HOME
What's New

 

Workshop Stuff

Workshop

Moving the mill

Spindle Noises

ShumaTech Digital Readout

ShumaTech DRO Continued

ShumaTech
DRO-350 Repairs


South Bend 9" Lathe

South Bend 405 Lathe Bench

Grizzly Mill Revisited

Surface Grinder Rebuild

Surface Grinder Continued

Grinder April 6, 2008

Grinder April 20, 2008

Grinder August, 2008

Grinder September, 2008

Grinder November, 2008

Grizzly G3103 Mill
Rebuild

Grizzly G3103 Mill
Rebuild - Part 2

Moving the Shop

Moving the Shop 2

Bringing Home a Sheldon 12" Shaper

Sheldon 12" Shaper 2

Sheldon 12" Shaper 3

Sheldon 12" Shaper 4

Sheldon 12" Shaper 5

Sheldon 12" Shaper 6

Sheldon 12" Shaper 7

Sheldon 12" Shaper 8

Sheldon 12" Shaper 9

Way Alignment Tool

 

 

Email Jim


 workshop.gif

Shumatech DRO-350 Repairs
July 4, 2009
When I purchased my DRO back in 2006, I also purchased a spare.  I was planning on putting it on a lathe, when one arrived.  I built and tested the second PC  board, but never got around to installing it on the lathe.  Fast forward to May 2009 when I purchased two more Jenix scales for the mill.  After three years of running the Chinese scales on the Y and Z axes, I decided to go all glass quadrature scales.  The mill's current Y scale, which was a 6" HF caliper had developed a bit of play in the gibs and needed to be tapped a couple times to center it or it would read a couple thousandths of an inch plus or minus the actual coordinate.  I could have probably pulled the scale apart and cleaned and lubed it, but I have grown fond of the Jenix I'm running on the X axis and the fact that it is always repeatable within its specifications. I consider the Jenix a step or two up from the Chinese capacitance scales.

The DRO had developed one problem in its display.  The center segment (labeled "G" on the display diagrams) of one of the Z readouts had stopped illuminating.  I planned to take care of this while I set up the box to handle another two Jenix scales.

Once I got inside the DRO and started tracking down the issue with the center segment, I found that the "G" segment of DS8A had burned out.  This proved to be a bit more of a problem than I expected as the MAN6940 Seven Segment LED Display is no longer a stocked item at my usual electronics suppliers.  I ended up ordering 10 replacements (630-HDSP-523E from Mouser), in case I wasn't pleased with just replacing the one with the problem.  I figured I'd just pull out the other DRO-350 board and swap it with the spare.  This would give me plenty of time to replace as many of the displays as I needed to get a nice looking read out.

What I didn't count on is that I would be careless and blow up the second board.  Yep, that's right, I plugged in the power plug to JP6 reversed.  As soon as I realized what I'd done, I knew that there would be problems.  There were.  I now had missing segments on the X display and nothing on the Y.  Nothing.  No r4 message at boot up, no no read out from the scale.  No nothing.  Well, that's not entirely true.  If I jumped across the number 12 and 13 pins of U10, I got the decimal points to light up.  However, that doesn't tell me enough to fix it.  Only the Z axis showed what it was supposed to, "nnill."

stech10.jpg
Tools used for the repairs. Digital volt/ohm meter, logic probe, circuit board vise and soldering station. My anti-static strap hangs from the clip, on the right.

I was in a heap of trouble.  I don't have a oscilloscope.  All I have is a decent DVOM and a logic probe.  The one thing I had going for me is that I had a working board to compare the blown one to.  That and all of the great posts by Lester and the rest of the guys who really know the DRO-350 on the Shumatech Yahoo group.  I had also purchased extra of all the parts necessary to build the DRO-350.

As I read through the posts, I found that I wasn't the only person to try and repair a DRO-350 without a scope.  It seems that without a scope, it makes the process of deducing whether a part is good or bad a bit harder.  Harder, yes.  Impossible, no - and since I am sure that I am not the only person to fry a board, I thought that sharing my method of repairing mine might help someone else in a similar situation.

The first thing to know about the display circuit is that each of the 3 LED displays per axis is actually two displays. Each of the 6 displays (2 X 3) is comprised of 8 anode segments (7 bars that make up the parts of the number plus the dot) and a cathode.  Each segment is given a letter designation. Letter "a" is the top horizontal bar, "g" is the center bar and "dp" is the dot. Pins 13 and 14 of each double digit display are the two cathodes.  Aside from the two cathodes which selectively complete the circuit though the ULN2803A chips , each of the other 8 segments in each half of the display are connected to every other same letter on the other 5 half displays.  This means that all Y axis "a" segments are on the same circuit and you should be able to use your continuity tester to show a completed circuit between all of the "a" segments.  This would include three number 11 pins and three number 16 pins of DS7A/B DS8 A/B and DS9 A/B.  If one pin lacks continuity, look for a broken trace on the board or a bad solder joint. 

