dgeesaman

iTrader: (7)

After reading about replacing the blown capacitors in Bose power amps, internet routers, etc I figured I'd open up my extra stock ECU for a look-see. It seemed like the 3k hesitation sort of followed this general pattern.

Interesting - one capacitor (C150, connects to ECU terminal 4F) shows up with 0.0Ohms resistance. Now I'm a mechanical engineer, not electrical, but from what I remember that shouldn't happen. 4F is the output which should go to the split air bypass solenoid. Some other caps tested out with a resistance of 300 to 3000Ohms, and a few with infinite resistance.

Am I testing these capacitors the right way? (My multimeter doesn't directly measure capacitance). What might normal be for capacitors like this? If I'm off-base let me know.

Comments welcome. Also, if you can identify what size/type cap to replace this with I'll go ahead and try it.

Dave

DaleClark

iTrader: (55)

Interesting theory! I do know from working on PC's that caps will many times swell when they head south.

From LONG time back on the big list, I remember people fighting the 3k issue with the local dealerships (when the car was still under warranty). I'm pretty sure ECU's were swapped out with brand-new ECU's from Mazda with no change.

I think it's a mix of a general programming problem with the stock ECU - maybe it's too sensitive to certain inputs. The PowerFC totally solves the 3k issue, part of the reason I bought one .

Dale

Gadd

That Is a ceramic capacitor and rarely go bad. to test them properly you need a "ESR" meter. but that one sould show infinite resistance. There is a way to test with a ohm meter but it's kinda crude. With the ohm meter at it's highest setting, touch both leads then reverse the leads and you should get a small "bump" on the meter.

dgeesaman

iTrader: (7)

Are you saying then that because it reads 0.0 Ohms on my DC multimeter, doesn't mean the capacitor is bad at all? My multimeter has a digital readout, so I doubt this bump would show up.

There are a couple other caps on that board of the same size/type that don't read 0.0.

Are there any other electronic components that might wear out like a capacitor can?

Dave

DamonB

In general capacitors have no or very low resistance. You're thinking of resistors. Capacitors store energy and typically function as a tiny battery.

You can't test that capacitor unless your Ohm meter has a setting to check capacitance (if you don't have an expensive meter you can bet yours doesn't ).

You also need to remove the capacitor from the board before you can test it because with the capacitor installed you have not isolated it from the other circuits on the board.

You weren't reading the caps, you were reading the contents of whatever circuit the cap you were probing was installed in. You have to isolate the cap first by removing it from the circuit. Just putting your meter on there is useless.

dgeesaman

iTrader: (7)

I thought about that possibility - if they are connected in parallel, it would exhibit that behavior. Which is why I asked.

FWIW, if I had bought one higher model I would have had capacitance capability.

Which brings the question: without removing them from the board, is there any way I can verify they are working? If I'm going to remove them, I might as well buy a complete set of replacements and replace all of the ones that might have deteriorated.

Do you have any suggestions how I might verify the components on the board?

Dave

DamonB

Ceramic capacitors are highly reliable. Unless the leads are broken or the cap is discolored or cracked you can be 99% sure it's a good cap. To be 100% sure you remove it and test it.

Ceramic caps are dirt cheap. You can replace every cap in there for $5 if you know where to get them but why bother? Personally I wouldn't even bother trying to troubleshoot that board unless I had a schematic because without a schematic I have no way of verifying how things work together. If I saw something obviously burnt or broken then sure, but just poking around in there without a schematic is pointless IMO.

dgeesaman

iTrader: (7)

I know in the case of the Bose amps and routers, it's those blue cylindrical electrolytic caps that went bad. They say the cheap ones dry out, and you can visibly see them deform at the top or bottom if they go bad. So the brown ceramic ones are probably all good - fine by me since there are so damn many of them.

I was figuring that it might be a good project on a snowy winter day. It would be really cool if I found that replacing a couple of them for $5 resulted in curing something like the 3k problem. And it seems plausible that since the 3k generally comes on gradually, is a transient problem, and occurs at the same RPM, that a bad cap would result in weak or delayed ECU signals. It's also the only thing I can think of that would be different between a brand new stock ECU and an old one that caused hesitation.

Since I already bought a fistful to try fixing my Bose amps in my Maxima, I might replace any on that ECU that I have spares for.

