How do digital sewing machines fail? Is there often a visuall clue on the circuit board?
Reply to Thread

MidAtlantian · Feb 6, 2019

I have often noticed that when digitally controlled appliances fail without an obvious cause, that the cause is often something relatively simple, like a dry solder join, or a burned out resistor or diode. Is that usually the case, when commercial appliance boards fail?

I am considering purchasing a "not working" digital sewing machine with a main logic board failure, on the chance I can repair the board. Does anyone have enough experience to ... make a guess at this. I do understand that I can only hope for, well ... statistical advice.

Thanks!

Bluejets · Feb 6, 2019

Could be anything from a broken wire to a so-called repairer induced problem and anywhere in between.
Basically much like how long is a piece of string.

MidAtlantian · Feb 6, 2019

Thanks for your try, but, for example a dc motor problem is likely - but never "certainly "- related to the brushes. If you've looked at enough motor problems you know that. If you haven't, its an unknown for you: a "piece of string" problem. I'm hoping to hear from someone who has played with appliance circuit boards.

dave9 · Feb 6, 2019

Appliance circuit boards is a category too broad to generalize about. Some appliances may stay always powered and be more subject to AC power line surges. Some handle water or cooling/condensation and could have eventual water damage. Some heat up so eventual heat damage. Some are just made to the lowest price point possible and could have the basic old capacitor failure issues that any SMPS powered device has.

Then you have each design in a class of appliances, having it's own weaknesses which means unless you are going to research failure modes most common to the particular make and model (or closely related models) of appliance, it is more of a take it on a case by case basis by observing symptoms.

Before buying it I would at least check to see if any parts are available. What is the evidence it has a "main logic board failure" opposed to a switch, sensor, power supply to the board, etc?

No it is not likely to just be a burnt out resistor or diode on the logic board. Possible, sure, but I doubt much in the way of power resistors under thermal stress and a diode failure would be more likely in the power supply to the board rather than the board itself which might be more subject to a transistor failure but again it would be an investigation followed by the evidence of missing functionality and measurements with a multimeter.

It seems a bit like jumping off a cliff to me to buy this widget without already having a good idea what is wrong then hoping we can remotely assist in repair, is a bit different when someone already owns a widget and they're already invested vs throwing it away.

Bluejets · Feb 6, 2019

MidAtlantian said:

Thanks for your try, but, for example a dc motor problem is likely - but never "certainly "- related to the brushes. If you've looked at enough motor problems you know that. If you haven't, its an unknown for you: a "piece of string" problem. I'm hoping to hear from someone who has played with appliance circuit boards.
Click to expand...

Well, you supplied next to zero info so you get what you came for, an answer as broad as your query.

MidAtlantian · Feb 6, 2019

Thank you Dave9. I guess I have been lucky so far. A friend's machine simply died, and it turned out to be a cold solder joint to the fuse. And I've seen solder issues in heating system circuits. But I do not have any broad based experience, which is why I had asked this question. I would assume the board was the issue if the motor engaged, and the screen was either off or not working correctly. But point taken, it is a bit silly - apart from an interesting challenge - to, um, look for trouble in this way. Checking component after component is tedious work, especially if chips are involved. But especially considering the shipping costs involved with a sewing machine.

73's de Edd · Feb 7, 2019

You could at least give the brand and model number of that unit .

MidAtlantian · Feb 7, 2019

73's de Edd. No. My question is really a more generic one, although I would probably only look at Husqvarnas or Berninas. I would wait for what I thought was a machine model I thought was good for me, and a chance to get it very cheaply.

MidAtlantian · Feb 14, 2019 at 4:55 PM

Progress, of sorts. I have a Husqvarna Viking 610 sewing machine, about 30 years old. It runs at full speed, even with the pedal disconnected. There is a transistor and a comparitor chip on the board. I think it failed without any "incident", Nothing looks bad, but some of the soldering joints look "redone". I'll assume that the comparitor or transistor is the culprit. All of the components appear to be OK. Is there anything particular I could test for?

73's de Edd · Feb 15, 2019 at 1:18 AM

Sir MidAtlantian . . . . .

