Let’s attempt to repair a Tandy 1000 ISA Slot and XT-IDE card
Вставка
- Опубліковано 27 тра 2024
- Two projects for the price of none!
Channel Membership: ua-cam.com/users/retrobitstvjoin
Patreon: / retrobits
Amazon: www.amazon.com/shop/retrobits
Twitter: / retrobitstv
Tandy 1000 Repair Part 1 - • Dead Tandy 1000 First ...
Tandy 1000 Repair Part 2 - • Tandy 1000 Part 2: Rai...
Kaypro 4 Plus 88 Testing - • Is this SWP CoPower 88...
Lee Smith’s Workshop - / @leesmithsworkshop
Hakko FX888D Digital Soldering Station - amzn.to/2TYYrbI
Hakko FR301 Desoldering Tool - amzn.to/3cP8VR3
Hakko 633 Soldering Iron Holder - amzn.to/3xiTJpj
YIHUA 959D-Digital Hot Air Rework Station - amzn.to/3ioSgXb
Kester Leaded Solder - amzn.to/3IIJz7s
SRA Rosin Paste Flux - amzn.to/3oBesE2
Chip Quik No-Clean Flux - amzn.to/438tSyD
Weller SM1 Solder Dispenser - amzn.to/3lFVXtC
INNOVA 3320 Auto-Ranging Digital Multimeter - amzn.to/3JezMnJ
MinusZeroDegrees - minuszerodegrees.net/xtide/XT...
VCF XT-IDE Rev 2 - forum.vcfed.org/index.php?thr...
Glitch Works XT-IDE - github.com/glitchwrks/xt_ide/
Thanks to our awesome Patrons and Channel Members for their support!!!
Peter (Melair)
Kevin Gonzalez
Jon Ingersoll
John DiLiberto
John Trammell
Paraffine
Pinay Collectibles - / pinaycollectibles
Josh K
Mitch Hoffman
Alex Smith
Vince Valenti
William Sullivan
Eric Stacey
Ryan Brooks
James Taylor
FULLSET
[mi.low]
Allister Fiend
Erik (goodjonx)
Pierre Kressmann
RobF228
Jim Leonard
Rayner Daniel Pedersen
Wolfgang Behrens
Eli
Thomas Jager
00:00 Introduction
01:50 Tandy 1000 Disassembly
05:06 Replacing an ISA Slot
09:53 Testing
14:23 A broken XT-IDE card
16:00 Reviewing the schematics
18:43 Visual inspection
20:41 Replacing SMT chips
27:34 Conclusion
*Retro Bits earns money from Amazon and eBay Partner Links - Наука та технологія
![Alive and Kick[start]ing: eBay Amiga 1000 Revived?](http://i.ytimg.com/vi/ML8LGzlJ38o/mqdefault.jpg)
Nice repair job.. Pretty sure the flux has nothing to do with the mask coming off, if you re-watch you will see that it comes off as you wiggle the desoldering gun tip around on the joint, therefore it would appear you just scratched off the mask with the tip. Not a major issue though, as long as you don't bridge any of the exposed tracks with the RF shield or when adding fresh solder.
Thanks! It's strange because I've never damaged any other PCB when using that pump before. Perhaps the mask on the Tandy is really fragile? The areas where I used the liquid flux showed more damage than the ones where I only added fresh solder. I wonder if there was a chemical reaction or maybe the flux is slightly abrasive? Of course, I couldn't see any of this while I was working on the part since I had no magnification and had to sit pretty far back so as not to block the camera. The camera has a much clearer/closer view of the carnage and I only really saw it upon reviewing the footage :( Fortunately no pins were bridged :)
@@retrobitstv If you have the PCB sideways there is less force on the pad while desoldering. Also if you have to solder small SMD there is solder paste with flux. It's a mixture of very small solder balls in flux. If you put a little on the pad's and place the IC on top and apply hot air the IC should flow in the right place. Ali sells practice PCB's for almost nothing.
You need to change the tip out. M P is 100% correct.
@@retrobitstv It could also be a combination of temperature slightly to hot in this instance with the motion. The only time I have seen similar is with older boards and the temp too high on the desoldering pump. It could just be the combination and weak solder mask. I usually do the same with the circular motion and rarely have a problem scratching off the solder mask.
Totally down for an XT-IDE episode!
That's the Tandy 1000A. The original 1000 has ROM version 1.00 and no 8087 co-processor socket.
