Getting a double clocking error w/at least ls273 and 164

[lacy]

Whoops. I made a wiring mistake in the last circuit. I have corrected it and now it is the way the orinial circuit was configured minus R4.

Here it is once more:

[Euler's_Identity]

Ok, you removed the double clock with that one.  (I also made the circuit "legal" by pulling up the input of the inverter and still it worked.)

But you've now shown me that the ls devices may have a problem, as, even though the standard TTL device did not show a glitch at its output, I did still get a glitch at the Qa output of the shift register.  Why?  Plus, the double hit returns when I replace the standard TTL with the ls.  This is weird.

Note:  I still have a blink in my Qa output red probe, which clued me to check the scope.   Sure enough the mystery glitch was still there.



 

[Euler's_Identity]

Yep, that's it.  It's the ls TTL.  I replaced the 74ls164N with a 74164N and there's no blink any more and output glitch is gone.

[sigh ...I spent so long working on this glitch...]

[Euler's_Identity]

I use Multisim like a breadboard, to prove concepts will work as planned (the best layed plans of mice and men anyone?)  I'm still hoping that someone's going to come along and say, 'EI, nope you're wrong this time.'  Because this is Multisim 10.1, and here I've got old ls TTL chips producing a glitch.  Please show me that this is just the result of my screwing up a simulation setting.

I did use the simulation fixer the other day for the first time, decided to see what it could do.  Perhaps, in "fixing" one of my blunders, it hosed up my settings in some way?  It would make sense that, after it's gone through its "fixing" changing things, it'd leave the changes intact lest the simulator burp again.  Hence, whatever was left over could be hosing me up in some way.

Anyone know how to reset to default settings?

 

[lacy]

Simulate>Interactive Simulation Settings Top right corner "Reset to Default"  Under the same dialog box goto Analysis Options and select Use Multisim Defaults.

This shoould get you back to all default simulation settings. I am going to try your LS chip theory with your schematics and see if I get the same results since I am already using the Multisim defaults for my simulations.

I will let you know if I see the same thing or not. If we both have the same experience then we can positively conclude your hypothesis about the LS components. I will get back to you with my results.

If you want me to tell you are screwing up I can definitly do that if it makes you feel better (LOL):smileyvery-happy:

Message Edited by lacy on 07-28-2008 06:24 PM

[Euler's_Identity]

I reset to default simulation settings (but digital setting is still "real".)  The problem is still the ls devices.

Due to the ls device for the shift register, you get a blink.  Due to the ls device for the inverter driving the clock input, you get Qa and Qb set from a double clock hit.  With both you get a blink and Qa and Qb set.

Remove one, the other, or both, and each shows its effect.  All evidence thus far says it's an ls simulation problem, that Multisim creates a glitch that otherwise shouldn't exist.

[lacy]

I will have to confirm your findings. I tried exactly what you tired and had exactly the same results. Everything works fine with the 74STD chips, but goes back to exihibiting the strange behaviour we havem both observed.

With this confirmed by 2 independent test, NI will have to further address this for us. Maybe they can let us know what it is about these components that are causing them to function incorrectly. Maybe they could provide us with a fix for this or a temporary workaround.

I don't see that there is anything that we can do on our end to fix this.

[Euler's_Identity]

Thanks for your help Lacy.

A possible interim alternative for ls is the 74hc series coupled with the 40, which is how my design has gone now.

As a bit of good news, and not intended to let NI(MSM) off the ls hook, through working through the bug I found a better and more elegant way via the 74hc and 40 series to do what I wanted to do. 😉

However, Lacy, it was good to see that you're still out there helping we the wayward travellers in Multisim land.  ...we haven't spoken in a while, and as busy as I am, I rarely get a chance to come to this forum aside from resolving an issue I'm having.

Changing the subject for a moment...