Dave

davesan1

iTrader: (4)

I agree with Damon concerning the need to remove the capacitor, (or at a minimum one lead of it), from the circuit board in order to make any reasonable measurement of either its resistance or its capacitance.

Two other comments-

1) Damon - Ref. your note "In general capacitors have no or very low resistance," I think you mean "In general, capacitors have no or very HIGH resistance." Just a clarification!

2) Dgeesaman - In my opinion, the possibility that a capacitor failure is in any way related to the famous hesitation problem is miniscule, at best! In real life, a capacitor that fails to a short circuit or a “low resistance” is more likely to cause major non-functional ECU problems, blown fuses, etc. A capacitor that fails to a "low capacitance" is very unusual, (especially the type of capacitor in your pic.), and, depending on the application, even then may or may not cause some sort of intermittent or permanent problem - tough to tell.

Concerning the hesitation problem, over a period of many years and two 3rd gens. I have tried to resolve this issue with no success. I have tried the various grounding fixes but there was no improvement. I am convinced that the problem is caused by a firmware fault in the ECU. I offer the following details to justify this belief. When I got my first 3rd gen. back in 1993, it did not have a hesitation problem. After several ECU swap-outs by the dealer over my first year of ownership, the hesitation suddenly started. I believe that during the swap-out process, Mazda had made changes to the firmware to improve the emission control issues that were causing problems back then, especially in CA. Over a period of 18 months around 1994 I had a long-lasting battle with the Mazda dealer and, eventually, with Mazda HQ in Irvine, but was never able to get Mazda to admit that a) there actually was a problem or b) that they had made a change that caused the problem. This was despite the fact that I was able to demonstrate to Mazda that if I disconnected the battery and allowed time for the on-ECU memory to lose its contents, on reconnecting the battery and restarting the car the hesitation would disappear. Then, over a short period of time, the situation would deteriorate and the hesitation would return. My conclusion was, (and still is), that the ECU has learning algorithm in its firmware that "adjusts" slightly to each individual engine, and that sometime in the first year of manufacture Mazda screwed up this firmware in a new release, causing it to "learn" incorrectly. I strongly believe that Mazda never addressed this problem. I am very familiar with how long it can take for designers back in Japan (generically) to learn of, understand, accept and eventually fix a non-critical or non-hazardous problem identified in a foreign market, (especially 10 years ago). I feel that by the time this problem got to the design team, (if ever), US 3rd gen. sales were stagnant and the issue was never addressed. The problem probably never existed in the European and domestic Japan marketplaces due to the much less stringent emission requirements in those areas at that time, so that there was no problem to address in the ongoing market after the last US version sales in 1995.

This having been said, in defence of the grounding fixes, it is quite possible that this firmware issue produces a marginal situation which could be pushed “over the edge” in either direction by changes to the grounding system of the engine electrical controls, and, in particular, the sensors. This would explain why, in some cases, the grounding fixes have been found to fix the problem.

The above, of course, just represents my analysis of my personal experiences, and could be wrong. The bottom line – I would still love to fix the darned problem!!!!

dgeesaman

iTrader: (7)

Thanks Davesan.

I realize the odds are pretty slim that this is 'the' answer that's been overlooked for years, but that's fine. I think I'll just replace any electrolytic caps that I have matching spares for, and leave all the ceramics alone.

From what I read (I wish I had an FD since 93), the hesitation usually begins after some mileage and never really goes away for good. Whether that's the firmware adapting or the electronics degrading, I dunno, but you gotta start with a theory, and I happened to be on the subject. I do wonder if your experience, which says that some ECUs have it, and others don't, might also be explained by a degradation problem. If they were swapping good ECUs and degraded ones, it would explain your black/white experiences. But in any case there are many variables involved, and without detailed design info nobody will probably figure it out.

Just another random thought: I couldn't find a place where the TPS signals for low-range and high-range switch over. If that takes place at 3k or thereabouts, it would make for some interesting possibilities.

Dave

davesan1

iTrader: (4)

Dgeesaman -

Hey - With the 7, all things are possible. In general, as long as you're very careful not to break the parts that are already working, it may worth a try.

