ST-RAM card: White picture with red border

[wolfman]

Hi @all,

I'm trying to fix my 2 MByte ST-RAM card in my recently fixed TT030.

Corrosion is removed and the card won't be detected at system start with a diag cart.
Only 2 MByte onboard STRAM are being detected and system boots fine in the diag menu.

After lots of inspection I found out that pin 37 and 38 of the MCU were shorted and I resoldered the pins.
It seems that there are no solder bridges left at the MCU.

I refittet the RAM card and booted the diag cart:
Sometimes the TT shows a white picture (size of the desktop) with a red border, sometimes it boots fine to the menu and the STRAM card won't get detected at all.

Whats going on there?

[frank.lukas]

Take a good Picture ...

TT_STRAM.pdf

The two 96pin VG connectors must not be oxidized. Not on the ST Ram additional card and not the two ones on the mainboard.

[wolfman]

Take a good Picture ...

IMG_20230317_080532_cut.jpg

IMG_20230317_080600_cut.jpg

IMG_20230317_081449.jpg

The connectors are immaculate!

IMG_20230317_080800.jpg

The solder joints of the connectors aren't pretty shiny, but there is continuity.

IMG_20230317_082715.jpg

I reworked some pins (37-44 or so) - not my finest work (my soldering iron is crap), but pads are connected and there are no shorts.

[frank.lukas]

Work with a lot of Flux ...

mcu.jpg

You can use DeoxIT 200ml D5S-6P on the VG connectors ...

[wolfman]

Work with a lot of Flux ...

You can use DeoxIT 200ml D5S-6P on the VG connectors ...

I used a load of flux but my soldering iron (340°C) has not enough mass to heat multiple pins at once.
These 2 pins in question are not joined: thats a residue of cotton swabs after cleaning the flux away with IPA.
Yes, I used Deoxit on these connectors.

Meanwhile I can't reproduce the white screen/red border anymore: the ST-RAM-card simply gets not detected.

[wolfman]

BTW: I would appreciate a recommendation for flux and a solderstation.

I have two types of flux:
- liquid (but it burns nearly immediately away)
- gel (very sticky but somewhat useful)

[stephen_usher]

Re: Flux.
I'm a recent convert to ChipQuik flux, namely this: https://uk.rs-online.com/web/p/solder-flux/1466656

[rubber_jonnie]

BTW: I would appreciate a recommendation for flux and a solderstation.

I have two types of flux:
- liquid (but it burns nearly immediately away)
- gel (very sticky but somewhat useful)

I'd suggest that the solder may be lead free and have a higher melting point than 340C, at least that is my own experience. I tend to use 370C for everything, but a lot of it is what suits your way of working.

Personally the two types of flux I use are Kingbo RMA-218 and for other jobs I use Topnik RF800. The latter I usually use for soldering sockets and ICs in as it gets into the vias nicely despite burning away quickly.

The RMA-218 is somewhere between liquid and a gel. I do have another little tub of thick gel flux, but it really is a bit too thick, and though it melts well, it's harder than with the RMA-218, which is applied using a syringe.

For surface mount chips, to reflow I'd tend to use a soldering iron and drag solder with a clean up using wick if needs be (I use Chemtronics stuff), but whatever you do you need to avoid putting too much heat into the board as it's liable to lift pads and traces.

Hot air I do avoid generally, as older 2 layer boards can scorch easily, though the RMA-218 does seem to help as it does form a bit of heat barrier when it melts. I mostly use it for thicker multi-layer boards that are a bit more able to resist the heat. Flow and temperature is pretty important for hot air.

Solder stations, well difficult to say without knowing what you want to spend. I have a relatively inexpensive Chinese dual soldering/hot air station. IIRC it came in about £100 -£150 and it does everything I need, and I usually work on 8 and 16 bit machines. To be fair, the soldering iron was very poor until I got a new set of bits for it, and after that it has been really good. I also have a desoldering gun that was also about £100, a Duratool one, but the same model is all over the place and rebranded. I still also have a 25 year old Maplin temperature controlled iron, and that is good for areas where there are large thermal masses.

The bottom line is though, whatever type of soldering you're doing, you really need to practice or you risk damaging whatever you're working on. The biggest mistakes I see people make are to linger too long or push too hard because they're not using enough heat causing pad/trace damage and to incorrectly assume all the pins of something are free and ripping the vias/traces up as a result.

It's all about practice, and it's something you learn. I still make mistakes from time to time, and then have to make repairs later.

It's worth remembering that different boards tolerate different handling, for example my Dragon 32 is extremely fragile when it come down to soldering, but my TRS-80 Model1 can tolerate quite a bit more and still come back for more.

[DoG]

Do you have CA400313-xxx or CA401059-xxx? The later model is a 2/8Mb board and have two jumpers (W1 and W2) for configuration. Since you haven't wrote anything about jumpers I guess you have the 2Mb jumperless ST-ram board. Otherwise it's just setting the jumpers correctly (W1 for 1-2, W2 open)

It could also be a ram-chip or more that is broken that makes it not detect the extra 2Mb.

I guess you should be able to use YARTT to pinpoint the not working bits?

[wolfman]

Do you have CA400313-xxx or CA401059-xxx? The later model is a 2/8Mb board and have two jumpers (W1 and W2) for configuration. Since you haven't wrote anything about jumpers I guess you have the 2Mb jumperless ST-ram board. Otherwise it's just setting the jumpers correctly (W1 for 1-2, W2 open)

It could also be a ram-chip or more that is broken that makes it not detect the extra 2Mb.

I guess you should be able to use YARTT to pinpoint the not working bits?

It is CA400313 without jumpers.
I use a diag cartridge for TT to rule out higher levels of errors (TOS/floppy/etc).
My assumption was to "find" the additional 2MB card in the diag menu and pinpoint via RAM test direct HEX-addresses for faulty RAM chips (if there are any).
But the system can't see that additional RAM card.

So I assume a bad contact at the connectors to the mainboard or a faulty MCU.
But I have no idea if I'm on the right track.