Making a New Coleco Adam Internal PSU

today we're gonna be taking a look at getting one of the lesser-known computers from the 80s working a calico atom this project is sponsored by PCB way but we'll take a closer look at that later the atom was released in 1983 as calicos first and only entry into the home computer market however the atom is significantly derived from the ColecoVision game system released in 1978 so much so that it's full title is the ColecoVision family computer system and the atom was available as an add-on for the ColecoVision that would plug into its front expansion interface the atom was designed so similarly to the vision that it's even able to play the same cartridges using the same controller albeit cast in white this time however don't think that the atom is just a simple reskin of the vision where the vision allows you to store your controllers the atom has internal upgrade slots it allows you to expand the memory from the built-in 64 kilobytes and to cassette Bay's that can be installed let's actually talk about that cassette drive a little bit because it is one of the more interesting aspects of the atom as well as the most significant issue it is not surprising that this computer from 1983 would use cassettes for software that was extremely common at the time however Coleco decided to make the cassette format for the atom proprietary with a cursory glance the atom digital data pack doesn't seem all that different from the Philips Compact Cassette but these are not able to be used interchangeably side B on the compact cassette is identical to side a however side B on the digital data pack has holes cut in the top these holes are meant to prevent you from being able to put in a standard compact cassette in your Coleco atom additionally the Compact Cassette has four sets of holes at the bottom the outer holes are used to pinch the tape to allow it to be pulled through the atom does not use these holes and only has two on the bottom this means that the atom digital data pack cannot be put into a standard compact cassette player okay so why not drill holes in your standard Compact Cassette then and try and stuff it into the atom well this right here is why pre formatted Coleco atom data packs are hard sectored they have a track marking embedded in the tape that keeps the data bits aligned with where they are supposed to be much like how modern high-capacity tape drives work today this means that pre formatting is vitally important if a magnet were to get near your cassette not only would it wipe out your data but it would also wipe out those hard sectors rendering the tape useless unfortunately the atom was also not a well-designed machine it was found on startup that the computer emitted an electromagnetic pulse that would erase tapes in the drives this would also happen if the tapes were stored too too close to the printer so later units had this sticker applied warning you not to leave your cassettes in the computer when you turn it on speaking of turning it on why don't we do that now well there's a reason I'm missing perhaps the most important peripheral for this system the printer the Calico atom was only available as an incredibly large box set that included the computer the keyboard two joysticks and the printer now because the atom would always include the necessarily large and bulky printer just because of the physical dimensions of paper they thought it would be a good idea to integrate the power supply into it this means without the printer you cannot power the Coleco atom because of this it is not uncommon to find atoms such as mine orphaned from their printers and unusable but I don't think that this should be the death sentence that it inevitably is in today I'm going to design a new power supply usable with the Coleco atom now thankfully there is a large crossover between vintage computer enthusiasts and electrical engineering enthusiasts and someone was able to get a hold of the original schematics for the Coleco atom and published them online and looking at what I believe is the power supply connector we can see that we have minus 5 minus 12 plus 5 and the other typical voltages found in an ATX power supply and looking around online I'm not the only one who's thought of this as there are power supplies that have been modified to have a db9 connect to plug into the atom however I feel there is room for optimization here and longtime viewers of my channel probably know where I'm going with this I created a similar power supply adapter for the IBM PC jr. and I'm going to take a lot of those design elements and apply it here making a power supply that can be attached to the side of the Coleco atom and run it off of 12 volts DC using a pico PSU however before I go firing up kaya designing a board for this thing there are a couple of things I need to verify I need to open it up and confirm that the power port is in fact j9 and that that is the correct pin out for it and that the negative power supply requirements are not too great for a Pico PSU especially since the cassette drives are integrated into the system I can see over here these wires running from this board with connector come over to here and then are labeled by color not by voltage but we have j9 back here and if I continuity test this j9 I can see that it goes over to the db9 connector on the side all right now going back and referencing that schematic I can confirm the pin out the pin eight on here is in fact ground pin five also connects to ground those two are shorted here so I'm sure if I go around and find chips I could find that the five volts 12 volts negative five volts all go to where you would expect them to go so I can confirmed the pin out of this port as well as the functionality of it so that's good now I need to spend a bit of time staring at the schematics before I can be totally sure that I can actually power this device off of a Pico PSU but before that seeing this empty space in here gives me an idea I could make the ATX power supply adapter internal but for now let me go peruse the schematic and see if I can figure out where