Wondering if I can fix this board myself.

[MRdolt2u]

Hi! I have a very limited understanding of electronics, almost entirely automotive in nature. I am trying to repair a PCB from an electronic racquet stringing machine. I blew something by inadvertently shorting the motor as the unit was powering up. I am inserting (I hope) a photo of the board below. It is not a large board but it's mostly greek to me.

The leads to the 12v DC motor are at the bottom of the board (2 arrows). I figured this would be the region of the board likely to be affected. None of the components on the board are discolored or burnt, but the resistor with the 1 arrow is marked 0.51 ohms. So, I did a quick google and figured this is a fusible resistor and sure enough, it was blown. Bought replacement and installed. While the board is powering up the control panel, the motor is still not getting power. Normally, you turn the machine on and the unit does a self diagnostic cycle and the motor runs forward till it trips a limit switch and then back and comes to rest. The motor works though.

I know the motor works because I have 2 of these machines. I am thinking that even with my limited knowledge I can compare the good board to the bad board component by component. But all I have is an old Radio Shack analog Multitester, so basically an ohmmeter. Is this feasible? Do I need a digital tester with diode and capacitor testing capability?

I would prefer to do any testing on the unit without powering it. I have already blown one board. If I blow both I am up a creek.

[james_s]

The picture is too dark and blurry to see what's going on. There's gonna be a transistor of some sort to control power to the motor, they usually fail shorted however it may have blown the bond wires inside it if there was enough current. You'll have to trace enough of the circuit to figure out how the motor is controlled.

[MRdolt2u]

Well, thank you James for taking a look. I wasn't expecting anyone to analyze the board. My limited experience with electricity and electronics basically boils down to testing for continuity, resistance, and live wires. So my question I think really boiled down to: Can you safely and effectively test semi conductors with an old analog ohmmeter or is it just going to blow more of them?

Now, for instance, I am posting another picture of the board (below) from a different angle since you mentioned transistors. I take it you think it's likely I have blown one or more of these? The 4 on the left are K2116's and the one all the way to the right is a TA7805S. I googled them and looked at the spec sheets but I don't understand any of it other than I've got 4 MOSFETS and 1 Voltage Regulator. What I am wondering (and this goes back to my initial question) is whether I can test these semiconductors with my old ohmmeter or will it fry them? I am especially concerned because I don't know the correct way to hook an ohmmeter up to anything but a resistor.

I have looked at some videos on youtube on how to test MOSFETS since I started this post and everyone is talking about setting up in diode mode so I am thinking at this point I need to purchase a DMM.

Anyway here is the pic.

[greenpossum]

7805 is a voltage regulator that in all probability supplies the logic circuits. Since the logic is apparently working, it's probably ok, but you can check for 5V on the output.

You'll need to trace out the circuit around the power transistors. You should be able to measure the voltages on their outputs with a voltmeter in various phases of the initialisation after that and go from there.

[Manul]

Yes, you can test them (not fully, but still) with any ohmmeter. Very often mosfets fail by shorting drain and source. Sometimes also gate and source. If there is quite low resistance (0-100 ohms) between drain and source - it is very likely failed. You can also try gate - source. No worries in damaging by measuring. You can in most cases test them while on the board, no need to remove.

But turn of power before doing that.

[james_s]

While it's possible to do with the tools, it sounds like you're likely in over your head. Unless you want this as a learning opportunity and are willing to invest more time than it would cost to pay someone more experienced to repair it, your best bet is to send it out.

[MRdolt2u]

While it's possible to do with the tools, it sounds like you're likely in over your head. Unless you want this as a learning opportunity and are willing to invest more time than it would cost to pay someone more experienced to repair it, you're best bet is to send it out.

Normally, this time of the year I would be working 70-80 hrs/wk and would not even have the time to post on a message board. Thanks to the pandemic I have time. It used to be possible to get repairs from the manufacturer in Japan but they stopped responding to my emails years ago. The unit is 20 yrs old now, btw. Point is, going forward, it will be useful to understand the electronics of this device because the manufacturer is leaving me on my own and it is aging. It is however better than anything being made today.

If I were to send this out, is there someone in particular you would suggest?

[james_s]

I really have no idea, I've never tried to find someone to repair things but you could post in the work wanted section of this forum.

If you do want to try to fix it yourself, we're happy to offer tips and hints, although it's very difficult to diagnose something remotely while also providing an electronics technician education.

[Renate]

I'll have to admit my ignorance, I have no experience with "electronic racquet stringing machines". 

So we've got a 6 pin connector that goes to the motor.
I'm thinking a center-tapped stepper motor.
But, as I noted above, I don't know what I'm talking about.

