By Ross Kuhre (Stock 93 MX-5, C package, 130k mi)
If you have an idle problem with the unmodified NA 1.6L engine, this article is intended to help you understand how the ISC (Idle Speed Control) system should work and how it may fail (my experience). If you isolate your problem to the ISCV (Idle Speed Control Valve) this article may also help you find replacements or in special cases, make repairs.
If you are just curious about how the ISCV operates and is constructed you can jump to the VALVE OPERATION section at the end of the article.
Abbreviations & Acronyms: It was my intention to only define them the first time they are used.
The ISC system has two components exclusively dedicated to idle air / rpm control. Those are the ISCV and the AV (Air Valve). Other components play roles in the idle rpm, including the IAS (Idle Adjustment Screw), which is part of the throttle body. This screw may or may not be under a “Blanking Cap”. The IAS is a simple air bypass needle that controls air in a passage around the TB (Throttle Body) butterfly valve. This path is separate from the ISCV & AV paths.
The Miata Workshop Manual has the following description with this diagram. “To improve idle smoothness, the ISC system controls the intake air amount by regulating (the) amount of the by-pass air that passes through the throttle valve. This system consists of the air valve that functions only when the engine is cold, the ISC valve that works throughout the entire engine speed range, and the control system.”
The “control system” mentioned in this description is primarily the ECU (Engine Control Unit). Normally, with input signals from Ignition rpm, Temperature, Throttle Position, and Air Flow sensors, the ECU operates in a control loop and adjusts several engine components including the ISCV. A signal to the ISCV from the ECU controls the amount of bypass air into the intake manifold and thereby the idle rpm. It is possible for bad rpm and temperature sensor inputs to create bad ECU commands to the ISCV. However, when the DC (Diagnostic Connector) TEN & GND contacts are shorted together for Self Diagnosis and troubleshooting the ECU control loop is opened. With the control loop open, the ECU ignores inputs from Ignition rpm, Temperature and Throttle Position sensors and sends a fixed signal to the ISCV. This should produce a stable idle value for troubleshooting & fault isolation. (More on that later) Here’s a Picture of the complete ISCV.
The other dedicated ISC component is the AV (picture later). The AV is a “Dumb” device. It is simply a wax pellet driven thermostatic valve. When working correctly it gradually closes its idle air bypass ports to the IM (Intake Manifold) as the coolant flowing through it comes up to the full engine-operating temperature.
WARNING: THE READER IS ADVISED THERE ARE HAZARDS PRESENT WHILE DOING THESE PROCEDURES. THE READER TAKES FULL RESPONSIBILITY FOR DAMAGE OR INJURY TO ANY PERSON OR PROPERTY INVOLVED.
I suggest you read completely through each section before taking any action.
Here is a brief description of my failure symptoms. I suddenly began having erratic idle rpm behavior. The symptoms varied from dying at idle, idling around 2k rpm and anything in between including normal rpm. The symptoms did not often remain the same for long periods. I found that if I stopped the car, turned off the key and restarted the engine, the symptoms would often change. The engine ran normally at highway speeds. In the end, I discovered I had 2 problems.
1. I had a dirty (Gummy) ISCV. My ISCV had not been cleaned for at least 70K miles.
2. I also had an intermittent & varying high resistance failure at the ISCV Solenoid (approx 60 to 180 ohms). After I cleaned the ISCV the dying at idle symptom disappeared. At first, solving the grunge problem led me away from the ISCV, thinking I had eliminated any problems there. Furthermore, my ISCV passed the “Click” test, which you may read about elsewhere. Therefore, I consider the “Click” test inconclusive. After the ISCV was cleaned my idle rpm values were high, unstable or normal. The dying symptom disappeared.
Here is my suggested troubleshooting approach: To Start, I will assume you had a normally running and idling engine. Suddenly, for no apparent external reason, the warm engine will not idle consistently at the correct rpm (800-900 rpm). Fight the normal urge to immediately adjust the IAS. This action might temporarily seem to correct the problem but will probably only mask the real problem. Find the DC next to the engine, above the upper shock absorber mount on the driver’s side of the car. See the picture below.
