octal latch circuit problem

[Bruce Ratcliffe]

I have trouble with a quiz show circuit (see http://kitsnspares.com/admin/pdffiles/SchoolCollege Quiz Buzzer.pdf). After spending time trying to get it working, I discovered that resistors R12-R19, which limit current to the SCR gate, are too big. 1000 ohms (in the article) does not work, 47 ohms (my trial and error value) does work. So I"m suspicious of the rest of the circuit! Nonetheless, I built it on a breadboard, works fine. So I designed and etched a PCB which worked fine too--for awhile. First, switches 1-4 failed to activate the light, then the whole thing stopped working. I'm wondering if I should build in some protection (caps or resistors?) to keep the IC (74LS373, octal latch) from getting damaged.

Thank you for any advice,
Bruce Ratcliffe
Edison High School,
Fresno, CA

 

[(*steve*)]

The first question is... Are you using sensitive gate SCRs?

Further questions include.. What load are you driving, and what is you power source (and how well regulated is it) -- or is it the one shown in the PDF?

What voltage do you measure across the supply rails?

Did the fault start with one channel and spread to the first 4, or did all of them fail at once (was it 2 failure events, or 8, perhaps in 2 groups)?

Are you using the LS part, or did you select another part (say 74HC373, etc)?

 

[Bruce Ratcliffe]

Octal latch problems

Dear Steve,
FIRST of all, WOW!--thanks for responding so quickly! I'll answer your questions in CAPITALS:
The first question is... Are you using sensitive gate SCRs?
I'M USING T106 SCR's. I WILL HAVE TO CHECK AT SCHOOL TOMORROW WHAT THE LETTERS ARE AFTER THE "106..."

Further questions include.. What load are you driving, and what is you power source (and how well regulated is it) -- or is it the one shown in the PDF?
THE LOAD IS A 6.V VOLT, 0.15 A INCANDESCENT BULB
THE POWER SUPPLY IS DEFINITELY NOT THE ONE IN THE PDF! THAT'S A LOT OF WORK! I'M USING AN APPLE ADAPTOR (WITH ELECTRONICS INSIDE, NOT JUST A TRANSFORMER 4 DIODES AND A CAPACITOR) I'LL HOOK IT UP TO A SCOPE TOMORROW AND SEE WHAT KIND OF RIPPLE IT HAS (THOUGH I'M EXPECTING, BEING APPLE, IT IS PRETTY SMOOTH...)
What voltage do you measure across the supply rails?
THE NO LOAD VOLTAGE IS 5.0, WHEN I PRESS A BUTTON, I'M NOT SURE--I'LL MEASURE IT TOMORROW IN MY CLASSROOM AND GET BACK TO YOU.
Did the fault start with one channel and spread to the first 4, or did all of them fail at once (was it 2 failure events, or 8, perhaps in 2 groups)?
THIS WILL BE A BIT LENGTHY REPLY: THE BREADBOARDED VERSION, AFTER CONSIDERABLE FAILURES, WORKS PERFECTLY--ALL 8 BUTTONS. THE SELF-DESIGNED AND ETCHED PCB (USING A REALLY OLD (READ: DOS) PROGRAM, EASY TRAX, THAT I DOUBT YOU HAVE THE FILE TO OPEN) ALSO WORKED FINE, ALL 8 BUTTONS, AT FIRST. THEN, AN HOUR LATER, HAVING DONE NOTHING THAT SHOULD HAVE DAMAGED IT, BUTTONS 1-4 STOPPED HAVING THE LIGHT GO ON WHEN BUTTON PUSHED. I TRIED TAKING THE 373 IC FROM THE BREADBOARD AND INSTALLING IT IN THE PCB VERSION, BUT IT STILL DOESN'T WORK.

Are you using the LS part, or did you select another part (say 74HC373, etc)? I SURE THINK I'M USING THE LS PART, BUT, AGAIN I APOLOGIZE, I WON'T KNOW UNTIL TOMORROW MORNING WHEN I GET TO SCHOOL. I'LL GIVE A MORE COMPLETE REPORT THEN.
BRUCE

 

[TedA]

Bruce,

I realize that you are already in for a pound, but I have to say that this design is fairly crazy.

The function can be realized with about half the parts, using no fragile ICs at all.

Since the SCRs provide the latching function, there is no need for a separate octal latch IC.

A couple of transistors, biased through a current sensing resistor, can turn on the beeper, while turning off the ( common ) power to the individual buttons.

No IC, no diodes, no regulated DC supply required.

Now, if you actually needed to use a logic IC, you would be well advised to protect each I/O pin that leaves the circuit board. Even 74LS parts are subject to static damage. CMOS ones, possibly more so. And you don't want the thing going crazy every time the water cooler compressor cycles.

TYN604 is an ST part. Data sheet says it may require 15mA and 1.5V on the gate to fire it. So the 74LS outputs may very well not be stout enough, no matter how low the gate resistor value.