First thing. Easy to do.
Use your ohmmeter on the motor end of that connector.
See what is connected to what.
I'm guessing that the 6 pins will measure like this:

Code: [Select]

1        2        3        4        5        6
  3 ohm    3 ohm   infinite  3 ohm    3 ohmIf that's correct, try measuring the board.
Measure pin 1 and see if it's directly connected to the center pin of one of the big 4 transistors (not the 7805).
Do the same for pin 3, 4, 6.
Can you see a transistor for each pin?
Check between pin 2 and pin 5, it should be a direct connection (you might also be able to see this directly on the board.)
If this is all correct, then we know what is going on and fixing should be easy.

Measure each of the six pins to the aluminum heatsink (is it grounded?)
Check with the probes in both polarities.

[james_s]

I assumed it was a DC motor, perhaps with a shaft encoder. It might not hurt to get a picture of the motor though.

[NivagSwerdna]

I blew something by inadvertently shorting the motor as the unit was powering up.

How exactly?

I think there is good advice already ^ but to add... likelyhood is an output transistor... measure the big ones for continuity between the pins and report back.

Are you happy to resolder one or many of the transistors?  Because that's probably where you are heading.

Hi res image of top and bottom would be fun to look at.

[MRdolt2u]

I'll have to admit my ignorance, I have no experience with "electronic racquet stringing machines". 

So we've got a 6 pin connector that goes to the motor.
I'm thinking a center-tapped stepper motor.
But, as I noted above, I don't know what I'm talking about.

First thing. Easy to do.
Use your ohmmeter on the motor end of that connector.
See what is connected to what.
I'm guessing that the 6 pins will measure like this:

Code: [Select]

1        2        3        4        5        6
  3 ohm    3 ohm   infinite  3 ohm    3 ohmIf that's correct, try measuring the board.
Measure pin 1 and see if it's directly connected to the center pin of one of the big 4 transistors (not the 7805).
Do the same for pin 3, 4, 6.
Can you see a transistor for each pin?
Check between pin 2 and pin 5, it should be a direct connection (you might also be able to see this directly on the board.)
If this is all correct, then we know what is going on and fixing should be easy.

Measure each of the six pins to the aluminum heatsink (is it grounded?)
Check with the probes in both polarities.

Thanks for the detailed instructions. I checked the resistance between pins 5 & 6 and got 15 ohms. I know these pins go to the motor which is not a stepper motor btw. The motor is regulated by a pair of optical encoders (I guess for forward and reverse).

Before you invest any more of your time (much appreciated btw) I want you to know I have a digital multimeter on the way and my first move when it arrives is to test the mosfets. As I said in my initial post I have 2 machines and therefore 2 boards, 1 good & 1 bad. If necessary I will test every component on them both to figure out what is damaged.

Thanks again, and I will definitely post my progress.

[MRdolt2u]

I blew something by inadvertently shorting the motor as the unit was powering up.

How exactly?

I think there is good advice already ^ but to add... likelyhood is an output transistor... measure the big ones for continuity between the pins and report back.

Are you happy to resolder one or many of the transistors?  Because that's probably where you are heading.

Hi res image of top and bottom would be fun to look at.

How exactly did I short out the motor? You want details? It's embarassing you know.

The long story is that the brushes had worn down to nothing. I found someone in China selling something with the right dimensions. It took months to arrive. The springs on the new brushes were too wide and the caps were too large to fit so I took my old brushes apart and soldered the old parts to the new brushes. Then I thought I should just check to make sure the motor was working. If I had been smart I would have mounted it to the housing but I didn't. And if I was smart I would have put electrical tape on the two solder joints I used to rejoin the leads where I was forced to cut them. I thought I could hold the motor in my hand, flick the on switch and then flick it off. Instead it jumped out of my hand like a king cobra and I guess when it went as far as the leads would allow, the two leads shorted where I soldered them.

As far as soldering the transistors, I don't feel that is beyond my abilities. I am more concerned about how long it will take to get them if they are indeed the problem.

Hi res photos I think would beyond my equipment. This is a two sided board though. There have been at least 2 posts where people suggested tracing the circuit. With the paths being on both sides I just find this is not something I can follow. But, as I said in the post above, I have 2 boards and I will figure this out.

[james_s]

As far as soldering the transistors, I don't feel that is beyond my abilities. I am more concerned about how long it will take to get them if they are indeed the problem.

Hi res photos I think would beyond my equipment. This is a two sided board though. There have been at least 2 posts where people suggested tracing the circuit. With the paths being on both sides I just find this is not something I can follow. But, as I said in the post above, I have 2 boards and I will figure this out.

Digikey or Mouser should have suitable replacements for the transistors, you can get them in a couple of days or less once you know what you need.