Warm up the engine to operating temperature. This value has always been near mid scale on my Temperature Gauge. Turn off the heater & A/C blower controls inside the car. If you’ve had A/C trouble you should probably confirm the A/C clutch is also not engaged. On a pad, note your current tachometer idle rpm value. If it is unstable note the range of values. With the engine still running at idle, connect the TEN & GND contacts in the DC (14 awg solid copper wire worked for me). Note the 2nd tachometer rpm value. Do not make an adjustment of the IAS at this time. If your engine will run smoothly with the throttle held slightly open but will die or not idle at 800 rpm or higher, move to the Cleaning part of this article.
Shorting the TEN to GND contacts in the DC enables the Self Diagnostic feature of the ECU. If you are lucky you will have a blinking error code being displayed by your instrument panel Check Engine light. Below is the table from the Miata Workshop Manual to decipher what these blinking error codes mean. For example, if you have an open coil in your ISCV you should see 3 long blinks followed briefly by 4 short blinks. This is the 34 error code, indicating a problem with the ISCV. This code will repeat over and over. I had no error codes displayed. Note the remark at the bottom of the table that says you will need to remove the Negative battery terminal to get this code to disappear after you repair a failure. I found this was not always the case. For example, when I removed the ISCV connector creating a “simulated” Open ISCV solenoid or broken wire from the ECU to the ISCV, I only needed to reinstall the connector and the 34 error code disappeared. However, I did have to remove the Negative battery cable to clear a 09 code that I set by removing the connector on the Water Thermosensor. Technically, when I “opened” the Water Thermosensor, I got what appeared to be a 19 error code (One long pulse followed by 9 short pulses). I have no explanation for this apparent table discrepancy but the reader might be on the lookout for other code mysteries like this. Obviously, since I had created this “error” condition on purpose and there is not a 19 code in the table, I disregarded the discrepancy. There may be a logical explanation and some reader smarter than I will probably inform us about that, later on this site. Sorry, I don’t have a comprehensive list of which codes are cleared without removing the Negative battery cable. Try the easy way first then remove the Negative battery cable if necessary. By the way, it does need to be the Negative cable. NOTE: Be sure to have the Positive battery terminal well covered as you may be reaching across the battery with a ratchet & socket extension to get to the Negative terminal. This process creates the real risk of shorting the battery. That is a dangerous event. I covered the Positive terminal and then still chose to use a short 10mm open-end wrench that was not long enough to reach from terminal to terminal.
If you do not discover any error codes continue from here. Check for any cracked or loose vacuum tubing that is connected to the IM. Check both ends of the tubing or, if the IM end is good, pinch off the tube temporarily with pliers to see if a change in idle rpm occurs. A decrease in rpm means air was passing through that line. In my case, this process revealed I had a small leak in my PCV valve. Unfortunately, after replacing the existing PCV valve twice, I found the original one had the smallest leak of the three. So, for the time being I chose to leave it in and temporarily accept a 50-rpm increase in idle speed. This leak was obviously not my core problem.
Next, turn off the engine and eliminate the AV as the culprit. Once the engine is warm the AV should close and allow no bypass air through it, into the IM. There is a workshop manual procedure to remove the AV, Freeze it, mark its shaft, then Heat it & confirm the valve shaft moves and is closed when Hot. If you choose to do this you will need to deal with the coolant that is in the AV coolant lines. I simply pinched off these tubes with small clamps or a small smooth flat jaw vise grip pliers. If you want to perform this test, install these clamps just far enough away from the AV to give you space to move the hose clamps back off the AV metal coolant tubes. Put a rag under the AV, as you will have a small amount of coolant trapped in the AV when you remove it. Because I had intermittent symptoms, I chose to reinstall the AV with a DEG (Dead End Gasket) to totally eliminate the chance that the AV was occasionally sticking and allowing idle air into the intake system (see the picture). NOTE: If you let the engine cool with this gasket in place, you will need to manually hold the Throttle open until the engine warms to operating temperature if you re-start the engine.
