Ye Olde Gigatron and Gigatronski

Using, learning, programming and modding the Gigatron and anything related.
Forum rules
Be nice. No drama.
walter
Site Admin
Posts: 160
Joined: 13 May 2018, 08:00

Ye Olde Gigatron and Gigatronski

Post by walter »

Hi all,

The Gigatron was designed to run with old TTL chips. The kits were supplied with modern, CMOS type chips, but the goal was to keep it TTL compatible. Marcel had already started sourcing some old TTL components, most notably an old 70ns SRAM. I've taken over that task, which has led to the creation of what I call "Ye Olde Gigatron". (And yes I know "Ye" was pronounced "The" and the word "Olde" does not exist.)

Not counting the clock and the 75HC595, all the ICs are old. The newest IC on there is the RAM, which is from 1991. All the rest is pre-1990, so over 30 years old!

I've thrown in a mix of stuff:
- 54LS, extended temperature range ("mil-spec")
- 74, standard TTL
- 74S, Schottky
- 74LS, Schottky low-power

Ye Olde Gigatron is a bit unstable and it consumes a whopping 1.2Amps. It requires a good power supply so I used my lab power supply. Also the clock did not work at first. The 74HCT04 provides a clock booster and that started ringing so I had to add a resistor. The clock circuit will not work with Schottky components, so I left in a 74HCT04. A 7404/74H04/.. will not work in a Pierce oscillator design.

The components used are:
6x 74163 (National Semiconductor, 1972)
4x 74157 (Stewart Wagner, 1975)
4x 74153 (Fairchild, 1976)
2x 74283 (Fairchild, 1978)
1x 74S240 (Signtics, 1979)
2x 74S153 (National Semiconductor, 1980)
2x 74S153 (Signetics, 1981)
2x 74S138 (Fairchild, 1982)
3x 54LS377 (Texas Instruments, 1982)
1x 74LS273 (Motorola, 1983)
1x 74LS244 (Signetics, 1984)
1x 74LS153 (Raytheon, 1985)
1x 74LS244 (Signetics, 1986)
1x 74LS32 (Motorola, 1986)
1x 74LS139 (Fairchild, 1986)
1x 27C1024 (ST Microelectronics, 1988)
2x 74LS273 (Texas Instruments, 1989)
1x 62256 (Micron Technology, 1991)

If somebody has older components than this, let me know :-)
yeolde1.jpg
yeolde1.jpg (2.77 MiB) Viewed 5797 times
yeolde2.jpg
yeolde2.jpg (3.8 MiB) Viewed 5797 times
at67
Site Admin
Posts: 647
Joined: 14 May 2018, 08:29

Re: Ye Olde Gigatron and Gigatronski

Post by at67 »

That's pretty awesome Walter, I have boxes full of old TTL boards and chips, I'll see if I can find anything older.
klf
Posts: 17
Joined: 10 Jan 2019, 22:48

Re: Ye Olde Gigatron and Gigatronski

Post by klf »

National Semicoductor's MM74C series has been introduced in 1976. So you could try a 74C04 instead of the '90s 74HCT04. However, it has 50 ns propagation delay and might be at its limit at 6.25 MHz.

The CD4069A from the very first 4k CMOS series is even older. -A is the unbuffered configuration, with only one transistor pair per inverter. Later -A got replaced by the buffered -B series (three transistor pairs per inverter), but the 4069A survived with the new name CD4096UB. This part works very well as a pierce oscillator.
walter
Site Admin
Posts: 160
Joined: 13 May 2018, 08:00

Re: Ye Olde Gigatron and Gigatronski

Post by walter »

Thanks. I found some on ePay. I got a CET (Chino-Excel Technology) CE4069UB from 1987 and a modern NXP HEF4069UBP.

Neither works.. The scope shows that it is oscillating, but the output voltage is very low. The propagation delay of these is about the same as the 74C04 according to the specification, but I doubt that would be the problem, as I see it oscillating nicely.

I really need a better scope to be able to do some troubleshooting..

Maybe it's just that the circuit needs to be different, to get the right boosting effect. I was hoping for a drop-in replacement.
walter
Site Admin
Posts: 160
Joined: 13 May 2018, 08:00

Re: Ye Olde Gigatron and Gigatronski

Post by walter »

Since I gave a talk about the Gigatron at Neuron hackerspace in Moscow, I wanted to make a Gigatron with old Soviet chips. Unfortunately, I did not get any from my new Russian friends :-) By the way, Pavel and others of Neuron Hackspace have a wonderful "Flipper Zero" kickstarter project.

But anyway, after a long time occasionally looking for Soviet TTL chips, I suddenly found https://evita.lt, a company in Vilnius, Lithuania, that has a lot of NOS Soviet components. I was able to order everything there. One chip appeared to be broken, but the company send me a replacement at no extra cost.

The chips were new. The newest of the lot is manufactured in 1992 (so strictly speaking, not Soviet but Russian!). They have a different colour than the IC's I'm used to seeing. Also on quite a few, you can see a piece of aluminum on the bottom. I've opened one up, to see that this is used to put the die on before the epoxy is added.

The Soviet chips also have a different pin spacing (2.5mm instead of 2.54mm). But I was able to (gently!) get them all into their sockets, even the larger ones. I did spend quite some time looking for Soviet replacements of the 62256 SRAM and 27C1024 EPROM, but don't believe they ever existed. The 75HC595 is also not available. A Soviet 74LS04 can be found, but it will not work in the pierce oscillator (same as with Ye Olde Gigatron). There is a Societ 74ACT04 version called КФ1594ЛН1, but that is pretty rare and I could not get one. Since it's used for the clock and not the CPU itself, it is not that much of a problem.