Do you not have a mobile phone made in the last 10 years or so? Most of them have quite good cameras, getting a high quality photo is mostly down to adequate lighting.

[Renate]

The long story is that the brushes...

Oh! It's a brush motor then.
I presume that it goes backwards and forwards?
So we're looking at a single H drive from the four transistors.

So get out your ohmmeter and see if we can identify stuff.
Measure from your new resistor to the middle lead (drain) of all four transistors.
(You'll have to try both ends of the resistor to see which gives the lowest reading.)
You should find two transistors that are directly connected, those are the "top" transistors.

Is the heatsink connected to the circuit? Or another ground?
Try measuring between ground and the right lead (source) of the four transistors.
You should find two transistors that are directly connected, those are the "bottom" transistors.

Then you've got the two pins that feed the motor (I think they are #5 & #6 pins?)
Each of those should connect directly to the middle lead (drain) of a "bottom" transistor and the right lead (source) of a "top" transistor.

If this all works out, you should now have a complete layout.
The four transistors are #5 top, #5 bottom, #6 top, #6 bottom.

If you find something without this symmetry, that's probably our problem.
You should check to see now to see if either motor connector pin is shorted to the (more active end of the) resistor or ground.

Next, you can do some diode tests.

[james_s]

A multimeter with a diode test function is very useful for testing transistors. It would also be good to know whether they're BJTs or MOSFETs which can be determined by the part number.

[NivagSwerdna]

If it was me.... and I knew this to be an expensive bit of kit... I would just change the four driver MOSFETs and see what happened.  Given the history of the short it is probable that one has blown.

I'm not the best person to advise... but maybe... SiHA180N60E which Mouser stock?

[SL4P]

I think we’ve almost reached that point where we ask...

Are you confident desoldering and removing the existing transistors.  Cooking the pads or board might lead to worse long term problems.

Any competent tech should be able to read the above, and save you a lot of pain within a couple of hours total.

[themadhippy]

Are you confident desoldering and removing the existing transistors.  Cooking the pads or board might lead to worse long term problems.

theres enough exposed leg on the existing transistors to snip em off well above the board and  attach the new ones .

[KL27x]

IME, you're probably looking at:

FETs are gone. Driver of the FETs also likely dead, since the control circuitry is on a 5V bus, and the motors probably 12+?

The key thing if this is me, is gate voltage. If the gates traces are doing what they're supposed to, then your uncle is Bob and you replace FETs. If not, you have to trace back until you find a signal, and then replace (or re-engineer) all the stuff in between.

It might go a little something like: snip out the FETs, and probe voltage of the gate traces while turning the controls of the control board. When the motor is supposed to start/stop moving or changing directions, the gate traces should change voltage. If this ain't happening, then don't put in new FETs. No point.

It looks like the gate drivers might be standard lego-brick buffer/logic IC's (or actual dedicated transistor driver IC, maybe even H-bridge-specific IC), so you might be in luck and be able to replace damaged parts. You can trace a double sided board using a continuity tester. You might find it handy to wrap a wire around your ground probe so you can solder it in place while you flip the board to find where the other side pops out. First time you run into a microcontroller, you're pretty well done with an easy fix.

AFAIC, if you can't find this drive signal, then there's nothing you can replace to get it back. Any damaged parts that may be found between this signal and the FET gate are most likely stuff to strengthen/sharpen the drive signal and/or for back EMF or whatnot. Voltage buffering, amplification, a diode here or there. This stuff quite often involves logic inversions, but you can cross that bridge if you get there. So that stuff can be re-made from jelly bean parts, with some homework. As long as you can find the signals what fires when the motor is supposed to do something. You can also work forward from the controls. If you can't find it there, doh. Start pulling suspected damaged parts that might be shorting the signal.

[SL4P]

theres enough exposed leg on the existing transistors to snip em off well above the board and  attach the new ones .

That’s a classy, reliable fix if I ever saw one. 

[Renate]

theres enough exposed leg on the existing transistors to snip em off well above the board and  attach the new ones .

That’s a classy, reliable fix if I ever saw one. 

If you cut off the old transistor's leads, it's easy enough to do this properly and solder suck the old bits out.
Yes, if you were experienced enough cleanly soldering it might make sense to just "shotgun" the four transistors.
But for a beginner if we can narrow it down to one transistor that might be helpful.

[SL4P]

ohhh ok.., !
I thought you were tacking the new leads onto the old stubs !   

[MRdolt2u]

Digikey or Mouser should have suitable replacements for the transistors, you can get them in a couple of days or less once you know what you need.

Do you not have a mobile phone made in the last 10 years or so? Most of them have quite good cameras, getting a high quality photo is mostly down to adequate lighting.