You can fabricate this temporary DEG from common gasket material as I did. Install it between the AV and the intake manifold. When you reinstall the AV be sure the factory O-ring gasket stays on the AV side of the DEG. It is not necessary to disconnect the coolant lines to make and install this gasket, as only air passages are open when the AV is separated from the IM. The four 8 mm mounting bolts have a recommended torque of 43-69 in. lb. While you have the AV removed for either procedure, clean the air passages with throttle body cleaner. NOTE: Remove the rubber O-ring gasket before cleaning the AV. Some cleaners may distort the rubber and cause it not to fit correctly. If the gasket is in poor condition and possibly leaking, replace it. If you have not localized the problem continue.
Leave the TEN & GND jumper IN. Do not touch, tap or disturb any components under the hood. Restart the engine. Note a 3rd tachometer value. Stop the engine, pause about 10 seconds then start the engine again and note the 4th tachometer value. While the engine is still at idle, remove the TEN, GND jumper. Note the rpm value again. Now you should have 5 tachometer readings. If readings 2,3, & 4 are not equal and/or not between the 750 & 1K marks on the tachometer, it is very likely either the ISCV or AV (if you did not isolate it) is the culprit. The first remedy I would try is cleaning.
CLEANING: A first non-invasive cleaning can be attempted by spraying throttle body cleaner (I used Berry B12 carb. & TB cleaner) into the air supply tube to the ISCV with the engine running. It is possible to remove the front end of this tube near the throttle body and seal the black molded plastic intake air chamber port with a cork. (see picture)
When spraying the cleaner you may need to keep the engine from dying by hand controlling the throttle control on the side of the Throttle Body.
After this external cleaning, remove the cork and reinstall the air tube to the ISCV. Now redo the idle check outlined above. If the second set of values are still incorrect but not the same as set #1 (at the same Temp.), I suggest you remove the throttle body, remove the ISCV and clean them directly with throttle body cleaner. NOTE: When using the cleaner do not spray the rubber O-ring gasket. Some cleaners can distort this rubber so it will not be reusable. You will need to remove only the TB end of the black air chamber tubing that runs from the air cleaner to the TB. The rest of the air inlet tubing needs to be loosened but not totally removed. TB removal requires the removal of only 2 bolts and 2 nuts (12mm). A 3rd bolt (10mm) holds a piece of the air tubing below the TPS (Throttle Position Sensor) connector. Disconnect the throttle control cable from its attach point on the side of the TB. This is done by manually rotating the throttle wheel to full open (CW). Then, with the cable slacked, remove the cable end through the slot on the side of the throttle wheel. You also need to remove the inlet air tube that is connected to the ISCV. When you get the TB free remove the ISCV & TPS connectors. If the vacuum line from the Cruise control is installed you will probably want the TB end of it removed also. You will need to disconnect the coolant lines attached to the ISCV. You can simply pinch off these lines close to the ISCV (as detailed for the AV) or drain coolant from the system until it is below the ISCV level. In either case there will be some coolant to catch with a rag from the hoses and valve when you make the final separation. Separate the TB and ISCV by removing the 3 screws holding them together. Watch for the O-ring gasket and set it aside. While you have the TB removed clean it carefully without removing the factory seal material, which is around the rim of the butterfly valve and seat. Confirm the Butterfly is not sticking and is fully closed at rest. This might be a good time to confirm the dashpot (see first Diagram) is not preventing the throttle from closing completely. It should cause the throttle to slow when it first makes contact but not keep it open after that initial delay. If the throttle valve remains open and the dashpot is not to blame, look for and determine whether the throttle stop screw is at fault. If so, adjust it to touch the throttle stop plate at the zero open position. I used a strip of printer paper and set the stop screw so the drag just barely let me pull the strip out without tearing it. Clean the rotating valve area inside the ISCV and confirm the shutter valve rotates freely but returns to a position, open about 20%. (see pictures in VALVE OPERATION section) While you have the ISCV removed, use an ohmmeter to measure the resistance between the pins on the connector. The Miata Shop manual calls for a value of 12 +/- 1 ohm. Remember to adjust for the resistance of your meter leads. If you have a resistance of 0 or below 11 ohms you probably have a short between windings in the ISCV solenoid. Double & triple check yourself here to confirm you are not shorting your leads as you probe the connector pins or have a defective meter. If you do indeed have a shorted or low resistance reading, I think you should look for a replacement ISCV. (See REPLACEMENT section) If you have a reading higher than 13 ohms you probably have an open wire or high resistance connection in the solenoid. If you can, use test leads with alligator clips to connect your meter to the connector pins. (See picture in REPAIR section) While you have this static connection, tap the black body of the ISCV around the coil and connector with the plastic handle of a screwdriver. If you don’t have the alligator test leads, try to clamp the ISCV in a vise (carefully!) so you can have both hands free to accurately touch the probes to the pins when you take resistance readings. If your resistance reading changes after a tap, it is possible you have a repairable problem. (See Repair section) This is where I found my 2nd problem. I had an intermittent & varying high resistance reading of about 60 to 180 ohms. Your readings will likely not be the same as mine or may even be “Open or infinite”.
