Hi fans -
first results from 'IWMless' beta testers are coming in. Here is a nice video made by Jav:
https://www.youtube.com/watch?v=NYFiMEVD79k
... which essentially shows how to remove the dead original IWM and how to put the 'IWMless' in.
It looks somewhat daunting despite he uses a semi professional desoldering gun, which has a hollow tip and an electrical air suction pump. But don't get frightened - a dead IWM can also be desoldered with a more brutal method, snipping off all the pins first and then desoldering the pins one by one, heating them with a regular soldering tip while pulling the pin with tweezers.
I use a truly professional grade desoldering gun which was very expensive some 40 years ago when I bought it, and had no issues with desoldering any IWM so far, but had issues on some other boards - it seems to depend on the size of the through hole and the thickness of the PCB, and how the IC pins were preformed before insertion / soldering. Optimized technique with the desoldering gun can help:
At 6:15 in the video, Jav lifts the PCB up to a slant angle for sucking the solder out of the pads, which is good (gravity helps you here). Even better: put the PCB in a bench vise so it is perpendicular to the table surface. This lets gravity help even more. You could try to keep the PCB flat and try to suck the solder out from below - even more help from gravity - but as you can't see where the desoldering tip is, this is not viable, at least for me, as I have no third eye growing out of my knee ;-)
I'm happy that Jav was able to boot this sophisticated desktop and that all the emulated hard drives (?) got mounted. Curious about which floppy emu he was using.
OK, this is the first report available from the 'IWMless' beta test phase, and it's even a video !
I'm quite astounded that while I was watching it, the view counter incremented from a tad below 300 to 555 ... and still counting.
Seems there is a lot of interest in the topic of how to replace dead IWMs.
I hope that the other beta testers also chime in and put their testimony and their findings here in this thread. A lot of testing work still needs to be done, especially with testing lots and lots of copy protected floppy disks. I'm quite sure that the 'IWMless' may be even more 'DISK II' compatible than the original IWM, and 'IWMless' also has a special DISK II compatibility mode, which can be activated by adding a switch, but so far this has not been tried.
I'm just in the process of building a second set of seven 'IWMless' beta test units, with a slightly different method to position the 'golden pins' and this will guarantee enough headroom below the stepper motor of the internal disk drive in the Apple IIc (the first seven units were provided with spacing washers, just in case, because headroom was tighter than expected due to a last minute engineering change, adding another resistor which was not planned initially).
- Uncle Bernie
P.S.: if 'linuxha' is right with his post here:
https://www.applefritter.com/content/happy-birthday-woz
... today is Woz' birthday. If this is true, and if you, Woz, are reading this, take this as a small virtual birthday gift from us - - - the importance is that now, with substitutes for all Apple custom ICs in the Apple II series being available, your Apple II will live on forever. The 'IWMless' was just the last piece in the puzzle.
I'm encouraged by the progress and looking forward to hopefully being able to help with the testing. I'm hoping the MMU that I sent to get your //e working again has arrived and helping out.
My apologies as I've tried to post this about 1/2 dozen tmes by pasting in text from something I typed elsewhere -unfortunetly everytime I did so the system gave me an error message saying it failed to post the message.
Anyhow:
Finally had a chnace to test the new IWMLess chip. For purposes of this first run UB provided a low-profile IWMLess socket into which you insert the IMWLess chip - which I understand in future version can be soldered direct to the board.
Preparing for the board:
For the record I did not destroy a working machine - I used a known IIc with a BAD IWM.
The instructions provided are spot on however I undewrstand that @Javier is creating a video which will fantastic, while I'm sure a novie can do this , without the write hardware and experience this process is risky. As UB points out the user needs to remove the current IWM chip which is soldered to the motherboasrd and not socketed. They key things to keep in mind here is to make sure you do not destroy the exisiing traces to the chip or the solder pads. Once the exisitng chip is removed, the process of inserting the new cocket and soldering that in place is fairly straight forward.
Installation
Installing the chip is straight forward enouch - assuming you've installed the scoket the right way with the notch in the same direction as the orignal and then placing the new IWMLess chip into the socket also notch matching with the original chip - doing this the wron way will fry the replacement.
Testing
First things first I powered up the machine without a disk drive attached and CTRL-RESET into BASIC, using the existing ROM - BASIC popped up and everything worked as expected - Ran a few BASIC programs with no issue.
I then installed the internal disk drive - given the new socket makes the IWMLess chip a little higher, we needed to first confirm that the new chip was not pushingagainst the bottom of the drive - in this case the tolerances were pretty good and we didn't have an issue (in the final version the socket will not be necessary but it may be appreciated by future users).
This is where I was having problems with my original IWM chip - however the IWMLess chip did not disappoint - the system came to life as I tested it with DOS 3.3 and PRODOS disks - no issues here - YAY!!
Next up trying some games (some of which were copy-protected). On hand I had original copies of Raster Blaster, Space Eggs , Ultima, and Wizardry - (UCSD Pascal) to name a few that I tested. All worked as expected which was again good news!
