Weller Soldering Iron WES51

[austfox]

I own a couple of the Weller WES51D units (digital display) so have been occasionally reading this thread. I figured your PIC code would be different from units with LED displays, so didn't contribute.

However, I was searching for something earlier this morning and stumbled across a WES50 unit that I purchased many years ago, possibly as 'not working'. I powered it on but got no response from the LED. I'll check the fuse later today to see if I can get it working.

Anyhow, do you know if there is any difference between the WES50 and the WES51? This one is marked as Made in Australia.

Does your board look similar to mine?

[bostonman]

Your board looks slightly different.

Looks like your board is missing the 5V DC regular (maybe it's on the other side?), and a fuse.

My advice is to extract the code from the PIC on both your solder stations if you can to have in case of a future repair since you have the resources on here. This forum is absolutely wonderful and like how people contribute their knowledge. Although I'm in this solder station repair for more than it's worth, I'm not only interested in repairing mine due to being stubborn, but, since I've had so much help, I feel it's the only way I can give back by uploading the code from a (hopefully) good one (I'll see once the one I purchased arrives).

UPDATE: I just got a notification that the solder station has shipped and should arrive by Thursday which means I should be able to remove the PIC over the weekend

[austfox]

Regulator is bottom left, pico fuse is on the other side of the board.

I'll have a go at reading the code and post it here. It will be interesting to see if it varies much, if at all.

I checked the fuse, and it is open, so I am suspecting a failure in the heating element (which also occured on one of my WES51D units some time ago).

[bostonman]

I missed the regulator. The board is very similar.

Have you tried searching for the schematics?

[austfox]

Haven't searched for the schematics for the WES50, but I just had a look on one of my drives and did find some schematics for the WES51D. The WES50, 51 and 51D all seem to share similar circuitry.

[austfox]

I replaced the pico fuse and all seems well with my WES50. Iron heats, melts solder, and cycles off and on. One thing I did notice, which seemed to concern you from your earlier posts, is that the green LED remains constant once power is applied.

From what I have read (at least for the WES50 model) is that a constant green LED allows use of the dial to set the temperature, and constant red is the locked in temperature, set with use of the lockout pencil (magnet). I do not believe the LED should be cycling on and off in relation to the iron heating or not.

You can also use the lockout pencil to recalibrate the tip temperature, which I assume is stored in the PIC's internal eeprom, and also reset it back to the factory setting, which I also assume is held in the eeprom.

I've socketed the PIC on my WES50, just need to get my PICSTARTER programmer working to dump the contents.

[bostonman]

Unless I'm wrong and my mind has reconstructed how my Weller works, the green LED should blink once temperature has been reached.

It should be solid green until temp is reached, blink once it's reached, turn off should the dial be decreased until the new lower temperature is reached, and then blink when the new lower temperature is reached.

If it doesn't blink, there wouldn't be a way to detect if temperature has been established unless you just assume.

[austfox]

Maybe the WES50 is slightly different, but there is no mention of a blinking LED for normal operation, only to wait 30 seconds after turning on until you start soldering.

Edit: Looks like they might function differently, the WES51 manual does state the LED should be off when it has reached the set temperature.

[bostonman]

I've socketed the PIC on my WES50

I forgot to mention that if your station is mechanically the same as mine, I believe the board will not fit with the PIC in the socket.

FYI, the hopefully working Weller is set to arrive Thursday. Unfortunately I may not get to removing the PIC until Sunday, but I'll try sooner. I also have to dig out the programmer, etc.. but this is somewhat of a priority for me because I've been without an iron for quite some time.

Also, if this one I purchased actually works, I'll have a real iron to properly solder the components I removed for troubleshooting on my broken iron (assuming the PIC solves the blinking LED issue).

[bostonman]

As stated, the Weller arrived yesterday. Today I set aside time to remove the PIC from the "good" iron.

Upon testing the solder station with the iron that came with it, the station wasn't working correctly and thought I got scammed. I tried my new iron and it worked exactly as expected.

The green LED remained on until it reached (I assume) the correct temperature, then began blinking. If I increased the knob (i.e. temp), the green LED remained on until the new correct temperature was reached. I decreased the knob and the green LED remained off until the new temperature was reached and it began blinking steady again.

I am attaching the original "bad" PIC file along with the new "good" PIC files.

I didn't want to screw with verifying or comparing, and hopefully Ian.M doesn't mind looking into comparing for me. Unfortunately the file sizes are identical and I'm hoping this isn't an indication that the PIC code is identical too.

