Page 1 of 1

crowbar circuit

Posted: 20 Nov 2015, 21:01
by Fonotec
As explained in the thread, Power Supply Trick, the LS-76 power supply creates several regulated DC voltages used by THE LATHE, including the +5 Volt section, for the logic chips (e.g., NAND gates). The +5 Volt section makes +5.6 Volts, when the power supply is unloaded, due to there being a silicon diode between the regulator's ground pin and Earth, but the +5.6 Volts gets loaded down to the safe target of +5.0 when the card cage snake is plugged into the PSU, due to a diode drop effect caused by the nearly 70 logic chips running off the same supply. None of the chips can take more than a maximum of +5.5 Volts (provided they are military spec), and most of them, being civilian spec, can take only as much as +5.25 Volts - and no less than +4.75 V - or else they won't work correctly, or could be damaged, if volts go too high.

The PSU has a a glass-housed fuse rated at 4 Amperes, which will, of course, blow if too much current is drawn. Under normal operation, the +5 Volt section draws a little over 3 Amperes, so, it's a good fuse value for protection. However, one of the common ways in which a 780x voltage regulator fails is to deliver an elevated output (voltage). In some scenarios, the current could still be under 4 Amps, even if the regulator output swings high. So, in order to protect the logic chips, in the card cage and in the control panel (e.g., the 7-segment LED drivers for the digital LPI readout), a good thing to add is a crowbar protection circuit. This circuit constantly measures the voltage and, as soon as it goes above the safe operating level for the IC components, determined by a (reversed) Zener diode, it turns on a silicon-controlled rectifier (SCR) that suddenly offers the output of the fuse a path of very low resistance to Earth, parallel to the logic circuit, causing the 4 Amp fuse to blow, since the 1 Amp 7805 regulator is being boosted with a parallel-connected, 7 Amp, PNP transistor (2N 5871).

For now, I've selected a Zener diode with a reverse breakdown voltage of 5.6, so, it will have a good chance of protecting the chips in case the 7805 swings its output high, but there is a little buffer between its turn-on point and the highest sustained voltage a civilian chips's Vcc pin can handle to prevent the unnecessary blowing of fuses in the event of any mild, and brief, surges at power-on.

The schematic, below, was made by slightly modifying another one, which was found online, that was drawn, originally, by someone named, Ravi.
Crowbar_circuit_LS-76_TTL.JPG
Crowbar_circuit_LS-76_TTL.JPG (33.26 KiB) Viewed 9550 times


Here's the surge tolerance of the selected Z5.6 diode.
1N5919BG_Surge_Tolerance.jpg
1N5919BG_Surge_Tolerance.jpg (97.52 KiB) Viewed 9550 times