Up and Running

[bugman]

Well, got it running last night. Have not completed the back panel yet, or tried the terminal interface. But the main board and control board are in and I put power to the mainboard. Seems to be operating OK. Am able to observe memory locations and deposit new data. Will try a simple program tonight.

I do have one concern though that may or may not be normal. When I initially power up the mainboard, all the leds come on for a bit. Then they begin to correctly reflect the memory location and data. But the panel does not become stable for about 10-30 seconds. So to re-cap, upon power up, all leds light up for about 5-10 seconds. Then the leds operate correctly, with a few (seemingly random) flashes of all leds on. Then it becomes rock stable after 10-30seconds. I notice that the effect is more dramatic if the computer has been sitting off for a while vs. turning it off and on.

I'm no expert, but it feels to me that this is either, normal cpu cylcing, or voltages becoming stable/capacitors charging. It does not feel like an intermittent connection because it doesn't last more that 30 seconds max.

Any thoughts?

[Geoff Harrison]

Does that happen when the halt/run switch is in the halt mode?

With mine, if I power up with the switch set to halt, the lights are instantly stable.

Geoff.

[bugman]

Yes, it happens when started in the halt mode. Do you expereince any delay at all when powering up?

[Geoff Harrison]

No, no delay at all. I'm not using the switching power supplies, but I doubt there would be any difference. When I power up there's no discernible delay before the the lights show whatever the switches are set to.

Geoff.

[bugman]

Any ideas where I should start to look for a problem? Some of the ICs were difficult to insert, and some of the leads had oxidation on them. I was thinking of starting by re-seating some of them.

[Geoff Harrison]

I was thinking of starting by re-seating some of them.

That's where I would start. Also, make sure your power supplies are not taking a while to stabilize, but that's a long shot. It might be instructive to pull the CPU chip and power it up. If the front panel is immediately stable then something is not halting the CPU on start up. If it isn't stable when the CPU isn't there then perhaps the front panel itself is flaky. You should be able to examine and modify memory from the front panel even without the CPU.

Edit: Well, maybe not. I checked that the front panel worked without the CPU before posting that and it worked fine, then I checked it again after and it doesn't work, so take that statement with a grain of salt. I'll try it again and let you know what I find.

Geoff.

[Geoff Harrison]

Well, I guess I don't know what I'm talking about. I can't get the front panel to do anything without the CPU now, so forget what I said.

Geoff.

[bugman]

Thanks, I'll do some troubleshooting tonight when I arrive home. I remember that my incoming 16 volts from the PSU is more like 14.7. I could also try tweaking it closer to the speced 16 volts.

I noticed that yours was over 16 volts.

[Geoff Harrison]

Ok, I knew I wasn't (completely) crazy. You can run the front panel without the CPU, but you have to pull BA (CPU pin 7) high. I removed my 6800 and connected CPU pin 7 to pin 4 (IRQ) which is tied high through a 3.3k resistor, and my front panel works fine.

Does anyone know why it worked for me once and then stopped? BA is driven through CMOS buffer PP, can a CMOS buffer whose input is floating drive its output high some times and low other times, at random?

Geoff.

[bugman]

I followed your lead in a different way. I plugged the 6800 cpu directly into the motherboard without the 32k expansion. All symptoms dissappeared, and it runs fine. So, its something with the cpu to expansion BRD connection, expansion BRD to motherboard connection, or some other problem with the expansion card. I need to leave home again, but I'll continue this evening. Any other thoughts based on this new info? I was carefull about seeing the pins making contact to the socket in the motherboard.

[Geoff Harrison]

Well that's encouraging. Unfortunately, I don't have a memory expansion board, so I'm not going to be much help there. Maybe Grant will have some suggestions when he logs in.

Geoff.

[bugman]

The board works by installing the cpu into its socket. Then the mini board has through connectors on the bottom that plug into the original cpu socket on the motherboard. If I am reading the info correctly, I have identified that pin 34 does not make the connection through the socket into the motherboard by using my meter. This would be easy to fix. Of course I need to confirm that this isn't by design. I'm reading the schematics to try and determine. I'm sure Grant will be able to confirm.

[Geoff Harrison]

Ah yes, that would be a Bad Thing. That's the Read/Write line from the CPU. I don't see anything on the circuit diagram that's not meant to pass through to the motherboard.

Geoff.

[bugman]

All that effort and I'm almost embarrased to say what the solution was. The socket into the motherboard from the mini 32k wasn't seated enough. The trouble is that you can't see the pins make contact very well because there is almost no clearance (Grant says this in the instructions). But, I had to use way more force to get a good seating than I'm usually comfortable with. Advice to anyone else, install the mini-expansion with the motherboard on a flat surface outside of the case.

What do they say? The obvious answer is usually the correct one?

The front panel and the terminal seem to be working perfectly.

Thanks Geoff, you led me in the direction of the CPU, which was the correct focus area.

[Geoff Harrison]

Great. And congratulations on getting it running. It's a fun little machine, isn't it?

Geoff.

[bugman]

Yes, its been awsome learning how to make it work through the front panel. I've built other kits before, but this gives me the vintage feeling better than anything else I have done.

