Finally Dremel 395 fixed

Thread Starter


Joined Jul 18, 2012
It was not running unless I pushed down hard on the switch and then barely did.
I took it apart, brushes worn and I thought switch. Got new parts and nothing.

So looking at stator, the tiny copper wire ends where they connect to the terminals, were breaking, copper wire metal fatique.
Only fix was to pull little connector and solder on some new wire ends,
Not worth the expense of getting a new part for it.

Other thing, putting it together it rattled horribly, the stator was moving around. I cut a new shim from top of cereal box and removed the compressed rubber shim, now its smooth and quiet like it should be. A rubber shim from an inner tube bike tire would be better.

It maybe my old switch was ok, $18 at Amazon... But it does work better now on partial speeds. The new brushes even though claimed 395 compatible, were too wide in one dimension, had to sand down to fit.


Thread Starter


Joined Jul 18, 2012
To attach new ends, I bent a tiny hook shape into old and new wire, then they sort of stayed together while soldering. Not much wire to work with running to the coils.


Joined Sep 24, 2015
I had the same problem - the coil wire just being long enough to make contact. With expansion and contraction the connection can break, leaving the tool dead. If it were the outer wire that could be unwrapped a single turn that would be OK. But it's been my luck that it was the inner wire that went open.

dammit im mad is dammit im mad backwards


Joined Apr 11, 2010
Intermittent contact is a bitch.

I had one of the first Digital Equipment Corporation’s 2060 mainframe computers. It would suddenly freeze after operating for several hours. We first suspected electrical noise. Concomitant to the installation, a medical company moved in below us. There seemed to be a correlation between shutting down the X-Ray machines and the 2060 freezing up.

That hypothesis proved unfruitful.

I was amazed at how the engineers debugged the problem. Seems back then, the computer’s timing was done with a spiral of two PCB tracks. The engineer would measure the pulses along the track and when the timings were synchronized, would place a solder blob between the tracks where synchronization occurred.

It was testing the timings along the tracks that the true problem was discovered. There was one track which selected the clock source, that had a nearly invisible crack. When the board heated up, this track would separate, the system would switch to an external clock WHICH DIDN’T EXIST!

A drop of solder across the crack and the 2060 ran fine for months, if not years.