Old heidenhain scale ratio problem

hi all,

Maybe you clever people can help. I've just got my hands on a well used but nice Bridgeport with a heidenhain 731b DRO on it. Problem is, the x axis readout is 25% out. I.e if I move the bed 10mm it reads as 12.5mm. The y axis is fine and if I swap the inputs for the scales, the fault remains on the x axis, so it's definately not the scales at fault.
I've had a look inside and can't see a pot or any switches that might be used to alter the scaling factor......and run out of ideas. Anyone got any thoughts?

What is your location?

Long way from you
In the UK.

You can still download a manual for it off the H-H site.

Manuals | HEIDENHAIN

Click on "Counters" then scroll down to 731B.I just had a look, it doesn't mention anything about having an adjustable scaling factor.

Regards Phil.

DRO stands for DIGITAL Read Out. I am deliberately putting the emphasis on the word "digital". Generally, digital devices do not have adjustment pots as they are most useful in analog circuits. The adjustment, if any is going to be in the software so look in any menus that it may have. chances are it lost the calibration factor when power was off for an extended period of time. You probably want to enter a factor of 0.80 or 80 percent. (12.5mm x .8 = 10mm).

If there is no software adjustment factor, then you may have to use some gears or timing style pulleys. Probably not a very practical type of solution. Post some photos if this is the case.

Thanks for the replies....I should have put in the original post that I'd checked the manual already with no Joy.

EPAIII, unfortunately I think it's too old to have any software option, (unless someone knows differently!) there isn't even an option to set a cf from the front panel. I guess you may be right with the extended down time it had. I forgot to mention the x axis didn't work at all when I first got it, I traced that problem to the mm/inch switch which I swapped out. Could it be related to that? I see what you mean about gearing it down, but seems a bit drastic, presumably it used to work fine, I'd love to resurrect it if I can.

If it is a downtime related issue, there must have been a method to calibrate the machine originally? Could it be a problem with a dried out cap or ancient resistor?

Thanks again for any ideas!

CHeck the wiring on the inch mm switch it seems the only answer

I recall the difference between reading an old 1000 line per inch scale and a modern metric one was about 25 percent

like your thinking.....I took a photo before took it out to make sure I rewired the same way, but maybe somethings gone awry where it attaches to the pcb. Stay tuned for more DOR action, if you can stand the excitement!

Set the switch to Inch and see what it reads
Also on HH DRO`s do not fit all HH scales
You have scales and DRO`s for 40 and 20 Mu scalesteps
But if thats wrong it eighter dubbles or halfs the numbers
Not 25% off

Peter from holland

Thanks Peter,

The problem only occurs on the x axis readout, both sliders work correctly when connected to the y axis, so it isn't related to incompatible scales. Interestingly it works correctly when set to inches......if only I worked in Imperial

gustafson said:

CHeck the wiring on the inch mm switch it seems the only answer

I recall the difference between reading an old 1000 line per inch scale and a modern metric one was about 25 percent

Click to expand...

I still think this is the right area, but having checked the switch and wiring it all seems as it should be......hmmmm.

So it must be related to the inch mm conversion if it works in one mode and not the other

perhaps the switch logic is incorrect

I can't help with the specifics of the problem, but used to be in the electronics industry a lifetime ago and have repaired this model, as mine was similarly not working on one channel when I bought the mill it's used on.

Just working through it logically I'd suggest it's unlikely to be the switch, if one channel is working correctly and the other is not. Without seeing the schematic (unavailable as far as I'm aware), the switch couldn't be ruled out completely as you said it was affecting the channel originally, but I doubt it.

Some good news and some bad news. The bad news you already have, the schematic isn't available, or at least nobody would give it to me. The good news is that the logic inside is all very basic with discrete logic chips that are easy to obtain and understand. If you can find an old-school electronics person retired or with otherwise copious amounts of free time, it's entirely possible to work through the circuit without a schematic. It takes patience, but the discreet chips make it possible to follow the logic. In my case I was able to compare a working channel to one that wasn't working and using the oscilloscope find the troublesome IC. In your case it won't be as simple as the channel is working, it's just not dividing correctly. I would go through the entire scale (if you haven't already done so) and confirm that the error is exactly 25%. As Machtool said, I'm not aware of any scaling corrections available, and as far as Im aware WYSIWYG. If the scale isn't linear after install, you need to install it better, as they're linear when they leave the factory.

