Dialog console pushbutton switches - are new switches available?

[wrench]

Well, that is an interesting question! Ordinarily I would say that it could be because the diode is there to prevent arcing. However, a whole sub-topic of this post relates to the cleaning action of arcing.... And that makes sense too. So, it is kind of like an EDM. They work on the principle of arcing but too much arcing isn't good.

Testing a diode is *very* easy so it doesn't hurt to just check them out. You will want one leg of the switch not connected to anything but then just use a fairly modern digital volt meter. They almost all have a diode test function. You won't be able to validate the reverse breakdown voltage but if the device conducts properly in one direction and not the other (which you will be able to easily test) then it is highly unlikely that your diode has failed. In fact, it is probably impossible...

 

[Mud]

My meter is a Fluke 12, and it does test diodes. I'll open the machine up and do some testing. There are diodes elsewhere in the machine built into a terminal block that I ran into when looking for a hydraulic shifting problem, Deckel seems to like diodes. Thanks.

 

[wrench]

Mud... Just a quick clarification... A bad diode will not cause the switch to not work. So if you are expecting to be able to search for a bad diode, find it, fix it, and then expect the button to work, that won't happen (unless it failed as a short, which is highly unlikely). The diode is more of a protection device in this particular scenario; it would only conduct when the switch is transitioning from closed to open.

Of course, checking the components is not a bad idea... I'm just trying to manage your expectations.

 

[Mud]

Understood. I'm wondering if a faulty diode would aggravate the problem. Primarily I wanted to know what to replace it with. In the hydraulic problem I jumped the diode out under DD's orders to see if the problem changed and the circuit still worked just fine. Here it seems to be an essential part of the circuit, at least on the switches I've worked on/looked at. On the switch in the photo, it conducts current to the contacts. On the toolchange enable switch it's connected to the lamp.

 

[wrench]

These are good questions and I hinted to some confusion in my last post. Although i can understand the concept of "cleaning action" via arcing, to be honest, to me arcing is still a less than desirable function and the diode is there to prevent just that problem. So (and, again, this is just my opinion), a faulty diode that is always open is not a good thing and would cause the premature failure of a switch. Of course, this really depends on what the switch is switching. Diodes are often added by manufacturers and may not even be all that useful for a control switch.

*IF* it were to fail as a short *and* the diode was across the switch (more on this in a bit) *then* that will make your switch look like it is always pressed. The important thing to realize here is that *one* failed diode could actually screw up the entire set of keys. This is due to the fact that generally switches that are non-critical (i.e. something other than an E-Stop or similar) are often arranged in an X-Y matrix. This is done to minimize the number of interfaces to the computer that ultimately reads the state. But, if you have a short, then that will affect the other switches too. I haven't looked at the schematic recently so I can't tell you what the Deckel looks like, but I'd bet money that it is using a matrix. I don't really recall what your failure mode is either, meaning one switch or the whole set of switches messed up... If it is the latter, then that is a significant piece of information and could point to either a switch or potentially the logic used to read the switch row/columns.

OK, with that in mind, you mentioned that the light has a diode across it. That doesn't make any sense, really. My guess is that you may be mistaken on that and should confirm. I say this because the whole point is to prevent arcing when a switch opens. A lamp is a resistive load that doesn't open so you don't have the same problem. So.... I'm confused.

Oh... And I'll give a plug for Deckel Doctor... He obviously works on this stuff all the time and has a very good understanding of the design and failure modes. Real life knowledge is always better than rusty theory. So, I'd give healthy respect to looking carefully into anything he suggests.

 

[Mud]

The diode isn't across the lamp, it goes to one of the lamp terminals. Not sure of the circuit, I'll look at it again, it may not even be relevant to the lamp. I do think many switches are daisy chained together by at least one circuit, lots of pairs of connections by white wires to side by side terminals.

BTW, where to get these diodes? Google had only turned them up at Mouser, Mouser has a 3500 minimum qty. order.

 

[Sean S]

BTW, where to get these diodes? Google had only turned them up at Mouser, Mouser has a 3500 minimum qty. order.

I'll take one if you buy

 

[Mud]

I'll take one if you buy

They are $0.187 each to me, I'll have to have $.22 at least

 

[Sean S]

They are $0.187 each to me, I'll have to have $.22 at least

Does that include shipping?