I found a company in Hong Kong that has stock of the original part. As a neophyte I feel ill equipped to select replacement parts on the basis of data sheets. If there was an online tool that spit out replacement part numbers when you input the part number of a discontinued item, that would be my kind of tool.

I can get up close to the board with a camera but then it's only possible to see a small section. Would this actually be useful?

[SL4P]

BTW when you do install replacement transistors, secure them to the heatsink BEFORE you solder the legs.
This eliminates stress on the parts for tiny misalignment etc.

[MRdolt2u]

I think we’ve almost reached that point where we ask...

Are you confident desoldering and removing the existing transistors.  Cooking the pads or board might lead to worse long term problems.

Any competent tech should be able to read the above, and save you a lot of pain within a couple of hours total.

My thinking was that I'd cut at least one leg with a pair of Xuron cutters and use a regular soldering station and a long nose plier to pull out the remnants of the leg(s), then clean up the pads with solder wick. In fact, if I get really unlucky and have to replace an IC, I think that rather than investing in a desoldering tool I would pull out my jewelers loupe and cut every leg and remove them one at a time in the same manner.

[MRdolt2u]

If it was me.... and I knew this to be an expensive bit of kit... I would just change the four driver MOSFETs and see what happened.  Given the history of the short it is probable that one has blown.

I'm not the best person to advise... but maybe... SiHA180N60E which Mouser stock?

It's certainly occurred to me that I can replace EVERYTHING on the board one piece at a time and eventually the problem will be solved. And I haven't ruled that possibility out either. 

Thanks very much for the suggestion of the SiHA180N60E. When I was looking online for the 2SK2116, I came across an online tool for finding Mosfet substitutions but you had to submit the performance parameters of the item in question and I gave up real quick.

[MRdolt2u]

IME, you're probably looking at:

FETs are gone. Driver of the FETs also likely dead, since the control circuitry is on a 5V bus, and the motors probably 12+?

The key thing if this is me, is gate voltage. If the gates traces are doing what they're supposed to, then your uncle is Bob and you replace FETs. If not, you have to trace back until you find a signal, and then replace (or re-engineer) all the stuff in between.

It might go a little something like: snip out the FETs, and probe voltage of the gate traces while turning the controls of the control board. When the motor is supposed to start/stop moving or changing directions, the gate traces should change voltage. If this ain't happening, then don't put in new FETs. No point.

Interesting and makes a lot of sense. I was sort of hoping to avoid working on the board with voltage applied but you make a good point. So, at the risk of sounding as ignorant as I am, I've got a red lead and a black lead. Which one to the gate trace and whither goes the other?

And THANKS!

[KL27x]

When measuring voltage, swapping the red and back leads only make the minus sign turn on or off. It doesn't really matter.

[NivagSwerdna]

google "testing a mosfet"... there are a few good videos on youtube

[MRdolt2u]

Just a quick update on a slow moving project:

It took about 10 days for the new multimeter to arrive. I tested the MOSFETS per a Youtube video and found that #1 and #3 were dead. The company in Hong Kong with the original 2SK2116 stock ignored my first RFQ so after waiting 4 days for a response I sent a second RFQ and waited another 2 days before giving up on them. Ended up ordering STP8NA50FI as a substitute from an Ebay seller in the UK. Ordered 4pcs and probably should have ordered more just in case I zap them for lack of an ESD mat.

I removed the 2 bad MOSFETs from the board and cleaned up the thru holes with solder wick. Also ordered what will amount to several lifetimes worth of heatsink paste. So I'm ready to go once the replacements arrive. My hope is that the fusible resistor I replaced blew in time to spare the rest of the board.

To KL27x: Good advice on tracing the circuit before installing the MOSFETs but unfortunately I couldn't trace a circuit if my life depended on it. So, I'm gonna install the MOSFETs and pray.

Will continue to post progress. Right now, it's up to Royal Mail...which is slow in the best of times.

Thanks everybody for your help!

[NivagSwerdna]

...in case I zap them for lack of an ESD mat.

IMHO Unlikely as long as you are not wearing polyester underpants.

[MRdolt2u]

IMHO Unlikely as long as you are not wearing polyester underpants.

Uh-oh....

