MicroMac Speedy for IIfx

[petteri]

I see there used to be MicroMac Speedy for IIfx but obviously it isn't anymore available. The only photos I managed to find are these:

https://img.aucfree.com/h431310795.1.jpg

https://img.aucfree.com/h431310795.2.jpg

https://img.aucfree.com/h431310795.3.jpg

It doesn't look particularly complex to recreate. Any information how it was installed on the IIfx logic board? Based on the legacy MicroMac website no soldering was required. I think there were some hook probes to be clipped somewhere.

 

[Phipli]

I see there used to be MicroMac Speedy for IIfx but obviously it isn't anymore available. The only photos I managed to find are these:

https://img.aucfree.com/h431310795.1.jpg

https://img.aucfree.com/h431310795.2.jpg

https://img.aucfree.com/h431310795.3.jpg

It doesn't look particularly complex to recreate. Any information how it was installed on the IIfx logic board? Based on the legacy MicroMac website no soldering was required. I think there were some hook probes to be clipped somewhere.

I imagine it will clip onto the pins of G3. The tin can clock near the processor. The clocks have an "output enable" pin, that lets you switch off the clock and enables you to inject your own clock instead.
I've designed something similar that gives you a choice between stock and an overclock, but I'm struggling to get parts due to the component shortage.

 

[Powerbase]

Didnt you have to cool the memory on an overclock? I know the Dash 30fx had a fan on it.

 

[petteri]

In case of using something like DS1085 frequency synthesizer, what are the requirements for the signal? Sine wave or something more squareish? And does it have to be 100.00MHhz or would something like 98.50MHz work? Or should I just order bunch of aliexpress oscillators in various timings and find the highest working one...

 

[petteri]

50MHz (100MHz oscillator) works fine, I have 70ns RAM. https://68kmla.org/bb/index.php?threads/macintosh-iifx-with-pronitron-80-21-v1-5.42667/#post-467388

 

[petteri]

I have been experimenting with Si5351 and SN74AHCT125N. The idea was to use Si5351 to generate clock signal and then use SN74AHCT125N to change its level more suitable to 5V circuit.

"The Si5351 is a versatile I2C programmable clock generator that is ideally suited for replacing crystals, crystal oscillators, VCXOs, PLLs, and buffers."

But so far I haven't managed to get a happy chime out of the IIfx (no sad mac either).

There are couple known issues. I don't have suitable oscilloscope to see how good the signal is and how it compares to the 100MHz oscillator. I have a hunch there are more harmonics etc. I'll ask around to find out if someone locally has such. But at least my DSO138mini displays reasonable square wave under 200kHz range

Another issue is there is need for a programmer every time Si5351 boots up. I use Arduino to initialise Si5351 but that takes a moment. I assume the clock signal should be available right away when 5V becomes available on the logic board?

 

[Phipli]

Another issue is there is need for a programmer every time Si5351 boots up. I use Arduino to initialise Si5351 but that takes a moment. I assume the clock signal should be available right away when 5V becomes available on the logic board?

I have some Si5351s on the side for a similar project. I assumed the machine would come up once the clock started.

Is your 74HCT125N fast enough? I was going to use a.... chip... that I need to look up, but I wasn't aiming for 100MHz.

You could possibly try to hold the computer in reset with the arduino until it is ready?

How do you have brownstone detection set on the AVR?

Removing the bootloader from the AVR and burning your code directly with an ISP will massively speed up how quickly your code runs.

 

[mg.man]

Removing the bootloader from the AVR and burning your code directly with an ISP will massively speed up how quickly your code runs.

How quick the code runs = "performs when running"? ...or how quickly your code starts executing from a cold start?

 

[Phipli]

How quick the code runs = "performs when running"? ...or how quickly your code starts executing from a cold start?

How quick it starts running. The bootloader waits on reset to see if a computer tries to reprogram it. Removing the bootloader means no USB or UART programming, but it runs your code near instantly, instead of waiting to see if it needs to be reprogrammed.

 

[Phipli]

Also, changing the brownout threshold to a lower voltage can make the MCU come up quicker.

 

[mg.man]

How quick it starts running.

That's what I thought. I'm after something that speeds up the code. So far I have my Mega2560 running at 20Mhz... which is still not fast enough.

 

[Phipli]

That's what I thought. I'm after something that speeds up the code. So far I have my Mega2560 running at 20Mhz... which is still not fast enough.

How come?

 

[Phipli]

You can use inline assembly in the Arduino IDE BTW.

 

[mg.man]

Rather than derail this thread... see - https://github.com/alexthecat123/ArduinoFile/issues/1

 

[Phipli]

brownstone

Brownout - stupid autocorrect.

 

[petteri]

I made some more experiments:
1. Ground the disable pin on the original setup and open it after pressing the start button. No happy chime. So the clock signal should be available right at the start.
2. Have the Si5351 and 74HCT125N running on a battery (common ground) before starting up. No happy chime.
3. Have the Si5351 running on a battery (common ground) before starting up. 74HCT125N gets powered from SCSI 5V so the signal should start flowing only after start.
4. Tried also with 80Mhz setup. No happy chime.

But no magic smoke and IIfx works fine after removal my equipment.

74HCT125N should be fine for over 100MHz, the max switching characteristic is 8.5ns. Arduino has the code loaded, so it is matter of boot loader executing and then running the short I2C init for the Si5351.

I also assume the phase of the clock matters at the startup. http://datasheets.chipdb.org/Motorola/68030/MC68030.pdf has signal flows starting from low.

 

[Phipli]

2. Have the Si5351 and 74HCT125N running on a battery (common ground) before starting up. No happy chime.

This could mean that the clock signal isn't in a form the mac likes.

I'm afraid you've really reached "borrow a scope".

In the meanwhile :
Double check your setup code.
Do you have a frequency counter to check the frequency is right?
Are you using the correct output on the 5351?

 

[Phipli]

Double checked my purchase history and it looks like I bought SN74AHCT1G125DCKT for level shifting, so we really were going the same route.

So at some point I do need to get what you're doing, but for my C650.

 

[petteri]

I don't have frequency counter but I used my RTL-SDR v3 tuned to 100MHz (or 80MHz) and I could see a spike in the waterfall display when Si5351 was powered and set to the corresponding frequency.

There is no 5V or 12V available on the IIfx power supply before startup?

My current plan is to have Arduino and Si5351 in a ready state but not emitting signal and figure out how to start signal in correct phase when 5V becomes available. Should be matter of couple lines of code (Arduino GPIO receives 5V -> send I2C to Si5351).

 

[Phipli]

There is no 5V or 12V available on the IIfx power supply before startup?

The soft power circuit is probably powered by the two batteries like on the older IIx and II. You could take power from there, but I'm not sure that the speed of initialisation is the issue. 100MHz is fast. How long are your wires? What does your circuit look like? Are you using any connectors?