A word about the capacitors -

The small ceramic ones in your picture are usually very reliable. In a (largely) digital design, i.e. the ECU, they are usually used for "decoupling" signals and power supplies. This gets rid of electrical noise that can be picked up in wiring harnesses and on +12 volt buses from the battery that get coupled in due to other things in the engine switching (usually high) currents elsewhere in the car and the engine. If these caps. go short circuit, (unlikely), the ECU will likely stop working in a significant way or a fuse will blow. If they go completely open circuit, (again unlikely), they will no longer perform their decoupling function and electrical noise could therefore be present on the relevant area of the 12-volt bus and/or the relevant sensor input signal(s). The result of this is largely indeterminate, but not generally desirable. However, the probability is very low, so I would not bother changing these components.

The larger "electrolytic" ones are less reliable, but still relatively rarely fail. In largely digital circuits, again, i.e., the ECU, they are largely used for "low frequency decoupling" of the 12 volt bus. They may also be used in other "analog type," (as opposed to pure digital), circuitry on the ECU board. A short circuit on one of these caps will blow fuses if it is on the 12 volt bus or almost certainly stop the ECU from working in a significant way if it is in analog circuitry. An open circuit on one of these caps may cause electrical "noise" issues if it is on the 12-volt bus or stop the ECU from working in a significant way if it is in analog circuitry. This type of capacitor can also have a (faulty) midway stage, where it shows a medium level resistance. This is not a good sign!

Particularly with Electrolytic caps, be very careful how you interpret meter measurement. With your meter on the ohms, (resistance), setting, when you initially connect the meter across the capacitor, you will see a low resistance which will then quickly increase steadily to a very high or "infinite" resistance within a few seconds. (This is because the battery in the meter charges up the capacitor as it does its measurement). The capacitor is now charged and when you remove the meter the capacitor may stay charged for 30 seconds or more. Reconnecting the meter during this time will give you confusing results, (likely will start at some medium level of resistance and then increase, as above. If you want to re-check the cap, it is less confusing if you first discharge the cap. by shorting out its terminals with a wire or screwdriver. Then you are starting from "ground zero" again. Also the behavior of the meter may be different depending on which way round the meter is connected to the cap terminals, in that the final resistance after the few seconds or so settling time may be higher one way round than the other. However, both ways round should be relatively high, (several tens of thousand ohms as a very general guideline).

Ceramics behave the came way as above, but they do not change behavior depending on the meter connection direction. Unfortunately, because of their relatively low capacitance compared with electrolytics, the initial low resistance phase is often very difficult to observe because it is of a very short duration.

Note that these measurements of capacitor performance are better observed with an (older style) analog meter, (with a real pointer that moves on a scale), as opposed to a digital meter. Also, at least one terminal of the capacitor must be disconnected from the board for any measurement to be meaningful

The bottom line here is that an electrolytic cap. can easily fool you into believing it is bad if you are not careful as to how you use the meter and interpret and interpret the results. The likely most reliable guide to a bad capacitor is visual inspection for "fluid" leakage around the terminal or case sealing area.

And lastly, be careful with any soldering. I'm sure you are aware of this , but don't use a very large soldering iron for soldering electrolytics. They don't like to get hot. Get the iron on and off the terminal as quickly as possible, within the constraints of a good solder joint.

For what it is worth, in my opinion, replacing the electrolytics has a significant risk of doing more harm than good, so be careful, and good luck

dgeesaman

iTrader: (7)

Well, it's a spare ECU, for one. They're worth about $40. And I intend to do a similar repair to replace the electrolytics on the Bose amps in my car, so I'll have some practice under my belt.

Thanks for the review of the cap measurement - it all makes sense but since I don't deal with this stuff every day it helps me a lot to know this stuff from an authoritative source. Rather than try to measure them, I think I'll try to replace them all (the electrolytics) if I can tolerate the work involved. Reading elsewhere on the internet, it seems there are many key performance factors in a capacitor, and some are more difficult to measure than others.

I bought some .075" soldering braid, 63/37 .03" No-clean solder, and a soldering sponge. I have a cheap iron that might be 20W. Would a 40W soldering iron be better?

Dave

davesan1

iTrader: (4)

20W is probably OK. More power risks overheating the cap. Practice on an off-board test set-up, (some old similar cap). The terminal on the cap. should allow you to melt solder on it after approx 1 to 2 sec. A bigger iron will melt it sooner but risks overheating the cap. A smaller iron requires longer to melt the solder, probably supplying more overall heat in the long run.

Good luck

Dave