a configuration program, as well. Fortunately, it doesn't have a clock module on it.
That Noctua fan isn't going to do much with like 90% of it restricted. I'd recommend a PCI slot cooler instead, it's a squirrel cage blower you can throw in an empty slot and moves a ton of air. It will prevent dead space between the cards where no air will flow.
That's a nice idea but MCA slots use a totally different bracket, it would require a custom solution. There's also the chance it would alter the airflow pattern and prevent all the components from getting cooled properly.
@@GGigabiteM There's enough of an airflow pattern that they put in air ducting. IMO the only things in there that could use cooling are the drives (and the power supply). Putting an exhaust in the rear might rob the drives of air. I don't think it would matter but it's worth checking temperatures before and after.
I think if it has automatic switchover, one cell instead of two might be just fine! There's probably no smarts there at all, just two cells connected via diodes against current flowing back into each cell. With the diodes terminating into the same voltage rail, the cell with the higher voltage will end up being the one to deliver current. They must have implemented that purely to provide more effective capacity and running time, as well as increase long term reliability in case of a single cell being a dud - if you use a battery holder or use a higher capacity cell like cr2032, you really don't care about that. Also something tells me you want a 3D printer eventually. Things like that baffle are way less work, they come out dimensionally accurate with no solvent/glue induced warping and no wobbly lines from hand cutting, and you can integrate further features like latches quite easily. Quick iteration time vs. sending it off for someone to 3D print for you across the globe is very nice too, since mistakes inevitably creep in, and rather than spend exuberant time avoiding them, you just test fit and try again.
Thanks! But don't be too quick to hack the Dallas if it's a DS1213. There are custom PCBs developed for it now by VintageSupportService. I'll let you know when they are done :)
Even if the color of that panel cover isn't perfect, it still looks like it could be something one could've bought back then to put a CDROM or tape drive or something in to that computer.
If those strips are polystyrene then the tape should have made it easy to use simple poly cement you can find at any hobby store, preferably the thin liquid stuff. Most modelers we use that and just brush it in gaps to attach the bits together. Again that is if it’s the material I’m thinking it is.
A tip that can be useful, if you use super glue or any other cyanoacrilate glue get cyanoacrilate activator, usually comes as a spray, apply it where the junction is and that thing becomes rock solid in a couple of seconds, it reacts with it, the cyanoacrilate gets burning hot (belive me, it's damn hot xD) and the glued area becomes 3 or 4 times stronger than with the glue only.
you can also use baking soda (not baking powder) for this. It's because cyanoacrylate reacts with organic matter, and baking soda is a crystalline organic compound with a very high surface area that highly accelerates the curing process. Also this is why you can get burns from spilling CA glue on your cotton t-shirt.
I'm pretty a sure they used to use small quantities of mercury in those old coin cell batteries. Not a great idea to heat them up and possibly release mercury vapor.
That is the same cdrom as the apple cd300 right? Those need to be recapped if you havent already. Enjoyed the epoxy removal. Nice to know we can salvage them more properly
Thanks, not quite, this is a Sony 6x read 2x write!. Yes, I have recapped it in a previous video. I have a CD300 and a bag of fresh caps right next to me actually :) We'll fix it in a future video.
It's cool how many clever ways people have hacked those potted ICs, but IMO, it's not worth the trouble to do any of it. They're still sold new, and they're trivially easy to make (PCB, pins, timer chip, crystal, battery.) Unless you're trying to McGyver something back together in a hurry, or just can't stand the idea of dropping a tenner on a new one, I just don't understand the fascination in reviving an old one. IMO, it deserves its rest. :-) But, to each their own! (EDIT: To be clear, this video is an exception, as we're uncovering a new module, and need to learn its secrets. >:-) But also Necroware's video was valuable as a proof of concept that begs the question -- is this unique to the compound Dallas used, or can this be used with other potted components that aren't easily sourced??)