 

[wrench]

I'd check eBay and DigiKey too. Unfortunately I checked my inventory and I don't have one... sorry! In a worse case scenario, we can likely replace this with another more readily available diode if you are having a problem finding that particular one. Do you know what voltage is being supplied to those switches? If it is 24VDC, I am actually a little surprised that they would be using a 62V zener.

Another option is to use: 1N5265 which is a 62V 0.5W zener

BTW, just to be certain this *is* a 1N4129, you should look at some of your other switches. You will likely see the full number

 

[Mud]

Here's a photo of a lighted switch, and a closeup of the diode. X1 and X2 go to the lamp, 13 and 14 are the NO contacts.

The reverse of the diode just said ITT, didn't photograph it.

 

[Mud]

One of a balkiest switches was the toolchange enable switch. On this machine it's on the console. Needed to be held down until it lit, sometimes pushed more than once, even sometimes wiggled while pushing down. So I took it out, hooked it to the ohmmeter and took a video of it to show here.
The video is pretty self explanatory, watch this behavior. The leads test at .1 ohm when connected together, there's nothing else connected to the switch except the leads connected to the NO contacts. I tried spraying it out with contact cleaner a few days ago, that made no no noticeable difference on this switch.

 

[Mud]

Boy i sure hope someone doesn't suggest "Stabliant" (sp) as an option.
Cheers Ross

You realize that's like saying "DON'T think of a pink elephant" don't you?
I hadn't thought of it until you posted that, so this is your fault.

Here's the 'Y'all ain't gonna believe this shit" dept.

I've never had an opportunity to test the magic potion before in a circumstance that could be quantified, and never in a switch-like device. I had the impression that it improved things on PCB contacts, but nothing black and white.
I couldn't disassemble the switch without taking the diode off, so I put a drop of stabilant on the little windows opening leading to the NO contacts and blew it in with an air nozzle. I attached a spare pushbutton and worked the switch as I blew it in.

Here's a video of the result.

Plugged it back together and the change was dramatically improved. The toolchange acknowledge switch was balky too, so out it came. Did the same treatment, tested with the meter and reinstalled. Started the machine and abused the switches a bit, trying to make them malfunction. Only the very fastest stab at the acknowledge switch made it miss, and only once at that, otherwise they both function perfectly, it's like a different machine. Raised eyebrows and fistbumps ensued.

All the other switches seem a little balky to some degree, except the E-stop, so I'm going to treat the rest of them and just see what happens over time, don't know what else to say about this.

 

[Sean S]

I've never even heard of the stuff :-/

 

[Mud]

Search for it here on PM, there have been discussions.

 

[wrench]

Supposedly "plates" when it is in the presence of an electric field. Something of a mythical panacea for all electrical woes that some feel is God and the other feel is snake oil. I'm not a chemist so I don't know but it seems more mythical than anything else. I know you, Mud, said it has helped but I have to wonder how long this will last. Maybe it is a mildly acidic formula that etches the contacts? Anyhow, it is supposed to be more for static connections and not so much for arcing contacts, such as a switch.

 

[Mud]

FWIW, I'm still looking for replacement switches and diodes.
And I tested the diodes, they passed.

 

[AlfaGTA]

You realize that's like saying "DON'T think of a pink elephant" don't you?
I hadn't thought of it until you posted that, so this is your fault.

Cheers Ross

 

[AlfaGTA]

Think there might be a run on "Stabilant" now.....
That is a convincing test, surprised me! Wonder if that will help the shifting by cleaning the relay contacts

John, are you sure you aren't getting a little kick back under the table....Ought to send that video to the makers...maybe they
will send you a case of the stuff.
Cheers Ross

 

[wrench]

I am rusty. I was thinking about this a bit more and the main place you will find diodes on a relay is across the coil, not the switch. The reason for this is that a coil is just an inductor and it want to keep flowing. When you switch it off, you'll get a spike and the diode limits this. In the Deckel case the side is across the switch which has no inductive component. So the designer must have pt the diode in there for whatever load is being switched. But I doubt much of an inductive lad exists there as these switches probably just go into a buffer to be read by the flu. In this case there s little use for the suppression diode. That means that the diodes may actually be used for a role other than suppression.

All of this is somewhat moot because they are unlikely to fail, something you have confirmed is the case