all it looks like just- five volts because even though this looks like a negative 12 and it's obviously positive 12 so it looks like there is no negative 12 which I can't remember is there actually negative 5 on the ATX spec hmm nope it is not so I'm going to have to go ahead and create my own negative 5 volt power supply great well looking more closely at the schematic I'm not really sure where the negative 5 volt rail connects because it is really low resolution and it's hard to tell the difference between a positive 5 volt rail and a negative 5 volt rail thankfully the atom technical reference manual tells you what the absolute maximum power draw can be and it needs to be able to power 200 milliamps on the negative 5 volt rail the negative 12 volt rail on the Pico PSU is only 50 milliamps and I don't like my chances of using something like an LM 70 905 to step that down it's probably not going to be able to deliver enough current so I'm going to have to put on my own inverting voltage regulator and honestly if I'm going that far I might as well do a 5 volt regulator as well because the Pico PSU would only be contributing 5 volts at that point so ultimately I guess my board is going to take a 12 volt input directly and output a positive 5 and negative 5 volt rail alongside passing through the 12 volts that should give me all of the voltages I need for the Calico atom and as long as you use say about a 3 amp 12 volt power supply with the regulator's I'm choosing it should be able to power it just fine now let's get on to laying out these schematic and circuit board [Music] okay the Adam internal power supply is complete I am just about ready to order this I just need to do one final check for things on the board just to make sure that everything is looking good now that the board design is finally finished I can go ahead and order PCBs and since PCB way is sponsoring this video I'm going to go ahead and do that with them but that's not the only reason I've actually used PCB way for professional boards I've designed in the past and their boards have always turned out great so the first thing I'm going to need to do here is put in my board dimensions and mine is about 79 by 61 and I'm going to get a quantity of 10 because they have a prototyping discount for a order of 10 boards that gets you a lot better of a deal if I change us to say 15 it jumps up quite a bit they want to do prototyping orders like this because they have panels that they'll run in quantities of 10 and it works out really well for everybody to do that then you can do different board options like the layer thickness of the fiberglass stack up of the thing is called the prepreg so yeah there's a lot of options I'm going to order mine in blue because I like that for prototype boards enig would be cool but it really is the price up one thing I noticed is their assembly service charge is actually really low so I'm getting 10 boards right now I'd put in 10 boards I have 23 SMD parts on here unique parts would be the db9 connector the barrel jack connector all that stuff I can just do that as hand soldering when the parts arrive and if I calculate how much that would add it's really not that much $30 charge to get assembled boards and I actually you bump up into free shipping with this okay if we take that away we get free shipping so it adds almost no costs it's $13 extra you get a 10 assembled boards now if you're ordering more boards the price break is even better but yeah that's a really good deal for the PCB assembly cost I am going to solder this by hand though because I want to check the different stages as I populate the board but that's it for now I'm gonna go ahead and order my boards after you place the order you'll be taken to a page where you can upload your Gerber's and then they get approved and the manufacturing starts so let's check back in a bit when I get my parts and the boards alright getting back to this a little bit later and I have my boards from PCB weight and parts from digi-key I ended up going with digi-key because there were some other things I needed for different projects that I can only find on digi-key so there we go but let's go ahead and take a look at the boards I have already opened this up and been in here and it seems like PCB way is celebrating five years of being in business and they've included some little goodies I'm not sure everyone's getting those or if that's because this is a sponsored video but cool and here we go the board's alright let's go ahead and take a look at these circuit boards they seem like they're pretty darn good to me this is a slight mistake that is my part I made a spelling error on there thought I'd corrected it the correct way with the foot print but it duplicated that so that's my fault but elsewise everything else looks really good I'm pretty happy with how this little thing turned out on there yeah everything seems fine there just a couple of little design flaws that are my fault like I didn't pay attention to how thick the touch points on are on here and I was hoping you could run wires through those and manually install this with soldering if you really wanted to but oh well right here the cable was actually internally going to plug in on the board so that the device can be installed inside of the atom if you don't have a actual standalone atom and you have the expansion module number three you'll be able to put a extension on here they make serial through extension and mail on one side female on the other and then you can stick this on the side of the unit I would not recommend using a extension cable just on the off chance that it's a null modem cable because that will actually switch the 12-volt in five volt lines that's those are the ones that are switched on this and that would that would do nasty things to your atom so you don't want to do that but okay let's take a look at how this is going to work so the first area on the board is this right here this is the power input so there is a five point five millimeter barrel Jack that will go on there and you will plug a 12-volt power supply into this there