Any chance that this is being the PCB that you are finding residing within that unit of yours . . . .if so the problem is probbaly the TO-220 cased power semiconductor device located at the far left center . . . in its being shorted.
Pass on to us, all of the numbering / I.D. being atop it .
I think that I am reading the dual comparator IC as being of an LM 193/293/393 series
Along with one or both of the discrete xstrs being BC547's ?

The units motors go fast - go slow circuitry . . . .

73's de Edd . . . . .

If we actually are what we eat . . . . then, I'm definitely . . . . FAST, CHEAP and EASY.

MidAtlantian · Feb 15, 2019 at 1:37 AM

Wow! Yes, that is exactly the board. I already ordered replacements for the transistor (IRF 620) and the comparitor (LM393M). I am absolutely impressed that you had one of these so readily to hand! It was on another one of these same boards (A friend's machine that stopped working and started this whole extended project) where the fuse holder literally fell off of the board while I was looking for a problem. Of course, you could imagine that the fuse fell off from my too rough handling, but soldering it back on solved the problem. I do not think I ever would have checked the whole board for cold solder joints! But I was also too lazy to even do basic voltage checks. In any case, thank you for your input!

MidAtlantian · Feb 15, 2019 at 2:15 AM

I had not even noticed those two transistors. There is a "CBC547B" and what certainly looks like an "8557", but must be a B557". Do you think there is any chance that they might be bad? It will take time for me to get whatever I need, so I'd order them just in case now. I am doing this without much of a kit, since I am sort of visiting. Oh and on the 220 case, there is another line "1R9S8805"

73's de Edd · Feb 15, 2019 at 4:21 AM

Sir MidAtlantian . . . . .

The most likely failure would be the failure of that units IRF620 Power FET, with a shorting of its drain to source . of .pin 2 to 3 as per

Or possibly the glass bead diode just above it, or the blue paint tipped disc ceramic capacitor / or /MOV, just to its right side . . . . . or lastly, the conventional black cased power diode down just below them . . . . with any one of them being shorted.
I would need a good photo of all of the foil side of the board to confirm any circuitry interconnects.

I suspicion that the LM393 would be the very last suspect.

The 8805 symbolization on the IRF unit is telling us that the part is being of a manufacturing / marking date of the 5th week of year 1988.
Thus, dating your machine as being of no earlier than that date . . .~30 years.

ALSO they seem to be design deficient /close cutting in their choice of that IRF620 unit with it only being a 200VDC spec unit, with upwards of 170VDC coming fron the units FWB power supply from the 4 diodes nearby clustered around the central filter cap .
I would go for its big brother . . .an IRF820. with it being a 500 VDC rated unit.

2SC547 and 2SC557 numberings. . . . go on the 99 and 86/100 percent probability of being good . . . small sig transistors.

https://alltransistors.com/pdfview.php?doc=irf820pbf.pdf&dire=_upd-mosfet

Thaaaaaaaaaaaasit . . . . .

73's de Edd . . .

A man should be greater than some of his parts.

MidAtlantian · Feb 15, 2019 at 4:57 AM

Right! Thank you for getting me to get down to doing some measurements. I am always uneasy doing that, since while I've designed and built lots of digital projects, I know very little about analog anything, and I am always unsure if I can do a valid measurement of a component while it is in a circuit. Anyway, I tried everything you suggested. I did not find any shorts, and the blue tipped capacitor measured within a nF of its 4,5nF rating. All of the diodes had the right forward voltage, and none had a backwards voltage, I am not sure what you meant by "MOV", just to the right side of ghe blue-tipped capacitor. I did check the 4 larger standard cased black diodes - theones you said were part of the power supply: a full wave rectifier. Maybe I have to hope it is the LM393.

Is it possible that one of the components is only failing when the voltage across it is much higher than the tiny voltage of the diode setting in a multi-meter?

Yes, I had guessed that the machine was about 30 years old, since I know the history of my friend's machine. But I did not know that production dates were stamped onto (some) components.

Having already ordered the transistor, I'll wait to see if I get this running, then replace it. It had never occurred to me to look at its rating, although I was always very conservative when I built my own stuff. It does seem that it cannot be so difficult. It is not exactly a complicated circuit. I emailed the people I got this from, asking how it went - like if it blew in an electrical storm, or something electrically "exciting". Hmmm. The plot thickens!