I got a white PCB from somebody in Hong Kong, which I used for the Гигатрон. I swapped SRAM in between pictures. It now has the same old SRAM as Ye Olde Gigatron.

Since I ran out of VGA connectors, I bought a cheap one from China, instead of buying from Mouser/Farnell/Digikey, which I've always done for the kits. Although cheap, it also is not as mechanically sturdy and.. quite a bit smaller, so it no longer aligns with the serial connector. Oh well.
IMG_20200817_153820.jpg
IMG_20200817_153820.jpg (3.64 MiB) Viewed 5611 times
gigatronski.jpg
gigatronski.jpg (1.19 MiB) Viewed 5611 times
The Гигатрон is not very stable, like Ye Olde Gigatron. Sometimes, it will work (I get the running lights) but there's no video output. The only possible cause, if you look at the circuitry, is that there is a problem with the HSYNC signal. I've hooked up my trusty old oscilloscope and found that with the soviet ICs, the HSYNC signal timing fluctuates too much. Pictures below, not as nice as the Rigol ones Marcel used to make though. First picture is a "normal" Gigatron, second one the Гигатрон.

IMG_20200812_195016.jpg
IMG_20200812_195016.jpg (2.72 MiB) Viewed 5605 times
IMG_20200812_195039.jpg
IMG_20200812_195039.jpg (2.54 MiB) Viewed 5605 times
klf
Posts: 17
Joined: 10 Jan 2019, 22:48

Re: Ye Olde Gigatron and Gigatronski

Post by klf »

walter wrote: 14 Aug 2020, 16:44 Neither works.. The scope shows that it is oscillating, but the output voltage is very low. The propagation delay of these is about the same as the 74C04 according to the specification, but I doubt that would be the problem, as I see it oscillating nicely.
Did you try to change the bias resistor? E.g. 10 MOhm or 4M7?

Another idea, in case you accept circuit modifications: bias two inverters with 1k resistors from input to output. Connect both in series. Connect the crystal from the first inverter's input to the second one's output. No capacitors.

This circuit was common in CP/M times. It works with bipolar TTL and LS-TTL.
alastair
Posts: 68
Joined: 10 Oct 2019, 14:28

Re: Ye Olde Gigatron and Gigatronski

Post by alastair »

The problem with the oscillator might be the resistor values. There is a HUGE difference between getting a CMOS chip and TTL chip to oscillate correctly. It's about 3-orders of magnitude! It takes about 1000x the current to get a TTL chip in to its linear region compared to a CMOS chip (think mA vs. uA).

Here is a Pierce oscillator I had working nicely with TTL. This one is about 5x the frequency as the Gigatron (F-series) so you should have no issues, but keep the 47pF capacitors for the 6.25MHz crystal.
osc.png
osc.png (12.83 KiB) Viewed 5584 times
walter
Site Admin
Posts: 160
Joined: 13 May 2018, 08:00

Re: Ye Olde Gigatron and Gigatronski

Post by walter »

I couldn't get it to work. The clock signal looked very bad on the oscilloscope. For Ye Olde Gigatron, it would be nice to use a 7404/74LS04. I even have a 74H04. But that's for when I find the time.

I don't know what changed, but suddenly the system is a lot more stable. I had it running with Mandelbrot for a day without any problems.
steve
Posts: 40
Joined: 08 Jul 2019, 19:40

Re: Ye Olde Gigatron and Gigatronski

Post by steve »

walter wrote: 17 Aug 2020, 14:49 The Soviet chips also have a different pin spacing (2.5mm instead of 2.54mm). But I was able to (gently!) get them all into their sockets, even the larger ones. I did spend quite some time looking for Soviet replacements of the 62256 SRAM and 27C1024 EPROM, but don't believe they ever existed. The 75HC595 is also not available. A Soviet 74LS04 can be found, but it will not work in the pierce oscillator (same as with Ye Olde Gigatron). There is a Societ 74ACT04 version called КФ1594ЛН1, but that is pretty rare and I could not get one. Since it's used for the clock and not the CPU itself, it is not that much of a problem.

I got a white PCB from somebody in Hong Kong, which I used for the Гигатрон. I swapped SRAM in between pictures. It now has the same old SRAM as Ye Olde Gigatron.

Since I ran out of VGA connectors, I bought a cheap one from China, instead of buying from Mouser/Farnell/Digikey, which I've always done for the kits. Although cheap, it also is not as mechanically sturdy and.. quite a bit smaller, so it no longer aligns with the serial connector. Oh well.

IMG_20200817_153820.jpg

gigatronski.jpg
Очень красивый!!! :)
(beautiful!!!)

With a friend we were just a month ago discussing to repair a broken ZX81 subsituting the Z80 with a T34VM1!

Anyway the white board really feels Russian style, but I think you should try to change also the other labels on the board. In case your new friend are not reachable I can try to help more than with my basic Russian, with some of my old russian friends!

удачной работы!
(the best translation I found for happy hacking! :P)
walter
Site Admin
Posts: 160
Joined: 13 May 2018, 08:00

Re: Ye Olde Gigatron and Gigatronski

Post by walter »

I had ordered some stuff for lb's expansion board. Instead of paying €20 for shipping, I decided to order some nice ZIF sockets to replace the big green ones, e.g. on Ye Olde Gigatron, so the shipping was free (or the ZIF sockets were free, depending how you look at it).

First, getting the old one out. Soldering level "guru".. I could screw off the top and desolder the individual pins, then assemble the socket again.
zif1.jpg
zif1.jpg (86.77 KiB) Viewed 3626 times
Then in with the new one:
zif2.jpg
zif2.jpg (56 KiB) Viewed 3626 times
Looks a lot better now.
Post Reply