If it all gets too hard, before tipping too much money in to it, the new ND780 is available as a replacement and will use those scales without any issue, you will just need to either re-wire the plugs or buy adaptors (the latter are a bit of a rip-off in my opinion).

Good luck!

Some more thoughts.

I also doubt that it would be that switch if one scale (the Y) works correctly.

Is the X scale the original or has it been replaced? If it was designed to work with one scale and another one was substituted, the calibration may not be correct.

Is it displaying EXACTLY 12.500 mm for 10.000 mm of travel? If so, then you are getting five counts where there should be just four. If this is so, perhaps a small circuit could be designed to drop one count out of every five. A PIC device should be good for this. If you can verify this exact error, then try to find a schematic I can look at and I will see what can be done.

EPAIII said:

Is the X scale the original or has it been replaced? If it was designed to work with one scale and another one was substituted, the calibration may not be correct.

Click to expand...

darkhorse0 said:

The problem only occurs on the x axis readout, both sliders work correctly when connected to the y axis

Click to expand...

The scales working fine when connected to the other readout.

machtool said:

The scales working fine when connected to the other readout.

Click to expand...

Agree, I don't understand why people immediately seek a bodge way to "repair" something rather than relying on basic logic. Gears, drop in boards, etc Unless there's clear evidence that some Einstein has been there before, the probability is that it was working before, hence there's a fault with it, and the solution is to find the fault. Not always easy of course, but that's why technicians get paid the big bucks! ... or not

Pete F said:

hence there's a fault with it, and the solution is to find the fault.

Click to expand...

I've got the solution. We rework all his engineering drawings so they are 25% smaller in X. Should work a charm.

Thanks all for your time and input.....I've searched for a schematic to no avail, as it looks like some of you have. Pete F has the solution I think, I need an old fart with a scope! Sadly a new readout is beyond the budget, so I guess for now I'll be old school with the dials. Or rewire myself to work in imperial......incidentally, would the fact that it works correctly in imperial help someone with the knowledge track down the dodgy chip?

Another thought....the electronic side of things is above my pay grade, but I can solder. Most of the chips seem to be available at negligible cost, is it worth just swapping them all out? Are any of these more likely to be the culprit than any others?
M9102 I assume some kind of buffer
LM339 Quad comparator
HEF4081BP Quad 2-input AND-gate
HEF4013BP Dual D-type flip-flop
HEF40174BP Hex D-type flip-flop
SCL4070BC ???
HEF4027BP Dual J-K flip-flop
HEF4011BP Quad 2-input NAND gate
SCM10007P ???
HEF4001BP Quad 2-input NOR gate
HEF4516BP Binary up/down counter
SCM1010BP ???

Thanks in advance..

Yeah they're all cheap as chips. Ha ha get it! Groan!

Rather than swapping them all out, just swap out the ones on that X channel. Not a terribly scientific approach but it may get you there. However there's no guarantee it is actually an IC. Again, have you checked the full travel and confirmed the error is both consistent and truly the ratio you quoted? That's not a moot point BTW and will help track down the culprit. If it's exactly that ratio it implies to me a logic fault, if it's a random error of about that, it could be anything. The guts of these things are pretty simple and they basically just count the pulses coming off the scale and convert them to a useful output.

One trick of the trade, when you're really up shit-creek with a busted paddle and NFI of where you're going, is to swap the components between a good and bad channel. I can't say I've ever done a whole board that way, but hey, I guess you have little else to go on so why not! I don't recall if that board is double sided, but if not, and you think you're up to it, you could swap the ICs between the channels until you see the fault swap. You're probably going to be investing a good few days in to this, so it's not going to be a quick exercise.

While you say you can solder, I've seen various interpretations of that statement too! Getting components out without damaging either them or the board is however not a trivial exercise for those not normally working in the field. Professionally we'd use vacuum pump desoldering stations (or I have a hand held version for personal use, still very expensive however). I'm going to go out on a limb and guess you don't have one. If you're simply replacing ICs the quickest way to do so is generally to snip off the leads, heat up the pads and pull the leads out. You can then easily desolder the through plated hole and have everything ready for the replacement. You won't have that option if swapping however so you want to give all this some very serious thought before diving in.