Right on. I discovered it was an "A" model after purchasing it on eBay and opening it up. It wasn't advertised as such but I'm happy to have the slightly updated model. I would still like to upgrade to a 1000SX though :)
Great job with repair! I usually stay away from desoldering guns because of risk of injury to the board. In future may want to stick to desoldering wick with plenty of flux, especially for more delicate operations.
Admittedly, my experience is still pretty limited but this was the first board I've worked on that was damaged using the desoldering gun. It's been a great tool but I will definitely have to be more judicious about where and when to use it in the future!
Nice Vintage Computer Federation mug!
Thanks! Got it at VCF East this year!
6:38 ugh! That desoldering gun nozzle makes damage to the solder mask. It's on the wedge of breaking traces continuity. Be careful.
Yea, I only saw that after the fact because I have to sit pretty far away from the part I'm working on as to not block the camera. The camera has a much clearer view of the work going on and upon reviewing the footage it looked much worse than what I saw in person! I have never encountered solder mask damage like this before when using the pump. The damage was definitely worse on the pins I applied the liquid flux to versus the pins I only added fresh solder to. I wonder if there was a chemical reaction or the flux is somehow abrasive?
Yes please build the xtide card!
years ago, i went to a church sale & they had a Tandy 1000 there they were salering for $5.00. it had the unit with the keyboard - printer - monitor. it looks like it was in good shape. but i didn't have the money to buy it.
Wow that would have been an amazing deal. These days, just the keyboard alone can cost $100 on eBay :/
When using HA for desoldering SMD components, you should use a flux... best fluxes for such things are jelly fluxes, with RMA-7 being my favorite.
Good tip, thanks! I will look into that.
Quality Tandy content!
Please, do the build of the card.
Thank you for the video!
I just love my Tandy 1000tx 🥰
But I just don't know why 😊
I love ur work and really enjoy ur vids, ur pretty good for a person like me who is super critical.
I appreciate that!
As is usually the case, the suspect chip has the finest pin pitch and is in a difficult spot.🤪
Tandy 1000 gang represent.
Have you checked the pins on the CF socket? They can push in, and it could be why your card isn't detected.
That is a great question! The answer is yes, I replaced the entire CF card riser with a brand new unit just to rule that out. Thanks!
@@retrobitstv After staring at the schematics for this exact board, and having recently built a PicoGUS with TSSOP parts, I would re-re-flow the 245's with copious amounts of flux. Barring that, U5*, U6, U7, U9*, U10*, U12*, and U14* are involved with IDE functionality. (U15 and U16 appear to be used for ROM access only, so leave those two alone.) I've starred the ones that appear to be directly connected to the ISA bus.
Great job on the drag soldering Matt, something I've been avoiding for too long perhaps.
Thanks! I took a little practice to get it right but with all that flux, cleaning up any mistakes was surprisingly effortless :)
For the XT-IDE I think my approach would be to manually test the address decoding logic and see if the circuit produces the expected results:
So, power the card, put the card's address on the bus, and see if the 688 responds by dropping the CD_IDE line (pin 19) to 0V. Go through like that checking the different parts of the address decoding logic, the IO read/write logic, etc. to make sure each one behaves like it's supposed to - at least in terms of DC performance. Keep on like that and you should find the problem, unless it's something that's marginally bad in a way that makes it work in a DC test like that but not when the bus is switching at higher speed.
Also, sanity check, try another slot other than the one you repaired. Could be that the memory I/O on the slot is working but I/O isn't.
Thanks for the suggestions! I did try another slot but I should also try a different machine entirely as well. I'm not well versed in the finer points of troubleshooting a device like this but I suppose I could set up a test rig to power it up and have access to the I/O lines. It's a shame I didn't learn anything useful in those engineering classes.
@@retrobitstv I only did a bit of electrical engineering stuff in school. I feel like I'm learning more now, though, messing w/ old computers and picking up info from other people's projects. Probably 'cause I'm more interested, and making something from it.
When you come down to it that addressing logic is pretty simple, like "if these address bits have these values, activate that circuit". Though of course there's a big difference between it working a few times with manual inputs and a meter vs. working reliably a 8-12 MHz...
Well done for giving the SMD soldering a go, you did great and very brave for doing it on camera, I think this will massively help others to give it a go too, bravo! I stared SMD by using solder paste, a template and a hot pan of sand! Moved on to fine solder tip 1 pin at a time under a microscope (which worked really badly) until finally I'm doing it your way, good flux and dragging using a chisel tip - only took me a year or so to get there! 😅 Off top of my head I'm wondering the XTIDE ROM has settings for a Debug mode which might give more information of what it's getting from the card.... Other idea is proving each pin on IDE connector to see if you're getting expected signals on the oscilloscope, during drive detection... Like a chip select line high, data ish, addr ish, etc... IDE is fairly simple signals and should look a lot like IDE signals.... GOOD LUCK!
