Diverter on JD H260 - switch problems

[KY Gun Geek]

Working through installing a diverter to work a grapple lid on a JD 5093e with an H260 loader. I have 2 issues that I don't think are related - slow bucket curl, and switch destruction. I have posted 2 threads to keep from confusing things.

Here's a link to the bucket curl thread:
http://www.tractorbynet.com/forums/hydraulics/375267-diverter-jd-h260-slow-curl.html#post4687660


I am on my 3rd switch and it has failed, too. They seem to fail after 20 - 25 presses. Valve is spec'd to draw 3A, I have measured it at about 2.6. There is a 5A fuse in line that has never blown. First and 2nd switches were rated at 3A, then 5A. Current one is at 8A. They are screw terminal, so no soldering (i.e. heating the contacts). Failure modes have been the same - switch becomes unreliable, then stops working all together. The valve has an LED that lights when voltage is applied with correct polarity - new switch starts working great, light is clear and bright, then doesn't come on everytime button pushed and LED flickers, then stops working. If I remove the switch and short the wire, valve works great. Coil does not feel hot - not holding switch down for extended period.

Here's my last switch:
Dorman Products - 84822

Any ideas? Are we dealing with "flyback" - a reverse voltage spike generated with the coil field collapses - see below. This valve is pretty common (bought from Surplus, link below) - see lots of folks here using it, anyone else have a problem? I haven't seen a mention of the LED here. Is this a new mod that might be causing a problem? Seems like the mod shouldn't make the spike worse, the LED would just be in reverse polarity and not do anything.

Called Surplus and they said if I sent valve back they would test it on the bench. They also said that given that I see a 2.6A draw, they probably would as well and call it good.

I guess I could test theory by taking switch out and shorting the wires. If there is a spark when I move the wires apart, then flyback is prolly the issue.

If you are interested, a little googling turns up this:
"Since an inductor (the relay coil) cannot change it's current instantly, the flyback diode provides a path for the current when the coil is switched off. Otherwise, a voltage spike will occur causing arcing on switch contacts or possibly destroying switching transistors.Feb 18, 2014"

If you just really have to study up:
Flyback diode - Wikipedia


Spec info just in case...
TractorData.com John Deere 593E tractor information

12/24 VOLT DC 13.2 GPM SAE 8 SOLENOID OPERATED DOUBLE SELECTOR VALVE

 

[BukitCase]

Don't need to follow your links, you're on the right track. Having spent the last 35 years of my working life doing industrial instrumentation/control, I've seen less than 1 amp inductive loads (mostly solenoid valves, just smaller ones than the diverters) take out 10 amp rated switch contacts in a few days til a clamping diode was added. Most of the time I got away with 1000 PRV rated 2.5 amp diodes; in this case, you'd probably want a diode rated at least the amperage of the coil, might wanna look for a 1000 volt @ 5 amps.

If you can't find one that high a rating that's still axial leads, it'd be worth trying a 2.5 amp rated but be sure you stay at LEAST 1000 volt rating (NOT a typo - but by now, you already KNOW that :=)...Steve

 

[KY Gun Geek]

I may be mired in group-think, but I haven't seen any mention of this being issue on this board, and there's lots of these valves discussed. Surplus didn't seem to know anything about it.

I'm also a little suspicious of the LED (and resistor) pinned across the terminals. Surplus doesn't show it in this instruction sheet (below). In fact, it could be pretty easy to get the polarity backwards and the LED wouldn't work - there's no note of it at all. Seems like a "quickie". I suppose for a short time the LED would prevent the flyback, but would prolly burn up quick.

Time to diode shop. I have have diodes used for just this purpose in electric golf carts (cuz I have golf carts). Those things are current and torque monsters, and they have a relay that will light you up if you don't have a diode on it.

http://www.surpluscenter.com/_MoreSpecs/i9-8602.pdf

 

[KY Gun Geek]

Well, I couldn't find it via search, but quite by accident, I found this. Looks like it has been discussed. In this case it was only a suggestion. Will read more and see how it turned out.

