Would you qualify these as broken solder joints? CNC Servo Amplifier Repair

[fastline]

I will try to first explain the issues in hopes it might help determine if I am on point here.

This repair is on a CNC servo motor amplifier. The motion controller is coming back with positional errors basically telling me the controllerr has commanded motion but not getting feedback from the position detection systems. This problem IS intermittent and the servo motor and feedback loop will run fine for several minutes, then simply error out. When that happens, power is forced off to the drive so the diagnostic LEDs won't help much.

I have done the proper troubleshooting to determine the problem is in the this amplifier.

The circuit is designed with a 220VAC to DC rectifier and then variable PWM is applied to the servo motor as commanded. The feedback comes back from the servo motor via a 'resolver" (not a digital encoder), and the drive is responsible for converting the resolver signal to a 6 channel digital encoder output. Either the resolver signal is being jumbled in the drive or the encoder output is being jumbled.

Due to the intermittent issue, I feel this is most likely a minor short or broken solder joint. Upon inspection, it is obvious the amplifier has been repaired extensively before and not the best work IMO. Several, if not all HCF4066 chips have been replaced years ago, among various other components.

Upon inspection, I am finding several through pin joints that look suspect. The board is a 3 layer board.

The pic with the header pin is power into the drive and the pin on the end is chassis ground that also goes through a series of caps. I am wondering if that could be causing noise in the drive and causing all this? The drive is otherwise grounded but you can clearly see there are a few pins that are fine and a few that are not.

The next pic is a couple of bridge rectifiers. About 4 of the 8 joints look suspect to me. thoughts?

Without any motion commands, at times, the controller has errored out and shut down with a small jolt of motion in the suspect axis.

NOTE: I have no schematic for this and cannot get one. These are obsolete and no one in the USA will even repair it. I have to hope for a simple find...

 

[MaxHeadRoom78]

I am guessing there is two loops here, the resolver back to the drive and then a pseudo incremental encoder signal is produced in the drive from the resolver signal which is then fed to the CNC trajectory controller.
What make of drive is it?
What is the type of motor, DC, AC or BLDC?
Is the command PWM or ±10vdc analogue?
If you have suspect solder connections go over them with a iron.
You could also test the encoder signal ouput.
Max.

 

[Nigel Goodwin]

Upon inspection, I am finding several through pin joints that look suspect. The board is a 3 layer board.

The next pic is a couple of bridge rectifiers. About 4 of the 8 joints look suspect to me. thoughts?

They look more 'starting' to be faulty, than having already failed - but (as already suggested) simply resolder any suspect looking joints.

 

[Tony Stewart]

Microscopic annular gaps around large mass parts often open such as relays on PCBs for Stovetops. THis requires a good eye or 10x lense.
Correction is to reflow the solder and add polyurethane to the body of the part to reduce micro-vibration. It's a mechanical failure mechanism of the crystalized solder.

But If that isn't the problem, I would be more inclined to think you have common mode (CM) noise injecting on the sensor lines and need to add CM chokes. CM noise is generated locally from the high Amp narrow current pulses the Rectifier Caps and also the PWM servo drivers which generate more EMI under load from the CNC operation.

He also had no problems with taking my suggestion to use CM chokes on all IO cables to sensors and no EMI problems.

A friend of mine has restored a couple Bridgeport CNC's by replacing the Servo electronics and interfacing it to a PC instead with AUtoCAD and CAD to CNC software. He found a few surprises like this on the 2nd unit that already had PWM servo's with high ESR caps. So getting an LCR meter is a good thought and checking or replacing all old electrolytics.

I consider "Low ESR" is when the capacitance * ESR < 200us so 1 Ohm*200uF or 1mOhm* 200mF.. General Purpose ones will be much higher which may be ok .. or not.

 

[fastline]

Well, as of right now, it seems the problem is resolved. I reflowed some suspected joints, cleaned other suspect areas with alcohol and a brush, air cleaned the whole deal, and reseated the daughter board that outputs the encoder data to the main controller. Did not miss a lick....

Now I have a CRT monitor down... Always something!

 

[MaxHeadRoom78]

Now I have a CRT monitor down...

Maybe time to convert to LCD!
Max.

 

[KeepItSimpleStupid]

FWIW: Freeze spray will highlight the possible faulty joints.

 

[fastline]

The only real issue in converting to LCD is space constraints. I would be about a 9" monitor. Unfortunately you can buy old 15" monitors but seems there is a LOT of gouging on the smaller ones. I can only find open frame units in 9" from industrial supply places for around $300. Ridiculous!

