Dear Fans -
My voltage increasing mods for all four candidate "Mean Well" switch mode power supplies ("SMPS") to power the unmodified Apple-1 announced in the old thread:
. . . are now completely developed and tested, and ready to be deployed in the field:
From left to right: "Mean Well" types RT-50B, RT-65B, PT-45B and PT-65B. Metal enclosures of RT series were removed for this photo. Replaced electrolytic capacitors are marked with yellow marker on top.
DIFFERENCES BETWEEN THE CHOSEN SMPS TYPES
RT types come in a perforated metal enclosure which was removed for the photo, but will be put back to restore electrical safety. This metal enclosure is grounded and prevents curious fingers from touching the "hot" sections of the circuit which carry the line voltage.
PT types are "open frame" ones with no such protection against electric shock. In the above photo, you can see the (smaller) heat sink of the input switching MOSFET lurking and if you touch it when the SMPS is powered up, it metes out an electrical shock. (Don't ask me how I know that ;-)
With the lowly 110 Vac in North America where I live it does not "bite" as badly as the 220 Vac use elsewhere in the world, but still, even 110 Vac could kill those who are faint of heart or less robust then me. So these "open frame" SMPS must be put into an insulating enclosure which still allows some cooling airflow. Such an enclosure could be home built. Or you could mount the SMPS in the wooden enclosure of the Apple-1, and add an interlock switch which disconnects the line voltage if the enclosure is opened. This is how the industry does it.
CHOICES AND DIFFERENCES
My personal favorite is the RT-50B because it's the smallest of the bunch, really cute, and it still can deliver much more current than any Apple-1 system should ever draw.
The PT-45B (which is built on the same PCB as the PT-65B but has lower height components) is the weakest and cheapest of the bunch but still can do 3 Amps at V1 (makes the +5V) and 2 Amps on V2 (makes the +12V). Even this is enough for a full blown Apple-1 system with 48K memory expansion, my color graphics card, and my Disk II compatible floppy disk controller card plus two typical 5.25" floppy disk drives. It could also power an additional small B&W monitor. But probably not a color monitor except for the really small ones. Or LCD ones.
The RT-65B has the same power and voltage ratings as the PT-65B, but has the metal enclosure. It can source 5.5 Amps at 5V which no Apple-1 system ever should use, as its LM323K is only rated for 3A. Various failure modes where the Apple-1 could suck more than the 3A its input rectifier diodes are rated for are possible. So these would die and typically turn into short circuits - a potentially catastrophic failure in case of transformer feed (the big blue Sprague input filter capacitor would explode). With SMPS feed, no issue there, and no damage other than the shorted diodes (and the part which drew the excess current). So you can see that using a stronger than necessary SMPS brings its own perils with it. The weakest one which can handle all the expected loads is the better choice here.
But in principle, any of these power supplies is fit for the job.
NO MODS TO THE APPLE-1 !
After having been modified, of course. Unlike with the "direct feed" method, no modifications to the Apple-1 are done, so the usual +5V, +12V and -12V output voltages of these SMPS do not have enough voltage headroom for the linear regulators on the Apple-1 itself. The mod rises the output voltages to +8.4V, +20V, and -15.6V, and this gives all regulators enough headroom. The LM323K "sees" +7.5V in this case. The drop depends on the particular type of the rectifier diodes in the Apple-1. Experimenters could turn down that voltage to 7.0V to make the whole thing run even cooler. But it runs cool enough with +7.5V . . . the typical Apple-1 witn no expansion cards draws about 1.5 Amperes on its +5V rail, so when fed with +7.5V, the LM323K only turns 3.75 Watts into heat. Compare that to the transformer feed where it sees ~10V. It's twice the heat (7 Watts) and so the heat sink gets much hotter.
ADVANTAGES OF SMPS FOR APPLE-1
Using SMPS for the Apple-1 has many advantages over use of transformers, other than keeping the heat down and saving energy (which is the same thing). These modern SMPS work all over the world, and they are inherently safe as they don't start to fry and give smoke signals like the transformers may do when overloaded. Making transformer based power supplies safe against overcurrent / overload is a very tricky task, especially when multiple transformers of different power ratings are involved. Split secondary windings complicate the task even more. Still, most Apple-1 builders I had contact with use the circuit seen in the Apple-1 user manual with the single 0.5A fuse, or with no fuse at all. The latter approach (no fuse) appears to be reckless, but analysis shows that this fuse can't protect these transformers from getting fried under all possible failure modes. So the fuse is almost useless anyways. And from a pragmatic standpoint, if the Apple-1 with such an unsafe transformer power supply may be run safely if under constant adult supervision, no exceptions. Which is advisable anyways - the more "vintage" the build is and the more 40-50 year old electronic components are in it, the more likely is some failure of such an old component. But it's the same story and the same warning message I wrote about several times over the past few years. I just repeat it here for newbies who just jumped into the Apple-1 scene.