Yeah, my only thoughts on tracing the problem would be some sort of debug ROM -- perhaps the XTIDE project has some way of setting it up so the port sends out specific signals that can be tested in some way. Or it might also be good to try plugging in a spinning disk drive and see if it even tries spinning up or not, since that connector appears easier to probe either through the right-angle pins on the connector or the secondary drive plug on the cable
Thanks for the suggestions! I may try solder paste and hot air next time and I definitely need some kind of magnification, even if it's just they eyewear type. I did some digging earlier into troubleshooting the XTIDE but I didn't find anything that was helpful to my case or mention of a debug ROM; that would be a nice thing to have! I may have to get smart on the IDE interface and start probing some of the lines as you suggest :)
I did try connecting a spinning disk but I never got any drive activity (light or head movement) but the drive spun up when powered.
@@retrobitstv nice try, sorry spinning disc didn't work either but I guess it helps to rule out any CF specific issues. I took another look at the IDE signals and it really is a very simple interface, mostly a subset of the ISA bus but with 2 Chip Select lines already decoded by the interface card for accessing the registers on the drive/CF. I'd be looking for activity on all the lines when trying to access the drives, I might start with the CS1-2 lines to see if they go low on accessing drive. Then take a look at data bus signals and the 2 address lines. Anyway it might give clues as to what's missing. No activity would be VERY interesting 😁 anyway happy hunting! Oh I just read the Wikipedia page on parallel ATA, and checked schematics on Serge's XT-IDE-LITE which is a nice simple adapter for 8bit ide CF cards only.
Oh yeah I totally hear you about not liking surface mount soldering. I tried replacing the video DAC on an Amiga 1200 board, without having the proper tools like a microscope, and it was a total farce. :(
I'd like to see you build the new XT-IDE card, because the more retrobits the better!
Soldering the SMD stuff gets a lot easier with not too much practice. You did great if this was your first try. ;)
You just have to be very careful to do a close-up visual of the pins, and ideally test for shorts, as you did.
In some ways, a lot of SMD parts can be easier to deal with than through-hole, especially when it comes to desoldering stuff later.
I used to just flood the pins with solder, then keep both sides (or all four sides) molten, then gently lift the chip off the board.
Now, I would almost always use the hot air station to remove chips. I just have a cheap Aoyue hot air station I got about 15 years ago.
Since the XT-IDE BIOS is showing up on the PC, the main address decoder U6 seems to be working.
But I would try replacing U4, and possibly U7 and U5.
One or more of their outputs might have got zapped somehow, and a lot of the pins control the drive directly.
I didn't realize the XT-IDE had to use latches (U2 and U3) to work with standard 16-bit IDE drives. I never really thought about it before. lol
IIRC, you can actually idendify the drive ID string, and send most commands just via the 8-bit part of the IDE bus.
But most (all?) newer IDE drives do need to use the full 16-bit wide bus for transferring the sector data.
I don't know if any of that helps? lol
It looks like the CF card slot is basically on the same bus (same control signals) as the "front" IDE connector, so we know it's an IDE bus problem on the card, rather than just one of the ports (front port or CF card riser) not working.
I would definitely try replacing U4 first, then U7, then U5.
It's not too easy to debug these unless you have a Logic Analyzer, or an o'scope which can do the slow-scroll thing, so you can see the very short Read/Write/CS/Addr pulses as the system starts up.
6:07 btw, when using the desoldering gun, if you don't press too hard, it will leave barely any mark on the board.
It could potentially be cutting into the surrounding traces as well.
(You can kind of "hover" the desoldering tip just above the board/via.)
Also, I found that if you can somehow hold the board horizontally when using the desoldering gun, it helps suck up more solder from the holes/vias at once, as the gun doesn't have to work against gravity.
And yep, the soldermask is pretty poor quality on some of these boards, and I guess can degrade a bit with age (don't we all. lol)
I was working on an ICL OPD / Merlin Tonto (Sinclair QL, basically) years ago, and it has a very similar-looking but awful brown PCB.
The soldermask on that would also fall off easily, and the vias/traces themselves were fragile.