MidAtlantian · Feb 15, 2019 at 1:52 PM

73's de Ed, ... Sir Edd?

Hmmm. I checked all of the transistors, and all of the capacitors (the meter I'm using had a capacitance scale curiously tucked in with the resistance options (like diodes and audio continuity). The MOV seems OK too. I can hope for the 393 being bad, but you are not optimistic about that. My other option is to take a hint from my encounter with the fuse holder's soldering joints, and resolder everything. However, I did all of this unpowered. Would it make sense to power it up and check for voltages. I suppose I'd see any solder joint problems that way. I could even take the time to sketch a schematic: absolutely a last resort for me. What are the three inductors used for?

MidAtlantian · Feb 16, 2019 at 1:09 AM

Sir de Edd,
I powered the board. The main transistor's gate is -12 volts relative to the source, and there is essentially no voltage between the source and the drain. It's full on. The voltage across (pin 4 to pin 8) the LM393 is 12v. Any suggestions where I should measure / check? I would assume that the circuit pulses the motor, but I don't know how.
Thanks

73's de Edd · Feb 16, 2019 at 10:53 PM

Sir MidAtlantian . . . . .

Now . . . . in my making a full implementation of your photos . . . GOOD PHOTOS ! . . . .
The black paired AC input inductors serve bidirectionally . . . both for filtering any possible sporaidic fast rising incoming AC high voltage line spike(s) and also back filtering the AC line feed from RFI-EMI-CRAP generated by this unit.

You unplugged the foot / speed control, so that sort of . . but not totally . . . eliminated the spring loaded foot control from having a mechanical breakage and its resultant swinging into full max . . . or minimum. . . . speed.

Your main point of interest would now be in referring to the mark ups now made and noting that the fused AC power in, feeds into the four 1N5400 series rectifier diodes which are configured as a FWB unit , which store up their produced DC into the BLACK central filter capacitor.
Looking at the top photo . . . the created RED ARROW DC flows down thru the machines motor and thru it and up to the Drain of the power FET and thru it, and out its Source for a ground return path to the power supply.
That is being . . . . . IF . . . . . there is being a variable width pulse train at its Gate, which is being provided by the two discrete driver transistors and the frontal variable pulse generator as being configured within the '393, in conjunction with feed back from the variable resistance of the foot / speed control.

I was kept from plotting the input of the '393 by not being able to see the left connector interfacings.
Currently . . . I'm suspicioning it to be the three little solder dots*** just below the massive black AC POWER connector at the board center left.
*** With them having a mating foot control plug going to the active 3 / or / 2 pins on the other side.

Another thing which could make the unit " dead in the water" would be loss of the sub power supply for the single end configured '393 and the 2 discrete small signal transisters . . . . . .
Those parts . . . . . I will just point out now.. . . textually.

BOTTOM PHOTO . . .

Going back to the now CONFIRMED BLUE disc ceramic capacitor which I had initially, optionally dual classified as also possibly ? being a MOV, instead.
That was just by my having read it at a sloped downward perpective, along with its marking being upside down. Which I then read as being . . . . .
Chinese/Korean/Japanese symbol followed by ? + ? + ? and then a final Chinese/Korean/Japanese symbol.
Or, if I now read again, I am now properly seeing a 4.7 nanofarad. So the unit is NOT being a Metal Oxide Varistor. But being the C element of a snubber network associated with the FET.

OK . . . now just above that cap is being a black diode which takes voltage dropped AC to convert to DC to feed to the BLUE cased E-cap at the top, which is regulated with the ORANGISH /YELLOW zener diode placed just below it.
That supply then feeds your single ended power for the '393 and the 2 discrete driver transistors.
Read across that zener and expect the DC voltage rated spec of the zener to be across it, and that being the sub power supply level / value.

If you want to get a jump on failure analysis, before your parts arrive, you can plug in the unit and reconfirm that it is still running at breakneck speed and then remove AC power and desolder and pull the power FET from the board and then try the board again and confirm that there is SILENCE . . . . with your having opened up the power loop, with no FET, now being in circuit.

Thaaaaaaaaaaaaaaasit . . . .

TECHNO REFERENCING ILLUSTRATION . . . ..

