Flux the chip up then heat it up with a hot air station using a circular motion making sure to only use enough heat to let the chip come loose. (I wouldn't be removing a chip like this with a soldering iron)

If you are careful with the hot air station then it shouldn't get excessively hot.

+heat from below

oh that takes me back... I used to do the same in the 2000s
but like others said, hot air is the best way
even a cheap station will do wonders in cases like this

rasz_pl wrote on 2022-05-01, 08:48:

+heat from below

Depending on how many ground planes the board has this may not do anything or take an excessive amount of heat.

TrashPanda wrote on 2022-05-01, 09:22:

may not do anything or take an excessive amount of heat.

Why? The board definitely will dissipate some heat, making hot air gun less efficient, and tempting user to raise the air temperature.
And if for bottom heat you use something that's not too powerful (I was using a spare heatbed for 3d printers), it just slowly preheats everything to a safe ~100 deg. The air gun would then add just enough heat to melt the solder, and overall thermal stresses would be minimal.

In my experience these chips are usually pretty resilient, you really need to go overboard with the hot air gun in order to damage them. Keep in mind these are designed to survive the factory soldering process where they are basically baked in an oven for a few minutes. As long as you keep the hot air gun moving around the chip to heat the legs evenly it has a very good chance of surviving.

Recently I desoldered a couple of S3 Trio64V+ chips from donor cards for the custom VLB card project by Madao and they both survived the ordeal. I just pre-heated the donor card around the chip, then set the rework station to 350C and after a while the chip came out without damaging any traces. Just make sure you don't use any force when removing it, although you may not care about the pads in the donor board surviving, you may also bend or even break the legs on the chip. It's better to wait a while until the solder on the joints is fully melted.

How much do you like the donor board? Cutting out the chip from the board leaves you a lot less ground plane pain. ... and if there's nothing else of value left on the board, taking a blowtorch to the backside and banging on it until the chip falls off can work.

As others have said, flux + hot air station is the way to remove SMD chips like the above. Plus can also be used to re-solder them more efficiently than an iron.

Thanks to everybody for your help!
I will dig out the donor-board as soon as I have some spare time and check it´s condition. It would be useful to do some basic tests to make sure the chip is in good working condition - this means to get the donor-board running first.
I will report here as soon as I have any news.

not more than others said 😀
BUT... a safely thing you can do, is put the donor card in oven for 1 or 2 hours at 50/60°C before desoldering.
why ? because of moisture. strongly heating old chip that have been in storage for long can have absorb moisture.
and when you strongly heat them, mositure turn very quickly into steam and... can go out of the chip like a gun bullet... and potentially damage it.

O.k. - yesterday I got hold of the donor-board. The BIOS chip was missing, so I gave it a try with the Symphony BIOS of my FIC 486SC-P. It started, but cache and RAM were not recognized correctly. I was unable to find an original BIOS online and now I knew the donor-chip was working so I decided to start the slaughter.

First I drilled 4 holes from the upside to mark the area to be cut out without damaging any legs of the chip. Then, 4 quick cuts later the chip dropped out:

Before, I made two diagonal cuts also to split the area in 4 quarters and each row of pins could be easily removed after heating the pins carefully with the hot-air station.
This is what I got:

There was very little heat needed for a short time only and no leg was bent!
Then all the pins had to be cleaned carefully and I applied a little fresh solder to all pins.
Finally the transplantation could take place and the result is as good as if I had used a new chip:

So far the system is working flawlessly. But I will have run tests for a while to make sure all issues have been solved.
For the first time the L2 cache waitstate can be jumpered for 0 ws now. This had resulted in loads of errors before.

