On Monday, March 3, 2014 4:31:25 PM UTC-6, Fabio Battaglia wrote:
I'm left with only one piece to install: the Q1 PNP transistor, which is leaving me some doubts.
Heres some data from my working boards.
For the PNP / NPN complementary pair I used 2N3906 and 2N3904 unlike the schematic. My magical component analyzer claimed betas around 300 for these guys which implies either the analyzer is broke or my memory is faulty as 300 sounds a bit higher than I recall decades ago. I recall "a hundred something" was about right for these guys in the old days. Hmm.
From memory the two families are about the same other than the 3904 family handles only about 2/3 the voltage and 2/3 the current of the 2N2222. Doesn't really matter because they're both overspec'd by a factor of ten or so, like 3/4 amp and 75 volts or something from memory. So either will handle 5 volts and a zillionth of an amp all right.
Anyway. For NPN 2N3904 the silkscreen is correct on my V 02 PCB and for a PNP 2N3906 you need to flip opposite the silkscreen. Flat side of the PNP facing the LEDs so to speak. Donno if the 2907 has the same pinout.
I used a 74ALS04 and 74LS123. "Wasteful" but I have a lot of AS and ALS laying around, a nice problem to have, so no point ordering a 74LS04. This probably has a minor effect on currents and voltages required thus resistance.
I used a 390 ohm NPN emitter resistor as per schematic and to get proper operation with my substitutions the PNP collector resistor aka R14 had to be adjusted to 3300 ohms. I fooled around at the bench sticking in various resistance in to get the behavior I needed. I don't remember exactly how I figured it, I remember guessing that as long as I ran above 1K I physically couldn't overcurrent the transistors and 22K would be ridiculously over so I did something like binary search until it worked.
I caught that previous post about using a screw mounted binding post in the ground and test and managed to purchase some that were just barely too large to fit next to the S100 connector. No worries I stuck another nut under to elevate the knurled part above the S100 connector so as to fit. Very tight fit indeed.
Instead of 180 ohm dropping resistors for the indicators I used 220, seems bright enough for me.
For "power bus OK" LEDs I used green (different colors have slightly different drops) and 15K instead of 20K for dropping the +/- 16 volts and 3900 instead of 4700 for the 8 volt. Your personal tastes in LED brightness may vary depending on the LEDs purchased etc. Again these values were "fool around on bench trying values till I liked the brightness". It was kind of a game to get the LED brightness on the 16 and 8 volt supplies identical brightness; I ended up pretty close.
I would imagine brightness perception depends entirely on the exact make and model of LED.
For V3 I'd suggest:
1) Move the logic probe input juuuust a 1/32 inch down, maybe even just a 1/64 downward so more binding posts fit
2) Change the collector resistor into a 820 ohm or 1K or so, with a 10 K pot in series and dial it in to work perfectly.
3) Label the pins on the transistors silk screen so if you substitute in something weird its obvious its "EBC" or "CBE" pinout or whatever.
4) Add binding post mounting hole(s) for +5V
5) Silk screen labeling of the S100 pinout. So right on the silk screen under pin 25 you see that S100 pin 25 is labeled as ...
6) Consider sticking current shunt resistors in line with power leads so you can measure current by sticking voltmeter across shunts? Might be handy.
7) Could you throw a LED on the reset line? Plenty of space on the PCB on that side. Some would call it "Bling" I guess.
8) DIP switch for config instead of jumpers? This is more of a personal taste issue than a "bug"
These are very minor suggestions, because the volunteers who developed this board did a pretty awesome job. Thanks!