I think they actually are twice that price, plus shipping. If it only takes ten minutes to repair. Why replace it? Another benefit is that you could potentially repair the module without even removing it from the mobo, if it's soldered. It's also very gratifying to peel that resin off :) I need to try this method on something more
Wow that was beautiful how you destroyed that epoxy, you gave this Dallas crap what it deserves. I remember I bought a new "replacement" Dallas many years ago for a commodore PC but I guess it never really worked or at least I never really used it, sure it is dead now. Did not know that they just bent up the pins, one could just attach a battery outside if you know where and maybe clip the leads to the original one. But anyway, not into that retro PC thing especially not MS-DOS machines. I have a commodore ac b3-a
Thanks. I have the very similar PC 35-III. I use a Dallas replacement in that machine at the moment, but I will hack and put the original chip back on that board. Now that we know how to fix them :)
The edges on the plastic piece are bugging me so much. If you take measurements of the piece, I will design and 3D print you one just to get rid of those edges.
Restoration video here: ua-cam.com/video/o6b9dE5wht4/v-deo.html
@Epictronics
Nice work! Very well done indeed!
Thanks!
Old trick - you don't need a buck convert to get 7V, just connect the fan between 12V and 5V lines! Back in the days you could even get brand-name Molex to 3-pin fan adapters wired this way in retail packaging. Having easy access to 7V is neat since many 12V fans won't start on 5V but will on 7V (they will often RUN on 5V but not enough torque to reliably start).
cool hack!
@@Epictronics1 Cool old hack, but dangerous for a power supply. While most PSUs have some way of short protection on each rail, next to none have protections for shorts between the rails. The way you did it is much better.
Do not do that.
Power supplies are not meant to SINK current on positive rails. It will "work fine" IF.... the lower rail has more load than the higher rail (to absorb the current you're sinking into it), and _nothing goes wrong._ But that does not make it safe for the fan or the PSU. It just means you're "probably" going to get away with it.
Just use a buck converter, or run it at 5v/12v.
@@nickwallette6201 5V runs the PC mainboard and cards it has massive load compared to 12v which mostly just runs some motors. It will always have greater load. spec show 5V load is 10x 12V load
@@nickwallette6201 I can't imagine a typical group-regulated power supply getting damaged by this. If you do the hack without any 5v load it will just raise the voltage of the 5v rail. This will tell the control chip to reduce the duty cycle (likely to zero) and the 12v rail will reduce and likely cycle or lock out. The exact same happens if you run the supply with only 12v loads, they need 5v load to run stable.
If your supply has independent rails, then your chance of damage is much higher.
Nostalgia overload for me. I remember testing and repairing these machines back in the early '90s. MCA architectecture was quite state of the art. Using edge triggering to speed up the I/O request process and so on. IBM still thought they could lead the market with their own designs but this is when the competition really started to hurt them. ESDI and MCA ideas were never fully adopted by the market. Instead the ideas were taken and redeveloped by the competition. The '90s saw a massive increase in technological development. Ideas came and went so quickly it was hard to know what to invest in.
You didn't get a Varta leak this time so you made a Dallas leak instead 😝
Seeing that battery juice fresh and bubbling, It really looks like it can cause serious damage. Nasty! :o
@@Epictronics1 Especially when it's hot!
That battery mod looks cool with those two batteries sitting on the side like that. Almost looks factory tbh.
Thanks. I rather like how it turned out too
I think you're best doing things the way you did, removing the card and getting the 165 code, then installing the new card. Skipping a step or even sometimes completing 2 or more changes at the same time would cause terror in these machines back in the day. Much better off going slow and steady, being sure to take every step in order. Plus it's super fun working on these old IBM machines. What a great system! I am so very jealous. Thank you as always for sharing this with us all!
Thanks!
I love the look of these old PS2 PCs.
The case looks OUTSTANDING. Great work on the crafting of the front panel cover. It looks amazing. Seriously. Nice job! Your work brings joy to my heart, and keeping these old machines going thrills me beyond imagination. I miss the 90s.
