Superbrain #3 : The Superbrain II board

[1024MAK]

Incidentally, you should see the total lash-up I have on my desk... this'll make you laugh!

Who said you shouldn't muck about with CRTs outside their cases?

Nice lash-up

But I think you are missing a vital item. Where is the cup of tea / soft drink?

Only people scared due to the internet being full of death/doom mongers have a thing about uncased CRT displays... Go back to only ten to fifteen years ago, and most TV shops that carried out repairs would have had many sets running on soak test with most of their cases / covers not fitted

No different to having an internal combustion engine running while not housed in a chassis.

Safe as long as you are not stupid or silly enough to put your bits in harms way

Mark

[PS I may be more than a bit biased, as in my job, a lot of the time we work on live electrical systems].

[leenew]

Any good for you Jon??
https://www.ebay.co.uk/itm/Green-Screen ... ctupt=true

Lee.

[jonb]

Good spot, Lee.. I'll think about it (having just bought a 9" B&W monitor)... If it was white or amber I'd snap it up!

[jonb]

Definitely on my way. Decided the breadboarded sync mixer wasn't cutting it, so did some seriously old school electronics on a bit of blobby board:

Top view, nothing to see here..

But underneath, it looks like a logic board from a 1960s IBM mainframe!

Not sure it works yet.


Edit: No, of course it doesn't.

[jonb]

Hmm, I think I fixed it but it looks like the GT65 has had it for now.

Drat, drat and double drat!

[jonb]

I spoke too soon, obviously.

Not only has the GT65 failed, but the LCD TV behind it just went POP and released some magic smoke.

On the underside, we can see that the tiny surface mount semiconductor, Q801 has completely exploded.

And the big transistor, Q802, is split in half!

Not having much luck today!


It might be worth repairing, if I could find out what Q801 is...

[jonb]

Some better news. I had another fiddle with the GT65 and it is working again.

Deflection is still bad, though, but for some reason I prefer it to the LCD. I also acquired a small 9" monitor from a well known auction site for the princely sum of £17 but it turns out to be not working well at all. A disappointment (thanks to a vague description that implied it worked) but it may be useful for parts. Or, I may get it working again (pigs may fly, too).

[jonb]

Hitachi 9" monitor repair thread: viewtopic.php?f=45&t=16381

Image generated by the SB II: download/file.php?id=43326&mode=view

Image generated by an Atom: download/file.php?id=43331&mode=view

Bearing this in mind, and the similarity between the images generated by the Hitachi and Amstrad GT65 monitors, I think the Superbrain's video signal is off slightly. I think I need a bit of advice on diagnosis. To recap: it provides separate H and V sync, plus luminance and I am using one of MNark's sync mixers plus my own PCW video mixer to generate composite sync. I can get an image on the GT65 if I fiddle about with the B+ power supply but I think that is luck, and the image is too wide. On the Hitachi monitor, prior to the repair, I could see the image albeit squashed to the top of the tube. A consequence of bad vertical deflection.

I could swap out the CRTC or other ICs that interact with the sync pulses.

I did notice one other thing. If I disconnect H or V or both sync signals from the mixer board it doesn't seem to affect the signal.

I can apply my scope to the H and V sync signals, but what should I look for?

[1024MAK]

At this link there is details of what a 625 line, 15625Hz, 50Hz (in other words, a PAL) video signal should be like. Complete with timing information.

Mark

[jonb]

OK. I've double checked the sync mixer and it looks OK.

Here is a scope trace. CSYNC is the mixer output and HSYNC / VSYNC are the inputs.

Y=CSYNC, C=HSYNC, B=VSYNC

Zooming into the VSYNC:

Y=CSYNC, C=HSYNC, B=VSYNC

Question - should the HSYNC pulses still be present in CSYNC while VSYNC is low?

[jonb]

More traces..

Y=Composite Video, C=HSYNC, B=VSYNC

The composite video signal doesn't look much like the one on http://martin.hinner.info/vga/pal.html

Let's zoom in..

Y=Composite Video, C=HSYNC, B=VSYNC

This is the start of a line. I'm a bit surprised that the pixels are not proper square waves but they do result in some display (which can be correctly resolved by the LCD monitor). I cannot see a back porch:

Y=Composite Video, C=HSYNC, B=VSYNC

..although the line duration looks to be correct (64uS according to the info site).

[jonb]

Ah...

After some adjustment (of the video mixer trimmer and H-HOLD) I seem to have a picture. Deflection is still way out though (too wide, too tall). Let's see what adjustments this monitor can make. Probably not enough, like the GT65 - I think the Superbrain's video output is to blame.

