Hi
I am in the process of building a replica of the Apple I, and am keen to understand what options exist for building the power supply.
I'd prefer not to build the original design, given I'm not particularly confident working with (high) line voltages. I'm also in Australia where we have 240v, so would need to modify the design anyway. I have also seen references to the 'direct feed' method in the forums, but have read that there are some issues with this approach.
From what I can tell, the best option is to use a switching power supply which feeds DC voltages direct to the main board, however there is no single power supply that is commercially available that meets the voltage requirements.
Given this, I have three questions that I would love some help with:
1) Is Uncle Bernie's modified version of the Mean Well PT-65B still the only single power supply solution?
2) If I went for a solution with multiple power supplies, what are the recommended DC voltages for each of the jumper pins?
3) Has anyone else used the multiple power supply solution and, if so, what specific power supplies (e.g., manufacturer and model #) have you used?
Thanks in advance!
Nick
Have you considered building the original Apple I linear power supply by rewinding the secondary coil of some existing transformers from another electrical device sold in Australia?
When I was a kid in the 80s for a period I was into building linear power supplies by rewinding existing transformers. I even had a Soviet-made winding machine almost identical to this one:
TransformerMachine.png
I am surprised that so few people are into doing this nowadays, considering how easy it is thanks to programs like Transformer Calculation:
transformer-ie-core-square-dimension.gif
1. "Direct Feed" still works, but it requires careful choice of the regulators which it bypasses, as some types of regulators do not like reverse bias conditions (output voltage larger than input voltage) and may get damaged if this happens. I found that most will "die" gracefully and not start to get hot and smoke. I provided my kits with carefully selected regulators which can sustain the reverse bias condition. Because it's almost impossible for "maverick" builders not using my kits to get these regulators, and some of those had bad luck (and dead regulators afterwards), I do not recommend "direct feed" anymore. NOTE: the reverse bias problem is known to exist in almost all of these early linear regulators of the LM323K, 78xx and 79xx families and it can happen even in a "normal" power supply circuit, depending on the size of the capacitors on the input and output side. This is why most later datasheets of these regulators showed an added protection diode between the input and output pins.
2. You can still buy suitable off the shelf transformers of current manufacture ... like the Triad VPP16-1900 and VPP28-1060. When you run these at 240 Vac your regulators on the motherboard will run a little bit hotter than normal, but it's not dramatic. You can try to find similar transformers with a slightly lower output voltage but if you go down too much, the regulators will drop out. Note that in many parts of Europe, they run the grid also at higher voltages than the traditional 220V to be able to pump more power through the grid without investing money into the infrastructure.
3. For the coolest possible motherboard regulators, use the PT-65B mod. But be aware that "Mean Well" has recently revised the PT-65 family to a new PCB using more SMD components. This means that unless you are a real electronics professional you will run into difficulties as I did not upgrade my modding instructions, and I'm not going to buy one of these newer PT-65 because after the extreme tariffs on Chinese goods and the rip-off by added "tariff handling fees" from the carrier one of them might end up costing $50 or more. Note that the PT-65C is easier to modify as it already has the correct +/- 15V conforming electrolytic capacitors and regulators on the output side. But you must disconnect a pulse transformer winding which "rides" on the rail for +5V and connect it to the GND/0V rail. This lowers the voltages for the +/-15V side. Only then you can crank up the +5V rail to +8V safely. I did not "invent" this alternate mod, it has been tried by another Applefritter member and it works.
4. Rewinding of transformers (as suggested by 'CVT' in post #2) often is impossible as the windings of modern transformers often are glued together / immersed in some expoxy resin and the core sheets are welded together. Which makes the transformer unrepairable and unmodifyable.
- Uncle Bernie
Thanks CVT, pulling apart an existing transformer goes beyond what I am comfortable doing, so will have to give this a miss
Thanks Uncle Bernie,
I think the PT-65C based solution makes most sense. Is there any reason I couldn't add a step-up converter to the +5V output to get the +8V that I need? This would obviously need to be external to the PT-65C itself, but that would be ok for me.
A DC-DC converter from +5 to +8 with sufficient current capability will cost a lot more (almost the cost of a PT-65C), add more conversion losses, and potential stability problems.
This is the only boost converter available from DigiKey that has the right specs: https://www.universal-solder.ca/product/automatic-dc-dc-buck-boost-sepic-converter-10a-1-30v/
Including DigiKey's markup, it costs $20.89, which is more than the Mean-Well PT-65C, $18.20.
It can sometimes make sense to use a downstream regulator to supply a missing voltage rail, like -5 V. The situation here is very different