Each segment is lit by a circuit completed from the plus side (+9 volts) of the wall wart flowing ultimately through one of the 150 Ohm resistors in R12, R13 or R14, then to  the lettered segment, then out the cathode on pin 13 or 14 and returning to the minus (-) side of the wall wart.  Between the resistors and the PIC, there are the display driving ICs that dictate when the segment will be lit. On the anode side, we start at the PIC microcontroller on pins  11 through 18, then to the 2 through 9 pins of the M74HC573B1R IC and out the opposite side on the 12 through 19 pins, then to input pins (1-8)UDN2983A and out through (9-16), then to the resistor pack. If you look at the circuit diagram, you'll see that the signal enters on one side and goes out the opposite side - straight across the IC.  Once you get to the display units, you need to read the pin numbers on the schematic as the representations of the displays are not sequentially numbered like the ICs.

On the cathode side of each display, the ending points are pins 13 and 14 of each half display chip.  Starting from the PIC on pins 5, 6, and 7, the signal goes to the 1, 2, and 3 pins of the M74HC238B1R IC, then out on the 9 through 15 pins and on to the 1 through 7 pins of the ULN2803A and out the13 through 18 pins to complete the cathode circuit on pins 13 and 14 of each display.  Confused?  Trace it on the display and microprocessor circuit diagrams which will make it easier to follow.

To add one more layer of complexity to diagnosing the display, it is multiplexed.  The data 0 through 7 lines carry the display data and the column 0 through 7 lines complete the circuit one column (each half display) at a time.  If I view the display through the viewfinder screen of my digital camera, I can see each column of the X, Y, and Z displays light brightly, one column in all three rows at a time, across the 6 columns of the display and seventh column, the indicator LEDs. With each column lighting for a fraction of a second, it tricks the eyes into seeing a solidly lit display.  Photos of the display sort of show this, but at the shutter speed I was using, it shows as two columns bright and the remainder a bit dimmer.

stech7.jpg
stech8.jpg
In this picture, we see the "ones" and minus sign columns lit.
In this image, we see the hundredths and thousandths (and 0.0005" dot) lit.
Each column is lit in sequence.  My camera's shutter speed is such that it is capturing about two columns in a bright, or power on, state and the rest of the columns are dimming. They will be refreshed on the next cycle.  The effect is more noticeable on the larger images.

I started with the PC board out of the box and held up so that the buttons would not get pressed accidentally. The only thing plugged into the board is power at JP6. Once plugged in, the boot message appears. In my case, only the top and bottom lines were lit and the characters in the X display were missing some segments. I left the boot up message on the screen and began my tests.

I looked at the schematic and found the VCC and grounds for each of the ICs and tested these with a volt meter.  With those are reading as they should, I grabbed my logic probe and started at the number 11 pin of the PIC (Data0). I checked it for a pulse. Good. Then on to pin 2 of U7, the M74HC573B1R on the X axis of my mill display. This display was mostly working, but missing the top center (a) segment all the way across the display. I got a pulse at the 2 pin, but only a low logic signal with no pulse on the output pin 19.  This was telling me that I had at least one bad circuit inside this IC and it needed to be replaced.  I then checked the input and output of U10. This is the same M74HC573B1R IC, but on the Y axis. Pin 2 showed a pulse, but pin 19 showed nothing. I checked it with the DVOM and it was reading a couple microvolts.  The Y display was entirely blank, so I wasn't surprised.  On to the working Z display's M74HC573B1R IC. Pulsed input, pulsed output. One of the three ICs checked out correctly, but this was only the first pin check.  With the information I now had, I knew that I had two bad ICs, but I went ahead and checked each circuit from the PIC to the output side of the 3 M74HC573B1R ICs.  The remainder of the pins on the X axis M74HC573B1R IC showed good.  The IC on the Y axis had 0 pin pairs that showed a pulse on both sides of the IC and the IC on the Z axis showed all pins good with a pulse showing on both the input and output side of the IC.