ZIP Chip compatible : YES
Wondering if this would work with my ZIP CHIP - I popped in z ZIP CHIP 8000 into the machine and everything again worked as expected except for the faster machine :)
One final note - I also upgraded the machine to a ROM4x (IIC) and did so without issue
More to come but overall things have been working out well Uncle Bernie - thanks for the opportunity!
Excellent news schafferm! I have been in contact with Uncle Bernie and since I don't have a //e to help test with I am expecting one of the test cards and an IWMless. I have several //e units both old style and Platinum and a number of different clones including Laser 128 to test with. If it performs well with all of these then I think it will be safe to conclude that IWMless is a great general purpose substitute for an original IWM for a LiRON type card. If we ever find someone with an original LiRON with bad IWM that would be the only better test.
For //c the results from both you and Javier are already encouraging. I'm pretty sure it's good to go for //c.
You can't use emojis. Any character above U+FFFE is prohibited (because the database uses UCS-2 without surrogate pairs).
Never knew this and surprised I didnt come across it here on the site - anyway who can I contact to delete my blank messages - I can't find a way to delete them?>
First and foremost, I would like to thank UncleBernie for creating the IWMless as it will save many //c machines in the future.
1- Careefully utilized solder wick to removed the old IWM. I have a Hakko and the //c boards are extremly delicate
2- Cleaned up board installed/soldered low profile socket provided by UncleBernie
3- Inserted IWMless chip/PCB into socket
4- Tested, and Voila it works!
3 - IWMLess.jpg
Next up...Reassembly and fitment results in conjunction with the internal 5.25" floppy drive. Stay tuned.
No need to be tuned. Since it is a partial emulator only it is competely useless here, and not only:
IMG_4513.JPG
Is that the board from a Unidisk 3.5?
I don't agree with the abrasive comment of 'transwarp2' in his post #9 above.
While it is true that the 'IWMless' can't be put into a Unidisk 3.5 drive, or any other target system needing the "FAST" mode of the original IWM, it is still possible to leverage the 'IWMless' for repair of dead original IWMs in these machines, and I wrote enough already about the why and how and all the very rational reasons behind this pragmatic approach.
Short, you can salvage a working original IWM from any Apple IIc or 'Liron' card for the Apple IIe, put it into the target system needing FAST mode, and then put an 'IWMless' into the parts donor to make it whole.
The rationale behind this approach is that Unidisk 3.5 drives, Apple IIc+, early Macs, and Apple IIgs having a dead IWM are too rare to make it worthwhile to develop an IWM substitute having the FAST mode. How many of these substitutes would be needed, wordwide, per year ? Three ? Four ? A dozen ? - - - nobody, absolutely nobody, will spend months of development and testing time to make a product with such a low demand.
And for the few cases where FAST mode is needed, use the organ donor trick, and put an 'IWMless' into the donor.
Keep in mind that some Apple machines have an IWM in the PLCC package, and this requires the development of an adapter. It could be done using castellated through-holes on an adapter PCB and a flex circuit which connects this adapter PCB to a DIL-28 socket, but this is ugly and not cheap. Especially if you need only a few per year. You could try to skip the adapter PCB and give the flex circuit a PLCC footprint, but then you need to work with solder paste and reflow soldering of the flex circuit into the target system without affecting other components nearby. Might need development of a special tool. How much do you need to charge for such a repair to recover your costs ? $1000 ? $3000 ?
All this has to do with economic viability and how much time and money somebody wants to spend on such a project. If I apply industry standard metrics on my 'IWM reverse engineering project' and the 'IWMless' effort so far, a real business would need to budget around $350000 to arrive at the same results as I have right now. If you could make a profit of $10 each IC, you need to sell 35000 pieces of such IWM replacement ICs to the Apple user community to just break even.
And I still have my doubts that the total market demand out there for such replacements would exceed maybe 50 or 100 pcs in total, over all upcoming years. I do have 144 pcs of these ispLSI1016 in TQFP-44 in stock and I would not bet that I could use all of those up to make 'IWMless' modules.
This is why "you can't always get what you want".
Especially true if you don't want to pay up for what you want. Frankly, I always wonder why there are small businesses and their websites where you can buy all these cute little gadgets for Apple II for maybe $20-$30. Such as the softSP card, or the MMU and IOU replacement modules which recently became available. I don't think that anyone makes any reasonable profit with this activity. So it must be a hobby, right ? And when it comes to hobbies, you can't ask for the moon.
- Uncle Bernie
Well, "transwarp" is always a negative nelly so I wouldn't take any criticism of his seriously. He complains a lot but he doesn't seem to be producing anything competitive to IWMless or most of the other things he puts down. At most it seems like he's pretty much just like that "retro devices" guy who apparently his whole business was making knock-offs of other people's designs. Not that there's anything wrong with making clones, but it's not necessarily the same as being the first to create something or solve a problem.