[jonb]

This is about as good as it gets. In order to shrink the image I had to adjust the B+ voltage (should be 12v, probably nothing like that now) as it has the side effect of zooming the picture (and changing the brightness). I've also adjusted V-height and Linearity, and H-width as far as it will go. Despite the picture, the image is not particularly bright or clear. upping the contrast or brightness results in blooming that blurs the image (and there is no focus adjustment).

[jonb]

Another minor update.

In order to get access to the SB II's serial ports, it's necessary to build a stand in for the "Intermediate Board", which is a small daughter card that provides the main and aux serial ports, as well as a place to mount the SB II CRT's brightness control. It connects to the 50 way IDC plug at the top of the main board (JP1), so I had to fish out the proper connectors. I built the thing out of prototyping board and point to point wiring and used an ATX DB9 connector. Only the Main port is implemented at the moment. After the usual faffing about with null modem adaptors and the like (not needed here as it turned out) and grounding the /RTS line on the main USART, I was able to use a USB-Serial adaptor to transfer files to/from my PC with Kermit.

So now the "lash up" looks quite a bit better:

Next steps will be to build a case for it in plywood, then see if the GT65 parts can be made to show a better image. I'll probably do the case in two parts; a base to hold the boards and PSU with a removable shelf for the monitor and HxC to sit on; then once I find a 12" monitor or get the GT65 working I'll build out the top part of the case that actually encloses it. I'm planning on making it look like a real Superbrain II (that is probably a bit too ambitious!) and I'll be on the lookout for a pair of full height floppy drives as well. I've got some BBC 1/2 height drives but they're only single sided so will not be able to read the QD spec floppies.

I have to say, this SB II main board project has been tremendous fun so far!

[jonb]

GT65 is kind of working now, thanks to some very helpful guys over at CPCWiki.

To reduce the width, it turned out that a reduction in one of the high voltage capacitors feeding the LOPT was necessary (wise? We will see!). Full details, if you're interested, here: http://www.cpcwiki.eu/forum/amstrad-cpc ... djustment/.

So now I can proceed to the "woodworking" phase, because of course it needs a nice box to sit in.

[mister35mm]

I have always wanted a Superbrain, or something similar.

[jonb]

There is nothing similar. It's my favourite CP/M machine.

But you might consider:

• A TRS-80 Model II (with 8" floppy drive) and Lifeboat, Aton or Pickes & Trout CP/M (I have one)
• A TRS-80 Model IV with CP/M
• A TRS-80 Model IVP (portable), again with CP/M
• Northstar Advantage
• Kaypro 2 (but has 9" screen)
• Philips P2000C (this has a built in SASI interface) (also 9" screen) (I have one)

The problem is they are quite rare and thus, valuable - to the right buyer.

All of these machines are supported by cheap-ish IDE implementations of one sort or another. For instance, I've just finished implementing uIDE drivers for the Superbrain. You can boot and run the whole thing without accessing the floppy drives. (uIDE is my own universal IDE interface for Z80 machines).

But if you really want a cheap CP/M fix, go for an Amstrad PCW. Or even better, build one yourself in a cheap FPGA board : http://searle.hostei.com/grant/Multicom ... gaCPM.html

Then make it go like hot snot off a shovel: https://www.retrobrewcomputers.org/doku ... jonb:start

CP/M at 50Mhz!! Whoosh!

Then make it look cool: viewtopic.php?f=41&t=12954&p=222030#p221457

[jonb]

So, after acquiring a lovely little amber monitor from a well known auction site for not much money, we have this:

Close up.

Mmmm lovely amber goodness...

I think there are still some adjustments needed (especially to the focus) but overall I'm pretty chuffed. I could probably do the uIDE drivers for it now (if I wasn't busy with the uRTC project)..

[SteveH]

You teasing us all Jon? What's this about a uRTC project?

[jonb]

Hi Steve

uRTC - Universal RTC for Z80 computers.

Fits onto the Z80 bus used by uIDE and attaches as shown using stand offs.

Like uIDE it is programmable for the base I/O address, but the RTC chip needs 8 of them as it has 8 registers. I am using the Superbrain 1 to develop the utility suite as my Amstrad PCW serial interface has packed in so I cannot transfer test code to it. There is a discussion over on the CPCWiki forums if you're interested : http://www.cpcwiki.eu/forum/nc100-nc200 ... ad-pcw-rtc.

(The machine featured in this thread is a Superbrain II which already has a battery backed RTC on-board.)

[jonb]

A little lockdown relief.

