Bosch relay coil suppression

Anyone ever take apart one of the common Bosch relays used in Ferraris?

On relay coils, a considerable spike voltage can be generated on deactivation as the coil field collapses (much the way an ignition coil generates high voltage on points opening). There are a number of techniques to suppress or quench this voltage spike, but most of them make the relay coil circuit polarized. (Aka: the coil winds up with a designated plus and minus, and you can't reverse them.)

The one option that leaves the coil unpolarized is a resistor in parallel with the relay coil. But suppression resistors increase current draw and heat dissipation in the coil circuit.

I've heard that some Bosch relays have this suppression built in.

I've also noticed that the Bosch relays draw a lot of current (160mA), and get uncomfortably warm to the touch.

I'm curious whether anyone has dismantled one of the Bosch relays used in common Ferrari applications and saw whether it had internal suppression.

(A polarized alternative uses a diode and zener diode connected in series, anode to anode, across the coil. I'd think this would be a far cooler running option, but the coil would have designated "plus" and "minus" contacts, making it less flexible in generic use. It would also be more expensive in mass production.)


Or does anyone know a lower current plug replacement relay? (E.g. higher impedance coil.)

I'm wiring up a "radio bus" relay for my 328 (to let the radio be ignition switched). But as I have to locate the relay along the dashboard wiring harness, I'd rather not use a space heater relay, as the relay will be active the entire time the ignition is on. I also have a Siemens and some "Omega" relays that fit the same socket, but they draw just as much coil current. (And also get hot.)

I have a Radio Shack relay with sufficient contact current capacity that only draws 40mA on the coil, but it's designed to mount on a PC board -- and I don't know how long a Radio Shack relay would last where it's active over long periods of time.

I'm also a bit worried about switching contact arcing. I'm driving the new relay coil off the driver's window circuit, which, in a 328, is a full 20A fuse off relay "q" which is switched by relay "u", with the ignition. (It's the one ignition switched circuit that seems to have spare current capacity, limited failure impact, and a presence right behind the center console.) But this means a cascade activation/deactivation as the ignition activates "u" which activates "q" which then activates the new relay. Without coil suppression, the new relay could impact the life of relay "q"'s contacts. (Relay q operates both power windows and the rear window defroster.)

 

The 6.3 mm blade connectors used in the car are back ordered to Europe, and it'll be at least another week before I get them.

So I've got lots of time to "over-think" it.

Heck, by the standards of the last guy who installed a stereo in this car, I'm way above and beyond. He just stuck a hidden switch under the dash to switch power to the stereo (and amp). (Forget to hit the switch when you park and watch the battery die.) He hacked off the car's original console connectors and just hard-wired the Alpine. He hacked up the wiring big time.

So here I am, ordering a factory correct connector to restore the car's wiring. (Except it'll be an 8-pin replacement for the original 6-pin connector, adding relay switched stereo power and an instrument lighting lead (so the front display will dim at night when you turn the lights on).)

Maybe I'm over-doing the installation, to make up for equipping a Ferrari with a cheap Chinese stereo. (But in the US, it's the only thing I can get (other than a DVD player) that will play mp3s from flash memory --- no moving parts. And I can put 30 CD's worth of music on an SD flash card that will fit in the "glove box". )

And a lot of thinking reduces the amount of hacking. The switched circuit (from F18) is on a 1.5BN wire directly behind the console, the instrument panel lead is the 0.8G wire at the clock right at the console. (There's also a switched lead at the console for the oil temp and fuel level gauges, but all of the car's instruments run off that circuit, and it's only a 7.5A fuse. F18 is 20A and only runs the driver's window. Since they tapped F19 (pass window) for the rear defroster, it seems there's spare current in the F18 circuit.)

Ferrari "over-designed" this car. Three (four counting the stereo) connectors and two screws, and the entire center console (clock, oil temp, fuel level, stereo, aircon vents) comes right out of the car. Since Ferrari was clever enough to design it that way, I'd like to leave that feature intact. The 328 is chock full of really clever design features like that. (The door latch adjustment is beautiful.) It's one thing I like about owning one. It's "elegant" engineering.

Or maybe I'm just feeling challenged to match that level of elegance in the after-market install.

I'm not too confident in the life expectancy of a Radio Shack relay in continuous operation. Bosch relays run hot, but most of the ones in the car have been there for 20 years already.

(Okay, that cheap stereo probably won't last more than five years. But with a proper DIN mount in place, the radio bus relay installed in the car, and the right connectors installed, the next replacement should be a snap: pop out the console, take it inside, snap the new stereo into the mount, wire new pins on the car connector, then drop it back into the car. Two hours, tops -- including coffee break.)

I only have the one spare Bosch relay. (The others were left over from the alarm system I had removed when it disabled the starter.) I did look under the cover, and didn't see any obvious resistor, but that's not entirely conclusive.

Point taken. But for external suppression, you already know the polarity of the circuit. Using zeners in internal suppression is more expensive than just slapping in a resistor. But that's Bosch's problem, not mine. (Bosch does make some polarized "suppressed" relays.)

I was just wondering if an internal suppression resistor was why the relays got so warm, or whether it was just a low impedance coil.

 

The factory connector has leads for the speakers, ground, and a constant (unswitched) hot lead. No switched lead and no lighting. The three gauges just above the stereo panel are tied into instrument lighting and have a switched circuit. But *all* of the instruments in the car run off that circuit (Fuse 21, 7.5A).

