Unreliable ROMs / sockets in BBC B

discuss both original and modern hardware for the bbc micro/electron
Post Reply
Adam James
Posts: 195
Joined: Tue May 26, 2020 2:32 pm
Contact:

Unreliable ROMs / sockets in BBC B

Post by Adam James » Wed Jul 08, 2020 9:01 pm

Hello,

I've been having strange problems when I try putting new ROMs / rearranging ROMs in the sockets of my BBC B.

It may well be that there are just arbitrary poor connection issues and I need to replace the sockets with better ones.

However, the symptoms seem to be very repeatable, and of a sort that may suggest something very specific is wrong...

Whenever I put a ROM into socket 13, it works, but whatever ROM is in socket 12 disappears from the system.

That's what it seems to boil down to. There may of course be other issues, but after a lot of confusion and fiddling around with ROMs, that is the one thing which seems certain.

I've tried lots of permutations and combinations of ROMs in various sockets. Word processors (just now), the Replay ROM, DFS ROM, BASIC ROM, and BooBip's SWRAM module. The above rule always holds true in every test so far, with the one exception that the SWRAM module, when in socket 13, appeared to work, just like all the other ROMs work when they are there, but on closer inspection there were problems: I could load the 8K WordWise into SWRAM and it seemed fine, whereas the 16K would eventually become corrupted.

The SWRAM module is fine by the way - now that it is in slot 12 and, in accordance with the rule in bold, nothing is in slot 13.

I just wondered if this pattern suggested anything to anyone, and there may be a simpler fix than having to replace all my ROM sockets!

Thanks in advance.

Coeus
Posts: 1659
Joined: Mon Jul 25, 2016 12:05 pm
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by Coeus » Wed Jul 08, 2020 9:10 pm

I'll be interested to hear what other have to say but to me that sounds like there is a dodgy connection of the socket for ROM 13. I don't think you need to rush to do anything with any of the others. As the corruption happens in the top 8K I would suspect the address line that distinguishes the top 8k from the bottom 8K is the culprit. I make that A13 so a look at the circuit diagram should give you the pin number to check.

As for why ROM 12 disappears, if the CPU is having trouble reading ROM 13 this can cause service calls that should return in a way that they are passed to lower priority ROMs to be seen by the OS as having been claimed by ROM 13 and thus not passed on.

Adam James
Posts: 195
Joined: Tue May 26, 2020 2:32 pm
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by Adam James » Wed Jul 08, 2020 10:12 pm

Coeus wrote:
Wed Jul 08, 2020 9:10 pm
I'll be interested to hear what other have to say but to me that sounds like there is a dodgy connection of the socket for ROM 13. I don't think you need to rush to do anything with any of the others. As the corruption happens in the top 8K I would suspect the address line that distinguishes the top 8k from the bottom 8K is the culprit. I make that A13 so a look at the circuit diagram should give you the pin number to check.

As for why ROM 12 disappears, if the CPU is having trouble reading ROM 13 this can cause service calls that should return in a way that they are passed to lower priority ROMs to be seen by the OS as having been claimed by ROM 13 and thus not passed on.
That all sounds very plausible, thank you:)

Is there a known way in which these things are likely to fail - e.g. is it likely to be the metal in one of the ROM socket 'holes' or the solder joint of the socket to the circuit board?

User avatar
Wheel_nut
Posts: 234
Joined: Wed May 01, 2019 1:46 pm
Location: West of Scotland
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by Wheel_nut » Wed Jul 08, 2020 10:29 pm

Adam James wrote:
Wed Jul 08, 2020 10:12 pm
Is there a known way in which these things are likely to fail - e.g. is it likely to be the metal in one of the ROM socket 'holes' or the solder joint of the socket to the circuit board?
You haven't said which Issue number Motherboard you have but there have been reports of Sockets with "thin" Pins on Issue 4 Boards which have severed below the Plastic housing and giving intermittent contact with the Board. I prefer to replace these with Augat branded Turned Pin Sockets though others prefer the Dual Wipe type.

You can check continuity with a multimeter while lifting the plastic body of the socket but this can be difficult to detect.
#1 BBC Model B Issue 7 + 1770 DFS + Dual TEAC Floppy
#2 BBC Model B Issue 7 + 8271 DFS + Dual Floppy + Speech
#3 BBC Model B Issue 7 + 8271 DFS + Cumana Single Floppy + PiTubeDirect on KenLowe's Tube Level Shifter

Adam James
Posts: 195
Joined: Tue May 26, 2020 2:32 pm
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by Adam James » Wed Jul 08, 2020 11:22 pm

Wheel_nut wrote:
Wed Jul 08, 2020 10:29 pm
You haven't said which Issue number Motherboard you have but there have been reports of Sockets with "thin" Pins on Issue 4 Boards which have severed below the Plastic housing and giving intermittent contact with the Board. I prefer to replace these with Augat branded Turned Pin Sockets though others prefer the Dual Wipe type.