Seems like surplus should be aware and or make a note on the data sheet.

http://www.tractorbynet.com/forums/...pple-hydraulic-setup-1026r-2.html#post2874456

 

[BukitCase]

Pretty sure I've mentioned the need for the diode in a thread or two, but don't remember which ones - but you're right - I doubt the LED/resistor is for anything more than an indicator of which button's pushed (or not) - and I bet golf cart relays really WOULD light you up with no diode, since they're usually at least 24 volts DC to start with - that should give one heck of an inductive kick - If you think about it, a car ignition's just based on turning OFF a coil, and even the older style with the right coils can put out over 50,000 volts (of course, I'm ignoring the fact that they're doing a lot of it because of the coil being an autotransformer)... Steve

 

[tre873]

You could always use the switch to control a relay. The coil side of the relay would pull way less than 1 amp through the switch. Most modern relays have a diode across the coil to prevent the voltage spike from damaging the switch.

 

[BukitCase]

You could do that, but you'd STILL need a diode clamp on the relay's CONTACTS or you'd just MOVE the problem to relay contacts instead of switch contacts - still, that might be preferable if the switch is more expensive or harder to get at than a plug-in relay... Steve

 

[tre873]

You could do that, but you'd STILL need a diode clamp on the relay's CONTACTS or you'd just MOVE the problem to relay contacts instead of switch contacts - still, that might be preferable if the switch is more expensive or harder to get at than a plug-in relay... Steve

I would have to disagree. Relays are designed to handle large current loads without causing pitting on the contacts. You can get 12 volt relays that can handle 20 amps without causing pitting on the contacts. You can go even higher if you spend a little more. You should be able to get a relay and connector at most parts stores for less than $20.

 

[BukitCase]

It's not the main current that's the problem with ANY mechanical contacts when switching an inductive load - it's the KILOVOLT or more of back emf that ARCS across the contacts every time they OPEN. For an extreme example, google "arc flash" ... Steve

 

[tre873]

I'm well aware of the voltage spike that occurs when a magnetic field collapses. The relay contacts aren't near as susceptible to detoriation as the switch contacts are. If the OP uses an automotive grade relay, he won't have anymore problems. Think about the number of cycles a typical automotive relay endures during the lifetime of the vehicle. Automotive relays have a suppression diode across the coil side only. You won't need a suppression diode on the load side. Think about how long a set of points would last on a gas engine. Points opened and closed 100's of times a minute and would still last 5-10 years under normal use.

 

[Industrial Toys]

I just showed a friend how to wire a four channel radio remote control to a hydraulic power pack. The relays claim 20 amps and it works for now, but I am skeptical. That is what the hydraulic shop offered him. I did add 1N4001 diodes to all coils. That is only around 100 PRV if I am correct, but I have never seen commercially designed stuff go nuts on the ratings for such protection diodes.

In general, most switch and relay contacts have the highest ratings for AC current. DC is MUCH harder on a contact.

I ALWAYS use an intermediate relay when designing such stuff and never put more then this low coil current through my control circuits. It adds complexity and cost but has served me well.

 

[KY Gun Geek]

Got my hands on another switch and some diodes today. Switch is same as above. Diode is 50v, 6A. The 50v is a little light, but I figured I would give it a try. I have some 1000v 10A coming.

Before I get to the install I did a little science experiment. I touched wires together (to simulate switch contacts) and videoed the arc. I then wired in the diode, and did the same thing. There is a noticeable difference in the sound, size, and intensity of the arc. Below should be some stills clipped from the vid.

However, again, the switch lasted about 30 to 45 minutes of me picking up brush (the job that is not getting done cuz I am fool'n around with the tractor). More details, and my next step in the next post...

 

[KY Gun Geek]

As stated above, wired in a diode and a new switch. Switch failed again in about 30 to 45 minutes of operation, maybe 25 or 30 cycles.

When I got the switch, I got an extra switch, and per the discussion above, I got a relay. My source for the diode is Radio Shack, and the switch and relay is Autozone. Diode is 50v at 6A, switch 8A. The diode is a little light for this duty. I have some 1000v/10A diodes, and a 5A switch coming from Amazon.