NO "DIRECT FEED" RISKS TO BE TAKEN ANYMORE
This new method to run the Apple-1 from SMPS avoids the risks of the "direct feed" method which bypasses the rectifier diodes and regulators on the Apple-1 motherboard, and opens the opportunity to destroy this SMPS modified Apple-1 by then inadvertently still plugging in a transformer based power supply.
The new method is even safe against plugging in the power supply cable turned 180 degrees - but unless the middle two MR500 rectifier diodes were removed from the motherboard, these would die. You see, I did my FMEA. And so I saw the opportunity to make the Apple-1 power cable plug-in process virtually foolproof by removing these two diodes, which, of course, is optional for non-fools. Still, it could be a nice to have additional safety feature if younger kids are in the house. These are known to be curious and to experiment with any new thing they might find ("open frame" SMPS without insulating enclosure are a no-no for such households, too, for obvious reasons).
OVERVIEW OF THE MODS
- the voltage regulator circuit is manipulated to allow an adjustment up to the desired +8.4 Volts. On the PT series, this is done by just cutting a PCB trace. With the RT series, no such cut, but a small SMD resistor is soldered over another one.
- the electrolytic capacitors which now will see higher voltages possibly exceeding their rated voltage are replaced with suitable types of a higher voltage rating. These "suitable types" must be specified for use in output filters of SMPS. Normal electrolytics can't be used, they might overheat and explode (opening their pressure relief valve, still can be violent). In the PT series, one electrolytic capacitor must be replaced. In the RT series, the same one and another one must be replaced (two in total). In RT-65B these replacements are of identical type. RT-50B needs one of this type but the 2nd replacement is a different type for space reasons (to fit into the crowded space).
- the overvoltage protection circuit is modified for the higher output voltage on the V1 output (change of one Zener diode) and an additional protection for the V2 output is added (adding another Zener and a small signal diode).
- the 7912 regulator is nudged into making -15.6 Volts on the V3 output rail. This is done by adding yet another Zener, and a trace cut is needed.
- the minimum load resistors sized by the "Mean Well" designers are removed, as with the increased output voltages they would bake the nearby electrolytic capacitors even worse than in the original, unmodified examples, which are bad enough already in terms of that heat. The are replaced by "bleeding" resistors of much higher resistance, which is nice to have to safely discharge the various electrolytics in case the Apple-1 is not plugged in. This measure costs only ~6 US cents.
- absolutely no mods to the "hot" section of the SMPS which carries the line voltage. All mods are done to the low voltage side on the other side of the insulation barrier.
TOTAL COMPONENT COSTS FOR THE MOD
Total costs for components (other than the SMPS itself) is between ~$2.50 (PT Series) and ~$5.00 (RT series). The diodes and resistors are penny items. The special SMPS rated electrolytics are the major cost factors.
Adding the mod to the PT series (the "open frame" ones) is ridicolously easy as it can be done without adding or removing any SMD components.
The mods to the RT series are more difficult as it involves desoldering and then adding a few SMD components. I do know that some Apple-1 builders were frightened (or overwhelmed) by the 0805 size SMD components in the Gen2 improved ACI. So modding the RT series is not for beginners.
The RT-50B (my favorite) alas is the worst difficulty, it has ~8 SMD resistors to be removed, and ~6 SMD components to add. And it has a lot of RTV goop squeezed between the output filter components which must be carefully cut out with a knife and then removed piece by piece. This is tedious and time consuming.
TIME NEEDED TO DO THE MOD
It takes me about 1 hour to mod and test a RT-50B and put it together again with its metal shell. To compare, the "open frame" PT-65B mod can be done in ~ 15...20 minutes and SMD components can be avoided entierly (although I did use a SMD Zener for the 7912 voltage boost, because I had them).
CALL FOR BETA TESTERS
These mods have been developed and thoroughly tested in my lab and they are ready to be unleashed on the world. I have no doubt that these mods are reasonably safe if done properly. "Reasonably safe" means that the modded SMPS are not any less safe than the original item. The enhanced overvoltage protection circuit I have developed is a major factor for that, and although it may be tempting to leave the added components out (and hereby save 15 cents), it would be foolish to be so cheap.
The sole remaining problem (and risk) is that my "mod manuals" (which are not yet written) may be faulty and / or confusing or not precise enough. So these must be beta tested, preferably by electronics professionals, who could spot stupid mistakes more easily.
BETA TESTS ABSOLUTELY NECESSARY
Without successful beta tests, I won't release any modding instructions, sorry. So this is where YOU, dear reader, may come in. If you live in the USA or in Canada and if you are an electronics professional, and want to volunteer as a beta tester for this great mod, please use the "send PM" button and tell me. You are welcome to post in this thread, too, of course.
I can't give you my contact information in an open forum. Did that once (early days of the WWW) and had to abandon my nice email address due to a flood of spam emails which stole my time. I don't want that to happen again. You sure understand.
ANY QUESTIONS / CONCERNS ?
Please just post them in this new thread.
- Uncle Bernie
P.S.: Beta testers of course will get my fully support to make their mods work. All I'm interested in is to assess whether my instructions are fit to be unleashed on the broader public.