The caveat to my solution is that the root cause of the failure is not addressed. If, as I suspect might be the case, the cause of the heat buildup is a lack of close engagement between the pins of the factory relay and the sockets in the fuse block, using a bad one as a plug, without improving the contact, may eventually lead to a failure elsewhere. When I did mine, I closed up the sockets in the fuse block with a small screwdriver to give a tighter grip on the mating pins.

It's just an under-rated relay. I installed a 40A generic one 4 years ago and no problems ever since.

In a fun twist of things...

So today I broke out some wire and connectors to make a temporary pigtail for a generic relay setup. I got the "harness" all nice and perfect. I then picked up my generic relay to check the pins against the OE green one and about punched myself making a facepalm....the generic relay has the same plug and fit right into the fuse box. I did take the cover off to compare it to the green one, and I found the construction to be a little less than optimal- like the wire for the coil being notably thinner.

Update- received the 2 new relays and neither are of solid construction...both are coils.

Took barely 2 weeks for one of the "new" ones to fail. When checking it out, I see the prong on the power circuit has melted the plastic where it meets the plastic housing. I know others have seen this too, but I'm surprised at this after such a short amount of time.

In another fun twist, I took the "old" one and plugged it in and it worked. WTF?

Pic of the "new" one with the melted plastic.

Looks like all the failures I had. I know how frustrating this is.

Try a JE16 relay. Supposed to be rated at 80a, and it's a direct drop-in part built for the Miata. Here's one for $9: Relay Denso 056700 8780 JE16 Used Warranty | eBay

Thanks Ian. I'll order that,but in the meantime I'll be swinging by the parts store to see if I can find a generic 40A+ relay to pop in there.

Found a generic 40A relay at AutoZone. Will hook it up to the harness I made and see if/how long it lasts. While trying to leave the parts store, the relay was acting up...letting the car starts, but dying 3-4 seconds after startup.

Dude you got something seriously wrong with your car. Every 1-2 months is frequent for these failures. Every few days? Serious short. Get your alternator checked at autozone or something.

Well, there was one relay that was old...like I don't know how old. That failed. I put a cheap Chinese foglight kit relay in there and it died quick...but to ve fair, before I put it in I checked its construction against the Imasan one and it was very noticeable that it was of lesser quality. Then when that died, I put the old one in and it worked again...but died later. Then put a new one in and that's the one that surprised me when it gave me an issue.

The odd thing is that ALL of them when "dead" still click/close the loop when the coil is powered up. Ive also noticed how hot they get when pulled after driving around...I wonder if its a heat issue causing them to distort in some way.

We'll see how the cheap 40A holds up while I wait on the above-linked Toyota 40A to get here. The only thing I've changed recently is the fuel pump, but that shouldn't be a problem.

But as I type that, I wonder- if my fuel filter was causing a resistance in flow and making the pump work harder (if it even does that), could the extra draw cause the circuit to run "hotter"? I know I've changed the filter not too long ago, but can't remember exactly when that was.

-double post

Bad connections can sometimes heat up causing issues. Corrosion is usually the culprit around here. Check the terminals of the fuel pump and the connector it plugs into.

Maybe you got some bad gas and it plugged the fuel filter?

I highly doubt that. I'd look at the relay terminals themselves, I be those are nasty after all these failed relays.

Bad connections should cause intermittent loss of power (maybe some arcing), not pull high current. But I'm not much for diagnosing electrical issues; just throwing out ideas.

Something is overloading those relays to cause them to heat up and fail. Would a bad motor in the pump cause it to draw more power?

Late to the party, but...

That is definitely not a solid-state relay, merely an enclosed mechanical relay.

Solid state relays, particularly those capable of dealing with significant DC currents, are both very large and very expensive. This is what a solid state relay rated for 20A DC looks like:

https://www.miataturbo.net/attachmen...1&d=1432212611

It measures 1.9" tall x 1.75" wide, and costs $65.73 from Digikey.



What you have in the picture there is a regular mechanical relay that happens to be enclosed in a plastic case, like this one:

https://www.miataturbo.net/attachmen...1&d=1432213737

That particular relay is rated for 30A, and costs $1.47 in quantity.