is a MOSFET on here that will be controlled by this switch I wanted to avoid going with a really expensive switch on this kit unlike the PC jr. kit so I went with one that was a lower capability it's not to have this cap on it but it looks like I didn't take into account the board size and the cap doesn't really fit well anyway despite being designed for this series of switch so that's that's frustrating so that's how that works this was an idea I had I don't really like the concept of antic waiting the printers entirely so we'll be able to put a pin header there and then take a serial header like you'd put on a motherboard and plug this in right there and then Dingle this up of the side of the case and then you should be able to use the printer still so this just passes through the data and reset lines on here all right the rest of this is divided up above the negative five volt regulator and down below both of these are buck switch mode power supplies now I've never laid out switch mode power supplies before so this is kind of new and exciting and scary to me I'm I have no idea if these are gonna work yet so this one should be able to deliver I think it's yeah five amps on the positive voltage side that's way more than enough I think it needs two and three quarters at most and then the negative side can deliver at one amp so that's five times as much as it needs so this should be good as long as it ends up working this is actually just a regular buck converter here and I'm using an in an inverting way it's kind of strange it's apparently really easy to make an inverting buck converter circuit so I'm curious if that's gonna work at all so but outside of that everything else looks good so I'm looking forward to assembling one of these actually I should see how it fits first and mmm I may have messed that up a little bit all right I put the screws in there the fits a little close to the edge on there but the power jack for the most part will fit in there but it's ein a little bit of an angle if I let me grab another one here if I extend out this on this side then this will fit just fine it really won't be an issue so there's nothing over there it's just ground plain and stuff so yeah on even these prototype boards I could fix that by just nipping out a little bit there and then scooting this over a tad oh yeah there we go I have a lot more wiggle room with the board now that's what I originally wanted all right that's the power switch circuit installed now I have this wire hooked up to the power input for the barrel jack this is the exact same pin out this way I can connect this to my bench power supply off to the side and not need to risk shorting out my big high amperage 12-volt power supply because when that goes south it's gonna go really really south all right there we go on at eleven point five and off at just a handful of millivolts I replaced our four which was 10k with a 1k resistor so the pull down effect is much greater and that seems to have solved the problem I guess 10k was just not enough all right so the power switch is a success so now for the next part and I don't know why I even thought about making it a question I want to do the negative 5 volt power section next because I really have no idea that's actually going to work or not so let's find out it's a lot of little parts to put on there to just take a guess and see if it'll work but let's do it yep that's a good-looking and doctor on there at turned out perfect I'm super happy with that wow it looks really good this is gonna be all slick when it's done if it works that doesn't work well yeah that'll suck I am getting 19 millivolts out of it 20 that's not good all right the problem was something I had suspected but couldn't find any real strong evidence about one way or the other before laying out the board I should have really accounted for this in the layout but oh well at least it was a pretty simple bodge wire the enable resistor for the buck regulator was tied to ground which is technically in a way high for a negative voltage regulator but it needed to be tied to be in but I am now generating 5 volts negative DC oh yeah just for the record though cleaning this was incredibly important or else the the bootstrap signal the frequency the that does the looping doesn't really get through all that well alright now if you're curious here's the output ripple and the frequency of the output that seems really low to me I'll have to go back and look at the datasheet see approximately what this should be but I'm getting about 50 millivolts this is 10 millivolts per division here for the output ripple that's not bad that's about what you'd expect I haven't tried loading this yet so I'm not sure how it's going to respond to that but I'll do that after I'm getting closer to putting this in the computer but for now I'm pretty happy with a 50 millivolt ripple I completely spaced on one aspect of this design and it's a slightly big problem the thermal pad for the 5 volt regulator has vias through it and I was going to put the soldering iron on the other side to heat those but no because I forgot to put a mask relief on this side so I could you touch it with the iron so it's gonna be really complicated to try and heat that up actually I have a potential alternative solution if I remove some of the solder mask going away from this I should be able to apply heat on the top side and have it flow underneath the part all right let's see how well this works this is going to be very strange that'll have to be good enough I can go back and pop it with the heat gun from the other side but now I can work on this all right and there we go there is the five volt output this is after a rigorous cleaning the entire board has been cleaned and I'd say that's pretty good that's a really low ripple voltage there again this is ten millivolts per division so that looks like about ten millivolts total of ripple again unloaded though and we are getting five volts out perfectly okay load testing this is slightly annoying because I don't have a programmable load but I have a couple of things that aren't bad here