Re: [CAD_CAM_EDM_DRO] My ebay $750 Retrofit non standard repairs

Dan Mauch wrote:

> While disassembling the Y axis I noted there was a BEI incremental
>encoder mounted between the ballscrew and the servo motor. I removed it and
>tested it. There was no output from channel A, B or the index. I called BEI
>tech support and relayed what I had found. They asked me how much current it
>drew. It was 70 MA. They told me that was too low and I would either have to
>buy a new one for $400 or send it in for repairs and that repairs could cost
>up to $200 if repairable. After thinking about it for a while I decided
>that I could easily replace it with a US digital encoder for a lot less. But
>I wanted to do a few more checks on the BEI encoder. I check quite a few
>items and then realized that it must be the light source in the encoder. I
>cut one of the leads and sure enough the encoder light source was open. I
>assumed it was an IR LED. I removed it and noted that this little light was
>only .125D and about .187L. I looked at the circuitry and was surprised to
>see no dropping resistor or the LED. It was wired directly to the +5
>terminal and ground. Upon closer examination I could see that there was a
>portion of a filament inside the miniature bulb. It was an Incandescent
>Lamp!.
>

Whew, that really dates that machine!

>I went through my catalogs and found that Chicago Miniature makes hundreds
>of different incandescent miniature lamps some of which were 5 volts. I
>then was able to find that Mouser electronics carried this replacement bulb
>for $1.60 . I needed to order some other parts for various orders I had so I
>went ahead and ordered it. When it arrived I installed it and powered it up.
>I could see the light on and when I connected my scope I could see that all
>the channels were working properly. These little lamps come with different
>filaments so you can get them with 10,000 -40,000 hrs of life.
>

This is an OK temporary fix, but I'd recommend you get a good IR LED and see
if you can replace the bulb before putting the machine into service. It
WILL
go again at some very inopportune time!

>Failure analysis of the Y axis was probably caused by the encoder lamp
>burning out. This most likely caused the Y axis to run uncontrollably into
>the
>column causing the Y axis servo amp to burnout, blow the fuse and fractured
>the ball
>nut flange.
>

Doesn't sound like good design! It shouldn't do all that damage,
especially when
bulb burnout is a guaranteed experience. (Allen-Bradley had very
elaborate safety
precautions on their servo interface. If the bulb burned out, the
differential
signals from the encoder would all go to zero, and if any channel of the
encoder
gave a non-complementary signal for more than 1 microsecond, the encoder
card would command e-stop. It also generated a velocity magnitude
signal from
the encoder, anc compared it to the DC tachometer magnitude. If the tack
magnitude exceeded the encoder magnitude by more than 10%, it would
command E-stop. Of course, the A-B 7320 was used on machines with much
greater danger potential than a small mill, and a servo runaway on a 50 foot
bed mill could kill several people.

Jon