Softjanitor, it was a pleasure to expose your incomptence several times. I know your profile by now -- probably because you have zero successful products you designed and you are manufacturing, the only thing you could do with your IQ in terms of hardware is to squawk and cheer frequently everywhere around this forum, hiding behind the agenda of opensoure/free contemporary replacements in order to save the apple2 world ;) Batman ;) You obviously envy, and you are sort of a begger. Sorry, I am not an opensource enigineer, and seems you highly dislike that. I doubt Bernie would slam you with the source files for his IWM partial emulator, too. Paradoxically the best IWM knock-off would have been an exact clone of Apple's IWM, or even SWIM, this had not been achieved. So far. It is a very pervert idea to sacrifice a fully working //c computer in order to obtain the IWM chip from there. On my humble opinion it is better to look for destroyed for scrapping macintoshies though. On another continent where original Apple2s are rare, far away from the USA, I managed to obtain replacement real IWMs for my damaged Liron cards I bought from Netherlands and UK once upon a time...
IMG_4519.JPG
This is a violation of section 3.a of the AUP. A second violation will incur a 30-day ban.
Back on topic of IWMless testing.
I've been running the IWMless test card side by side with a "real" LiRON card for several days now and other than finding apparently some spurious memory issues on one of my //e units that appears to be resolved by re-seating the chips, it has been running reliably.
The only major testing I think that needs to be done is with a real Apple Unidisk 3.5 drive, which I don't have. Other than that, it is looking like IWMless is good to go as a replacement for an IWM on a LiRON card or in a //c. And regardless of nay-sayers, to me that is a major accomplishment.
Thanks to uncle burnie,
I have been able to successfully repair another IIc.
I am grateful to Bernie and the other mad scientists who share their work with us, and allow the hobby to thrive and are so generous of their time to help others.
Kindest regards
Pocoaust in Australia
I promised Uncle Bernie to do a fitment video of the IWMLess when reassembling the //c case. Well here it is;
https://youtu.be/Ix0u3-2eUAQ?si=jxA4G9r5u8KrsPNQ
In post #17, "6502Retro"wrote:
" I promised Uncle Bernie to do a fitment video of the IWMless ..."
Uncle Bernie comments:
Thanks for the video, it is very valuable for demonstrating the "mechanical challenge" of using the 'IWMless' in the current form (with the "golden pins"). I was aware of this problem since the beginning of the project, and did design the PCB such that it would fit into an Apple IIc even if an IC socket is being used. But it always has been a tight fit and there were concerns about the gap (or lack thereof) between the top side of the 'IWMless' and the bottom plate of the stepper motor of the floppy disk drive. This concern was unavoidable because I did not have the original blueprints of the Apple IIc with the mechanical tolerance numbers / specifications. So all I could do is to take measurements from the few Apple IIc specimen I had and hope for the best.
Use of a socket was deemed to be required for the beta test units of the 'IWMless' - just in case issues with compatibility would require edits to the logic within the CPLD, and the 'IWMless' units being sent back to me for reprogramming. Once this concern is gone, then the 'IWMless' could be soldered in directly, without use of a socket. And then there would be plenty of headroom.
The big question is how many prospective users of the 'IWMless' would want to go for the "direct solder in" solution. It may be a bit problematic / frightening if the usual industry standard square pin rows would be used - these are very difficult to desolder, so replacing a blown up 'IWMless' would require special desoldering equipment the typical hobbyist does not have. With improper tools (such as the ubiquitous spring-loaded piston type solder sucker) the attempt to desolder certainly fails and may lead to damage of the Apple IIc PCB (desoldering the original IWM is easy as its pins just can be clipped off at the IC body and then the pins can be desoldered one by one without any special tools, and the holes can be cleaned out using solder wick).
So the use of square pin rows may not be the preferred solution (for most), but it is the cheapest solution. Assembling the 'IWMless' with the "golden pins" used in the beta test units is extremely expensive as it takes ~30 minutes of hand soldering work to install them, and this requires highly developed soldering skills. But such an 'IWMless' could be installed either with an IC socket or without any socket, and even when being soldered in, it could be desoldered easily, just by heating up the "golden pins" one by one and pulling them out to the top using tweezers as long as the solder joint is molten.
The IC socket, it used, also is very expensive. It is a special "ultra low profile" socket type ED90218-ND made by Mill-Max here in the USA. Digikey rings up a price of $4.48 a piece. Normal, cheaper sockets can't be used due to the headroom issue.
All these are just mechanical fit / costs / serviceablity tradeoff decisions. It has nothing to do with the digital logic design as such. I always try to find the cheapest solution possible that still works, which would be the use of these square pin rows which are so difficult to desolder. Maybe there should be two versions of the 'IWMless', a cheaper one with these square pin rows and a premium one with the "golden pins" ?
Seems that some market research needs to be done ... feel free to comment which version you would prefer, considering your own desoldering skills and desoldering equipment (No, it's not worth to invest in a $800 professional desoldering station to save maybe $10 off the 'IWMless' installation - consider you first need to blow up such an 'IWMless' to need it, if the nasty square pin rows were used).
Comments invited !
- Uncle Bernie