I have confirmed the CRT is the issue and the tube is not even glowing. CRT makes several chirps when power is applied. Board is not easy to get to and I don't see any obviously failed capacitors.

 

[Tony Stewart]

CRT's vary in cost usually with Bandwidth, ranging from 4MHz (old TV) to 250MHz, 150MHz nominal
PC CRT's range in bandwidth from 100MHz to 250Mhz typically depending on pixel rate.

 

[Tony Stewart]

The only real issue in converting to LCD is space constraints. I would be about a 9" monitor. Unfortunately you can buy old 15" monitors but seems there is a LOT of gouging on the smaller ones. I can only find open frame units in 9" from industrial supply places for around $300. Ridiculous!

I have confirmed the CRT is the issue and the tube is not even glowing. CRT makes several chirps when power is applied. Board is not easy to get to and I don't see any obviously failed capacitors.

Sounds like the power supply section has failed.

 

[fastline]

Here is what I found so far. There is obviously cracked solder joint on the flyback transformer but does not look totally hosed. I tend to think that is not the issue as it would have probably had other symptoms.

Due to the whining and the area where it is heard, I was looking in the PSU area of the board. I ESR tested the power capacitors and there are a total of 4, 2 each in parallel. Each is testing around 6-8ohms. These are SME 200V, 100uf caps. I would expect to see them below 1 ohm? Especially testing a pair in parallel!

Thoughts?

 

[MaxHeadRoom78]

The crt filament is often fed from an overwind on the flyback, no Horiz. out no filament.
Max.

 

[Nigel Goodwin]

Due to the whining and the area where it is heard, I was looking in the PSU area of the board. I ESR tested the power capacitors and there are a total of 4, 2 each in parallel. Each is testing around 6-8ohms. These are SME 200V, 100uf caps. I would expect to see them below 1 ohm? Especially testing a pair in parallel!

High voltage capacitors don't have low ESR values, so those values are probably perfectly normal - I can't be specific, because you don't bother ESR testing capacitors of those voltages (as they won't be switch-mode ones, which are where you get ESR failures).

 

[Nigel Goodwin]

The crt filament is often fed from an overwind on the flyback, no Horiz. out no filament.
Max.

Not just 'often', pretty well universally - although there might have been a few pre-1960's TV's where they didn't (but once the technique was created, then it rapidly became universal).

 

[Tony Stewart]

Here is what I found so far. There is obviously cracked solder joint on the flyback transformer but does not look totally hosed. I tend to think that is not the issue as it would have probably had other symptoms.

Due to the whining and the area where it is heard, I was looking in the PSU area of the board. I ESR tested the power capacitors and there are a total of 4, 2 each in parallel. Each is testing around 6-8ohms. These are SME 200V, 100uf caps. I would expect to see them below 1 ohm? Especially testing a pair in parallel!

Thoughts?

ESR * Cap value = T,, my rule of thumb is anything < =200us is low ESR. Polypropylene are normally low ESR but motor caps are not.
so 100uF*8Ohms = 800 μs is OK here as this RC "rule of thumb " value increases slightly with higher voltage.

So definitely cracked solder joint is an issue.
Clean the flyback with low residue isopropyl alcohol while you are at it and use Silicone RTV if necessary to prevent arcs, as dust accumulation can cause corona.

Then add Polyurethane sub-floor adhesive around the base of the transformer on board to prevent vibration. THis comes in a caulking gun tube(cheap) and hardens to prevent vibration. I always have a tube on hand for applications like this. COmmerical Mfg use white polyurethane or similar in power supplies for big caps and coils.

 

[MaxHeadRoom78]

Not just 'often', pretty well universally - although there might have been a few pre-1960's TV's where they didn't

Agreed but we are also talking of CNC monitors here, and I have found a wide mix, nothing really common.
Max.

 

[fastline]

Well.... Before I checked back, I ended up removing every single electrolytic from the board. Approx 15 of them. Used the ESR meter many times and nothing was reading reasonable. I even checked against new but not 'low esr" caps and values were quite different.

Here are some values with ESR meter at 100khz

50V, 47uf = 51 ohms
25V, 100uf = open
160V, 10uf = open
10V, 2200uf = 8.1 ohms
16V, 1000uf = .25 ohms
16V, 1000uf = 2.6 ohms (different one)
160V, 100uf = 6.1 ohms
200V, 100uf = 7-16 ohms (power supply caps, leaking starting on bottom, not swollen)

Due to the low cost and fact they are 20+yrs old, I am probably replacing a few as preventative.

For comparison, I tested some 400V, 470uf caps I have and they test at .1 ohms. They are not rated as low ESR.