73's de Edd . . . . .

From my wifes perspective . . . IF YOU WANT BREAKFAST IN BED . . . then sleep in the kitchen.

MidAtlantian · Feb 17, 2019 at 5:52 AM

Sir de Edd,
Thank you again, I won't really have much time before Monday. But I disconnected the drive transistor, and the motor is no longer powered at all, as expected. The voltage across the zener and also between pins 4 and 8 of the LM393 is 12 V, and the voltage across the smaller standard diode located between them th efield effect transistor and the zener is about 3.5v.

In what way could the LM393 chip fail to give the results I have? Anything I could test without a scope?

I see - sort of - that the zener circuit creates the control voltage that ultimately goes to the gate of the FET. I don't really understand how the comparitor fits into the design, except that it must be the source of the pulsing of the voltage applied to the motor, and probably controls the pulse width ?

The three suspicious solder joints next to the AC connector are not connected to anything. There are several similar unconnected solder joints on the board. I know that the models 620 and 630 both have functionality that is selected by electrical switches, so these must be options not used for this model.

When I did measurements of the two small transistors with a diode tester, there were no shorts between any of the terminals in any direction: the connections were either open or about one diode's voltage drop between them.

I'll try a bit more on Monday.

73's de Edd · Feb 17, 2019 at 6:46 AM

Sir MidAtlantian . . . . .

In following relevant circitry traces, had those 3 unconnected "dud" solder domes been electrically interfaced, that would have been my guess for the inputting of the speed control , as the left side of the '393 is where the time constant parts of the circuitry are located .
You gots the unit in hand, for examination . . . . here I am 1,473 miles away and olefactorily challenged, deaf , senseless, blind . . . . and DUMB.
Can you filll me in on specific connections where the speed control wiring ties into circuitry ?
At least you now know that you might have a FET that is breaking down D-S on voltage application or might have a constant HIGH drive level present on its GATE.
To somewhat confirm that, replace the original FET into the pcb and lift the two right side leads of the two carbon film resistors located just above the FET GATE feed trace, that feeds it drive..
Then Retest.
If no wild motor action, the fault is suggesting of a gate drive circuitry problem.
If the motor runs wild . . . a possibly of voltage induced breakdown of the FET D to S.
Also, there is some possibility of the FET electrostatically self triggering, when being with a floating GATE connection.
To confirm . . . . power down and temporarily solder tack in an ~10k resistor between the FETS GATE to SOURCE .
Re test.

RE . . .
I don't really understand how the comparitor (sic) (comparator) fits into the design, except that it must be the source of the pulsing of the voltage applied to the motor, and probably controls the pulse width ?
YOU got it ! . . . . The speed control resistance tells it to create a constant stream of narrow width pulses for slow speed . . .stream of medium width pulses for medium speed . . ya da ya da ya . . . . . . a la . . .

https://www.google.com/search?newwi...73.339.5......0....1..gws-wiz-img.Lig0eE29XqM

73's de Edd . . . . .

If they don't have chocolate in Heaven . . . . . . I ain't uh goin'.

MidAtlantian · Feb 20, 2019 at 12:17 AM

Sir de Edd,

First, you were right about the three dud solder joints near the edge of the board being a part of the speed control circuit. This is a model 610. The better model, 630 has an axtra switch that sets the machine to run at half speed. I strongly suspect that the three duds are the ... expansion holes for the other parts of that circuit.

Yesterday, my electric kettle failed, my laptop computer failed with an irrecoverable logic board failure. And I failed: a nasty stomach virus knocked me down. But as Monty Python put it, I ain’t dead yet. But that is only one of three that survived.

I have begun to try to come up with a schematic, but I do have some additional hints. Replacing the FET did not change anything. Then I removed the LM393, and tried the live circuit. The motor still ran at full speed. I put the chip on a base, so I can remove it if I need to. But, with or without the LM393, the speed controller does not work. I admit that I never realized how tedious it is to work backards from a circuit board to a schematic! But I will keep at it.

How do digital sewing machines fail? Is there often a visuall clue on the circuit board? Reply to Thread

Want to reply to this thread or ask your own question?

How do digital sewing machines fail? Is there often a visuall clue on the circuit board?
Reply to Thread