so let's start off by powering a Raspberry Pi from this so I have let's start this out here the fluke 287 running a current sense through there all right the 34401a measuring the voltage and the oscilloscope measuring ripple AC so we can see the PI's pull in about 60 milliamps it's really not all that much okay let's step it up to the next most hungry device this weird Chinese cheap power supply and we're getting a hundred milliamps current draw it's holding around four and that's bouncing around all over the place I didn't notice that's the first time next-largest loads see how that stays consistent here a bigger battery bank and this one is consistently 200 milliamps that is all over huh the ripples pretty good but let's get to the last one that draws the most current and has the worst ripple by far this rechargeable Logitech G 700 s Mouse and if we turn this on yet 320 milliamps current draw this is really bad ripple we got about 50 millivolts of ripple here and voltage is going down quite a bit there so that's great next let me try the negative 5 volt rail I'm just gonna swap the two power connections coming off of the board here because theoretically that should be exactly the same that should be capable of delivering an amp so we'll see how that goes I'm gonna go ahead and summarize the negative 5 volt rail test and post here because the USB cable I was using for all this wasn't in the best of shape and it was changing the results and the ripple changed significantly with load for the negative 5 volt rail so we'll take a look at it again later well okay I think this is validated enough that it's time to stick it in the atom and see what happens I don't even know if my atom works yet okay here we go I'm gonna plug in power I have no power light on this I guess I could have done that um ok it should be off and I guess it's not oh it was in the open that was working dude it works it's not great I'm connecting to the monitor oh there we go it works the power supply works the atom works oh that's awesome ok after a bit more testing I can save for rather certain that the atom is working I am just really shocked and thrilled about this so yeah this kind of really rocks here I'm not sure if this controller is fully functional I don't have a lot of Cliquot fission experience do you have to push both buttons to activate because that's what it's taking maybe I'm just pushing one rather hard cuz I'm pinching more but anyway yeah this is pretty much working now so I don't have sound connected in any way ah no there we go so alright that was just one button but this is feeling pretty darn good here so I need to resolve that power button issue because I don't know what's going on there but yeah let's get started on that it's clearly working it just is also not really working that should be off right now actually but it's not welcome to the next day and I'm still trying to troubleshoot the power button and I just figured it out alright so here I have cut a trace on the board between the barrel jack and the MOSFET so I could verify that the current was actually flowing through these switches in the on position up here it's negative just disregard that we're flowing 45 milliamps through which side note this thing's running on almost no current at all right now anyway turn it off off zero milliamps now it's pretty low current but it's not running on nothing well here we can see this light alright that's running just fine and I put that light there because we need to see when the unit turns off when there's no screen because that's the problem if I unplug the monitor the LED turned off I am NOT going to plug it back in because I did it once and there were sparks the power supply I'm using is somewhat of the culprit here I have my fluke 287 off to the side and continuity mode all right barrel Jack on the power supply third pin ground it's grounded hmm those of you who know how not to fry your oscilloscope know where this is going ground on the power three prong power cable going to the barrel on the monitor also ground so it's conducting DC ground through the video input so yeah the switch actually is working if I weren't using this power supply now we're using an isolated to amp power supply let's just say I probably have one around here but you know the idea it's just one of the wall work type ones not a big beastly one like this it would have worked normally so it was actually a good thing I was using his power supply because it demonstrated a potential flaw on this design I must switch to high side switching for this because this is this is a I'm not gonna be able to fix this there's just no way so yeah the switch does work but it goes through the video port so that's an interesting development okay let's give it power all right now actually this by 12 volt power supply has a fail mode here for some reason here is actually this here's the ripple coming out of that right now so there's 50 millivolts 200 millivolts or ripple if I unplug and plug it back in BAM 50 millivolts ripple that's a lot better though my 12 volt power supply it kind of sucks but anyway so this is without the tape drives no controllers no cartridge no keyboard anything connected to this so there should be a a medium sized load for this so the positive 5 volt rail is 5 volts solid and pretty solid negative 5 volts ok let's look at the ripple on these so negative 5 is let's see it's dropped us down to 10 millivolts that's 20 millivolts per division Harris we're getting about 80 millivolts ripple that's not tremendous that's a very interesting curve there too well right I think this is about the maximum it's gonna draw for negative I don't think it's gonna change all that much let's check out 5 volts 5 volts that's it's rock solid that one's not having any problems at all that one has a little bit better filtering on it I could probably add a couple more filter caps to this it may need just more bulk output I don't know yeah I think I'm going to try putting the keyboard and tape drives in here I'll tape drive I only have the one and see how it does then and see if I can load a cassette