You can check continuity with a multimeter while lifting the plastic body of the socket but this can be difficult to detect.
Thanks. My motherboard has a number 7 on it, though there's a label covering up that area. I presume the label is covering the word "issue"! Picture attached.

Once I've got a broad range of opinions and things to try, I'll take a deep breath (I have a soldering iron and a multimeter but find electronics scary) and dismantle everything to perform whatever checks are recommended. I'm hoping I'll only have to do this once:)
Attachments
board.jpg

Adam James
Posts: 195
Joined: Tue May 26, 2020 2:32 pm
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by Adam James » Wed Jul 08, 2020 11:42 pm

Kazzie, I'm just adding a quote from you from a different thread where I raised possibly-dodgy ROM sockets as a side-issue (a couple of weeks old now and took me ages to find!). At that point I hadn't noticed the pattern, but the info in your quote will be very helpful when I finally take my motherboard out and start checking stuff!

I want to collect all info in this thread, hope this is not against board rules / too confusing!
Kazzie wrote:
Sun Jun 14, 2020 7:55 pm
This sounds very much like a case of worn ROM socket contacts. The socket is making marginal contact with the ROM in it: if all pins are making contact it'll boot, if not, the CPU will get garbage when talking to the ROM, and crash. On those instances when your machine boots, any loss of contact will crash the machine in that state (though the video circuitry will still be flashing the cursor, so you may not notice at first).

If you want to test this hypothesis, you could use the continuity function of a multimeter to check for continuity from the shoulders of your (BASIC) ROM's pins to the shoulders of the CPU: by going shoulder-to-shoulder you'll be testing the contact in the socket. Test each of A0 to A13 and D0 to D7 between the two chips; the most obvious way of finding which pairs of pins to check would be by looking at a circuit diagram (e.g. this one) to check where A13 etc. are on each chip. Do bear in mind that things may flex a bit as you poke them with a meter probe. ;)

Kazzie
Posts: 1781
Joined: Sun Oct 15, 2017 8:10 pm
Location: North Wales
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by Kazzie » Thu Jul 09, 2020 8:27 am

Hi Adam.

The investigation you've summarised in the first post is top-notch work, well done.

The suspect pin, pin 26 or A13, of the sockets for ROMs 12 and 13 (IC52 and IC58) are connected together, and are then connected via jumper/link S32 to either 5V or line 13 of the Address Bus. Setting this link to 5V is intended for using smaller 24-pin ROMs in the sockets, which would expect a power supply where the larger ROMs have A13. As a result, it's invariably set West. Setting it East with 28-pin ROMs fitted would tie input A13 high, enabling only the top 8K of a 16K ROM, or having no effect on 8K ROMs. If you have a pair of 8K ROMs (possibly on 27C64 EPROM chips?) you could try those in those sockets with S32 East, and see if they work properly. (Replay Mk2 won't as it's a 16K ROM, as are the 1770 DFS and the BASIC ROM.)

There is the outside chance that using a soldering iron to reflow the solder on the back of the A13 pin might be enough to settle the metal contact in a better position, and solve the problem (for the time being at least).

I recently replaced a 28-pin socket on my Issue 7 board because of poor contacts, but in my case it was the socket for IC73. My big ATPL Sidewise ROM board fits into it (an the OS socket), and some of the socket contacts were getting very tetchy, requiring regular reseating of the board.

My socket was a single-wipe one: looking into the holes from above there was metal to be seen on one side of the hole (the inner side) but not on the other. The vacant socket in your photo looks very similar to the ones on my board: they may also be single-wipe ones. With a twin wipe socket there's a second sprung contact on the other side of the pin that cal "fill in" if one contact is slightly loose; in a single-wipe socket there's no fallback.

Trying to clear all twenty-eight holes of solder with just a soldering iron and solder sucker/wick is fiddly job. But while working on removing my socket, I found that the plastic portion was only weakly connected to the metal pins (the heating from my soldering iron as I was sucking up solder from the rear of the board may have loosened them). I was able to lever the plastic socket off the board with a flat-head screwdriver, while leaving the remaining pins soldered to the board.
IMG_20200623_1242282.jpg
After that, I was able to remove the pins and clear the holes individually, by standing the board vertically, applying the soldering iron to the underside and pulling the pin out with pliers from the top side. That was much easier than trying to get all 28 holes clear before moving the socket.
IMG_20200623_1249265.jpg
If the plastic portion of the socket doesn't lift off freely, you could use a small pair of cutters on the plastic to cut it into two halves, or even smaller chunks.

When it comes to fitting a replacement socket, in your case I'd specifically suggest a dual-wipe socket as opposed to a turned/machined pin socket. While the latter are considered to be higher quality, they only accept the round pins of ICs, but not the blade-shaped pins that are (probably) on the base of your Replay ROM carrier.