ADJUSTING THE IAS. If you have cleaned & checked the TB, the ISCV, and the AV per the preceding steps and have found and corrected a problem, I would now reassemble the parts & adjust the IAS to 850 +/- 50 rpm with the TEN & GND contacts shorted in the DC. Ideally the rpm will be the same (850 +/_ 50 rpm) with or without the TEN & GND jumper installed.
If you localize your problem to the ISCV as I did, you have several options.
1. You can search the net or contact dealers for new parts. I did a search on the net and found the cheapest new ISCV I could find cost $514.00. If this seems reasonable to you, order and install that ISCV. OH yeah, and send me your old one!!
2. You can also do a search of salvage yards on the net. I found several sources for a used ISCV. All the used sources I found sold the ISCV as part of the Throttle Body. This might be a way to pickup a spare TB if you are into stocking “spare” parts. The prices I found for this used part ranged from $75 to $250(+ shipping ). Some sites showed a 30 day guarantee on their parts. I would try to get this guarantee specified for the ISCV. Some places have a disclaimer that excludes Electrical parts and this would likely be in that category.
3. You can also go to local Salvage yards to search for a used unit. I could not find any used Miatas in my area salvage yards. However, I found a “You Pull It” yard which had an early 90s era (sorry exact yrs unk) Mazda Protégé with the engine dumped in the rear seat. Earlier, on a Miata.net forum I had read a post that said the ISCV was interchangeable between the NA Miata and some Protégés. As noted in that post the Protégé ISCV has different cooling tubes. I worked around this by buying new 5/16th ID tubing and running longer coolant line(s) to/from the ISCV. The used Protégé ISCV was available for $20 + tax. I think I got lucky!! Before I pulled the ISCV I took a resistance reading and found a solid 12 ohm +/- 1 ohm value at the connector. I suggest you also ask any mail order source of used parts to take this resistance measurement before you have them send the part. I pulled the used unit and also took the O-ring seal and the in/out coolant tubing. One of these coolant lines was useable in my Miata for the line between the ISCV & AV. The used unit I bought needed to be cleaned of Gummy carbon before it functioned properly. I installed the used unit. (ISCV to TB screws, 25-35 IN.-lb) & (TB to IM bolts, 14-19 ft-lb) The gasket between the TB & IM was broken and needed to be replaced. I found one at my local Auto Zone, P/N 60874 for $2.99. (I have no local Mazda dealer)
ADJUSTING THE IAS. If you have cleaned & checked the TB, the ISCV, and the AV per the preceding steps and/or have found and corrected a problem, I would now reassemble the parts in reverse order. Start the engine and warm it up. Adjust the IAS to 850 +/- 50 rpm with the TEN & GND contacts shorted in the DC. Ideally the rpm will be the same (850 +/_ 50 rpm) with or without the TEN & GND jumper installed. You should be ready to go zooming, unless you need to clear an error code or take out the trash!!
REPAIRS ??: Now the fun begins!
Having an “abnormal” curiosity and having repaired other potted modules in the past, I decided to attempt the repair of my “bad” ISCV. Later, I also decided to disassemble the unit to learn how it operated and was constructed. I already had a fully functioning replacement installed so there was no real pressure. Following is a description and pictures of that repair process. NOTE: This process can very easily result in an ISCV with an open resistance solenoid. This should not leave you any worse off than your failed condition. However, if you are not prepared to replace this unit with one from another source do not proceed. Furthermore, if you have any hope of using the ISCV again, do not disassemble it as I did. There are virtually no repairable parts inside the ISCV body. Instead, see the VALVE OPERATION section where I have already disassembled the valve for our enlightenment.