Since the last photo, this SB II board has been packed away. Well, I say "packed away", but what I really mean is "stuffed into an antistatic bag and leaned against a bookcase in my study" as I have nowhere safe to store it. However, yesterday I took it out and connected it up again as you saw in the pictures. The purpose of this was to sort out a Gotek FDD emulator for it that I'd bought a while back and just dumped in the bag with the main board.

So, following Chris Cochrane's excellent guide on YouToobz : https://www.youtube.com/watch?v=-K31S2xqZIk I reflashed the Gotek with FlashFloppy firmware. I found the ST Microelectronics download page for the Flash Loader Demo software was broken, even after registering for the download, so grabbed it from somewhere else. It took several attempts to reflash the Gotek - in the end I managed it at 9600 baud. Then I fitted a small OLED display and piezo buzzer and it worked first time.

Gratifying..

I wanted to check the RTC on the main board but the boot image I have on the Gotek doesn't have the correct utility (TIME.COM). I have a SB II boot floppy here but as my SB II board doesn't have an actual drive I had to transfer it from the other Superbrain via Kermit. I'd normally do this by sending to my PC, swapping the serial cable over, then send to the SB II but thought I'd try it directly SB to SB II using a null modem cable. Happy to report that it, too, worked first time.

Finally, I have begun the next project, and that is incorporating uIDE drivers into the SB II BIOS. I've dumped all the memory to a text file (including the BIOS) and got it transferred to the PC. I'm now disassembling and commenting it. So far - and surprisingly - it is different to the Superbrain BIOS, so there will be a fair amount of work to do. I'll need to do the same to the boot / floppy ROM as well... but it's all good fun, right?

[jonb]

And now some more lockdown relief: uIDE drivers for the Superbrain II.

http://www.vcfed.org/forum/showthread.p ... post620734

I now turn my attention to making some sort of case for the Superbrain II board. Can't leave it on my desk, naked. It'll get damaged.

The Superbrain case a very particular shape made out of some sort of moulded plastic which Intertec said was recycled. It is tough as old boots and not particularly heavy. The form is sculpted with compound curves and as such it is distinctive to the brand / model. The front panel (around the screen and floppy drives) is a separate piece, but the rest of it is a single moulding. There is a second moulding on the base (you can just make out that sliver of grey bottom right) and they fit together precisely to leave a clean join line. However, the shape itself isn’t precise – you can see where there are inconsistencies of form, such as the radius of the top left and right corners of the hood. Almost like the mould was created from a hand made buck.

So how to reproduce it? I had some thoughts...

• Use thin plywood sheets and bend over formers with heat or steam. This would give an approximation of the hood (the bit in white on top) and maybe even the keyboard surround. The keyboard bezel and screen / disk surround would be separate. With clever cutting the front overhang can be incorporated, but the top edges, while curved, wouldn’t drop down towards the back as you can’t get a compound shape by bending wood.
• Again use plywood, with wooden inserts wherever there is a compound curve. The idea here is to build a frame from wood that follows the edges of the case, and shape it where necessary to achieve the correct outline (like a wire frame model I suppose), then infill with ply panels. There would need to be a recess for these, and the whole thing once complete would need to be filled and spray painted so as to look just like the original.
• Instead of wood for the frame, use 3d printed plastic. I’d have to get me a 3d printer and learn to make 3d models. Expensive, but it’s a hobby I am looking into anyway.
• Use the existing case as a buck to make a mould out of fibreglass. This approach will probably be the easiest, but how do I protect the original from damage?

Any other ideas?

[jonb]

Thread resurrection... The Superbrain II is stirring....

I've been exceptionally lucky. I bought a 3D printer thinking I could print a new case for this Superbrain II board, and fit all the gubbins inside it, but the problem is the SB case is full of compound curves that are hard to model in 3D if your are a novice like me. So I posted a wanted thread on VCF for a Superbrain case on the off chance, and some very kind soul replied saying he has a spare (really, he has two Superbrains but wants to keep one machine plus the mainboard for the other as a spare), and he has let me have it, along with a PSU and CRT / LOPT / Video board. In other words, everything I need (apart from the floppy drives) to turn my SB II main board into a fully fledged Superbrain.

Because this case and PSU are from an early Superbrain (not II), it will not just connect to my SB II board. Intertec updated all the connectors in the II, so I will need some sort of adapter. The Superbrain PSU board also carries the two serial ports as well as the video signal, and a single 22 way ribbon connects it to the Superbrain main board. So I will need to have a little adapter that takes that 22 way cable and breaks it out into a 7 way power connector, 7 way video connector and 50 way expansion (IDC) connector (because the serial lines come out of the SB II's expansion port).