Yeah, the 90s were super fun. (Actually 1987 to 2001-ish) Computers were changing and evolving incredibly fast compared to the 80s and the 2000s. Love this era of computers too
Dallas chip is difficult to open and replace the battery. This heat treatment should be remembered the next time repair an old chip :)
19:00 Two batteries is only to extend the life of it. You don't have to install two batteries since now it can be changed easily. Just install one and change it next time it is flat. You can also install a pin header and plug in a small battery holder that contains 2x CR2032 in parallel (+ to + and - to -, same voltage, double the capacity).
Ok, cool!
To avoid those "dents" on a PCB while desoldering, I usually use couple mm wide capton tape taped down very close to and along the row of pins. Both sides of the row on the PCB's side I work on. So hot nozzle does not directly touch PCB surface, but only the solder joints.
That's pretty clever. I wish I knew about this trick earlier :) My new nozzles have arrived now. They are small enough to only cover the solder pad, so they don't scratch the PCB
I wish caddy loading CD drives had become the standard. It would have been so nice and professional looking to keep a collection of discs in caddies you could just pop in at any time instead of bare discs… maybe in another timeline it caught on…
So many disks would have been unscratched
The only problem with this is that we would be missing a lot of manuals. jewel CD case have a insert for a manual in them and if every big box had a caddy instead of a jewel cd case we would be missing a lot of Manuals and CD Keys. I suppose manufactures could print the CD Key onto the caddy. But other than that I 2 would of loved for the Caddy Loading CD Drives to have taken off :)
Yeah, they're neat, and I enjoy using them in old drives, but their days were numbered from the start.
For one, they're very expensive. This wasn't SUCH a concern when CD-ROM discs were new and very expensive. But, in the days where mass-producing discs was cheap, it wouldn't have made sense to keep spending so much money on complex shells and elaborate loading mechanisms in the drive.
For two, they wouldn't have worked at higher speeds. They were always going to be limited to the early, slow drives.
But how cool is it that all those manufacturers agreed on a common caddy design? It blows my mind that there weren't six different variations that were totally incompatible with each other.
It is amazing that it took so long for someone to try depotting the dallas modules with heat only for it to turn out to be super effective and easier than trying to drill or dremel your way in.
Yeah, this is quite the discovery!
The work on the timer chip looks neat.
Thanks
Awesome job on that Dallas chip!!!
Thanks :)
I got from a liquidation at a old factory a box with a lot of those long pins sockets. Apparently they were used for wire-wrap circuits. I used to cut theirs legs when I needed to use them. I also used to make small circuits (spectrum and MSX modules) with sandwitched pcbs and those sockets came in handy by passing from a pcb to another, given both structural strength and electrical connections.
I think they are still available!
@@Epictronics1 They are available, but are way more expensive than the regular ones.
Considering that for a long time, computers weren't really color coordinated pieces of artwork, and were more variations of beige, I think your cosmetic panel turned out well. Perhaps, the fact that it's a different shade even makes the overall presentation a little less boring. Personally, I might have tried to find some sort of functional accessory to fill the bay, whether it's a slot fan, a 5 1/4 / 3 1/2 combo drive, or a 6x 5 disk CD changer. My machine has a memory card that's similar to that, except it uses 20 pin SIMMS and it's EISA based. It has a configuration program, as well. Fortunately, it doesn't have a clock module on it.
That Noctua fan isn't going to do much with like 90% of it restricted. I'd recommend a PCI slot cooler instead, it's a squirrel cage blower you can throw in an empty slot and moves a ton of air. It will prevent dead space between the cards where no air will flow.
That's a nice idea but MCA slots use a totally different bracket, it would require a custom solution. There's also the chance it would alter the airflow pattern and prevent all the components from getting cooled properly.
@@eDoc2020 There is no airflow pattern in that case, it's almost entirely dead air.
@@GGigabiteM There's enough of an airflow pattern that they put in air ducting. IMO the only things in there that could use cooling are the drives (and the power supply). Putting an exhaust in the rear might rob the drives of air. I don't think it would matter but it's worth checking temperatures before and after.
