"Spicy O'clock" project has started

JDW

Administrator
Staff member
Founder
Sep 2, 2021
1,547
1,353
113
53
Japan
youtube.com
I'm proud of you, son.
you saved the little guy!
Drake you truly gave Philippe a "leg up" in his repair! :)

@iPhil64
Great job!
I've done chip repairs like that, and depending on the chip it's not too fun to carefully scrap away the black body of the chip to reveal more of the leg stub in order to solder something to it. In my case, I desoldered the entire broken off leg, such that I could resocket the chip. I show that in my video here.
 

Kay K.M.Mods

Active Tinkerer
Sep 23, 2021
295
647
93
Tokyo
www.kerosmm.com
It's still in the testing phase, but we're designing a Spicy O'Clock! Very Hot version. It has an amazing silicon oscillator that can be tuned from 18MHz to 120MHz.

It was a use that was supported by 1/2 clock until now. For example, if we wanted a 68040 to go to 50Mhz, it had amplified it by a factor x2 with the clock driver, so Spicy was actually generating a 25MHz clock. However, in machines like the Quadra700, the oscillator that has twice the clock is reduced to 1/2 and extracted to the CPU, so if you want to go to 38MHz, you need a 76MHz oscillator.

The theoretical value can be extracted up to 120MHz, so it can support up to 60MHz CPU clock.
What this means is that it will be possible to support 68030 machines such as MacIIfx and PowerBook170... Of course, it can also be applied to some PPC-based accelerators and other devices.
スクリーンショット 2022-11-03 19.18.42.png
 
Last edited:

Drake

TinkerDifferent Board Vice-President 2023
Staff member
Sep 23, 2021
447
782
93
Whoa @Kay K.M.Mods! 😃 Now that gets my attention! 🤩

Can the 60ns VRAM and OS-CON caps still support frequency ranges up to 120 MHz with the 68040 running at 60 MHz?
Good question, There's something to investigate on the data sheet side regarding access, read/write timings. When selecting the ram chips we knew the current configs were good for ~50mhz but obviously other elements get in the way of pushing this limit.
 
  • Like
Reactions: retr01

Kay K.M.Mods

Active Tinkerer
Sep 23, 2021
295
647
93
Tokyo
www.kerosmm.com
Whoa @Kay K.M.Mods! 😃 Now that gets my attention! 🤩

Can the 60ns VRAM and OS-CON caps still support frequency ranges up to 120 MHz with the 68040 running at 60 MHz?
@Drake is right. Thank you.

All operations can be messed up if the timing is not right. For example, the serial stops working around 45Mhz.
Although it is a theoretical value, an individual SMC-VRAM chip can operate at 55MHz.
 
  • Like
Reactions: retr01

JDW

Administrator
Staff member
Founder
Sep 2, 2021
1,547
1,353
113
53
Japan
youtube.com
Can the 60ns VRAM and OS-CON caps still support frequency ranges up to 120 MHz with the 68040 running at 60 MHz?
Impossible to know because only Kay has one of those new overclockers right now. But I suspect VRAM speed may be a limiting factor at some point.

I thought it has been the limiting factor in my own overclocking attempts beyond 50MHz, but now I have discovered a 68040 cooling issue. I am filming that now for possible release as a new YouTube video, so more on that later.
 
  • Like
  • Wow
Reactions: iPhil64 and retr01

iPhil64

Tinkerer
Apr 5, 2022
122
86
28
France
Hello all !

So after replacing all caps on my 660AV with OS-CON, and also receiving the MC88920, I started again testing today.
First good news, the Mac still boots up with the original 25 MHz. At least, recapping (for the second time...) went good.

Now I need to figure out what's the upper limit for the Spicy.

@Kay K.M.Mods Quick question, since you're working on a "Super hot" version, do you think you could design a more universal one; that could go to lower frequencies (in my case starting at the 12.5 MHz).

And yes, I do remember, I can and will use an external resistor. I'm just trying to clean out the build.

By the way, I'm curious to see which wires / connectors you use. The quality is good, and I'd like to make shorter versions.
 
  • Like
Reactions: nottomhanks

iPhil64

Tinkerer
Apr 5, 2022
122
86
28
France
Guys, OS-CON rules.

I now seem to be able to get a stable 33 MHz, which I could not use before.
I will reassemble, plug devices (that made the Spicy 660AV crash before) and see how stable it is.

Also I will start to look at CPU temps and cooling, but since I have an RC33, I'm not too worried for now.
 

Kay K.M.Mods

Active Tinkerer
Sep 23, 2021
295
647
93
Tokyo
www.kerosmm.com
Hello all !

So after replacing all caps on my 660AV with OS-CON, and also receiving the MC88920, I started again testing today.
First good news, the Mac still boots up with the original 25 MHz. At least, recapping (for the second time...) went good.

Now I need to figure out what's the upper limit for the Spicy.

@Kay K.M.Mods Quick question, since you're working on a "Super hot" version, do you think you could design a more universal one; that could go to lower frequencies (in my case starting at the 12.5 MHz).

And yes, I do remember, I can and will use an external resistor. I'm just trying to clean out the build.

By the way, I'm curious to see which wires / connectors you use. The quality is good, and I'd like to make shorter versions.
Bridge 1K OHMS instead of SL jumpers. It is possible to set below 15MHZ, so I think you can drop it to 12MHz. GOOD LUCK!!!
 