It really didn't make much sense to trying and follow the circuits to their next step until I fixed what I knew (or thought I knew) was wrong.  To remove the two ICs, I used a hobby knife with a newly sharpened #11 blade.  I placed the side of the blade flat against the IC with the sharp edge between the first two pins of the IC.  By rocking the knife toward, then away from the pins, I used pin 2 as a fulcrum to cut pin 1.  Then pin 3 as the fulcrum to cut pin 2.  If I had a pair of diagonal cutters small enough to fit between the pins, I could have used those, but my smallest pair of cutters were too big for the job.  Once I had cut all but the two end pins from the IC, I clipped the last pins with my diagonal cutters and removed the plastic portion of the IC.

stech1.jpg
stech4.jpg
The leads for U10 have been clipped and the IC is ready to be removed.
The new IC sits on top of the old IC. The new IC has its pins pretinned.

I then used my soldering iron and a pair of tweezers to remove the pins from the circuit board.  The next step was to use some rosin coated desoldering braid to soak up most of the solder and clear the holes.  The last step was to wash the area off with isopropyl alcohol and dry it with my heat gun.  Soldering the new chips in was a bit tougher than it was when the displays weren't on the board.  I made this a bit easier by tinning each of the legs of the IC before inserting it into the board.  I also left a little "blob" of solder at the top of each leg so that a quick touch with the iron would melt it and flow down the pin.  I think this helped to get enough solder on to each pin quickly and avoid overheating the IC.  With some talk about the possibility of the through-holes not being connected from top to bottom, I wanted enough solder to solidly connect each pint to both the top and bottom of the circuit board.  Also, be wary of your soldering iron tip.  It is very easy to touch the iron to the displays and produce melt marks. I put a piece of aluminum duct tape over the sides of the displays to help me avoid making a mess of the plastic. I ended up with only one small burn that won't show when the case is back in place, so I consider myself lucky.

I plugged power back into the board.  This time I had all displays working for the boot up message.  I was quite pleased with myself until I plugged in the scales.  The second from the right half display on the Y axis (DS6A) wasn't showing anything. Nuts!  Looking at the schematic, it appeared that the only thing that could cause this would be the U12 ULN2803A IC wasn't getting a ground to pin 14.  I confirmed this by checking the input #2 pin (pulsed) and output on pin #17 showing nothing with the logic probe and microvolts with the DVOM.

I replaced the ULN2803A IC and rechecked the display.  Everything worked with my Chinese capacitance scales.  However, I could only check 2 of the axes at once as I had been running a Jenix on the X and it was still on the mill on the other side of the room.  I did have a 12" Chinese caliper that I could wire up, but I still wanted it to be able to be used as a free-standing caliper.  I decided to add a 0.1 µF and a 100 µF cap in the battery compartment, then add a MTA-100 4 pin header to the scale's header.  I'd then be able to plug either a scale cable or battery in that spot.  This would allow me to use the caliper for testing a DRO or as a stand-alone caliper.  Not a big modification, but it works like a champ.

stech5.jpg
stech6.jpg
stech9.jpg
Caliper circuit board with MTA 4 pin connector attached.
Battery plugged in with a female MTA connector.
Unplug the battery and plug in the scale cable

I added the two additional QCC-100 dongles and checked again.  Now I had the X and Y display working and no Z.  Not even a zero on the Z display.  I swapped the QCC-100s and the problem moved with the QCC.  With the logic probe, I checked the output pins at 6 and 7 on the little PIC.  No pulse.  I had one extra PIC for the QCC-100 and checked that on my PIC programmer. It checked good, so I desoldered the old PIC and soldered in the new.  8 pins is about the extent of my luck in getting an IC out in one piece without ruining the pins or getting the IC too hot and killing it.  Same issue, no pulse and no readout on Z.  After checking the traces for the tenth time, I figured it had to be the oscillator, but no, the two spares I had made no difference.  I was getting a bit frustrated, so I ordered a new QCC-100 kit.  When it came after a few days, I had been thinking about the problem and reading the posts on the Shumatech Yahoo group.  I believe it was Lester who had stated that he had gotten a bad batch of oscillators.  I was now pretty sure that this was my problem too.  I tried the newly purchased oscillator in my broken QCC-100.  Success.  It turned out that over half of the oscillators I had purchased a few years ago must have been bad.  What are the chances of getting 3 bad ones in a row?  Probably about as good as winning the lottery.  I've looked carefully at the parts and they look fine.  I guess it might be possible that they were mis-marked, but whatever the reason, it is not the type of thing that one would expect.

