[urq] Mac 02 repair - How to.

Steve Eiche seiche at shadetreesoftware.com
Mon Feb 26 13:47:45 EST 2007

FWIW, yes, I used to repair these and still could, but honestly, these 
are a PITA.  Not because of the ECUs per se, but because I end up being 
on the hook for troubleshooting other CIS issues, like vacuum leaks, 
pressure issues, sticky metering heads, etc.  Soooo, because I really 
never want to mess with another MAC-02, here you go:

First off, make sure that you actually have an ECU issue.  First make 
sure that you have NO vacuum leaks.  A pressure test is best, but the 
quick test is to pull the dipstick.  The engine should stumble, but 
continue to run.  Then pull the oil filler cap.  The engine should die. 
  If it doesn't you likely have a vacuum leak which MUST be fixed before 
you should mess with the mixture setting.  Often, a mixture setting that 
is not correct will cause the duty cycle of the valve to be out of 
range, perhaps because someone compensated for a vacuum leak by 
adjusting the mixture, and once the leak was fixed the mixture is too 
rich.  Sometimes the valve will run when the engine is cold, but once it 
goes open loop it will stop.  If this happens, your mixture may be out 
of range of the control.  If EVERYTHING else is OK - no leaks, correct 
fuel pressures, good O2 sensor, then you probably have a failed ECU.  It 
is pretty common.

Parts for MAC-02 ECUs.  First off, IME, the actual output driver usually 
doesn't fail.  The small signal transistors that drive it usually do. 
These are components Q8 and Q6 on the board.  I believe that one handles 
the warm up and the other operates with the Lambda regulation, but I 
don't have the schematic in front of me.  You should replace both of 
them.  These are standard NPN transistors, but I bought a bunch of the 
actual 2SC458s just because I like to keep it exactly as is was.  I 
bought them from www.bdent.com years ago, but you could also get them 
from matelectronics.com.   You can use something like a 2N2222 as long 
as you make sure you have the collector, base and emitter in the right 
place as the 2N2222 may not have the same pinout as the 2SC458.  Note 
that the PCB is marked.  As for the main driver Q7, which is a 2SD1087, 
they are long gone.  I have seen a few places that list them, but once 
you try to order them they cannot be had.  I don't see any reason that 
you couldn't substitute another injector driver, like those used in more 
modern engine computers, but I haven't had the need.  Anyway, with all 
of the engine swaps going on, there are a lot of good MAC-02s floating 

As for other failure items, #2 on the list is the pressure transducer, 
which have been NLA for about 20 years.  The most common failure mode is 
  0V output, in which case the timing is very advanced (the ECU thinks 
that the engine is under full vacuum - no load) and will ping badly 
under boost.  IMO, this is often caused by resistor or diode mods put on 
the output of the sensor, which puts too much load on the output, which 
eventually fails.  They can also fail with a full scale 5V+ output, 
which will cause the engine to start but the fuel pump will cut as the 
ECU thinks that there is too much MAP (overboost).  These are simply 2 
bar MAP sensors with a 0-5V output, but an 8V supply.  You can 
substitute the MAP sensor from a later MAC-1X ECU, just grab 5V for the 
supply off the the PCB instead of the original 8V supply.  The range, 
offset and linearity of these are close enough that you won't notice 
operational issues.  I think that a 2 bar Bosch MAP, like from '91 200q 
which I have a TON of, would also work, but the sensors from the MAC-1X 
ECUs are nice because you can use the vacuum lines and fitting from the 
MAC-1X box in the MAC-02.

Some tips on soldering on these old boards.  Number 1, do NOT use solder 
wick to remove components.  This puts too much heat onto these delicate 
old boards (or ANY PCB IMO), and WILL result in lifted pads, etc.  Use a 
vacuum type desoldering tool and add new solder and flux before you suck 
out the old solder. Wiggle the pins a bit with the tip of the iron as 
you heat it to make sure that the pins are free.  Number 2, If you don't 
get all the solder the first time, add more solder and do it again. 
Minimizing the amount of time you heat the PCB is critical.  You should 
not be touching the PCB with the iron for more than 3-4 seconds. If done 
properly, the component will simply lift off of the board without 
needing to be pried off.  If you are prying any harder than is necessary 
to free the chip from the conformal coat, you will likely damage through 

EPROMs.  The ancient 2532s are long NLA.  You can find them from 
electronics places that have parts for old video games, but you may need 
to buy a few to find one good one.  These generally aren't as robust as 
new IC designs.  If you really want to, you could make an adapter for a 
more modern EPROM, like a 2764 fairly easily.  The pinouts for both 
EPROMs are easy to find on the web.

Sockets.  If you don't want problems, use ONLY machined pin sockets with 
gold contacts.  ZIF sockets are not suitable for vehicles, or any 
equipment that vibrates.  If I see these in ECUs that I work on, they 
are the first things that I remove.  Personally, I can't think of 
anything other than a programmer that I would use a ZIF socket in.

OK, that is about all there is to tell.



Mike wrote:

I believe Steve Eiche was at one point and may still be?

regards, Mike

Jeff wrote:
I have a Mac 02 that I suspect has issues with the frequency valve 
driver circuitry.  I have heard of people repairing these in the past; 
anybody still doing so?

Anyone have a known good unit they'd care to sell?


More information about the urq mailing list