Some pics below of the diode install. Had to get after the connector with the dremmel - the diode is pretty big. The diode is installed so that it is reverse biased when battery voltage is applied to the side with the band (red wire is +12v when switch is pressed, black wire is grounded)

Given the difference seen in the arc, was really surprised when the switch failed again. I thought the diode had failed open (it is not up to the spec discussed above), and that caused the switch to fail. I tested the diode with my DMM (it has a diode test setting). Diode tested as open when reverse biased, and shows 5287 when forward biased. I'm not sure what this number actually means (diodes have a voltage drop and small forward resistance, so I'm assuming this is just how the meter reads that). Anyway, the key was that the diode I had installed and all the other diodes in the package (there were 5 total) read almost exactly the same. So they were all good (or bad) but the installed one tested the same as the unused ones. I must say I rigged up a little circuit to test the diode (diode and a resistor) and it worked as I would have expected - current flowed in forward bias, did not flow in reverse bias. I don't trust gravity right now either.


I did a test rig of the relay and it worked fine. The switch current will be 0.14A according to my meter, so let's hope it can handle it. I will leave the diode in the diverter solenoid to give the relay contacts as much help as possible. So tomorrow I'll install the relay and try to move some slash!

Will probably switch to the higher rated diodes when the arrive next week.

Any other ideas are welcome at this point. If the relay fails, I'm stumped. I have been searching this and a couple of other boards, and while I see people discussing using a diode, and a few saying they did it, I don't see anybody with this level of problem killing switches...

 

[KY Gun Geek]

Think about how long a set of points would last on a gas engine. Points opened and closed 100's of times a minute and would still last 5-10 years under normal use.

There's more going on there.

Points also need a condenser (capacitor) or they burn in a hurry.

 

[BukitCase]

"Points also need a condenser (capacitor) or they burn in a hurry."

Exactly - been a long time, but I used to remember which way it went; if you got the wrong capacitance (either from some parts guy picking up the wrong box, cap failure, etc) too MUCH capacitance would pit the points in ONE direction, too LITTLE would pit them the OTHER way. The RIGHT amount for the particular coil, and the "tank circuit" is BALANCED (minimum current OUTSIDE of the "tank circuit") and the points last a LONG time.

Same effect (and for the same reasons) as welding - material deposition follows electron flow. DCEN (STRAIGHT polarity) gives more deposition but less penetration because the negative side is ALSO heating up more, although that's MOSTLY because the MUCH smaller cross-section ELECTRODE is providing all that current while the WORK can (usually) spread the flow out over a much larger area.

Not sure if these would be available LOCALLY, but we used to use "snubbers" in a lot of places - looks like a capacitor of sorts, but contains a cap AND a resistor in SERIES - placed across an inductive load along with the (paralleled) diode, they can reduce the inductive effect so the diode doesn't have to do it all.

BTW, it's possible that the reason your diode tests OK but didn't SOLVE the problem - MIGHT be because your diode is rated for enough CURRENT that it doesn't fail PERMANENTLY when the counter emf spike hits it - under some conditions, semiconductors can "break down" and allow reverse current but then "heal" themselves if it's short term enough and doesn't generate enough heat -

If you have access to a scope you might be able to do a more quantitative experiment (actually put NUMBERS to each scenario), that'd make it faster to "dial in" a fix for your specific situation - my own projects involving solenoid diverters are too far down the (Looooong) project list to be of much help to you now, but I'm pretty sure that when they reach the top of that list there WILL be more info forthcoming (assuming my Tektronix scope still works, it's been decades)

Hope SOME of the above helps, good luck on your "quest for (NO) fire" :thumbsup: ...Steve

 

[KY Gun Geek]

Relay wired in - set up worked longer than the previous 45 minutes. Will give it a bit of a work-out tomorrow (2 or 3 hours).

 

[tre873]

Relay wired in - set up worked longer than the previous 45 minutes. Will give it a bit of a work-out tomorrow (2 or 3 hours).

Glad you got it working.

 

[KY Gun Geek]

About 10 hours of run time and the relay seems to have solved the problem!

As many folks are talking about diverter valves, I hope this thread helps!