This is not a bad thing. Mechanical relays, all else being equal, are astoundingly reliable devices which are used in all manner of safety-rated applications ranging from spacecraft to nuclear reactors to passenger elevators every day.




There is absolutely no direct correlation between current-carrying capacity and whether a certain relay happens to be solid-state or mechanical.

At high loads, in fact, solid-state relays tend to produce considerably more heat than their mechanical counterparts. This is because, like all semiconductors, they have a relatively high on resistance. The unit I pictured above has an on resistance of 0.039Ω, meaning that at 20A, it will cause a voltage drop of 0.8v, and generate a massive 15 watts of heat. (for such applications, the relay is required to have a heat-sink attached.)

A typical mechanical relay, by comparison, will have a contact resistance which is about an order of magnitude lower (just a few milliohms) and will therefore generate far less heat in operation.



Again, we're speaking only to the physical relay itself, and not to the plug-in terminals on the bottom of the relay / the socket into which it plugs. From the pictures I've seen in this thread, it certainly appears that this is where the problem is occurring.




Unrelated, but cute; Here is a picture of a cat named Zener standing atop a giant wooden replica of an NE555 integrated circuit:

https://www.miataturbo.net/attachmen...1&d=1432213937

this this this this this

I dunno, maybe it's cause you guys have uber high mileage cars, but I've never ever had to replace any relay on any nb ever

Thanks for all the info/clarification Joe.

Curly- the terminals on the box look OK, there are some markings on the connector that has been overheating.


Here is the temporary 40A test rig...(I will be insulating/covering the exposed areas on the relay before installing)

<p><br /><br />I'm going to re-state and reinforce what Curly said, as I'm not certain it's been understood fully.<br /><br />1: It appears that the damage is concentrated mostly around the area of the contact point between the relay terminal and the socket into which it plugs.<br /><br />2: There doesn't appear to be any damage inside the relay itself, nor is it common for relays to generate a lot of heat internally.<br /><br />3: It is, however, fairly typical for heat to be generated at the interconnection point between two devices (such as between the terminals of a relay and a socket) when a large amount of current is required to flow across a marginal connection (one which is loose, corroded, undersized for the application, or at which insufficient contact area exists.)<br /><br />4: The male pins on the bottom of the relay are not readily deformed, and certainly not in a dimension that would be critical here.<br /><br />5: By contrast, it's entirely possible for the female socket terminals in the fusebox to spread out, which would contribute to the failure mode noted in point 3.<br /><br /><br /><br />So, with reference to the temporary test rig&nbsp;you've pictured:<br /><br />1: If the thickness of those crimp-on male terminals is greater than that of the terminals on the bottom of the melted relay, then this will improve the contact at the socket in the fusebox. If those female terminals are in any way degraded, however (spread out, corroded, etc), then this is at best a temporary fix.<br /><br />2: If the thickness of the aforementioned terminals is not significantly greater than those on the relay, I would expect heat to continue to be generated at the contact point.<br /><br /><br /><br /><br /><br /><strong>TL,DR: The relay itself isn't likely the root cause of the problem. </strong><br /><br /><br /><br /><br /><br /><img alt="" src="https://www.miataturbo.net/attachment.php?attachmentid=140207&amp;dateline=14 32235318" /></p>

I

To quote an earlier post of mine:

One other concern: With bad connections at the the main relay the ecu will see a lower voltage and ramps up the alternator voltage to compensate. It is already high.......

Apologies for resurrecting an old thread, but that seems helpful in this case. One suggestion in this thread was to use the NA's JE16 relay instead of the NB's B5B4. I've had quite a bit of trouble with both the JE16 and the B6S8 before them in my NA, which makes me think they're not better than the B5B4.

My car blew a relay yesterday and I'm looking for a replacement. I'm hoping to find something that won't keep failing. Any ideas?

For example, has anybody tried the NF02 relay from an FD RX-7? 93-95 Rx7 Main Relay (NF02-18-821A). The relay physically looks similar to the JE16 but I'm suspecting Mazda could have specified a higher rated relay in those more powerful cars. That would address Reverant's comment that the NA and NB relays are underrated and can't handle the loads we need.