So I am working on a 24 volt starter from an old International track type dozer that has been in working condition. To start out with when I hit the starter button nothing happened. I suspected a poor battery terminal connection but that turned out to not be the case. - What happens next is this: I hit the starter button, the starter takes off, but I did not have the throttle open (fuel is cut off) so I release the button but the starter keeps turning. Ugh! Not about to pull a battery cable off on a 24 volt system drawing this kind of current. After about 20 seconds it quits. I notice one of the battery terminals had dripped some molten lead off so I assume the connection is broken and I safely disconnect the batteries. I had wondered if low system voltage had caused the solenoid to stick so I go about repairing the battery cables and hook it back up also having replaced the starter push button switch and installed a battery disconnect. I try it again and sure enough the starter continues to run after I release the button. I quickly pull the disconnect and naturally the starter comes to a stop. But yet again when I reconnect the disconnect the starter does not turn and when I hit the starter button the same thing happens again. The starter turns and releasing the button does not stop the starter. It seems to be latching itself on. - Fast forward to having pulled the starter off the machine and disassembled it in the shop. If find the inside of the starter looks excellent. The brushes and commutator are in fine shape and everything is very clean. However, inside the starter solenoid there are some heavy contacts that are quite worn. For those who know, I don't need to explain how this works but for those that don't I will explain it. Solenoid current causes a magnetic field that pulls on a plunger which pulls a rocker arm which slides the gear ahead on the starter shaft that engages the ring gear on the flywheel. The other end of the solenoid pushes a conductive disk against two heavy contacts. One contact connects to the battery and the other connects to the starter motor itself. When the solenoid plunger moves it completes the circuit causing the starter motor to turn. - So I just so happen to have a pair or contacts on hand and the disk to replace the worn ones with. I had assumed that the old contacts were sticking as they were badly burned. Guess what? No change in operation. I tested it with a battery and pair of jumper cables. The starter latches itself on. - I had a little experience in the past with starter solenoids but kind of rusty on how they are wired. Here is the deal as far as I can remember: There are actually 2 coils. One is a holding coil and the other is a pull in coil. The pull in coil draws a fair amount of current and the holding coil considerably less. The holding coil is wired how you would expect it. One end of the coil is grounded and the other connects to the starter push button. If I recall correctly the pull in coil is wired one end to the starter push button and the other end to the starter motor terminal. In other words, the pull in coil is bypassed when the plunger pushes the conductive disk against the contacts that complete the circuit to the starting motor. Very clever. - So my mind is spinning now as to how this can work. Once the starter is engaged wouldn't there be feedback through both coils that would cause the thing to latch on? What is it that prevents this? Unless I am not remembering this correctly it would seem that how this starter is acting is exactly how I would expect it to yet I know there are millions of starters out there designed like this that work flawlessly. At this point my assumption is that there is a short between the two coils. Why the thing does not latch in normal operation I assume is a combination of cancelling fields and creative positioning of the coils. - @jim hardy this sort of thing seems to be right up your alley. Maybe not on an International TD20 dozer but I would guess you have had your share of starters apart. - Any ideas? Anyone?