This is real MacGyver work. Kudos to your patience! 😀

so you took a fully working board and cut it to pieces to fix another one? seems dumb

I never even tried desoldering a single pin like those on a Am386DX-40 to disable it while people can replace them entirely.. I wish I had more soldering skills and tools. 🙁 😁
Not to mention to change modern DDR3 modules on a OrangePi SBC to increase ram size.. I'd like to try that but I suppose it's too much difficult.

maxtherabbit wrote on 2022-05-18, 14:38:

so you took a fully working board and cut it to pieces to fix another one? seems dumb

Not necessarily, if you´re 100% sure which one you want to keep and if you have been searching for a new (NOS) replacement chip for 5 months with zero success.
Unfortunately this chipset is quite rare and there are just a few mainboards using it.
And I think this beauty is worth it:

Still wondering what to do with the donor board - maybe I´ll put it up on Ebay as "not tested" for $ 300.00 😉

Or I could harvest the KB-connector, 8 SIMM sockets, the power-connector, the other two Symphony chips and 8 ISA sockets.

majestyk wrote on 2022-05-18, 15:34:

Not necessarily, if you´re 100% sure which one you want to keep and if you have been searching for a new (NOS) replacement chip […]
Show full quote

maxtherabbit wrote on 2022-05-18, 14:38:

so you took a fully working board and cut it to pieces to fix another one? seems dumb

Not necessarily, if you´re 100% sure which one you want to keep and if you have been searching for a new (NOS) replacement chip for 5 months with zero success.
Unfortunately this chipset is quite rare and there are just a few mainboards using it.
And I think this beauty is worth it:

486SC-p_tot1.JPG
Still wondering what to do with the donor board - maybe I´ll put it up on Ebay as "not tested" for $ 300.00 😉

dead_dead_good.JPG

Or I could harvest the KB-connector, 8 SIMM sockets, the power-connector, the other two Symphony chips and 8 ISA sockets.

it's your property and you're free to do what you want of course, but still seems ridiculous to me

why not wait until a board ravaged with battery damage comes along to lift a replacement from? or simply remove the chip using chipquik so as not to destroy a perfectly functional PCB? just so many better options

(I have a young micro MB with the SYM461 chipset, they are cool I agree)

It seems I have yet another case of a defective Symphony SL82C362.
It´s on a FIC "386SX-20/25S":

The mainboard POSTs, video, keyboard etc. are working, RAM (8MB at the moment) and Cache (128K) are detected correctly, but no matter what super-I/O I try it cannot boot from floppy or harddisk.
"Diskette Boot Failure - Insert BOOT diskette in A:"

So I probed all the address-, data- IRQ- and other lines at the ISA-bus with an Ohmmeter. All of the have the typical resistance against ground (red probe grounded), so they are at least connecte to a working input or outlut stage. But there´s one exception - DMA Request Line 2 (pin6) has infinite resistance. It´s connected to pin 83 of the bus-controller chip SL82C362SX, the trace is not interrupted.

Would this problem fit the non-working floppy symptom?
And is there any difference between the commonly used SL82C362 and the - quite rare "SX"-version on my mainboard?

Not to criticize your approach, but there's a few methods for removing surface mounted components that are cheap, easy and nondestructive. Just for future reference...

As easy as hot air can be, there's an even easier and cheaper method for removing surface mounted ICs. It's called ChipQuik. Here's a 39 second video that explains it well. It's extremely easy to use.

858D is an affordable (sub $50) hot air station that works really well. "858" is a generic model that many faceless brands copy. I have the CO-Z brand 858D that I got from Amazon and it works well. You can find many reviews of the 858 on YouTube, and I would suggest watching a few. Here's a random video of removing a QFP chip using hot air. There are many hot air tutorials on YouTube, and I would again suggest watching a few of them to get a good idea of technique and settings. But in short, using hot air to remove surface mounted stuff is usually very fast and easy. And it takes very little practice to get the hang of it.

Thanks for the info.
Maybe I shouldn´t have reopened this thread from more than a year ago. At the moment it´s all about the question if the odd DRQ2 line behaviour is caused by a defective bus-controller.

As for removing chips it depends on the situation. When a known defective chip has to be removed, cutting the pins with a dremel and removing the remnants with a standard soldering tip is done in about a minute. It´s fast enough and the surrounding components are not exposed to heat (you must take care of the dust though).

If the chip will eventually be used again, because it´s not sure if it´s defective, I use my regular hot air station for that.

When it comes to the Symphony Haydn II chips, you have to be aware of their extreme rarity plus they are prone to fail often and easily. This is the reason why I took extreme care and even applied a "destructive method" when removing the chip from the donor board.