Dead DRO

[Steve RowbothamParticipant @steverowbotham77083]

In the midst of milling a cylinder casting for the Halls Rotary engine on my Warco WM16-V mill last night the DRO readout suddenly and very annoyingly (as in I’ve lost my datum) expired. The 2 mains fuses tested to be OK so I opened up the unit to find the internal power supply is working fine and all internal connections seem good, so sadly it appears that the PCB has failed.

The failed unit is a 2-axis Ditron D60 (badged as Warco) supplied by Warco several years ago, but they don’t sell it any longer. Warco supply a modern equivalent but a) it’s currently out of stock and b) it costs £189.  I have found the actual Ditron unit online but cost is still £150+.

Before ordering I’m just wondering if anyone has encountered this before and has found a lower cost alternative, or even has a Ditron D60 DRO readout they don’t need ??

 

 

 

[Dave HalfordParticipant @davehalford22513]

Does the display have a ribbon connector? If so they can loosen.

[peterhodParticipant @peterhod]

 

This LED type is better. They are £115 on eBay much better functionality with visual display of functionality, hole layout etc. I’ve had the ones like yours and sold them on. There seem to be 2 types of wiring for the chinese scales. If you are lucky they will just work if not you will need to swop some wires on the plugs. I have a diagram of scale plug wiring somewhere from when I upgraded to the LCD ones.

[KenLParticipant @kenl]

I’ve had a similar problem before with an older Warco fitted DRO – always in cold weather!

If I warmed up the unit with a hair drier it started working again.

 

Last winter I removed the cicuit board and went over all the joints with a soldering iron to get rid of any dry joints and haven’t had the problem yet this winter. Having said that I’m tempting fate aren’t I?

 

Worth trying?

[Steve RowbothamParticipant @steverowbotham77083]

Hi Dave, thanks for the suggestion. It does indeed have a printed ribbon connector as you can see at the bottom of this pic, I checked it is firmly mated.

On power-up it goes through a self-check routine and the key-press beeper beeps as the display cycles round, but it isn’t beeping either, so it’s not just a display problem.

 

[Steve RowbothamParticipant @steverowbotham77083]

 

 

 

[Steve RowbothamParticipant @steverowbotham77083]

Thanks Peter, will have a look at the LCD devices as an alternative.

 

I will have ago at warming it up Ken to see if Hypothermia is the problem, if so it will be another one down to Rachel Reeves! The PCB is all surface mount so challenging to touch-up the solder joints.

[SillyOldDufferModerator @sillyoldduffer]

What are the symptoms Steve?  By ‘expired’ I assume totally dead, no smoke, scorch marks, and nothing lights up. If there’s any sign of life, please say what it is.

The most common reason electronics fail is mechanical!

• Condensation
• Corrosion due to condensation
• Loose connections or cracked circuit board tracks due to warming and cooling or vibration.
• Dirty switch and pot contacts and pots.
• Dry joints
• Damaged insulation
• Overheating
• Assembly errors.

Quite often, simply reseating plugs and sockets will fix, ideally with a squirt of switch cleaner.  Corrosion requires more aggressive cleaning.  Also check screw terminals and look for loose crimps.

Make sure the electronics are dry and well ventilated.

Dry joints and cracked traces are easy to fix if you can find them.  A freeze spray helps isolate problem areas.

Unfortunately, some mechanical failures result in fatally overstressed electronics.  That’s hard to fix without a circuit, a reasonably good understanding, and the right equipment.   For that reason when my stuff fails, I run through the basic mechanical checks and, unless the problem is fairly obvious, I replace rather than repair.   It’s because my time is valuable and I don’t enjoy repair work.  I assume Steve doesn’t have any test gear, in which case explaining how to diagnose and fix even a mild electronic problem is unlikely to help.

Dave

[Steve RowbothamParticipant @steverowbotham77083]

Thanks for your very helpful response Dave, the unit has expired as in completely dead. It was working, then all went blank, but no drama as in smoke or a bang. The internal PSU is working and delivering 5v DC to the main PCB, and all connections seem good. On switching on it does nothing, it used to run a self-check cycling the LED displays and beeping as it did so, but now nothing. The PCB appears to be in good condition as can be seen in my earlier post, no damp or mechanical problems, and the environment in the workshop is good – RH approx 60 and temp never falls below 10C which is fine for commercial electronics. Given the PCB is full of surface mount devices I think the only repair feasible would be a new PCB, which would likely be impossible to source, and as you say time is anyway better spent doing enjoyable stuff. But I do need a working DRO, so if no unwanted units are offered I will order a replacement as suggested by Peter.