  • Like
Reactions: iPhil64 and JDW

Kay K.M.Mods

Active Tinkerer
Sep 23, 2021
295
647
93
Tokyo
www.kerosmm.com
Guys, OS-CON rules.

I now seem to be able to get a stable 33 MHz, which I could not use before.
I will reassemble, plug devices (that made the Spicy 660AV crash before) and see how stable it is.

Also I will start to look at CPU temps and cooling, but since I have an RC33, I'm not too worried for now.
IMG_4804.jpg
Use male and female pin headers. Bridge at 1Kohms.
IMG_4805.jpg
I think it goes down to below 15MHz.
 
  • Like
Reactions: iPhil64 and JDW

JDW

Administrator
Staff member
Founder
Sep 2, 2021
1,547
1,353
113
53
Japan
youtube.com
@alxlab

A fellow member of one of the vintage Mac FaceBook groups I participate in very graciously donated to me one of his heatsink clips that was printed by an associate of his, based on your design. It arrived today and I shot some photos...

tempImager5m44H.png tempImage6GK1Zy.png tempImagecHW24K.png
tempImagerl7z7E.png tempImageHojGh3.png

He said he used PRUSA PLA filament and printed it on his mk3s printer. The quality of the print seems to be quite nice.

I am totally ignorant about 3D printing, so please excuse this most basic of all questions, but when checking specifications of PLA, I see "heat deflection temperature" cited. What does that mean? Is that the temperature at which the plastic begins to soften? If so, then most PLA would begin to soften at about 55°C, and I suspect that would be an issue because I have measured up to about 60°C on my gold aluminum heatsinks before (when secured with 2-sided thermal tape).

My concern is that if the filament used doesn't allow for at least 65-70°C heatsink temperatures, then the plastic may soften and not push down on hard enough on the heatsink, which could impact heat transfer through whatever thermal paste is used.

Would you happen to have an opinion on this?

Many thanks for your amazing work on this 3D design by the way. I am just curious about filament recommendations, and whether or not this PRUSA PLA printed clip is adequate for a 50MHz overclocked full FPU 68040 in a Color Classic Mystic, which again I have measured temperatures of around 60°C on in the past.
 
  • Like
Reactions: retr01

AvadonDragon

Tinkerer
Jan 27, 2023
52
50
18
Tennessee
@JDW I have a fair amount of experience with 3D printing and PLA. I can say PLA is definitely unsuitable for the temp ranges you might see on a 68040 heatsink. It will almost certainly deform and fail to maintain good contact with the CPU. I recently had several prints exposed to around 65°C and they deformed terribly. Most people can also print with ABS and it should tolerate up to around 100°C. Other less common filament types may be even better suited. I would guess carbon fiber reinforced nylons are probably the best option.
 
Last edited:
  • Love
  • Like
Reactions: retr01 and JDW

retr01

Senior Tinkerer
Jun 6, 2022
2,473
1
793
113
Utah, USA
retr01.com
As you can see, @JDW, in my annotation, the green highlight shows the safe temperature ranges followed by orange highlights when the material starts to deteriorate. Finally, the red highlights worse deterioration progresses as the temperature rises.

1679414063281.png
 

AmigaOfRochester

Tinkerer
Oct 31, 2021
23
36
13
Yeah, that member was me. Something that may change factors, is the person I made these for is using a fan directly attached to the heatsink. I can't imagine the heatsink would stay that hot, maybe something you could look into?
I don't have any ASA or ABS as I generally don't print that, but I might grab some and reprint some for them just in case.
 
  • Like
Reactions: JDW and retr01

Paralel

Tinkerer
Dec 14, 2022
115
47
28
As you can see, @JDW, in my annotation, the green highlight shows the safe temperature ranges followed by orange highlights when the material starts to deteriorate. Finally, the red highlights worse deterioration progresses as the temperature rises.


I would actually adjust that last section with ABS to match that of the ASA. If you look closely at the 4th piece of ABS, there is dimensional loss.
 
  • Like
Reactions: JDW

JDW

Administrator
Staff member
Founder
Sep 2, 2021
1,547
1,353
113
53
Japan
youtube.com
I would actually adjust that last section with ABS to match that of the ASA. If you look closely at the 4th piece of ABS, there is dimensional loss.
For an overclocked LC575 motherboard application, even when used within the smaller case confines of a Color Classic, I have not measured heat much beyond 60°C. Of course, I do have a fan added, and that tends to keep temperatures in the 50's or lower.

Based on that, I would see PLA is questionable because there is a huge gap between 20°C and 70°C in the photo posted by @retr01. PETG would seem to work fine, because I honestly don't think it will get into the 70°C range, and I have my full FPU version 040 overclocked to 50.0MHz. Then again, some older mask versions may get hotter. Not sure. But with a fan? Again, that is an unknown.

ASA and ABS would be idea choices to play it safe.

But again, I am brain dead when it comes to 3D printing. I merely have pondered this issue for a long time since the clip was first announced. And also, in the recent TinkerDifferent livestream video, I remember somewhere toward the middle @Branchus mentioned something about the filament used and heat given off by the heatsink. So it isn't just me who is pondering this topic.

All said, you want a strong clip that will keep the heatsink clamped down hard on the 040 and not loosen over time due to heat from the heatsink, which could compromise the thermal paste, possibly opening up gaps that would drop the thermal transfer a little or even a lot.
 
  • Like
Reactions: retr01