I now had my spare DRO working, but I still needed to fix the one bad display on the DRO on my mill. I decided that I needed the DRO working now, so I ended up swapping the repaired board into the old case.  I had purchased the extra Jenix connectors so that I could plug the Jenix cables straight into the DRO and spent some time adding the QCC-100 circuit boards inside the DRO case.  Once I had this done, I needed to decide whether I would try to run the 3 Jenix scales from the 5 volt voltage regulator on the DRO-350 or add a regulated 5 volt wall wart to power the scales.  I re-read Scott's post on the situation and decided that since I was using a 9 volt regulated wall wart for power, I would probably be safe with running everything off the one power supply.  Just to be on the safe side, I milled up a large heat sink that would attach between the two regulators on the circuit board.  On the side of the U17, the regulator that supplies the 1.5 volts, I made a thin plastic insulator and used a nylon 4-40 bolt to attach the heat sink.  On the side of the 5 volt regulator, I added some good heat sink paste and a metal 4-40 bolt.  This is a really big chunk of aluminum and it soaks up a lot of heat.  As you have probably figured out from my description, the idea was to not have the bases of these two regulators be joined electrically.  Had they both been single voltage regulators with grounded bases, it would have not mattered, but U17 is an adjustable regulator and grounding the base in this circuit would have been bad.  As it turned out, the big heat sink seems to work well and after a full day in the shop with the DRO on, the side of the case isn't even warm. 

stech2.jpg
stech3.jpg
A healthy block of aluminum makes a good heat sink.
Heat sink, QCC-100s, and the Jenix connectors just fit.

I spent a couple days working on the mounts for the Jenix scales.  I planned to keep the locations of the new scales in the same places as the old ones, but both mounts were going to require new holders fabricated as I chose an 8 inch scale for the Y.  An 8 inch Jenix is actually 14 inches long. This is about twice the length of the capacitive scale it was replacing.  The extra length would come very close to the end of the knee, so I needed to shim the scale a little further out so it wouldn't interfere with the body of the mill.  It turned out that I only needed about an extra fraction of an inch, so the scale's mount is still very robust.  I used a 0.25" thick by 1.00" wide length of 6061 aluminum flat stock as a mount for the scale.  The holes at each end of the scales are slotted so that I could tweak the scale to be parallel with the knee ways.

The Z scale was a bit harder to mount.  The side of the mill column is angled at about 3°, so I had to take this into consideration when building the mount.  I initially made up a piece of 18 gauge sheet metal as the connector between the scale head and knee.  I used gage blocks between  the sheet metal and scale to check that the scale was as parallel to the Z ways as possible.  I used shim stock to achieve parallelism and then pulled the mounts and used a sine table and gage blocks to measure the length and angle of each mount.  Once I was happy with the measurements, I milled the mounts. I bolted everything back together using the sheet metal as the connector and rechecked the scale head to make sure that it ran parallel to the scale. Once I was satisfied, I used the same 0.25" thick aluminum plate I had used elsewhere to make the connector between the head and the knee.  There is no flex or deviation from parallel that I can measure.

jenix1.jpg
jenix2.jpg
jenix3.jpg
jenix4.jpg
Jenix scale on the Y axis is a tight fit.
At about 14" long it spans the whole knee.
The Z axis mount keeps the swarf away.
The armored cables are cut to length.

One last consideration for the Z axis scale mount was that the open side of the scale needed to face away from the flying swarf.  While I do have a rubber guard that keeps the chips away from the Z ways and extends far enough out on either side of the ways to prevent a lot of metal from flying in the direction of the scale, it won't catch all of the chips.  Due to the way the head was mounted to the scale and that I needed the cable to exit toward the top of the mill, I would need to mount the scale with the front side facing the mill's body.  I considered taking the scale apart and seeing if I could reverse the direction of the head, but decided that I didn't want to take a new scale apart.  I built the mounts to conform to the plastic end caps on the scale and it seems to work just fine.  To keep the armored cable from flopping around, I cut a disk from sheet metal and mounted it on the lamp mount.  I also added an angle adapter to the lamp base to fix the bothersome way the lamp was originally mounted.  Both the angle adapter and the disk worked out quite well.  The weight of the armored cable keeps it tidy and the lamp no longer gets in the way of my cuts.  Once I had all the scales in place, I shortened the cables and resoldered the plugs back in place.  The most difficult part of this little task was getting the shrink tubing the right size to fit inside the plugs and not shrink while I was soldering the connections.  This was one of those tasks where it would have been helpful to have a couple more hands.

So after about a week and a half of working and waiting for the QCC-100 parts, the scales were installed and the DRO had all its digits again.  The 2 new scales seem to work just as well as the one I have had on the X axis and having the 5/10,000 indicator actually flash between each thousandth is nice.  I have the DRO filtering set to 0 and there's not a hint of jitter.  Quite a nice upgrade.

I am not an electrical engineer. I believe my description of the way the circuit works is accurate, although a bit incomplete. If you find errors in my descriptions or diagnostic steps, I'd appreciate hearing about it.