[Joseph Noci 1Participant @josephnoci1]

Since all appears dead, most likely it is the u Processor not doing it’s job.  Straight failure of such device is rather rare, and often can rather be attributed to associated components – often the crystal. the crystal has 2 pins, one each going to a pin on the uProc. Also from each pin to ground will be a capacitor – in your case a small surface mount device. if one of these has failed – cracked, end not soldered properly – no oscillation, and the uP is dead. Worth checking that out under a magnifier. Also worth removing the PCB from the front panel – 4 screws – and checking the underside as well. Most times the PCB mechanical mounting stress causes small SMD devices to fail – crack, or break a poor solder joint.

The ‘suspicious’ label refers to the ribbon cable bulge – does not look kosher, but that cable is possibly only the keyboard interface.

Partially dead is often not worth repairing, esp if old, as it means finding replacement IC’s etc – totally dead is sometimes easier as most times it is a single component sitting in a critical path, ie, its failure kills all, so, PSU, Crystal oscillator, uP reset circuit, etc

 

[Dave HalfordParticipant @davehalford22513]

Steve,

You need to match the pinouts, there is no standard wiring for the D types. Some like the link list them.

[SillyOldDufferModerator @sillyoldduffer]

I like Joe’s suggestion – the crystal!  Easy enough to check if you have an oscilloscope or a radio that covers the range.  Another way is there may enough RF coming off a crystal pin to rectify the protective diodes in a analogue multimeter.   Putting one lead on results in a reading, so the crystal is working.

Decided to test this on the Arduino Nano I’m using for the Flashing LEDs topic.  Oh dear, I’ve half killed the Nano, probably by shorting two pins together – clumsy repairmen are a menace!

Power OK, oscillator still running, onboard serial IO LED flashing as expected,  but USB and the hour, minute and seconds LEDs are all dead.  Pretty sure this Nano is done for.   Condensation or a sliver of swarf can do the same sort of damage.

I shall have to replace the Nano…

🙁

Dave

[Steve RowbothamParticipant @steverowbotham77083]

Hi Joseph, many thanks for your really helpful response. I suspected the oscillator, it looks OK physically, I am currently trying to scrounge an oscilloscope to see if it’s running. The ribbon is just for the keypad as you thought, I initially thought it had delaminated but it seems it was made like that. I will check the oscillator if I can borrow an oscilloscope, and if its not running will call it quits – I need to get milling again!

 

Dave thanks for the link, I did wonder whether there is standard pin-out for the transducers / D-types as there are no pin designations on anything. If I do have to change display units hopefully I can just connect up to the currently installed devices as they work very well and it would be a shame to disturb.

[Steve RowbothamParticipant @steverowbotham77083]

Dave (S.O.D) looks like our posts crossed in the ether!! Sorry about the Nano 🙁

[Joseph Noci 1Participant @josephnoci1]

Steve.

Dave’s idea of using a multimeter is not good…That usually stops the oscillator and leave the uP pin at a logic level that you the interpret as ‘working’..

You could check as well what goes on under the crystal – in production the SMD devices are normally fitted first. When they do the leaded devices, esp if done manually and the solderer is a little judicious with solder, it wicks thru the hole and makes a dome on the other side of the pcb – under the crystal. Few mnfrs fit insulators between crystal and the board. If a ball of solder touches the crystal case, it’s sort of OK, but if if both pins have balls, the crystal  be shorted. Measure with an ohm meter between the ends of capacitors CJ1 and 2 – they appear to be the two caps connected to the crystal. One end of the caps should go to ground and be common. The other ends should not be connected to each other..

[Joseph Noci 1Participant @josephnoci1]

Dave – sorry, isee you said -one lead- so no DC would be measured – but no AC either as the osc stops..

[Steve RowbothamParticipant @steverowbotham77083]

Dave, no shorts around the crystal. Will let you know what happens if I get access to to an oscilloscope

[howardbParticipant @howardb]

I purchased a Chinese made 2 axis x and y axis DRO system from Vevor.fr a couple of years ago for very little money ,identical to the keyboard and display in the photo, and adapted it to my Warco WM 16 B mill/drill using my own version of the mounting hardware.

When not in use I always unplug the DRO system from the mains supply as we have mains spikes here in France from storms, lightning strikes on medium voltage lines and general EDF inability to manage the voltage, frequency and stability of the grid.

So, if I have a problem as described, it’s due to the hardware, not the grid.