[james_s]

You don't have to go out of your way to find a 2SK2116, glancing over the datasheet shows that it's a relatively mundane N-channel mosfet, there are hundreds of different parts that ought to work fine.


https://www.digikey.com/products/en/discrete-semiconductor-products/transistors-fets-mosfets-single/278?k=mosfet&k=&pkeyword=mosfet&sv=0&sf=1&FV=69|411897%2C606|378999%2C608|268887%2C608|269287%2C608|269986%2C608|270422%2C608|280734%2C608|283252%2C608|283516%2C608|283855%2C608|284432%2C608|284742%2C608|284746%2C608|293735%2C608|293752%2C608|296560%2C608|297340%2C608|297668%2C608|297674%2C608|304939%2C608|72473%2C608|72508%2C1291|413240%2C1291|413246%2C1291|413248%2C1291|413268%2C1291|413270%2C1291|413271%2C1291|413272%2C1291|413273%2Cmu450V|2068%2Cmu480V|2068%2Cmu500V|2068%2Cmu520V|2068%2Cmu525V|2068%2Cmu530V|2068%2Cmu550V|2068%2Cmu560V|2068%2Cmu600V|2068%2Cmu620V|2068%2Cmu640V|2068%2Cmu650V|2068%2Cmu700V|2068%2Cmu750V|2068%2Cmu800V|2068%2C-8|278%2C612|195828%2C612|199438%2C612|199800%2C612|213451%2C612|215953%2C612|217218%2C612|218572%2C612|218577%2C612|221362%2C612|221374%2C612|223927%2C612|225219%2C612|236942%2C612|239177%2C612|240212%2C612|240222%2C612|241255%2C612|241490%2C612|243238%2C612|244404%2C612|246709%2C612|247486%2C612|255987%2C612|255989%2C612|255993%2C612|255995%2C612|256006%2C612|257980%2C612|257982%2C612|258783%2C612|259940%2C612|259941%2C612|261070%2C612|262419%2C612|262421%2C612|262428%2C612|264181%2C612|264969%2C612|273159%2C612|273166%2C612|273172%2C612|275107%2C612|275110%2C612|275954%2C612|277515%2C612|277517%2C612|277522%2C612|279776%2C612|286370%2C612|286385%2C612|286388%2C612|286394%2C612|287918%2C612|288607%2C612|289921%2C612|290893%2C612|290895%2C612|290896%2C612|290899&quantity=&ColumnSort=0&page=1&stock=1&pageSize=500

[MRdolt2u]

You don't have to go out of your way to find a 2SK2116, glancing over the datasheet shows that it's a relatively mundane N-channel mosfet, there are hundreds of different parts that ought to work fine.

Too late. Already ordered. Found the STP8NA50FI on a datasheet of Mosfet cross references. Went with that. Thanks.

[NivagSwerdna]

You don't have to go out of your way to find a 2SK2116, glancing over the datasheet shows that it's a relatively mundane N-channel mosfet, there are hundreds of different parts that ought to work fine.

Too late. Already ordered. Found the STP8NA50FI on a datasheet of Mosfet cross references. Went with that. Thanks.

The STP8NA50FI is a bit underated current wise (4.5A) compared to the existing (7A)... I preferred my suggestion a few posts back.... SiHA180N60E (12A)... you can lead a horse to water...

[MRdolt2u]

The STP8NA50FI is a bit underated current wise (4.5A) compared to the existing (7A)... I preferred my suggestion a few posts back.... SiHA180N60E (12A)... you can lead a horse to water...

Interesting. I could have purchased the STP8NA50F which is rated at 8A too. There's a company in Belgium with stock at a very agreeable price. But the list of cross-references I found for some reason did not list the STP8NA50F as a suitable replacement for the K2116, they only listed the STP8NA50FI. And since I am not coming at this with any real understanding of what the difference is... I played it safe.

What performance shortcomings am I likely to encounter using a MOSFET that is "under-rated current wise"?
Will I blow fuses? The machine has a 5A fuse.
Will I blow the Mosfets again?
Will I be unable to pull string at high tensions? The Mosfets control the forward stroke of the stringing machine which is the power stroke. The reverse stroke simply returns the carriage to the starting position.

And are there no dangers in using something that is rated higher than the MOSFET I am replacing?

BTW, this is a stringing machine in action.




 

[NivagSwerdna]

What performance shortcomings am I likely to encounter using a MOSFET that is "under-rated current wise"?

Overheat and failure. 
You might get away with it... the old one might have been overrated.  But you are gambling.

I would have got mine from DigiKey, Mouser or Farnell.... I don't have experience with fake MOSFETS but I think they are not unknown.

PS
I play tennis... if you lived nearer I could have bought over my rackets and traded you for a re-string. 

[james_s]

Cross-reference lists can be notoriously unreliable, generally they are a good starting point but they are no substitute for carefully studying the datasheets of the original part and the one you are considering and then understanding enough about the application to determine whether the replacement is truly suitable or not. In almost all cases it is a much safer bet to go with a mosfet rated to handle more current than the original rather than less. 

[MRdolt2u]

Unfortunately I come at this with not only no experience but no understanding. When I was a kid my father showed me how to replace the fuses in our utility room and I was instructed to never replace a 10A fuse with a 15A or 20A so I took that sort of mentality into this decision.

