Whether intended or not, I am receiving bad service

[Euler's_Identity]

Patrick Noonan,

Thank you.  Through the self-extraction file, I was able to install the update, v10.0.1, over my old installation.

I tried out my circuit, but it wouldn't simulate.  However, through starting and stopping the working circuit (_3) many times, I have a clue about what could be at least aiding Multisim's tendency to prevent simulation. 

I have found that I have accidentally created a random number generator, at least it appears to be random.   If you start and stop that working circuit over and over, each time you start it you will find that the +I mux bank address is always 00h but the other three are seemingly random numbers.  In fact, I documented the numbers generated, over 55 iterations, below.

13, 12, 2a, 05, 2a, 23, 2e, 26, (07, 33, 33), 3c, 2a, 08, 06, 00, 20, 1a, 2d, 04, 08, 0c, 26, 06, 1a, 34, 27, 2a, 3e, 07, 0d, 15, 3b, 35, 29, 0a, 29, 0e, 11, 27, 11, 3e, 3c, 21, 20, 19, 2a, 2b, 08, 28, 14, 16, 1e, 15, 22, 13, 1d

(The () one shows that the mux's had varying states for that iteration.)

I restarted the program, reloaded the circuit, and then did it again, until I ran into a new number. After four iterations I generated a new number.  Therefore, with each attempt at simulating the circuit, a new starting state is generated, which, unless planned around, may be the source of errors in Multisim's engine.  And the sequence has not repeated!

Here's the question, with a flip-flop-based integrated circuit simulation, how does Multisim determine the initial conditions of the flip-flops?  For example, I simulated a simple r/c CMOS gate-based oscillator some time back using Multisim, and I found it imperative the initial conditions be forced in order for the oscillator to start and run.  Maybe this idea plays into latest particular circuit problem.  What if the circuit were amended to force a particular initial condition?  Somehow the software must be deciding the initial state of those flip flops.  Also, this might go somewhere in explaining how sometimes it simulates and sometimes it doesn't.  Yes?

(I think I mentioned this in a prior post when I said "float," but only yesterday did I get a chance to run the above described test.)
         

[lacy]

Good dedction Euler. This coming up random never had entered into my thinking when I was anyalyzing your circuit. I have had this happen with other IC like the 4017 where the outputs would come up with random one high on start-up. You can try this but I don't know if it will work. Try setting the simulation setting to set to zero. This may do the trick. If not, you may have to find a way (at least for simulation purposes) to make the IC's reset on power up.

Great Job in your analysis!

Have a Nice Day

[Euler's_Identity]

With respect to the original circuit, I tried "Interactive Simulation Settings," but before I could set to zero it came up with an improper start error window.  That's all I have the time for for the moment.

[lacy]

If I read your post correct, then I may have something to try for the random thing. Each one of the IC's have secondary enable. One iC it is label OC and the other is labled G (negative logic). Try putting a 1K reistor to ground  tie a small value non-polarized capacitor to it and back to VCC. Tie the Junction of the OC and G pins to the junction between the two.You would have to do that on each IC (I know it is a lot of work) I tried this on bank 2 and it seemed to work (no guarentee though).

This is the only way I can see to reset these when power is applied. There may other ways, but I just don't see it right now.If anyone of the NI team members have a better solution then we would definitely like to hear your input.

Have a Nice Day

Edit: I would try the set the zero first,(once you can get it simulating) as this would be the least amount of work. I have mine set this way and the banks appear to come up withthe correct bank selected and the only PIN Adreess on would be the first one.

Message Edited by lacy on 09-19-2007 09:50 PM

[Euler's_Identity]

Lacy,

It's the little things that were never tended to that drives one nuts.  Do you know that they still didn't fix the info for that octal latch?? Anyway...

Just for the heck of it, I tied all the 'oc's high via +5v sources.

Good news: The simulation began.

Bad news: Multisim stopped responding when I shut it off, and I had to terminate the process to close the program.

But, at this point, I'm not going to have any more time to fool with it for the next few days.

[Euler's_Identity]

One thing that would've went a long way, I think, concerning this bug...

Why in the world did Multisim not come with at least SOME CMOS analog multiplexer?   That is strange!

I need a large multiplexer, so I figured the largest I could get was a 16 channel.  But, since Multisim has NO multiplexers (not even digital that I can tell!), I built the 16:1 analog mux.  I wouldn't be surprised, assuming Multisim ever gets this problem solved, that the solution involves replacing the muxes in some way.  (I haven't heard a thing in some time now.  I'm thinking they're baffled and hope I'll just go away.)

Incidentally, if you need a 16 channel mux, there's a 4067, which I'll try to model.  Then again, that'll require MOSFET's, and my experiences with Multisim's MOSFET's hasn't been the greatest either.

You see I'd have no problem with a minimum of parts if the program were just dependable, as I could just build my own models as subcircuits.

Oh well, that's all the time I have for now. 

[lacy]

Multisim does have CMOS Multiplexers (at least in my version), but I haven't seen a 16:1. There is an 8:1 and it's part number is 4512BD. This is the only one that I can find that would even remotely be close to what you are needing. The rest seem to be either 2:1, 4:1 or decoders. The only problem I see is that these multiplexers are not bi-directional. I believe they only take the inputs and send it to the output, but you can't put something on the output line and send it the other way as it appears that yours is doing.

Maybe this one, if you have it, can help you in some way.

Have A Nice Day

[lacy]

I have found another part that looks real interesting as far as your application is concerned. This may even simplify it . Look under MIXED>ANALOG SWITCH and look at part number ADG4068BP. This is an 16:1 analog switch. I am sure that this component would be bidiirectional. Take a peek at it an see if this would be something that could substitute for your multiplexers.

If this would work, it would greatly simplify your design and would probably simulate a world better than it does now.

Let me know if this is what you are shooting for. There are more there, so you will have to look at them at see which one could possible work.

Have A Nice Day

[Euler's_Identity]

Aha!  I thought it quite odd that Multisim didn't contain a single mux.  As it turns out, thanks to you, I see what they're calling an analog switches'  library actually contains mux's!

An analog switch either acts alone (each sw has its own ctrl line) or is ganged (a common control line), but there is no decoding (# chooses switch).  A mux decodes.

Yes, I will try to redo my circuit using one of those mux's.  I'd prefer a 4000 series CMOS, but AD's ok.  (I prefer commonly had parts for the most part.)

It'd be interesting to sub in the new part and rebuild with the new part (2 ckts) and see if they both work, as it may yield a clue about where things went wrong -- somehow I killed the multimeter, and I still want to know how it happened to avoid doing it again.

Thanks for the help!
 

[lacy]

"An analog switch either acts alone (each sw has its own ctrl line) or is ganged (a common control line), but there is no decoding (# chooses switch).  A mux decodes."

I am not exactly sure what you mean by this statement, but the ones I see have address lines to select a particular switch to activate and a master enable to select the the chip itself. If you reverse things and run the input signal to the output then the switch selection process determines which line it will go to i.e. 0-16.

I would call that a form of decoding. My definition of decoding may be incorrect and if it is maybe your could elaborate on that when you get a chance. You could e-mail me with this information so we can discuss this off forum so that we don't violate any forum rules.Just whenever you have the time.

I am glad to be of assistance to you. I am also putting these in your circuit to try it for myself.

Have A Nice Day