[Ian.M]

The so-called 'good' files don't contain any valid PIC code.  The CONFIG word does match the suspect one I was working on for you (0x3F6C), so we can be reasonably certain the good PIC isn't code protected (and the program memory area of the HEX file is blank (0x3FFF), not the scrambled values characteristic of PIC12/16C era code protection), so I suspect your programmer had problems reading the PIC, namely it failed to enter programming mode on the first pass.  Possible intermittent contact with the socket?

[bostonman]

Thanks for looking at this so quickly. Next time I’ll check the buffer.

I went through you know what  and back to remove this PIC. It had solder on the IC side of the board, up to the body (the assembler must have used a roll of solder to install this). I couldn’t remove the solder after a few attempts and resorted to cutting the leads before any more heat damaged it. i soldered fly leads into a socket, cut them shorter, quickly applied heat to each fly lead to the PIC lead, and then the programmer kept popping out the socket. I had to put that socket into a larger one - obviously aligning the pins.

Due to the headache, the moment I saw it finished reading, I let out a sigh of relief and didn’t think to check the buffer.

I apologize for wasting your time with a blank file. I’ll try again tomorrow or Sunday - this time checking the buffer.

[bostonman]

Out of curiosity, can one (or more) pins not making a connection cause the code to be blank or does a possibility exist the PIC got damaged due to heat?

I assumed the programmer would have told me if it's not making a proper connection on a pin or pins, however, maybe it's more thorough on a write cycle than reading.

[Ian.M]

Its a baseline PIC, so doesn't have a chip ID so no the programmer wouldn't tell you it failed to read.  If the PIC doesn't enter programming mode for whatever reason, it fails to clock out the program memory word contents when requested, and the programmer 'sees' whatever level the ICSPDAT pin is at as the memory contents, in this case it was floating high or pulled up so read back as all '1' i.e 0x3FF a blank memory location (rather than all '0' i.e. 0x000 NOP).

[bostonman]

In the previous "good" PIC code I failed to save the buffer (text file), so I'm uncertain if comparing this buffer to that buffer would prove I was able to 'read' the PIC correctly.

I'm uncertain what to look for to prove I got a valid read off the PIC this time, so I apologize if this new code (see attached) is also empty.

Hopefully it read correctly because I made sure to get a solid connection. If it failed to read this time, I will fear the PIC got damaged due to excessive heat while unsoldering because I got a much better connection in the programmer socket.

[Ian.M]

That looks like a good read.  It differs by one bit from the bad one at word address 0x98:
    movlw 0x50 ;bad
    movlw 0x54 ;good

That's in the mains frequency detection and selection code, where its checking the frequency range for 50Hz.

Its also got a different OSCAL data word, but that's expected.

Unfortunately its a '0' to '1' bit change and EPROM programming without erase only goes the other way so you can't patch that bit in-situ in your bad PIC.

[Andy Watson]

That lowers the detection frequency from 48.8 to 46.5 Hz - which makes sense. However, I don't think that is the show stopper, esepcially since we're talking about detecting 60 Hz (I assume). Perhaps the OSCVAL is too far out ?

Since it's already broke would there be any harm in patching-out the frequency detection and fixing it to 60Hz? I believe this could be achieved with a few well-placed NOPs (all zero ?) instructions.

[bostonman]

Unfortunately its a '0' to '1' bit change and EPROM programming without erase only goes the other way so you can't patch that bit in-situ in your bad PIC.

Is there a need to fix the bad one? I'd rather burn this code to a blank PIC, however, I believe you needed to do something to it first.

I still have two, or maybe one, blank PICs, so if this 'good' code can be patched to burn to a blank, that will hopefully solve the issues with my iron; providing I can replace the components and any broken wires that were tinkered with during the troubleshooting. Also, anyone who needs the PIC code will have it for this needs in the future.

Plus, the PCB is relatively simple, so people can build their own solder station if they desire.

EDIT: I also need to burn a new PIC because I had to cut the legs on the 'good' one (not to say I can't just solder it to the board using fly leads. In any case, even if I didn't have any blank PICs (which I do), I plan to buy more anyway once I have the ability to burn good code onto them and confirm it works.

That lowers the detection frequency from 48.8 to 46.5 Hz - which makes sense. However, I don't think that is the show stopper, esepcially since we're talking about detecting 60 Hz (I assume). Perhaps the OSCVAL is too far out ?