I'm finding myself having to possibly go into uncharted territory (for myself) in Multisim with this latest project.  In the past, using the largest workspace, I've always stayed on the same workspace and connected wires where they went.   However, I'm now finding myself on the verge of multiple linked workspaces.   I'm guessing this is done via Multisim+busses??   That said, assuming you have the time and desire, could you upload an example of a schematic that uses multiple workspaces for me, where, say, a batch of analog and digital lines leaves one workspace's schematic and mates to another schematic, just to show me (and others?) how it's done?  Also, (changing the subject again) what would you recommend for mixing both long and short timing?  What I mean is say you have one section whose timing best relates with microseconds while another is based in seconds.  What would you recommend for observing the simulation of their interaction?  You see I have a very slow (relatively speaking) sequencer that activates very fast measurement circuits.  The two must work together.  How would you recommend that be done?  For example, I've noticed that RC-based timing isn't accurate in Multisim -- I expected the sequencer oscillator to run much slower than it actually did, while my measurement op amps screamed as was expected.  How do I synchronize the two very different timings?  ...yeah, I know.  I don't ask much do I? 😉
     

 
 

[lacy]

I will post you something tomorrow on the Multi-page. It really is not that hard once you get accustomed to it. You just have to remeber to name your connectors to something that makes it easy to troubleshoot later.

On the timing issue. I am not sure if you are familiar with the Analysis options under the Simulatate tab. They would be AC Sweep, Parameter Sweep, Transient Analysis, etc. This sounds to me to be something that one of these can handle. Since you have 2 different timings here the time-domain (simulator run switch) analysis would possibly introduce errors into the results. I don't know that it would, but is possible.

These Anaylsis take it out of the time-domain and posts the results either in tables or in graphs simular to the O-scope. It is actually much faster (almost instant in most cases).  The thing is that you would have to go in and manually set your parameters in order for them to display a result. This could be a Node Voltage, Current, or even a mathmatical expression. You would have to examine these to see which one may be sutiable for your particular application. I believe that this would posssibly take care of the dual timing problem you describe, but nothing is ever guaranteed. Also, I am not as versed as I need to be on all these Analysis. I have used a few of them, but I don't use them on a regualr basis so you might have to consult the manual or some of the tutorials  on the NI site in order to get a good grip on them.

I hope tyhis helps and I will be back tomorrow on the Multi-page.

[Euler's_Identity]

Yeah, you're right.  As a truthful plug, I just get so used to using the scope with my circuits that, frankly, I largely don't even consider the analysis going on behind the scenes.  That's what I like about Multisim: it feels like a breadboard and a bench.  But yep, you're right.  Sometimes the number crunching is the only way to go, namely when there is a large timing difference such as in this case.  A standard transient analysis around the time intervals of interest should work just fine.

Thinking back many years ago, I remember my introduction to Spice analysis.  I thought it was the most amazing tool.  One just described the nodes of a circuit, set up the time interval and source, and then walla!  Out came tables and graphs from those tables.  Then came the GUI years later.  Now you could draw a schematic, the computer would create the nodes behind the scenes, and then out came the numbers again which became graphs.  It wasn't until I was introduced to Electronics Workbench sometime after, through returning to school, that I really felt at ease with simulation.   It's a weird feeling.  Sure I'm simulating.  After all, I'm looking at a computer screen, just like those many years ago with Spice, but it's different, so different.  It's now a bench and a breadboard ...until I run into a problem like this one.  To me that is the key selling point of Multisim.

One of the guys I work with happened to come by while I was working with Multisim on my laptop.  He asked me to look at a node.  I instinctively grabbed the scope, made the connection, clicked the power button, and the circuit came to life.  Frankly (he's still doing the very expensive Spice and placing probes thing you see), I threw him for a loop, like into a new paradigm.  Then, I showed him a schematic from a project just earlier tonight.  He asked me what CAD I was using.  I said 'CAD?  No, this is my simulator.'  Once again, he was stunned ...and I could swear jealous. 😉

True, Multisim does have its problems.  That I can atest to; this forum has quite a few words from this author.  But I swear, even with its bugs, it is the greatest simulator I know of.  It's just this simple, boils all down to this.  If the simulator, even though you're looking at a computer screen, causes you to forget you have analysis tools you had years prior because you're so used to using the scope now, that my friend is a great simulator.  It truly is an electronics' workbench.

Ok, the commercial's over.  🙂