basic I don't even know if that cuz that's good because these can toast it on startup if you leave it in there but let's see okay this is just gonna be straight up power on test to see what happens so here we go didn't seem to try and access the tape drive but it may have some kind of well I think it does actually have tape detect let me throw the basic cassette in there and see what happens when I reset computer all right smart basic and I cuz that's not doing great in there I need to stop it that pork is said I'm pretty sure it wasn't like that when I put it in there great okay after much careful finessing it is rewound I'm going to go ahead and close it off hand put all the screws back in how in the world did that happen okay I'm not taking any chances and I rewind that sucker by hand so all right here we go let's try this again it's actually on the spool this time that's terrifying whoa it's supposed to be superfast Wow yeah it didn't seem very successful okay I have no idea I think I have one more program Buck Rogers what's the packing I can try loading that and we can see super game pack all right that's uh this has to rewind how do you rewind these in the atom I mean there's got to be a way is it going to detect that it's at the end automatically rewind it and then do it again allows let's uh let's start this and while it's doing that I'm gonna check the manual yeah just rewound it actually well at least that's fancy so okay then I don't Bachrach it's working sweet so am i basic cuz that's toast awesome of course it is but hey that means that the cassette drive works with the power supply so I'll take this as a win nonetheless it's disappointing that this is toast but I think there might still be some of these out there so I'll just need to try and track one down all right then actually you know what I'm going to load Buck Rogers one more time and I'm going to have the multimeter set to measure the min and Max of all the voltages it reads that way I can see if there is a voltage spike that is significant on 12 volts when it stops running the motor because it's supposed to have a separate inductive load and I don't know how bad that actually is so let's see well I didn't see anything but it's also possible the spikes happen so quickly that it's just very difficult to see I'll probably just put the flyback diode on there for all of the ones that I ship out but it may or may not be necessary it's really low cost for a peace of mind so it's it's worth it one final test here is I'm getting ready to button this thing up and call it done the power supply working outside of the computer as it would have to be used with an expander module three this is very very finicky this is really I'm gonna unplug this from this end actually this is really crammed up against there you would want the extension the db9 extension that I mentioned to get enough clearance for that you could probably if you bought this as a kid I'm not sure if I'm an offer this is a kit yet it could be possible to extend this out by not putting it all the way flush to the board but then if you used it internally it would be a lot more prone to cracking the solder pads during install so yeah well I think that pretty much wraps up the first revision here of the Coleco Atom internal power supply I'm not going to be shipping out these boards as is I think that the low-side switching was just too bad of an idea I've actually already started working on a second revision that has high side switching and I will be getting those boards produced soon and ordering new parts to replace the end channel FETs that I should not have purchased in the first place but overall I'm very very happy with how this project is turned out I mean I have a functional Coleco atom now and I don't have to have a giant printer hanging off the side the power supply ripple voltages seem pretty reasonable to me I have on the second revision put additional filter caps on the negative 5 volt rail and I tested it on this one and it about halved the ripple voltage on it so that's going to be a big improvement I also change some of the physical location of them on here to a more suitable point so it should be overall much better on the second revision and was already not too bad on this one one more thing the power supply cap that didn't fit should have been correct I was actually right the company that makes these East which has a series called LC and these are a part of it and that switch is a part of it and even on their own website they say it's a part of it but not all parts in that series are compatible so what's the point of having that be a series they need to break it up into different series so anyway um I don't know what I'm going to do to replace those I'll probably will just super glue these on I'd kind of like this switch but I'm also thinking about getting a smaller plunger cap so I don't have to route out around the switch because it's really difficult to route next to the power jack which I also move forward actually on the new board I'll cover the new changes when those get in here I've been working on this video for nearly a month now so this one really needs to wrap up but one more thing before we go I do really want to say thank you to PCB way for sponsoring this video and I genuinely mean that it's really helping out with getting these parts made this was a much much more expensive kit than the ATX 2 PC junior ever was just because this is a real proper circuit design so I'm very grateful for that but that's it for now stay tuned for the update on this which will be much much shorter this video has turned out incredibly long but that's not surprising considering how I showed everything so if you enjoyed this I do more stuff like this if you want to support the channel I am on patreon but for now that's it I'll see you next time

Channel: Tech Tangents

Views: 50,845

Rating: 4.9450498 out of 5

Keywords: AkBKukU, main, computers, EE, Repairs, 1980s, Coleco, Adam, ColecoVision, power, supply, psu, new, replacement, printer

Id: i4nYqfZlnMY

Length: 38min 40sec (2320 seconds)

Published: Sat Nov 16 2019