Since it's already broke would there be any harm in patching-out the frequency detection and fixing it to 60Hz? I believe this could be achieved with a few well-placed NOPs (all zero ?) instructions.

I am dealing with 60Hz.

[Ian.M]

I've stripped the OSCCAL so you can now use it to program a blank PIC following the workflow I gave in reply #66.

I suggest you try the newly programmed PIC in the working station to confirm its good before further experimentation with the bad station.  If all is well program another one for the bad station.

[bostonman]

Great, thanks!

I have a bunch of stuff to do today, however, I'll try getting to this as soon as possible. If not today, hopefully tomorrow.

[bostonman]

I referenced your previous post. Before I program a new one, I wanted to confirm something.

The steps you gave show to use Motorola as file type. I thought this was an Intel file type?

[Ian.M]

The image was only to show which dialog and what checkbox. 

Yes, select IntelHEX, and clear the checkbox!

[bostonman]

You're not going to believe this.

I placed the blank PIC into the programmer, did a blank check (checked "OK!"), performed a 'read', selected the HEX file you provided, selected Intel, the blank check box was already checked, and programmed.

It programmed in 01:70.

For laughs I did a blank check and it showed: chip address 00000000, chip data 2855, buffer address 00000000, buffer data 2855, and showed 'error' for the blank check (as I assumed since it's now programmed).

I also 'verified' and that check "OK"!.

Soldered it onto the board, turned on power, the green LED came on, the soldering iron heated, and the green LED never turned off. I touched solder to the tip and it melted indicating it's at least 360 degrees F, turned the knob to minimum, and the green LED remained solid green. Increased the knob, let it heat more, turned knob to minimum, and the LED remained solid green.

Thankfully before turning off the programmer I saved the buffer and info that was read back from the newly programmed PIC (see attached) should you decide to look at it.

[Ian.M]

That's *TWICE*.  I now suspect that there is some process that must be run after programming possibly in a test jig to initialize the PIC's 16 byte I2C EEPROM die.

Try:

RESET STATION TO FACTORY DEFAULT SETTINGS
With station turned Off, adjust temperature set knob to 600°F (315°C). Apply the Lockout Pencil to the ESD symbol on front panel and turn station On. Remove the Lockout Pencil and the procedure is complete. Any tip temperature offset programmed above will be reset to nominal factory settings.

@Andy Watson: You've dug into the EEPROM code further than I have.  If the above doesn't help, any Ideas?

[bostonman]

I've been hesitant to respond because I don't want your request for Andy Watson to provide his input to get buried, but I've been giving thought to this.

First off, I did consider doing a 'factory reset', but I didn't want to perform any additional steps to avoid throwing anomalies into the work you've put into helping me.

Thinking out loud about this from a production point of view, each PIC needs to be programmed at the factory individually before installing. Rhetorically asking, what additional step would the manufacturing team need to take that would be acceptable since it's going to add time to production?

* Removing the PIC from the programmer and sticking into something else to run an additional step could be a possibility, but adds an additional step and time. I'm almost ruling out this based on additional production time.

* Installing a temporary jumper or some sort of physical step on the PCB in order to activate the PIC also seems unlikely since that is time consuming.

* Assembling the whole unit and testing it before packaging for shipping seems more logical. At this point it seems more likely a step could be easily implemented to do a factory reset (although why bother if they can add that to the code); or the additional step could be a particular temp setting with a magnet that tells it to read and set the line frequency.

Initializing the PIC somehow seems logical because it's ironic it seems to work correctly but doesn't blink. When you compared the old "bad" code to the new "good" code, the detection frequency was lower. If Weller builds these for 120 and 240 (60Hz and 50Hz respectively), maybe it needs to be "told" which frequency to look for. The transformer in my unit is strictly for 120, but the 4000-02 is 240 with the same 24V secondary. Weller could build the same solder station that works off the same DC voltage and they'd only need to swap transformers. Now the problem is how does the PIC know to clock off 50Hz or 60Hz; and maybe this is the missing piece of the puzzle.

Also keep in mind, my iron had an open thermocouple (didn't know this initially and started off measuring the circuit thinking it was the TRIAC, capacitor, or 5V regulator) and believe the station was working correctly, but the LED stopped blinking because I may have shorted two pins while taking measurements (although I can't say for sure this is what caused the LED to not blink). Maybe I got the PIC into the raw factory mode where it doesn't know whether to clock at 50Hz or 60Hz.