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