Remember, my failure symptoms were a varying high resistance condition and usually changed as a result of my tapping the ISCV solenoid module.
From past experience I suspected the problem might be a bad connection or broken wire at the rear of the male connector pins. Therefore, I decided to delicately grind into the rear of the connector, planning (ok, hoping) to locate the coil wire connections without destroying them. I used a Dremel tool with a 3/32nd spiral grinding bit.
The first thing I did was decide how far back from the edge of the connector to start grinding. I measured the inside depth (Popsicle stick pic) from the outside end of the connector edge to the bottom of the pins. This was about ½”. I added ¼” for connector pin potting material in the body and decided the rear of the pins were “probably” about ¾” from the outside edge of the connector. (Sorry, these pictures were taken after the process, as I did not know at the time I would write this article)
I was unsure how the wires exited the back of the connector pins so I started grinding about 1” back from the outside edge, taking small amounts of material at a time and watching closely for different material. I used a 10X magnifying visor for better visibility. This process takes patience and a little luck. When I found a cavity or change in material type, I took a small metal pick and removed bits of material to investigate before grinding further. After I exposed both wires, I was able to determine using the ohmmeter which one was at fault. I found the inboard “connector pin to wire” connection had failed. By grinding and picking my way down to the brass pin, I was able to clear space on the back of the pin and solder the coil wire back to the pin. This connection may have originally only been crimped. For that reason, I have a hunch many Miata ISCVs fail this same way.
This was the resistance reading after the repair was completed. Keep in mind I have about 1 ohm resistance in my test leads. Therefore, the actual reading is about 12.2 ohms.
All that was necessary to put this ISCV back in service was to seal the “disturbed” area with High Temp. Black Silicone sealant. The resulting ISCV is probably “Better Than New” as a used car salesman friend used to say. NOTE: There is no need to remove this Solenoid module from the body of the ISCV to attempt this repair. It is only removed in these pictures because I had removed it as part of the following article section.
VALVE OPERATION / CONSTRUCTION:
This valve is a rotary solenoid with a flat coil return spring. The armature of the solenoid is pressed onto a rotating shaft carried by two precision bearings. Between these two bearings is what I will call a cylinder segment plate or shutter. This shutter sits behind a matching curved window in the body of the ISCV. This shutter adjusts idle air/rpm by rotating clockwise (CW) (viewed from spring end of shaft) off its stop and metering bypass air into the IM. The (ECU) controls the current through the solenoid coil and thereby the CW rotational force applied to the valve shaft. The return spring opposes this CW rotating force with a CCW force. Where the two forces equalize is where the shutter sets in the window and determines the amount of idle air bypassing the TB butterfly valve. When no power is applied to the solenoid (connector removed) the shaft is rotated CCW by the return spring and the shutter rests against a stop screw. In this “no power” or at rest position the valve allows “approx” 20% of the air it would pass if fully open. This explains why removing the connector at idle normally causes an increase in idle rpm.
Here are a series of photos of the separate parts.
This is the ISCV body with all parts removed. The top of the body in this picture is where the solenoid module is nested and attached with 3 screws.
This is the ISCV shaft. From the left, the primary parts pressed or welded on the shaft are:
The highly magnetic solenoid armature cylinder, next, a short brass spacer, next, the small bearing, next, the shutter valve plate (spaced off the shaft), next, the larger bearing (labeled Japan NTN 635Z) and then a small brass ring pressed on the right end.
This is a view of the “bottom” or hidden side of the solenoid module, which mates to the metal valve body. You can see the coil location and the stator arms that bring the magnetic field from the coil to the armature core on the shaft. Barely visible in this picture is a rubber gasket that goes around the outer perimeter on this surface of the module and seals it against the metal valve body. NOTE: Three screws attach this module to the metal body. Only one of these screws is visible when the unit is fully assembled. The other two are hidden inside under the Return Spring. You can see where they pass through the metal stator bars in the picture above.
