Retrofitting a 1986 Maho MH400E

In a differential signal the two inputs are both float (none of them is connected to ground for noise reduction purposes). But with no signal earth connection between the two circuits the inputs may drift and one of them may rich the core voltage. When this happens the voltage stops to drift because the output circuit sets the difference between the two inputs, but the circuit wants to drift feather so we have an oscillation with the voltage between the two inputs, this oscillation the 7i77 think is a move signal so it in order to correct it is actually moves the axis.
This is my theory (hypothesis to be scientific correct) of the problem.

hanmon Thanks for the photos an your thoughts about tantalum caps. If I remember correctly there are also some tantalum caps used in the Indramat 3TRM servo drives. Using barrel wire termination looks great ,to me. I think it is the small details that really pay off in long time usage.
Am wondering what computer mother board and it's power supply that you are using?.
Also am interested in what you are planing for as a operator control panel?

Thanks for your 2 cents

---

Hope you continue
Bob

Here is something that can be done by using a slotting head.----
Internal Spiral Cutting www.practicalmachinist.com/vb/deckel-mah...iral-cutting-359503/

Bob

I got a start on the troubleshooting yesterday. Switched the EXE's back to original config. Y on the 19A1 EXE, and Z on the 602 EXE.

There are a few different rabbit holes to explore:
- EMI from the 5V chinese
- EMI from the 12V ITX PSU
- Failing capacitors in the Y EXE channel
- ??? Floating single ended signals?

I'll try and work through them systematically. First up - the capacitors.

After letting the 3D printer run unattended, where it got lonely and rebelled...


...finished wiring up the battery pack with 4x 2600mAh NiMH's. The are a few years old and only delivered 4.8V at idle, dropping to 3.8V powering the MAHO's 5V rail, measured at the exe. None of the exe's output a valid signal, and the MAHO E-stopped as soon as motion was commanded. If I need to move on to the EMI rabbit hole, I'll get some fresh batteries.

Switched out the battery pack for my hacked ATX PSU. It delivers 4.8V at idle, which is near the bottom of the Heidenhain spec. Once connected, X and Z were stable, and could be homed, but Y still E-stops as soon as motion is commanded.

The current draw of the whole 5V rail is 1.03A at start up. (Oops polarity reversed.)


Hanno, what sort of current rise would be expected in a failed capacitor case?

The next step will be to return the system to known start condition of using the chinese switching PSU, just to get Y working as before. For this the ammeter will be removed from the main rail supply point and wired to only measure the current to the Y EXE. Then run the circle drawing gcode, which draws one pattern, waits 15mins, draws the next , waits etc. And log the current flow during that test.

Then let it cool down again, and repeating the test for the 12V rail. This may be a likelier target If 5V supply to the 19A1 EXE is only passed through to power the encoder bulb, but the signal processing electronics are powered by the 12VDC.

Mark

J Green wrote: Here is something that can be done by using a slotting head.----
Internal Spiral Cutting www.practicalmachinist.com/vb/deckel-mah...iral-cutting-359503/

Bob

Another procedure which is about 1000x easier with a four axis CNC machine.

Hanno, good choice on adding the 7i74. Much cleaner and easy hook up.
Mark

Mark
If you have another Meter , you can also measure the (AC scale on meter ) ripple value.. As in DC voltage ripple .
Although the meter may not be sensitive enough for high frequency .
isn't it HanMon ? my bifocals must be clouded - yup that's it !

Bob