This is what I came up with:

Adapter schematic

Adapter PCB

There's a spare 50 way socket for onward expansion, a uIDE socket so I can move that board off the main board and recover a Z80 shim, 2 power sockets, 2 floppy power sockets, a fan connector and the 22 way ribbon connector. This'll make it easy peasy to integrate the SB I parts with the SB II board. I ordered a set of three from OSHPark at great expense, so if anyone wants one do let me know!


Meanwhile, I decided the keyboard needed a clean. Just the usual pulling the keys, wiping the key switch stems with Armor All to help them move smoothly, wash and dry the keys and refit. I actually thought these were museum grade key tops as the board hasn't really been used (it was supplied as a service part to an Intertec UK service engineer who goes by the name of Talywain round these parts), but on closer inspection I can see some discolouration. Oh well, let's call it "patina".

The next consideration is the lack of floppy drives. I do have one Tandon TM100-2 here which I thought was broken but it turns out it is good. This will be fitted as Drive B:, with a Gotek as Drive A: for convenience. Don't forget the machine has uIDE so the floppies won't see that much action anyway. Obviously a Gotek needs some help to fit in the huge space vacated by Drive A:, so I had a look on Thingiverse and found a couple of models I could chop and hack together to create a printable part to mount the Gotek at a good position in the Superbrain case.

I printed it in black PLA and fitted the Gotek PCB. Then, a trial fit inside my existing Superbrain (in the right hand bay).

I think that will do nicely. The new case arrives in five days so I will spend the time between then and the adapter arriving from OSHPark checking its PSU and video boards out. More soon!

[jonb]

So, to the new case. Or rather, "old and a bit ratty case".

Front of the case, looks proper ratty.

Rear of the case, looks OK

Screen and drive bezel.

But the keyboard bezel is bad, needs a respray

Tube and support chassis after a clean.

The video board needs a scrub. Not sure if it works yet.

To my amazement, the PSU appears to be working (with dummy loads on +5v and +12v rails). Does need a clean and recap though.

So I think we're off to a flying start with this one... even the "Mini Brute" 115v AC fan works (first time I've seen a SB with a working original fan). Colour me amazed!

[jonb]

A pretty full day of cleaning boards, recapping the PSU and reassembly.

Trial fit of the PSU and mainboard.

A temorary solution to connecting the main board's video lines to the PSU board.

The video board and tube reinstalled in its chassis.

Coming together...

Now, while this unit is booting from the Gotek board inside that 5.25" carrier case (which means the PSU is working, supplying correct voltages etc), the screen is not displaying anything. It's getting power - I can see the tube heater glowing, but there is nothing showing on the screen. I double checked the video connectors, all good. So it looks like another video board issue, but with no obviously damaged components to guide me. I've checked for dry joints and replaced the bipolar capacitor (actually, that my be the problem so next job will be to swap the old cap back). To rule out the video mixer on the power board, I will swap in my working video board tomorrow.

So close!

[leenew]

Coming together quickly now.
Well done so far!

Lee.

[jonb]

I tried my other video board in this case and the screen came to life, so it's this video board that is dead. As expected, really. Then, a proper setback. One of the PSU transistor heatsinks grounded against the chassis (none of it was screwed down) and the PSU promptly died.

Here we go again... groan!

[leenew]

Gah! Talk about speaking too soon

[jonb]

You know me mate... one step forward, two steps back.



I have given up for the day. Back to the PSU tomorrow. It’s probably something silly like a fuse. The board has several of them, though they look like tiny resistors with white ceramic unmarked bodies. Could be tricky to source.. we’ll see.

[jonb]

So, the short occurred between the grounded chassis (the folded gold anodised corner) and the heatsink of the nearest transistor. I did hear a spark as I moved the base of the machine slightly. The design is slightly different between the two PSUs:

The one I already had has a T03 packaged power transistor - a Motorola MJ3030 - and the heatsink has fingers to increase the surface area for good heat dissipation..

On the PSU I just zapped, it has a different package for the transistor - TO220 UMT 13006, and the heatsink is just a folded piece of flat aluminium. This is what touched the chassis and caused the failure.

It is labelled Q1 on the PSU schematic: download/file.php?id=28726&mode=view

If the collector is connected to the tab as I guess, then something like R1 may have burned out.. but I'm guessing! Or maybe the 3006 is dead? I can hear the PSU tweeting every 1/2 second as it tries to start up - it's a similar sound to when there is no load.