These are the major parts that make up the ISCV. They are oriented as they would be if they were assembled. NOTE: The white ring on the end of the valve shaft is the Return Spring. See the next picture for a better look at this spring when it is removed from the shaft. Also notice this view looks into the “Return Spring Set Screw Hole” on the outer circumference of the black solenoid module. I strongly suggest that this screw not be turned. There is surely a factory pre-set tension applied to the return spring. That correct pre-set tension will likely be lost if this screw is backed out even a few turns. That pre-set CCW tension would be critical to resist the CW rotation of the valve shaft when current is applied to the solenoid by the ECU.
This is the Return Spring and the white ring that is molded onto it’s outside end. The silver component is a return spring to shaft adapter that slides onto the valve shaft. Notice the offset slot is where the inside spring end sets when installed on the shaft. Also be advised there is an internally toothed spring washer that goes on the end of the shaft to hold the items in place.
Here are a few other Misc. views of the ISCV parts.
This view is looking into the end of the shaft hole where the larger shaft bearing seats on the side of the body opposite where the solenoid module sets. Also notice the Shutter valve Stop screw appearing in the left side of the hole just beyond the bearing seat. The shutter plate rests against this screw when no power is applied to the solenoid and the return spring pre-tension sends the valve to this location. Also notice visible on the other end of the shaft hole is the smaller hole where the shaft and armature exit the body. Remember that the smaller bearing is on the shaft on that side of the armature and essentially seals off that upper hole.
This view is looking into the two rectangular air outlet ports on the side of the ISCV that mate to the TB. These ports are isolated from one another and the coolant passages (far left) by the special O-ring gasket sandwiched between the ISCV and TB. In the back of the right port (the Warm idle port) you can see the shutter valve installed (see armature at top where solenoid module would be attached). The valve in this picture is resting against the stop screw out of sight on it’s right. Notice the valve is actually open in this position approx 20%. This is the no power, failed open coil, or connector removed, position of the valve. Note that from here the valve shutter can only move/rotate to the left (CW, from above) and must pass through the actual valve closed, no air position to get to the normal operating positions where it operates when driven by the ECU. If the engine dies at warm idle and the ISCV is dirty, the shutter is probably stuck in the closed area.
This picture shows the valve rotated CW off the stop screw and past the fully “closed” position. This position is much as it might be during normal operation.
The left air passage is the Cold idle port and leads through an isolated passage in the TB and IM to the AV and back into the IM, causing the idle to be higher during warm-up. This port is not controlled by the shutter valve but should only have air flowing through it until the AV warms to operating temperature and fully closes.
This is a shot of a few parts that were not shown or not clearly visible in other pictures. Upper left is the retaining spring washer that is installed next to the return spring shaft adapter on the valve shaft. Next, to the right is the spring to shaft adapter. This shot shows the offset slot where the inner return spring end sets. Also on the opposite surface is visible a larger center slot that mates with flat surfaces on the sides of the valve shaft at the spring end. Below that is the rubber perimeter gasket mentioned but barely visible on the solenoid module. To the left of that is the special O-ring gasket that is sandwiched between the ISCV and the TB. You can see that the air outlet passages (cold & warm) are isolated from one another.
What may be of interest is that the two left hand coolant ports are not isolated from one another. If you look above at the picture of the metal body you’ll see the coolant outlet tube is connected to the lower hole. The coolant comes through the valve body and exits at the upper hole. No isolation is apparently needed because the coolant only makes a “token” trip into the TB before immediately coming back into the ISCV and passing out the lower tube to the AV.
The written references I used for this article
were the 1993 Mazda MX-5 Wiring Diagram and the 1993 Mazda MX-5 Workshop Manual. I also want to thank Bill Strohm
as a considerable resource for answering questions during my troubleshooting
phase and during the writing of this article.
Also, there were numerous Miata.net forum posters that influenced my
successful outcome. Kudos to them all!
Comments from Jur:
The article on the ISCV is very clear and helpful. it helped me to get to the culprit of *very high* CO on mu 1991 NA 1.6. that was caused by clogged coolant lines that run through the ISCV and the cold idle valve. maybe you could add that as a point of interest. these cars are getting older and clogging of thes very thin lines might occur more and more.