Basically, if I pulled out a chess set and told you the names of all the pieces and the way they move, you would have as much understanding of chess as I have of electronics. There are only a couple items on this PCB I can't name and I have a rudimentary understanding of their function but beyond that I haven't a clue. I know what little I do know from working in the stockroom of a company that made power supplies...20 years ago.

While I am waiting for the Mosfets to arrive I went through all the IC's on the board with my multitester, noting all the pins that gave me a 0.000 reading or an OL reading. I should clarify: I checked all the IC's whose datasheets indicate that the bottom pin on the left side is marked as ground. I then checked the good board and found that I got the same readings on the same pins. So, in total I checked 21 IC's with no problems so far. I am also confident in the EPROM since it was not in the machine when the failure occurred. It's a spare and I put in after the unit failed in the hopes it would magically solve the problem.

[james_s]

A fuse is a safety device, it's designed to quite literally be the weakest link and fail before the wire does if the circuit is overloaded. Replacing a fuse with one that is of equal or lower rating is safe, it's still the weakest link. Replacing the wire after the fuse with one that is capable of carrying less current is NOT safe, now that wire may fail before the fuse opens to protect it.

The mosfet is more like a wire, you don't want it to fail to protect the fuse, you want it to be as robust as possible so that the fuse always blows before the mosfet is damaged by an overload. The reasons to not use the highest current mosfet you can find are mostly a matter of cost, and the fact that much higher current mosfets can have other different characteristics that may not work properly in a given circuit.

In this case there is not a lot of risk in trying a weaker mosfet, you'll just have to hope the original parts were sized larger than they really needed to be to do the job. Worst case is likely that they will fail and put you back where you started, it's not going to catch fire or anything.

[NivagSwerdna]

...I was instructed to never replace a 10A fuse with a 15A or 20A

Likewise don't replace it with a 5A fuse!
Larger is not always better... if you go too large you might find other parameters such as gate capacitance come in to play.
If I was you I would place a DigiKey order.

[MRdolt2u]

Thanks guys for the explanation! Will order from Digi-Key!

[james_s]

Likewise don't replace it with a 5A fuse!
Larger is not always better... if you go too large you might find other parameters such as gate capacitance come in to play.
If I was you I would place a DigiKey order.

Replacing a 10A fuse with a 5A fuse is not dangerous though, it may blow under normal operating conditions but it's not going to be a safety risk like replacing it with a larger fuse.

The gate capacitance is one of the "other different characteristics" I mentioned. Going a bit higher should be generally ok though, especially when using a more modern part.

[MRdolt2u]

Well, I was really hoping that my next post was going to be a celebratory one but here I am again scratching my head and wondering where I've gone wrong. The Mosfets arrived today from Digi-Key. I figured, before I install them I should check them with my new multimeter. Consulted a couple youtube videos for method and connecting my positive lead to the drain and negative lead to the source, got a nice "OL" reading in diode mode. Then keeping the negative lead on the source I briefly touched the positive lead to the gate to put the Mosfet into the "on" state. And still I get an "OL". All 4 of them. Are the Mosfets not working? Is my new multimeter not working? Seems like a simple enough test and yet...

[james_s]

This is not really a reliable way to test a mosfet, the gate capacitance is very small and even touching it with your finger can discharge it. They're brand new parts, you can safely assume that they're fine, don't worry about trying to test them.

Personally when I test mosfets I just verify that they are not shorted, 99.9% of the time when one fails it fails shorted between all three pins.

[MRdolt2u]

Well, I installed them last night and the unit still doesn't work. Unfortunately, I am not going to find time to pursue this. I am already buried in racquets after reopening my shop last week. I do want to thank everyone here for your thoughtful suggestions and if I ever get this thing to work again I will be sure to post. Thanks again everybody. This has been an education for me.

[NivagSwerdna]

Keep us updated; good luck with your work. 

[Zoli]

Well, EPROM.
Looks like the board has one; the life expectancy of the content is around 10(ten) years.
Before the "lifespan" is up, read, erase, reprogram, to have a working unit for the next 10 years; otherwise, bad luck - or not?
If the original manufacturer is still around, maybe you can "extort"* a copy of the EPROM from them, by buying an a second stringing machine from them, of course, ONLY if you need one.
Anyway, EPROM corruption can explain a lot the problems with the machine - blown mosfets, random moves, wrong tension - choose or add what's fit your situation.
As sidenote, I'm talking from my own experiece: the company I'm working for, had a couple of EPROM based products; when they come back fo repair, we've checked the age of the unit
before to proceed to other troubleshooting; 95% of the problems were solved by this approach; typical lifespan of those products were around 15 years.