I think if it has automatic switchover, one cell instead of two might be just fine! There's probably no smarts there at all, just two cells connected via diodes against current flowing back into each cell. With the diodes terminating into the same voltage rail, the cell with the higher voltage will end up being the one to deliver current. They must have implemented that purely to provide more effective capacity and running time, as well as increase long term reliability in case of a single cell being a dud - if you use a battery holder or use a higher capacity cell like cr2032, you really don't care about that.
Also something tells me you want a 3D printer eventually. Things like that baffle are way less work, they come out dimensionally accurate with no solvent/glue induced warping and no wobbly lines from hand cutting, and you can integrate further features like latches quite easily. Quick iteration time vs. sending it off for someone to 3D print for you across the globe is very nice too, since mistakes inevitably creep in, and rather than spend exuberant time avoiding them, you just test fit and try again.
Yeah, I'll get a printer eventually for sure. Hopefully, I'll find an original bezel someday too!
I never had any specific attachment to IBM machines, but that is a handsome one indeed. Nice work.
Thanks :)
Great info about that Dallas chip! You totally helped me figure out something about the Acculogic RAMpAT! Cards I have. And the bezel looks great!
Thanks! But don't be too quick to hack the Dallas if it's a DS1213. There are custom PCBs developed for it now by VintageSupportService. I'll let you know when they are done :)
Even if the color of that panel cover isn't perfect, it still looks like it could be something one could've bought back then to put a CDROM or tape drive or something in to that computer.
Dallas: when you're dealing with badderies
At least it's a little bit better now and a slightly more elegant fix :)
First PCB draft sent! Thanks for the video.
Damn, you're quick! They look sweet!
I know nothing about nothing but how impossible would it be to create a Powerleap PL-IP3/T adapter ?? :)
Very nicely done, and good to see the machine not hang when booting ... thankyou ...
Thanks!
If those strips are polystyrene then the tape should have made it easy to use simple poly cement you can find at any hobby store, preferably the thin liquid stuff. Most modelers we use that and just brush it in gaps to attach the bits together.
Again that is if it’s the material I’m thinking it is.
I woulda stuck a bit of double-sided tape under the breadboard to stop it sliding everywhere. :)
Yeah, I realized while editing that I could have just taped down the breadboard to the bench!
Hi Epictronics, Nice mod and repair from the IBM PS/2 Model 80. Greetings from Steven from the Netherlands
Thanks!
A tip that can be useful, if you use super glue or any other cyanoacrilate glue get cyanoacrilate activator, usually comes as a spray, apply it where the junction is and that thing becomes rock solid in a couple of seconds, it reacts with it, the cyanoacrilate gets burning hot (belive me, it's damn hot xD) and the glued area becomes 3 or 4 times stronger than with the glue only.
you can also use baking soda (not baking powder) for this. It's because cyanoacrylate reacts with organic matter, and baking soda is a crystalline organic compound with a very high surface area that highly accelerates the curing process. Also this is why you can get burns from spilling CA glue on your cotton t-shirt.
@@Nukle0n Yeah, I have used baking soda in the past mostly for filling, but the activator does a much better job :)
Yes, this is a sweet trick I regularly use. However, I use regular Sodium bicarbonate. I wasn't aware of a special activator
@@Nukle0n I have fixed so many things with this method!
Interesting. I thought the activator stuff was just water as that glue is activated by moisture. Good to know!
you've already got 7 volts on molex between 12v and 5v pins
Back in the days we lowered the fan voltage to 7 volts by connecting between 12 V and 5 V rail
This is an old hack, but it is not a good idea. See the longer thread here about this.
I'm pretty a sure they used to use small quantities of mercury in those old coin cell batteries. Not a great idea to heat them up and possibly release mercury vapor.
With the resin, i wonder if you can't heat it locally with a soldering iron to crack it? Just a hypothesis, haven't tried it myself
Nice Spock Card ❤ , ebay lottery was good too you. I got a Z690 mobo with 3 dead ram slots 😂
They should be fun to replace :D
Very cool stuff!