The practical answer for getting something working might be to bridge across each of the 8 latches in the IC, and tack solder a transistor circuit to take care of the remaining logic.

Each switch can drive the SCR gate through the existing resistor. ( You might want to add a small cap from gate to cathode to suppress firing on transients and RF.)

Toss the voltage regulator. In its place connect a resistor that gives the desired lamp voltage. The voltage across this resistor can be used to turn-on a PNP beeper driver. The output from this transistor can then control a second transistor that removes power from the common line to the buttons, as soon as the beeper begins to beep.

On second thought, you could put the current sensing resistor between the negative supply and the SCR cathodes, and use a single NPN transistor to drive the beeper and kill the power to the buttons.

If you want to be very careful, you might put a bit of resistance in series with the common line going out to the buttons, along with a couple of diodes on the transistor collector to limit the voltage swing to the power rails.

Ted

 

[(*steve*)]

TedA gives very good advice,

I was going to suggest a similar approach (without the IC), but I think it's worth finding out why it doesn't work (or why it failed) first.

If your power supply is outside of 4.75 to 5.25 volts then the LS TTL chip may get upset. I don't think they fail until something like 7V though (not something I've tried and I refer to this from a vague memory of the absolute max voltage on the specs).

Note the link above. Reading the specs is always helpful.

TTL (including LS TTL) is pretty poor at sourcing current (better at sinking). TedA point out this, and the specs confirm that you can expect 2.6mA from the output when high (vs. 24mA when low). Note that you can get a little more especially because you're not concerned with the output voltage being valid.

The main problem "solved" by the 373 is the triggering of multiple outputs, but this problem can be solved in other simpler ways (again, as suggested by TedA).

Here is a really cute approach.

 

[Bruce Ratcliffe]

Dear Ted Octal latch problems

Hi, Ted,
Thanks for your detailed advise. I smiled at your opening sentence ("...into it for a pound"), and want to praise your prescience. I AM into it for probably a kilogram at this point, having designed, exposed, etched, drilled and populated the board all that mess goes on. Between that fact, and the fact that it DID WORK PERFECTLY at first, I'm inclined, for the moment, to troubleshoot my circuit before charging off in a new direction. (I've already done that for four earlier didn't-quite-work quiz show circuits, over a span of 20 years...)
Bruce

 

[Bruce Ratcliffe]

Octal latch problems

Dear Steve,

AS I said earlier, I'll get back to you when I get to my classroom tomorrow, but I must say that the circuit you directed me to at the end of your letter (just to show you how long I've been on the quest for the holy grail of a quiz show device), is one I've already built! It worked great, and was VERY simple. Its weakness did not become apparent, until, to my horror, during the contest in my science class, I noticed that sometimes two teams' lights would both be on. I think something (stray capacitance?) kept the voltage from falling fast enough after the first button was pressed, so volts was high enough for a second button to register, and turn on its light. Luckily, the kids didn't think it was anything important. I knew differently!
Bruce

 

[Bruce Ratcliffe]

Update on Quiz show

Dear Steve,

Sorry to take so long to get back to you--it has been one of those days^3. I can finally give complete answers to all your questions from yesterday.
1) YES, i am using a sensitive gate SCR, the T106B1
2) The power source is that Apple device I mentioned, Specifically, from the label:
IMATION
Switch Power Supply P/N 52/0000/7302/4
Model DSA-S15-05 A
Input 100-120 V~
output +5V DC 1.5 A
Level 3
It seems well-regulated. I tested volts, got 5.13, which never wavered even 0.01 volts when I pressed the button which activated the piezo buzzer, though the light doesn't come on.

The supply rails volts are 5.13
YES, I did get the actual LS part, as in SN74LS373A, not anything else.

Hope this helps.

Bruce

 

[Bruce Ratcliffe]

quiz show update

News flash! The output from the 373 going to the 47 ohm resistor that feed the SCR gate is a constant .08 V, both before button is pressed and after (when the piezo is merrily buzzing.)

 

[(*steve*)]

Disconnect it (the output) from the SCR (and everything else) and measure it again. At least one output must be high if the buzzer is sounding

I suspect it's not happy though.

You know, I'd be tempted to try replacing the 74LS373 with a 74HC373. The output current from the HC version is significantly greater.

My reluctance springs from 2 sources. Firstly the HC variant is static sensitive (not that it will matter once it is connected). The second is that I don't think your circuit should have killed the LS chip.

 

[Bruce Ratcliffe]

Quiz Show

Dear Steve,

I don't think the LS version is being damaged by my PCB because when I take the 373 off that and put it on the breadboarded version, it works like a champ. I'll try disconnecting the output from the 373 going to one of the SCR's (by unsoldering the 47 ohm resistor at one end) and test what voltage comes out then. I'm still stymied because, initially, the PCB worked flawlessly. (?)
Bruce