*Well, you can request a copy of the EPROM; but to dig out a copy from the archieves(if they are still around) will take time(which isn't free);
and since this is a custom request, it will be quoted accordingly; BUT!: if you agree to buy a new stringing machine, the pressure from the
salesperson suddenly will pull the copy out.

[james_s]

10 years? That may be the guaranteed minimum but I have loads of EPROMs that are 35-40+ years old and still going strong. Early arcade games have loads of them, I've encountered bad ROMs a few times but it's not one of the more common faults I see.

[MRdolt2u]

As I said way back in post #39, the EPROM currently on the board was not in the board when I shorted the motor. It is a spare and I installed it in the hopes that perhaps it would return the board to functionality.

Meanwhile, I ordered a book from Amazon called "How to diagnose and fix everything electronic". When I will get the time to read it is the next question.

[NivagSwerdna]

Probably revert anything like the EPROM swap... after replacing the MOSFETs what is the behaviour?  Even if it doesn't drive the motors... does it live?

[MRdolt2u]

Probably revert anything like the EPROM swap... after replacing the MOSFETs what is the behaviour?  Even if it doesn't drive the motors... does it live?

When powered up, what should happen is that all the LED's on the control panel light up, the digital display should light up, the motor should cause the carriage to travel from it's normal resting position to the end of a full forward stroke, then return to it's normal resting position and chirp 3 times to indicate it is ready for use.

Instead, when I flip the power switch to ON all the control panel LED's light up and the digital display comes on and the unit chirps 3 times.

So, everything is normal except no motor action.

[KL27x]

This is what I understood all along when I made my first post, reply #19.

The next step I would do is to trace what connects to the FET gates. The FET gates (IIRC on an H bridge) will be connected in pairs.* So there will be two semiconductors driving these two pairs. (Maybe the same chip/driver driving both pairs). There will also be a series resistor in between this driver and the FET gates, which you can probably just ignore. Find that part/s. The driver. If your lucky, you can buy this part.

At rate, you can remove it, and you can probe all the pads on this part while turning on the machine. You will hopefully find that the input to this semiconductor is still alive. That there's one of these pads/traces that changes voltage when the motor is supposed to run this way, then that, on startup. If that's not there, then you continue working back. This working back from here might be impractical unless you can find the datasheet for this part, so hopefully you find the datasheet and find live signals on the input pins (or pads, after you remove this part).

I suppose it is possible that the main brain/micro is still running, but the output pins that drive these FETs got burnt. Then you're basically stuffed. I've never seen this happen. The micro always dies, too, IME, but I've heard of it. 

*Edit: I rescind. I refuse to look at an H bridge schematic/tutorial. But logically speaking, I believe there should have to be 4 separate control signals, one for each FET.

This would be something I could use my multi channel logic probe for, one probe soldered to each FET gate, so see which half (or both) of the bridge isn't receiving gate drive when the motor is supposed to be moving. Then same on the FET driver chip's (most likely a single IC with like 20 ish pins) inputs. 

[james_s]

I wonder if the motor power is supplied by a separate rail? Maybe a fuse or other component has failed when the mosfets shorted cutting power to the motor?

I would expect there to be some sort of protection because when the mosfets in a H-bridge short you get shoot through and a direct short across the supply.

[KL27x]

Yes, 99% different voltage rails. Reply 19.

1% to the fuse. We all want to just replace a blown fuse. But it almost never happens.  But for sure, check your rails. And check at the source of your 2 Pfets of the Hbridge, to make sure they're tapping the rail.

Since the motor isn't moving in either direction, the power supply could do that. But I would still put the money on a single driver IC for all 4 FETs that got toasted.

[NivagSwerdna]

So, everything is normal except no motor action.

So now it's down to checking some voltages.... So either the MOSFETs aren't being driven or there is no power to switch.  You now need to trace the route for the Motor or at least measure some voltages on the MOSFETs. (They are nice and big so measuring voltages on the pins should be do-able).  I'll have to check back on the thread but I thought all the MOSFETs were identical?  In an H-Bridge there might be 2 P's and 2 N's.... or 4 N's and a gate driver?    Is there only 1 motor?

and probe voltage of the gate traces while turning the controls of the control board.

^ this

[KL27x]

We're too lazy to look it up, lol.* I assume 2 topside are typically P, the 2 lowside are typically N. Technically the high side tranny's could be N, and the lowside could be P. But that would be extra cost and work, in most cases, requiring higher voltage rail/pump to drive a topside N, and a negative rail to drive a lowside P!

*Seasoned vets know their RAM isn't big enough to load this kind of info until it's time to use it.

[james_s]

Yes, 99% different voltage rails. Reply 19.