Thanks!
It's nerve wrecking when you trying to fix something but there is a chance of Effing it up worst in a worst matter.
Luckily, I think we did pretty well today?
@Epictronics1 You done great. I how ever always have 20% chance of success and an 80% chance of failure.
Nice job very elegant.
Thanks :)
That is the same cdrom as the apple cd300 right? Those need to be recapped if you havent already. Enjoyed the epoxy removal. Nice to know we can salvage them more properly
Thanks, not quite, this is a Sony 6x read 2x write!. Yes, I have recapped it in a previous video. I have a CD300 and a bag of fresh caps right next to me actually :) We'll fix it in a future video.
@@Epictronics1 ah! I've done so many of those. Always a pain
Pair this machine with an IBM Model M and it screams "I'm getting SERIOUS work done!"
Could you 3d print the missing plastic parts ?
I'm saving up for a printer. Next week we are going to fix another PS/2 with printed parts from PCBWay!
@Epictronics1 Always find your videos interesting- look forward to viewing it.
@@larryk731 Thanks!
Prints will be available soon! A viewer is sending in a bezel. I'll measure it and make all the info available
It's cool how many clever ways people have hacked those potted ICs, but IMO, it's not worth the trouble to do any of it. They're still sold new, and they're trivially easy to make (PCB, pins, timer chip, crystal, battery.) Unless you're trying to McGyver something back together in a hurry, or just can't stand the idea of dropping a tenner on a new one, I just don't understand the fascination in reviving an old one. IMO, it deserves its rest. :-) But, to each their own!
(EDIT: To be clear, this video is an exception, as we're uncovering a new module, and need to learn its secrets. >:-) But also Necroware's video was valuable as a proof of concept that begs the question -- is this unique to the compound Dallas used, or can this be used with other potted components that aren't easily sourced??)
I think they actually are twice that price, plus shipping. If it only takes ten minutes to repair. Why replace it? Another benefit is that you could potentially repair the module without even removing it from the mobo, if it's soldered. It's also very gratifying to peel that resin off :) I need to try this method on something more
Good work as always.
Thanks :)
You should have socket it.
Not sure it would fit. It was already two sockets high
Wow that was beautiful how you destroyed that epoxy, you gave this Dallas crap what it deserves. I remember I bought a new "replacement" Dallas many years ago for a commodore PC but I guess it never really worked or at least I never really used it, sure it is dead now. Did not know that they just bent up the pins, one could just attach a battery outside if you know where and maybe clip the leads to the original one. But anyway, not into that retro PC thing especially not MS-DOS machines. I have a commodore ac b3-a
Thanks. I have the very similar PC 35-III. I use a Dallas replacement in that machine at the moment, but I will hack and put the original chip back on that board. Now that we know how to fix them :)
Great
Thanks!
@@Epictronics1 thanks
Cr1616 batteries maybe fit ok?
As long as you use two 3v batteries It shouldn't matter
So, can we remove this Dallas Chip and connect one battery direct to clock chip?
From what I could read in the datasheet, it seems to need two batteries for operation
@@Epictronics1 wow, how weird :)
@@Epictronics1 I say one batt is worth a try.
Imma double your watch engagement and chat/comments too ❤ 😂
19:00 - wow, that sounded dramatic Hope you have some more life left than the coin cells, my man
lol. I hope so too!
The edges on the plastic piece are bugging me so much. If you take measurements of the piece, I will design and 3D print you one just to get rid of those edges.
Thanks, I appreciate it. I may have an original cover on the way. If I get it, I may need some help to create print files to share in the community
That's awesome using hot air to fix Dallas chip
You have to be quite petty to invest so much extra work (and money) in making your clock-chip un-servicable.
I don't get this. Apparently they still make these dallas chips and you can get similar ones anyway.
This is a special! It's not made anymore. Other Dallas chips are still made but quite pricy...