I'll let you judge whether this sort of operation is something you'd feel comfortable doing. If not, it may be possible to find someone nearby to help you do it. Whereabouts are you based?
BBC Model B 32K issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM, ZIDEFS with 512MB CF card
RiscPC 600 under repair
Acorn System 1 home-made replica

User avatar
1024MAK
Posts: 10102
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by 1024MAK » Thu Jul 09, 2020 8:42 am

CDBDB42D-1E22-4B9D-A0A6-70C397ABEEB7.jpeg
BBC Model B sideways ROM socket signals
Pin 1 is not connected (N/C).
Pin 20 (/CE - chip enable) is also known as /CS (chip select).
Pin 22 (/OE) is output enable.

Mark

johnkenyon
Posts: 245
Joined: Wed Jul 20, 2011 3:21 pm
Location: Coventry
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by johnkenyon » Thu Jul 09, 2020 10:42 am

Kazzie wrote:
Thu Jul 09, 2020 8:27 am
After that, I was able to remove the pins and clear the holes individually, by standing the board vertically, applying the soldering iron to the underside and pulling the pin out with pliers from the top side. That was much easier than trying to get all 28 holes clear before moving the socket.
My technique is to have a spot of solder on the iron tip to ensure a good thermal contact, then heat the end of the contact where it curves down then lift the contact out of the board. All done with the board lying solder side down, on the bench. A pair of tweezers or a hook soldering tool in my left hand can be used to persuade any legs out.

Once all the legs have been removed, I then add a little solder to any holes that aren't completely clear, then use a solder sucker to give me 28 completely clean pins (usually done from the solder side as the individual pads tend to be bigger)

(For an encore, when removing a duff soldered IC, I use side cutters to disconnect each leg close to the IC body, remove the body and use the same technique as above to remove each leg individually)

shakesc
Posts: 44
Joined: Tue Apr 24, 2018 3:34 pm
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by shakesc » Thu Jul 09, 2020 11:01 am

desoldering braid is my greatest find :D

So much easier than solder suckers (manual and electric). Just wish I tried it sooner :lol:
BBC B+ 128K

Kazzie
Posts: 1781
Joined: Sun Oct 15, 2017 8:10 pm
Location: North Wales
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by Kazzie » Thu Jul 09, 2020 11:06 am

shakesc wrote:
Thu Jul 09, 2020 11:01 am
desoldering braid is my greatest find :D

So much easier than solder suckers (manual and electric). Just wish I tried it sooner :lol:
I've tried it many a time, but have yet to get good results myself: I keep turning back to my solder-sucker.
BBC Model B 32K issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM, ZIDEFS with 512MB CF card
RiscPC 600 under repair
Acorn System 1 home-made replica

User avatar
KenLowe
Posts: 1230
Joined: Mon Oct 18, 2004 5:35 pm
Location: Scotland
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by KenLowe » Thu Jul 09, 2020 12:08 pm

Kazzie wrote:
Thu Jul 09, 2020 11:06 am
shakesc wrote:
Thu Jul 09, 2020 11:01 am
desoldering braid is my greatest find :D

So much easier than solder suckers (manual and electric). Just wish I tried it sooner :lol:
I've tried it many a time, but have yet to get good results myself: I keep turning back to my solder-sucker.
And I've landed on a combination of solder sucker, followed by heat gun if I don't want to damage the component. Or a pair of snips and pliers to pull the legs as shown earlier in this thread, if the component is going to be replaced.

Adam James
Posts: 195
Joined: Tue May 26, 2020 2:32 pm
Contact:

Re: Unreliable ROMs / sockets in BBC B

Post by Adam James » Thu Jul 09, 2020 12:48 pm

Kazzie wrote:
Thu Jul 09, 2020 8:27 am
The suspect pin, pin 26 or A13, of the sockets for ROMs 12 and 13 (IC52 and IC58) are connected together, and are then connected via jumper/link S32 to either 5V or line 13 of the Address Bus. Setting this link to 5V is intended for using smaller 24-pin ROMs in the sockets, which would expect a power supply where the larger ROMs have A13. As a result, it's invariably set West. Setting it East with 28-pin ROMs fitted would tie input A13 high, enabling only the top 8K of a 16K ROM, or having no effect on 8K ROMs. If you have a pair of 8K ROMs (possibly on 27C64 EPROM chips?) you could try those in those sockets with S32 East, and see if they work properly. (Replay Mk2 won't as it's a 16K ROM, as are the 1770 DFS and the BASIC ROM.)

There is the outside chance that using a soldering iron to reflow the solder on the back of the A13 pin might be enough to settle the metal contact in a better position, and solve the problem (for the time being at least).
Thanks for all that info and the photos. Hopefully re-flowing the solder at the correct pins will sort it!

The thought of trying to replace my ROM sockets makes me feel as physically sick as the thought of jumping out of a plane would make some people feel:) If that turns out to be what is needed I may just accept that I need to leave socket 13 empty, or as you say I may see if anyone local can do it.
1024MAK wrote:
Thu Jul 09, 2020 8:42 am
CDBDB42D-1E22-4B9D-A0A6-70C397ABEEB7.jpeg
That's great, thanks.

Post Reply

Return to “8-bit acorn hardware”