Unfortunately my first 80486 prototype is gone out for fabrication David. It uses a 78H05, but a very good point, the next one will have a jumpered 3.45V supply. BTW, I assume that is just to the CPU itself. Is there a special need for dual voltage buffers around the chip or for example 1K resistors on input lines and the likes.
From: n8vem...@googlegroups.com [mailto:n8vem...@googlegroups.com] On Behalf Of David Fry
I've got some 1uf Tants here so I'll use them in positions C51 & C74 taking care to observe polarity etc...
wrt the LM323k on the 80386 board I managed to get a genuine LM323K device from Farnell UK for UK £20 (ouch!!) though stock numbers are dwindling even at that price.
I did try the ebay route and paid UK £6 for one and surprise, surprise it turned out to be fake (Ov output), I did get a full refund when I challenged the seller.
My genuine LM323K from Farnell outputs 4.95V under no load test which is fine by me.
The LM2576 switching regulator that you mention I have not used myself but would be perfect for the 80486 board and beyond, using the adjustable version will give the flexibility to cater for various different 486 core voltages using a jumpered approach like on a true 80486 motherboard, whilst minimising heat build up.
If you remember, some time ago I mentioned the AMD 5x86 P75 chip which was the fastest 80486-like processor produced and fully 486DX2 compatible, this one runs on a 3.45v supply at a supply current of 950mA approx (at 133Mhz, 4 x 33), based on conventional linear regulator technology you would already need to dissipate 4.3W at the regulator before you start adding the surrounding support chips (8v - 3.45v * 0.95A), a switcher regulator is the way to go !!.
Out of interest what is your initial target 486 chip version ?? (DX2-66 or otherwise)
The 32MB SMD on board SRAM board looks neat but would be some way down the line for me, way too many things to get sorted before then, but I appreciate the insight on what is coming down the line from your workshop, very interesting stuff :-)
On Friday, October 31, 2014 9:37:28 PM UTC, David Fry wrote: