Toggling a latch

Terry Pinnell · 2004-12-09T00:56:00+00:00

I'm getting unreliable behaviour from my Curtain Controller. Is therea design flaw in the first part of the circuit, shown here please:I've not shown the subsequent sections as they are working OK. (Forthose interested, X & Y go to exclusive OR logic, then outputs fromthat, together with X, control the motor power and directionrespectively).This section I've shown appeared to be OK too until recently. But nowthe 'toggle' button is not operating if the curtain is in its closedstate. (I may have swapped the 4001 while looking at another problem.Can't think of any other changes.)Anyway, I'm beginning to wonder if it ever *should* have worked! Is itOK to send pulses simultaneously to both Set inputs of these twobistables? The idea is that, whichever is not already set will now beset, giving alternate opening or closing of the curtain. Thesimultaneous RESET works OK, and that seems identical IMO. And thediodes from pins 4 and 11 ensure both bistables cannot be in a Setstate simultaneously. Also I *am* getting a +ve pulse when I press thetoggle button. And, as mentioned, that is closing the curtain whenthey are open; it's just the reverse that's no longer working.-- Terry PinnellHobbyist, West Sussex, UK

T

Terry Pinnell 21 years ago

re

formatting link

Both you and John have been generous with your advice so far, thank you. I hope you're both still patiently taking an interest, as this issue is still puzzling me.

For the last day or so, the toggle has returned to working reliably. I've made no changes. This design does have the merit of simplicity, (as well as being something I thought of myself ), so although I'll almost certainly change it in the course of a rebuild, I'd still like to satisfy my curiosity as to whether or not it *should* work. I think I'm close to resolving this.

Attempts to simulate have so far been unsuccessful. Digital mode in CM is no use here, as it won't accommodate diodes, resistors, caps, etc; and so far in analog mode I can't get outputs to change state at all! That's a side issue I'll also explore, as I *should* be able to simulate it somehow.

As for the real circuit, from the resting state with both latches off, immediately following the 'simultaneous' pulses reaching pins 1 and 8, there is presumably a very brief period when one latch is on and the other off. (They cannot both be on together, because of diodes from output of one latch to the reset of the other.) If that happens to be the 'wrong' latch, the curtain will try to move against the existing limit. That definitely used to happen on occasions; currently it's not done so on my last dozen trials. So, because the toggle button will still be closed, it seems to me that it will *then* set the correct latch. That was the simplistic thinking that prompted me.

However, I've just realised there's plainly at least one other factor that must affect the behaviour: whether or not the limit switch is closed. Sometimes when the curtain reached a limit, it would maintain the microswitch closed. Sometimes not. Another 4 trials a few minutes ago showed that, for the time being at least, it *is* keeping the appropriate switch closed (at either limit).

With the curtain closed, I manually forced the 'Close Stop' switch open. Now Toggle did *not* work. It started motor in wrong direction, i.e. continuing to try to close it, and I had to use Emerg Stop.

If my analysis is correct, then the inherently uncertain mechanics means this approach is indeed unreliable.

-------

I'm also thinking of returning to battery operation of the motor (3 NiCads). During opening and closing, my '13V' line drops from 14.1 to

10.0 V. That can't help stability of the logic sections. My main gripe with the battery approach was that, after a failure (almost certainly due to the issue discussed above), one of the three would get reverse charged. So, after removing the stall condition, my built-in 35 mA trickle would not restore them. (Had to nurse them back in the workshop, a PITA.) But if I can electronically remove the stall, that problem should disappear. Of course, I'd have to prevent the stall repeating indefinitely, but that's another matter...

Terry Pinnell Hobbyist, West Sussex, UK

Vote

F

Fred Bloggs 21 years ago

Normally TOGGLE means you can make X/Y go from 1/0 to 0/1. Your circuit cannot under any circumstances do this. If a limit switch is clearing an output, then your so-called TOGGLE input will set the other one. It is important that your limit switches stay closed.

You need to realize that your dual NOR-latch circuit should be abandoned, and then go with any of the numerous suggested circuits in which stall is impossible- unless a limit switch fails.

Vote

T

Tony Williams 21 years ago

Hello Terry,

Suggestion..... unless it has already been made.