1% to the fuse. We all want to just replace a blown fuse. But it almost never happens.  But for sure, check your rails. And check at the source of your 2 Pfets of the Hbridge, to make sure they're tapping the rail.

Since the motor isn't moving in either direction, the power supply could do that. But I would still put the money on a single driver IC for all 4 FETs that got toasted.

It's almost never *just* the fuse, but if the mosfets fail then it will blow the fuse or cause a fusible resistor to go open. Occasionally a PCB trace or via can act as a fuse.

[MRdolt2u]

So I promised y'all a progress report and here I am with an update.

Before I started this thread I read the forum guidelines and they suggest including in your original post everything you have done
thus far. In my own case, I think the only thing I had done was to replace a fusible resistor that had blown. During the pandemic shut
down, I replaced the mosfets with the ones suggested by NivagSwerdna and purchased from DigiKey. Eventually the mosfets I ordered from
some company in Belgium also arrived  but I didn't even bother opening the packet. I just set them on the side with my lifetime supply
of heatsink paste.

I purchased a couple books on Electronics but neither of them was particularly helpful. Basically, what I wanted to know was this: I
have a good board and I have a bad board. I have a DMM. How do I proceed to compare on a component by component basis the in-circuit
readings of each and every lead on both boards without ending up with two bad boards? But the shutdown here ended in May and since
tennis is one of the few sports you can play while socially distant I was back to being buried in work and PCB repair got put on hold.

Eventually, with the return of winter I have had some spare time so I started testing components with the DMM in diode mode.  I tested
every pin on every IC and diode on the board and compared the readings to the good board. No discrepancies and no damage to the good
board.

I start watching youtube videos about PCB repair. Mr Carlson's Lab especially. In one of them he says something that piqued my
interest. He says: "The reason they socket these things is for a reason". I take note. I have two socketed IC's...an EPROM and an
EEPROM. And as I read up on them I am alarmed to find that they don't have indelible memories and they are 20 years old. I invest in a
TL866 II+. I read both chips and archive the files. I try to read the EPROM that was actually in the machine when I shorted the motor
but the TL866 detects a bad pin! This worries me, since the EPROM is located well away from the section of the board where the fusible
resistor blew. Lots of stuff between them.

Initially, I thought this PCB was pretty complicated but with each passing day I realize that it's actually not a particularly complex
board. I'm not suggesting that I understand how it works just that there's not much to it. And I am slowly learning what each of the
components is. So I am no longer looking at the 2 bridge rectifiers and wondering what they are, or the crystal oscillator, or the
handful of SIP's. I start to contemplate randomly replacing things and seeing if it solves the problem. I start eyeing the
electrolytic caps. I make a list of parts and start looking at pricing. I stumble across an Amazon ad for a kit of electronic
components. I start looking at capacitor kits. Then I realize that one of the capacitors I am thinking of replacing is actually a
varistor. And it is sitting right next to the the fusible resistor I replaced. I focus on this varistor as the most likely candidate
for replacement. I scour the internet looking for this varistor and of course the only place that has stock is in China, so I figure I
will just get something off Amazon that is somewhat close and see if I can at least get the machine to run. But I read a post from
someone who says if a varistor is blown, you probably wouldn't even be able to read the part number off it. And mine looks fine. I am
looking at it with my jeweler's louple and it's fine.

But as I am examining the varistor with my loupe I also get an up close look at the fusible resistor I had installed last year when
this all began. I notice it has a small longitudinal crack. I try to test it with the DMM while still in circuit but can't get a
reliable reading. I desolder one lead from the board and test the resistance. Blown.

I replace it. I re-install the board in the machine, re-attach all the cables, plug the power cord in, say a brief prayer and with all
my fingers crossed and holding my breath I flip the switch and the friggin thing starts like a champ. And it only took a year.

[james_s]

Wow that's great news, thanks for the update.

It's not often that a brand new part is bad but it can happen. It's possible the resistor got dropped somewhere along the way to your hands, or maybe you wrenched on it a little too aggressively while installing. Either way you found the problem and got it going.

Yes EPROMs do fade over time, I've had several fail when they reached about 30 years old. One one case I got lucky and was able to read the contents in the programmer but at full speed in the video terminal it went to it would fail. I read it and then re-programmed the same chip and the machine came to life. Then I saved the original file in case it ever does fail.

[NivagSwerdna]

Good job.  I'm sure those new MOSFETs will enjoy their life stringing racquets.   

[BrokenYugo]

As a side note your "bad pin" EPROM might be OK. The socket on those TL866 programmers aren't super high quality (A real 3M socket like that retails for as much as the whole programmer), mine has a flaky pin if I center the chip, works perfectly (read and program) if shifted to either side.