I really want to easily retain the ability to enter and execute programs through the front panel only. I was thinking of installing a small toggle switch on the back plate to allow the switching of no terminal - terminal setting. From what I'm reading, it is necessary to have that set correctly to execute a program from the front panel because it then knows to start execution at 0000 as opposed to the monitor.

Do you concur with this? Is there another way to get the same result?

[Geoff Harrison]

I like the front panel too, but it doesn't really have the functionality to do anything useful. It is missing the examine_next, deposit_next, and start_at_address functions that are pretty much essential, and are available in the PROM monitor. It's handy to have a front panel to debug the hardware when things aren't working, and in that sense the 680 front panel is more useful than the one on the 8800, which requires a functioning CPU to do anything.

Geoff.

[Grant Stockly]

Ah yes, that would be a Bad Thing. That's the Read/Write line from the CPU. I don't see anything on the circuit diagram that's not meant to pass through to the motherboard.

Geoff.

Thanks for helping so much Geoff.

The R/W line does not go directly from the 6800 to the motherboard. The GAL decodes the address. If the 6800 wants to READ from the flash or sram, then it puts the external Altair 680 bus in a WRITE mode. If it was in a read mode, then the data buffers from the expansion card memory and the 680 motherboard buffers would clash! :(

I know that isn't the cause of the problem here, I just wanted to explain it...

[Grant Stockly]

Have you used basic? Or VTL-2?

I have never used the assembler / editor yet. I use a software assembler on a Windows computer and upload the paper tape image file.

The memory upgrade board is a pretty nice spoiler... A few seconds to boot basic vs 13 minutes with an audio cassette player. :) Have you tried your upgrade board out yet with the updated flash monitor? I tested every single upgrade board by booting and loading basic.

[bugman]

I havn't played with VTL-2 yet, but I'm looking forward to it. I have used the new flash monitor and loaded basic. The convience is nice. Now I get to finish assembling the back panel tonight. I didn't want to button everything up until I was sure everything was OK.

[Geoff Harrison]

The R/W line does not go directly from the 6800 to the motherboard.

I see. In fact I see now on the circuit diagram that the processor & motherboard R/W pins are not connected together, I didn't notice that yesterday. Sorry about that Bugman, I'm glad I didn't send you off on a wild goose chase trying to fix a problem that didn't exist.

Geoff.

[Grant Stockly]

I think 39 of the 40 wires on the two sockets are connected together. Any time you see a signal of the same name, they are of the same "net" and are connected.

[bugman]

No harm done. The main thing that tipped me off was when I removed the CPU, the front Panel looked exactly like my intermittent problem. Hence the CPU had intermittent connections which turned out to be a partailly seated socket.

I assembled the back cover last night. It was at that time I realized how compact this unit is. It was a tight fit! I decided to install an AC fan (despite popular suggestions!) I realize it probably isn't needed for stable operation. But I want to be able to pass this on to my kids so I'm thinking more in terms of component lifespan.

I'll start playing more around tonight. I'm anxious to use the assembler because as far as I can tell, it's the only way to output new paper tape dumps. I also want to familiarize myself with the VTL-2 and basic memory usage. With other similiar projects I have been able to identify where the data is located so I can do a raw memory save of a basic program rather that a cut and paste listing. For other computers, the paper tape read is easy to configure. But I have found that because basic and such are interpreters, they can't handle the normal transfer speed of a cut and paste directly into basic because there is a delay at the end of each line while the computer accepts the new line of code. The result is a slow transmission using that method. However, this may not be the case with the 680!

Lots of fun to be had.

Thanks for a great kit.

[TomL_12953]

But I have found that because basic and such are interpreters, they can't handle the normal transfer speed of a cut and paste directly into basic because there is a delay at the end of each line while the computer accepts the new line of code. The result is a slow transmission using that method. However, this may not be the case with the 680!

I don't know about the 680 version of Microsoft BASIC but in the early 8800 versions, you can set NULL 2 or higher before writing a tape so that there will be some space added after each line. You then set the same NULL number to read the tape back. This gives the processor time to enter each line. In the later versions of MS BASIC, NULL isn't necessary. BASIC can read in the file at full speed (110 baud - Wow!) from paper tape.

[bugman]

So far in my experimentation I am finding that I must set a line delay of about 500ms in hyperterminal to send a basic listing to microsoft basic. Not optimal, but it isn't too bad. Anything less causes corruption and the program is not entered correctly.

[bugman]

TomL,

Are you saying there is a paper tape command within basic? For the 680, all I can find is the csave command used in the cassette version for cassette saving. I am not aware of any paper tape writing routine in 680 basic. So, I have to either paste a text listing into basic to load, or type "list" and buffer capture for save.

[TomL_12953]

Are you saying there is a paper tape command within basic? For the 680, all I can find is the csave command used in the cassette version for cassette saving. I am not aware of any paper tape writing routine in 680 basic. So, I have to either paste a text listing into basic to load, or type "list" and buffer capture for save.

There's no paper tape SAVE command in the 8800 versions of BASIC. If you type LIST, then turn on the Teletype tape punch and press Return, the tape will be punched and contain the listing with an additional OK at the end. This can then be read into a BASIC session but will give a Syntax Error because of the OK. The entire program will be in memory, though.