Each Limit switch should hit *two* R-S flip flops.

The CLOSED switch should unlatch power to the motor, and pulse the motor polarity flip flop to OPEN.

The OPENED switch should also unlatch motor power, but pulse the motor polarity flip flop to CLOSE.

This deals with unreliable Limit signals and the START button is only required to pulse the motor flip flop into the ON state.

Two relays perhaps and if one of them was a maglatch relay it would remember the last signalled state even through power downs.

Tony Williams.

Vote

T

Terry Pinnell 21 years ago

Thanks Tony. My circuit effectively achieves the same, but by a different method.

formatting link

Note first that, because of the diodes, the R-S flip flop outputs X and Y cannot be simultaneously high at any time. They are then EXORED in

formatting link

and drive the two relays in the way you described.

Eventually I expect to get rid of the limit switches anyway, and use current limiting (as discussed up-thread). But meanwhile I'm keen to get a definitive understanding as to whether my simple 'toggle' works in theory as well as it (usually) works in practice!

Terry, West Sussex, UK

Vote

T

Terry Pinnell 21 years ago

Not sure I follow you. That's precisely what it does now, in the circumstances I described, namely when the limit switch has been closed and kept closed.

That seems to say we *are* in agreement, despite your comment above. Just to make sure we have this straight, I repeat: X and Y are reversed when I press my toggle button, providing the limit switches stay closed.

Or does the confusion stem simply from the fact that in

formatting link

I've *drawn* both limit switches in an open state?

Agreed, in the present context, yes. They need to be kept closed for Toggle to toggle. But I assume you don't mean that in a more general sense, for any circuit? Assuming I keep the limit switches at all, I only need one pulse from them when the curtains reach their extreme positions. (In reality, despite my bounce protection, I probably get several, but that should be irrelevant.) So IMO a circuit like my present one that *requires* the limit switches to be kept closed is inferior to more tolerant circuits that don't.

Or the much simpler approach I'm thinking of using current limiting, removing limit switches from the equation entirely!

Terry, West Sussex, UK

Vote

T

Terry Pinnell 21 years ago

I think I now understand where we were at cross-purposes on this. I'm using 'toggle' to mean something different to you. I mean that, if the curtain is closed, pressing Toggle 'toggles' it open. And if the curtain is open, Toggle toggles it closed.

You are of course quite right that it does *not* make X/Y go from 1/0 to 0/1. What happens is that it goes from 0/0 to either 1/0 or 0/1. Whether it does it in one step or two (the first being very fast) I don't yet understand. But it certainly 'toggles' in the sense I've been using (providing the limit switch is closed when Toggle is pressed).

Agreed.

I was plainly wrong there. The F/F outputs are not reversed, the

*action* (since the last press of Toggle) is reversed.

Obviously irrelevant.

Would still appreciate your clarification on this point please.

Terry Pinnell Hobbyist, West Sussex, UK

Vote

T

Terry Pinnell 21 years ago

I finally managed to simulate the present toggle button circuit

formatting link

and the result (showing the same behaviour as the actual circuit) is here:

formatting link

Incidentally, for anyone interested in the Spice details, Y displays an 8 us excursion to +/- 400 uV at t = 2.0 s

Also, here is the 'regular' simulation I showed earlier, with some explanatory notes:

formatting link

And the Emergency Stop button also simulated correctly.

With that sorted, I'm going to experiment with my current limiting ideas for a while.

Terry Pinnell Hobbyist, West Sussex, UK

Vote

F

Fred Bloggs 21 years ago

Without this the circuit has multiple states: OPEN limit switch means either under/over travel past the switch position- not really a definitive position indicator. You have just eliminated the cheapest one-bit non-volatile memory element you can buy. And how did you work your LDR control into that circuit?

Why don't you work on simplifying your limit switch triggering. I can't tell what you have going on there from that photograph. Can you draw how that is supposed to work. It looks like you have some thick length of spliced cord that snags some forked shaped thing to pull on the lever switch...If you make that lever vertical and add a cheap guide (eyelet) for the cord so that the thick section is pushed horizontally enough to close the switch, there will be no overstress- and absolutely no precision in predicting overtravel is required. Last time I checked, an eyelet is cheaper than a relay.

Vote