With the relay drawing 2W by itself, I wanted to get the current from a circuit that had spare, and wouldn't impact much if the fuse went anyway. The wire to the driver's window is right behind the console in the wiring harness, but you have to poke around a bit to find it in that big bundle.

I've been brushing up on my basic field stuff. (It's been three decades since field theory class, and even then they mostly went straight to the Maxwell equations.) You can get the same magnetic flux on one fifth the current by running five times the turn density. But then you have *25* times the inductance (and 25 times the collapse voltage) -- so you'd really need suppression. You'd also reduce your time constant, so the relay would operate slower.

Going from iron to a permalloy or mu-metal core would also boost the flux density quite a lot, but I haven't seen the hysteresis plots for those alloys, so they might not be practical for relay coils.

For some of the relays, switching time would be significant (e.g. fuel pump safety circuit when starting). So I guess the high current isn't because of coil voltage suppression, but instead of -- and to make a faster relay.

Update: it also seems that Bosch sold their relay business to Tyco.

I've been carrying around the dead Alpine for a few years. But when the old (undocumented) alarm system went bad, the easiest fix was to tell the shop to yank anything the factory didn't install. Fixing up the stereo connector is just finishing the job. And my current commute has a lot of idling at traffic lights in it, so a working stereo wouldn't be superfluous. But the noise levels are why I'm not worried about the fidelity difference between CDs and MP3. And the Alpine died on a pothole, when it threw the CD at me. So running without moving parts seems like a good idea.

 

I was contemplating converting the CIS to a custom EFI system. But then it occurred to me -- who would fix it? I become interested in "originality" when it comes to documentation. This is the issue with a lot of after-market alarm systems. They use "security" as an excuse not to document the system. But then, when it goes bad, nobody can even find the parts, much less diagnose or repair it.

But mostly I'm getting the factory original connectors because they're a lot better than the stuff I can get at Radio Shack.

(I was going to use one of the 9 position round pin connectors, but even on the bench it's nearly impossible to make a clean crimp on the pins and have them lock into the shell properly. So trying to connectorize the bare wires in the car would be a mess. The factory connectors use standard quarter inch (6.3mm) blade "quick disconnects". It's the plastic shells that have to come from Europe.)

It's a DC circuit at steady state. But during the transients when the relay switches, it becomes a "step response" circuit, where "AC" type factors are significant. (Similarly, the edge effects become significant on a finite length solenoid.)

Technically, the term is "self impedance" of a solenoid coil. It affects switching time and the amount of energy stored in the coil at steady state (which is what produces the voltage spike when the current is suddenly switched off).

No worries. I know I don't know anything about the new alloys (iron or air where the choices when I studied Faraday and Coulomb). But permeability is a factor in the flux equations, and I just noticed, in my reading, that mu-metal (mostly used as magnetic shielding) has 100 times the permeability of iron. (There's probably some in your disk drive.)


Yes, in practice, I'll just be using what's on the market. But I wanted to brush up on the basic physics, and try to understand some of the design decisions behind the stuff that is on the market. (Yes, I am a nerd. )
At least I didn't design in an $800 hammer.

 

Since I have no children, when I back out of the garage, there probably won't be a tricycle sitting in the driveway.
But I still look.
(Only two or three times in almost four decades has there been a tricycle in the driveway. )

If I had the parts on hand, I probably would have plunked in the relay without worrying about it, too. But I have time to check. But less from a feeling that it's needed than from nerd curiosity. (And a vague feeling that a hot relay is "inelegant".)

Perhaps I should have called this thread, "Why do Bosch relays take five times the coil current as other types?"

And I think the answer relates to why you usually don't have to worry about coil voltage spike suppression with Bosch relays.

 

While a diode alone will clip your voltage to 0.7v, it can have a huge impact on the release profile of the relay, and, depending on current through the contacts, can even cause the relay to hang in the active/closed position.

It can benefit a normally closed contact relay, as the slower return speed will reduce contact bounce on the rest position stops.

A zener diode in series with the back biased diode (anode to anode) will allow a higher voltage surge, but will result in quicker release on normally open contacts.

However; it would appear that Bosch relays depend on high current through a low impedance coil, so they likely generate less voltage surge than a low current coil using higher turn density to achieve the same magnetic flux. (Flux is proportional to turn density; impedance (L=N(d(flux)/dI) (and emf = -L(di/dt)) is proportional to the square of the turn density.) So I'd be more worried about suppression on a 30mA radio shack relay than a 160mA Bosch one. (Although with lighter contacts and weaker springs, the RS relay probably runs lower total flux, too.)

Originally, I was wondering if the Bosch drew so much current because it had an internal suppression resistor across the coil, but it appears that the coil itself draws that much current.

Hopefully, I'll get the connectors in the next couple of days, and I can get back to installing instead of "over analyzing".
(But the refresher I plunged into on the coil physics was interesting. )

I'm still stuck using a self heating relay -- nobody seems to make low current relays in that form factor.

 

Interesting reading: http://relays.tycoelectronics.com/appnotes/

(I mentioned before that Tyco took over Bosch's relay business.)

 

Ask the manufacturer: http://relays.tycoelectronics.com/appnotes/app_pdfs/13c3264.pdf

This was intended to be a legitimate thread. When nobody else had the answer, I looked it up and supplied it for anyone that might be interested.

I didn't set the level that others are content to rest upon.
And I don't accept responsibility for any annoyance people may feel at someone else going the extra distance.

The herd doesn't set the standards. Each person sets his own.


The people who pull into the left lane to block the Ferrari could give any number of reasons why a Ferrari "isn't relevant" to driving.
Just a different set of standards.