[SOLVED] problem with 74LS85 comparator

[maark6000]

having a strange problem with a 74LS85 comparator. I'm running a binary clock on the A inputs, 0000 thru 1111, and on the B inputs I just have a fixed number. A=B goes high at the appropriate time... but also 2 before. So if I have my B input number at let's say 13... the A=B output goes high at 11 and 13. if I move the B number to be 14, then A=B goes high at 12 and 14. I tried swapping out the chip, the next chip did the same thing. It's driving me bananas... I've quadruple checked everything. Anybody ever encounter that problem?

Mark

 

[j33pn]

For the specific example you've given with 11 and 13, A1 and A2 or B1 and B2 could be swapped.

Oops, didn't read the whole post. Are you allowing some time for the output to settle?

 

[dick_freebird]

If this "clock" is a ripple counter, there could be enough
timing skew in the different bits' outputs that transient,
invalid codes are presented to the magnitude comparator.
This would not be the case with a synchronous counter.

 

[maark6000]

the "clock" is a 74LS193 counter, which is synchronous. it is clocked by my BK precision signal generator, set to a square wave, about 10 Hz, from the TTL/CMOS output. What's funny about the fact that the A=B output goes high 2 before is it isn't a transient... it's not some little spike, it goes high for the entire time the clock is outputting that number. I'm going to isolate the clock and the comparator tonight and see if somehow other ICs in the circuit are causing the strange behavior (other ICs being a 7485, an OR gate, and a hex inverter).

 

[dick_freebird]

Like maybe LSB+1 is stuck? Then you'd have two "valid"
compare-codes.

 

[maark6000]

well, i'll never know what I was doing wrong. tonight I ripped out all my jumpers and started over. I did do something I think was smart... I set up my clock and my comparator, and applied power to make sure it was working without any other parts of the schematic attached. I had the A>B, A=B, and A<B outputs lighting up LEDs. Then I kept the power on while I continued to connect up the rest of the ICs. That way, if something squirrelly started happening, I'd know exactly the moment it happened and with which connection. Don't know if that would have been a smart move with CMOS chips, but with rugged ol' TTL, it worked great. Circuit runs as designed. Thank you to all for your help. Cheers!