Non-Debounced Vacuum switch is ruining my relays! Scope pic shown

[iso9001]

So... I have an electronic vacuum pump on a diesel vehicle because on it's own the engine does not generate enough vacuum to run the brakes. So, the pump is wired up to a relay that has a vacuum switch on the control side. When the switch gets to 22" of vacuum it shuts the relay off by interrupting ground on the control side. Ignition powers the 12V high of the control. Simple enough....

My switch is WAY too sensitive. It's just a mechanical diaphragm, so it has no deadband/hysteresis. I'm having a problem where the relay is clicking on and off rapidly when the switch makes the connection, the switch itself jutters, then the relay does which further makes the switch signal oscillate. After about 1-2 seconds of random on/off, the relay stabilizes.

The issues is that this is wearing out my relay extremely quickly (obviously). I think it's arcing initially, but that quickly leads to latching. I need some way of smoothing out the signal or otherwise debouncing. I'd like to accomplish this without a circuit board if possible as this is under the hood and I'd rather not have a hokey looking box, I wasn't the tech that did the original install, but I have to fix this in the most professional way possible. Whatever fix I can find, has to be solid for years - I do not want to see this vehicle here again.

I've considered a capacitor on the line, but mildly concerned about the initial draw and I'm not certain what value would be most appropriate. Is this something that a pass filter would be acceptable for? Maybe I need a few components and a transistor to debounce? Does anyone make a special relay or other part for this?

Here is what the scope of the signal normally coming to negative pressure looks like:
**broken link removed**

 

[crutschow]

I think you will need a box. How "hokey looking" the box is depends upon what box you use. You should be able to eliminate the switch bounce with a diode, a resistor, and a capacitor. The best values for those depends upon the relay current. Can you determine that or measure the relay coil resistance?

 

[iso9001]

Automotive relay, so iirc I think the coil resistance is about 75ohm. There is a 1M across the coil shown inside the relay.

Well, diode, resistor, cap I can handle. It's adding a transistor that I was getting hesitant about because I'd also probably then want/need to make up a pcb and do a few of these.

 

[MikeMl]

How about a 555 wired as a retriggerable 1 sec delay circuit? The relay will pull-in the first time the vac switch closes, and stay pulled in while the vac switch does its bounces, and finally drop out about 1 sec after the vac switch remains open.

 

[crutschow]

Automotive relay, so iirc I think the coil resistance is about 75ohm. There is a 1M across the coil shown inside the relay.

Well, diode, resistor, cap I can handle. It's adding a transistor that I was getting hesitant about because I'd also probably then want/need to make up a pcb and do a few of these.

With further though I believe you just need a resistor and capacitor. So try a 1Ω resistor between the relay return and the vacuum switch. Add a 220μF 20V capacitor between the relay return and ground (+ to relay).

With that circuit the vacuum switch closing will rapidly discharge the capacitor through the resistor to turn on the relay (the resistor limits the peak current). When the switch opens from the bounce, the capacitor will continue to sink the relay current until the next switch contact, which will remove any charge on the capacitor accumulated on the capacitor during the bounce. This process continues until the switch finally stops bouncing and stays closed.

 

[iso9001]

With further though I believe you just need a resistor and capacitor. So try a 1Ω resistor between the relay return and the vacuum switch. Add a 220μF 20V capacitor between the relay return and ground (+ to relay).

With that circuit the vacuum switch closing will rapidly discharge the capacitor through the resistor to turn on the relay (the resistor limits the peak current). When the switch opens from the bounce, the capacitor will continue to sink the relay current until the next switch contact, which will remove any charge on the capacitor accumulated on the capacitor during the bounce. This process continues until the switch finally stops bouncing and stays closed.

Ok.... So....

**broken link removed**

Or did I get that wrong ??? ... Wait, I did didn't I... That cap should be on the + side of the relay... I'll fix.

 

[iso9001]

**broken link removed**

... Wait, that can't be right either. I'm not the only circuit using that 12V IGN line, this could screw other things up I suppose, but I think I'm missing something anyhow.

 

[KeepItSimpleStupid]

Thinking out loud.

1. Would you consider an electronic sensor?
(I'm only mentioning this because it's an option)
2. Would a Vacuum resovour help in this situation?

 

[alec_t]

You had it right in post #6, except that I think you'll need a higher cap value (~1000uF)

 

[iso9001]

I'll take a look at #6 and a larger cap, maybe I'll meet in between.

KISS, on #1, yea, that would be great if I was running a micro in this application, but I'd like to avoid extra parts like the comparator I would need at the minimum. #2, there is already a very large reservoir. I think the switch is a little too sensitive, but this is the second type of switch we've tried with the same issue.

 

[crutschow]

Yes, your post #6 circuit is what I meant. I chose 220uF since the discharge time-constant (switch closed) of 220μs and a charge time-constant (switch open) of 16.5ms seemed a good compromise for the bounce times shown in the scope photo. But you can certainly play with the cap size to determine the optimum value.

 

[iso9001]

Ok, haven't tried it yet, but I did go and pick up the components. I'm curious the hour life of the 220uF or 470uF 35VDC caps I got at 14V and upto 250ºF will be.... ? Like maybe I should make two of these 'kits' to wire in, one for when the first dies.

Then the next issue, the only 1ohm resistor I could find was a wirewound 1W, it's huge. I picked up a coupe 10ohm 1W which are a ton smaller but I don't know what that would do to my relay's activating.

 

[crutschow]

Ok, haven't tried it yet, but I did go and pick up the components. I'm curious the hour life of the 220uF or 470uF 35VDC caps I got at 14V and upto 250ºF will be.... ? Like maybe I should make two of these 'kits' to wire in, one for when the first dies.

Then the next issue, the only 1ohm resistor I could find was a wirewound 1W, it's huge. I picked up a coupe 10ohm 1W which are a ton smaller but I don't know what that would do to my relay's activating.

Yes, the life of those caps at 250°C could be limited. Could you mount the box in a cooler location?

10Ω is a little large to place in series with a 75Ω relay. It may not pull in properly.

Where are you located? Radio Shack has 1Ω, 1W resistors available to order.

 

[Mosaic]

He said 250 F.

125C rated caps are required for auto applications.

 

[iso9001]

He said 250 F.

125C rated caps are required for auto applications.

Yea, I'm just not sure how long that rating will effectively last. 2000 hours gets me probably two years in this vehicle, then the customer will surely come back with another fused relay! :\

I put a 440uF and 1ohm in and the vehicle has shipped. The scope shows a much cleaner on/off now, so I'm hoping that will do it. I was going to have one of the guys put dielectric grease inside the relay to further prevent arcing, but sort of forgot. That's the next step if I hear from him again. But I think this ought to do it for now.

 

[crutschow]

He said 250 F.

....................

Oops. I'm too used to thinking in centigrade.

 

[crutschow]

..................................

I put a 440uF and 1ohm in and the vehicle has shipped. The scope shows a much cleaner on/off now, so I'm hoping that will do it. I was going to have one of the guys put dielectric grease inside the relay to further prevent arcing, but sort of forgot. That's the next step if I hear from him again. But I think this ought to do it for now.

Don't put dielectric grease near the contacts. It won't prevent arcing and, if some gets on the contacts, can actually cause the contacts to degrade faster.