So I don't think it's just the pressure you're applying with the desoldering gun, it's probably the quality of the PCB, too.
What size/diameter is that solder, btw? It looks like maybe 0.7mm, maybe smaller?
I would try using say 0.5mm for most stuff, as it's easier to control the amount, and easier for smaller SMD stuff.
(I haven't used anything below 0.4mm for many years. Even for soldering TQFP chips with fine pins, I just do the drag soldering thing.)
Cool thanks for the tips, I will take a look at those components next. I would like to try and get my scope on them but I am a total novice and access is limited!
I am using Kester .8mm leaded solder. Yea, I have never damaged a pcb with my desoldering gun before so at the very least it's a combination of the solder mask quality and that.
Hey great job and another excellent video. Surface mount stuff is not my favorite either. Good practice for you though. ;-) One thing that really blew me away is that this is episode 86! The time has flown by since episode 1. Wow. Thanks for the great channel and content. Keep up the good work.
Crazy, huh? Glad you're still enjoying the content 3 years on and thanks for all of your support!
Great job on the ISA slot replacement! As for the SMD based XT-IDE card, perhaps a simple voltage test might provide some insight. Also, those CF-IDE adapters sometimes go bad without warning. I've gone though a handful of them in only 4 years. It might be worth trying a new one or just a plain old school hard drive without the adapter board attached.
Btw, I would love to watch a soldering montage on that THT XT-IDE card with some music similar to what you chose to close this episode out on.
Thanks for the suggestions and I will definitely do an XT-IDE build in the future :) I did try testing with a brand new CF adapter board and also a spinning disk but still nothing :( I verified the +5VDC from the card to the CF riser was good, but I didn't test for any other voltages on the card itself.
The amount of tip damage you did the the solder mask with your desoldering tool was grievous. The solder mask on those older tinfoil layer pcb's are very fragile, especially now after 35 years of age. I'd recommend using a board heater, get the pcb in the whole area up to around 38-40c first. This will enhance your efforts. Mix in a wee bit of fresh 60/40 and use a very small amount of rosin type flux instead, brushed on nice and thin. Lay the desoldering tip cone atop the joint and let gravity alone do the work. As soon as it liquifies, hit the suction till the tip slides down to the base. If solder remains, flow some fresh solder in and repeat.
Thanks for the tips! I will try that next time and be more careful. Still lots to learn! For what it's worth, I've never had that happen to any PCB I've worked on up till now. The camera shows the damage in perfect clarity, but in my defense I couldn't really see it as it was happening due to how far away I have to sit not to block the shot and the fact I'm not using any magnification.
@@retrobitstv folks commenting that you scraped off the solder mask are correct of course, but don't be too hard on yourself -- that board is like old C64 boards which have the entire surface tinned underneath the varnish. When you rework a board like that my experience is the solder on the traces tends to liquify underneath the varnish, and then you have a weak crust floating on a liquid and it's just exceptionally fragile. I don't know that it's particularly important either though except for aesthetics, since the solder also protects the underlying copper from corrosion.
The name solder mask doesn't even really apply here in fact, because since it's applied after the solder, it's not masking anything out.
Someone please get this man a plastic spudger
You scraped off the solder mask with your desoldering pump.
This. Those tips can be sharp and putting downward pressure while swishing round the pin will scrape off the mask and in time also lift/scratch off the trace.
Yea, it's strange because I've never damaged a PCB with the pump before. Either the solder mask was really fragile on this machine, or there was some kind of reaction with the liquid flux I was using because the damage was concentrated only around the pins I added flux to and not the ones I only added fresh solder to. Weird.
@@retrobitstv The board looks similar to early c64 boards, with the ground plane being bumpy due to solder resist being applied after wave soldering. In my experience, this type of solder resist is pretty weak and really only serves to stop oxidization.
@@retrobitstv some solder mask is thicker or more robust than other. But, I could see in your motion that you were pressing on the board a bit, probably trying to get good heat transfer. Not a big deal, though, as it’s super easy to re-apply either solder mask or clear nail polish. I’ve done the same, when I have damaged the solder mask, myself.
This is why I am not fond of pumps.
Is ide even compatible on the that Tandy I think uses the older standard.
The Tandy has no built-in hard disk interface at all, so you can add any ISA controller you want. It does have a built-in floppy controller. When I bought it, the machine came with an installed 'hard card' but the drive had failed so I removed it entirely. The 1000 is an XT class machine, which is exactly the type of system this XT-IDE card was designed to be used in!