Pull the plug out when you are not using it is my advice.

So far – no problems

 

[DiogenesParticipant @diogenes]

FWIW (if it’s worth anything at all) – if you don’t get a result from checking components on the board.

As this essentially sounds like a power failure with no bang or smell, I would revisit all the really obvious stuff between the mains socket and the supply to the board..

Dead boards – seen a couple

Iffy crimp connections/cheap switch ill-behaviour – dozens & dozens & dozens

[Joseph Noci 1]

On Joseph Noci 1 Said:

…

Dave’s idea of using a multimeter is not good…That usually stops the oscillator and leave the uP pin at a logic level that you the interpret as ‘working’..

…

Yeah, which is why tested it and reported a bad outcome!   Didn’t stop oscillator in this case, but I blew some internals by clumsy prodding.

I’ve used the trick successfully in the past on bigger crystals without evil results. RF must be present if a meter needle moves with only one lead connected, the assumption being a meter test lead isn’t long enough to load a crystal enough to stop the oscillator.  We all know about assumptions though, don’t we?   Let my shipwreck be your sea-mark!  (A Dutch proverb if my brain is working properly.)

Avoid.

Dave

[Steve RowbothamParticipant @steverowbotham77083]

I have managed to get use of an oscilloscope, the crystal is indeed oscillating at 8MHz, though I don’t know if the amplitude of the oscillation is OK at 0.33 V peak – peak? This is measured with AC coupling from ground to a pin of the oscillator. The microprocessor is completely dead.

 

 

[Nigel Graham 2Participant @nigelgraham2]

There’s mention of cold but I’d think damp the more likely environmental hazard. Modern electronics should be reasonably happy in humid conditions, and distilled water is of low conductivity, but the connectors’ contact surfaces could corrode enough to upset things.

[Joseph Noci 1Participant @josephnoci1]

The oscillation is good.

The CPU looks like an STM32F103 – the reset pin is on pin 7, right next to the crystal connections ( 5 and 6) . Check that the pin is Hi when power is on, and try short it to ground momentarily ( before the bofs get upset with this since the reset driver might be an active device , a quick short to ground won’t kill anything…).

It the CPU does not wake up, then try to scope the pins on the attached E2Prom – U2 ( actually an AT24C01 or 2).

Pin 5 = SDA and 6 = SCL

Look at pin 6 = the CPU would very soon after reset try to read data from the E2Prom and generate clocks on pin 6 to do this – You may need to play with the scope trigger to see the clocks..they would probably be around 1us period clocks, and probably in a group minimum of 8 clocks.

The E2Prom stores your settings ( inch/metric, etc) and would be read to restores these settings.

Pin 8 of said chip should be 3.3v ( I think you did say you checked for correct voltages on the board?)

Since you have the scope, it is worth looking at other CPU pins to see if it attempts at all to access display drivers, etc – The STM32F103 data sheet, page 26, shows the 48pin device pinout. Look on the PCB to see which pins are connected and scope the pin after a reset..Sometimes best to set the scope in ‘normal trigger’ mode, ie, not auto trigger, set for 1V div, DC coupled, set the trigger line midway – 1 to 2v above ground, and fiddle..

[Steve RowbothamParticipant @steverowbotham77083]

Many thanks Joseph. The processor is marked as ARM APM32 F103CBT6, but the OSC pins are 5&6 and the NRST is on pin 7 as you say. The reset is always high and grounding doesn’t bring it to life.

U2 is marked as Atmel ATMLH718, which I think is equivalent to AT24C01 but have struggled to find a data sheet.

The SCL line (pin 6) just goes high on reset, no sign of any clock pulses. The odd thing is that the 3 address lines (pins 1,2, &3) are all connected to ground?

 

 

I have indeed checked the supplies on the chips and all seem fine.

 

The CPU does seem to be completely dead unfortunately.

[Joseph Noci 1Participant @josephnoci1]

I have a data sheet if you would like – can email it to you – but if there is no CPU activity it won’t help…

The address lines all at gnd is ok – they are used to set the I2C address of the device – allowing 8 devices on the same I2C bus.

The ARM number is the same as the STM32F103. I use the STM devices a lot and have never had a processor just die on me, or at least that I have not killed…So I am still suspicious of the failure you have – so unlikely for the processor to die like that.

I suppose a workshop is not a great place for these electronics – the measly switchmode supplies that power these things are not good mains borne spike suppressors and powering off AC motors does cause a little havoc on the mains.

Not much more I can advise…Sorry.

[Author]

Posts