English Amiga Board


Go Back   English Amiga Board > Support > support.Other

 
 
Thread Tools
Old 25 January 2021, 20:02   #1
Sim085
Registered User
 
Join Date: Apr 2009
Location: N/A
Posts: 795
How does this run at 14MHz?

Some time ago, while browsing the Amiga Hardware Database I came across the ICD AdSpeed;

http://amiga.resource.cx/exp/adspeed

What I cannot understand is how does this run at 14MHz without any crystal? And if it uses the crystal on the A500 itself then howcome this runs at 14MHz while the CPU that comes with the A500 runs at 7MHz? At the end of the day aren't both connected to the same socket? I know for sure the CPU does nothing in this regard because I have a 12Mhz 010 which still run at 7Mhz. Therefore the AdSpeed board itself must be doing something to increase the 7MHz to 14Mhz... However cannot understand how this can be done without a crystal.
Sim085 is offline  
Old 25 January 2021, 20:30   #2
zipper
Registered User
 
Join Date: Mar 2004
Location: finland
Posts: 1,671
I think it just doubles the main 7.16 MHz. Seems it just uses looping data from cache but can't access the bus faster :
Attached Thumbnails
Click image for larger version

Name:	ScreenHunter_614 Jan. 25 21.26.jpg
Views:	57
Size:	72.6 KB
ID:	70603  

Last edited by zipper; 25 January 2021 at 20:38.
zipper is offline  
Old 25 January 2021, 20:39   #3
hooverphonique
ex. demoscener "Bigmama"

 
Join Date: Jun 2012
Location: Fyn / Denmark
Posts: 1,244
One solution could be xor'ing two clock signals together, where one is 90 degress out of phase with the other, but I don't remember if such signals are available at the 68k socket.
hooverphonique is offline  
Old 25 January 2021, 20:46   #4
Sim085
Registered User
 
Join Date: Apr 2009
Location: N/A
Posts: 795
I belive what it says is that ChipRAM is accessed as 7MHz while the onboard RAM can be accessed at 14MHz. If it access the onboard RAM at 14MHz then doesn't that mean the CPU is also running at 14MHz?

Quote:
Originally Posted by zipper View Post
I think it just doubles the main 7.16 MHz. Seems it just uses looping data from cache but can't access the bus faster :
Sim085 is offline  
Old 25 January 2021, 20:54   #5
LaBodilsen
Registered User

 
Join Date: Dec 2017
Location: Gandrup / Denmark
Posts: 128
https://amiga.resource.cx/photos/pho...res=hi&lang=en

The Amiga 500 onboard crystal is running at 28.37516 mhz. Zoom in to the highres picture to ser it.

Last edited by LaBodilsen; 25 January 2021 at 21:02.
LaBodilsen is offline  
Old 25 January 2021, 21:05   #6
Sim085
Registered User
 
Join Date: Apr 2009
Location: N/A
Posts: 795
Yes, but the CPU runs at 28.37516 mhz division 4 = ~7MHz. If you connect a 12MHz 68010 to the motherboard without additional hardware this still runs at ~7MHz. So the AdSpeed must be doing something to re-double this to ~14MHz.

Quote:
Originally Posted by LaBodilsen View Post
https://amiga.resource.cx/photos/pho...res=hi&lang=en

The Amiga 500 onboard crystal is running at 28.37516 mhz. Zoom in to the highres picture to ser it.
Sim085 is offline  
Old 25 January 2021, 21:14   #7
roondar
Registered User

 
Join Date: Jul 2015
Location: The Netherlands
Posts: 2,792
I'm not an electrical engineer, but as I understand it you can use circuitry to double the speed of clock speed signals. They probably used something that achieves that. But what and how, well... I don't know
roondar is offline  
Old 26 January 2021, 09:04   #8
cpiac64
Registered User
cpiac64's Avatar
 
Join Date: Apr 2010
Location: Italy
Posts: 1,019
on aminet you can find the schematics for run at 14mhz
cpiac64 is offline  
Old 26 January 2021, 18:27   #9
Exodous
Registered User

 
Join Date: Sep 2019
Location: Leicester / England
Posts: 113
I don't know if this is how it's working, but the 74F86 is an XOR gate.

Take the oscillator input from the CPU socket to one of the XOR gate input pins.

Take the same oscillator input through a diode, which adds a delay, to the second of the XOR gate input pins.

With the appropriate diode, you get double the frequency at the output of the XOR gate.

In best ASCII tradition:

Code:
--- Osc In ---.--->|--- XOR IN 1 --\\--.
              |                    ||   >--- XOR OUT = Double Osc In
              \-------- XOR IN 2 --//--'

>| is the best I could do for a diode.
Exodous is offline  
Old 27 January 2021, 01:06   #10
Rotareneg
Registered User
Rotareneg's Avatar
 
Join Date: Sep 2017
Location: Kansas, USA
Posts: 226
A few inverters in series gets the job done too:

https://www.falstad.com/circuit/circ...XKNoJhPQEAJMQA
Rotareneg is offline  
Old 27 January 2021, 17:49   #11
nogginthenog
Amigan

 
Join Date: Feb 2012
Location: London
Posts: 1,062
Many years ago (1993?) I was living in the NL. I bought a 14MHz board from a guy at a local computer fair. Worked but the system clock run twice as fast!

Any ideas? Was it overclocking the CIA timers?

No idea where that A500 or the board are.
nogginthenog is offline  
Old 27 January 2021, 20:44   #12
hooverphonique
ex. demoscener "Bigmama"

 
Join Date: Jun 2012
Location: Fyn / Denmark
Posts: 1,244
Quote:
Originally Posted by nogginthenog View Post
Many years ago (1993?) I was living in the NL. I bought a 14MHz board from a guy at a local computer fair. Worked but the system clock run twice as fast!

Any ideas? Was it overclocking the CIA timers?
Yes, this is a common problem of simple 14MHz "accelerators". The CIAs are clocked by the 'E' clock signal from the 68000, which is doubled when you double the cpu clock. The 'E' clock is something like the cpu clock divided by 8 or 10, can't remember exactly.
hooverphonique is offline  
Old 27 January 2021, 21:20   #13
Rotareneg
Registered User
Rotareneg's Avatar
 
Join Date: Sep 2017
Location: Kansas, USA
Posts: 226
It looks like E clock output is connected to a trace that comes out from under the PLD, so maybe they're halving it there, which would only need a single flip-flop to do.
Rotareneg is offline  
Old 27 January 2021, 21:53   #14
Sim085
Registered User
 
Join Date: Apr 2009
Location: N/A
Posts: 795
I have read this elsewhere. Can the CIA chips not get the E clock signal from the CPU and instead have this provided from the original 7MHz - slowed to whatever value required by the CIA chips?

Quote:
Originally Posted by hooverphonique View Post
Yes, this is a common problem of simple 14MHz "accelerators". The CIAs are clocked by the 'E' clock signal from the 68000, which is doubled when you double the cpu clock. The 'E' clock is something like the cpu clock divided by 8 or 10, can't remember exactly.
Sim085 is offline  
Old 28 January 2021, 11:22   #15
hooverphonique
ex. demoscener "Bigmama"

 
Join Date: Jun 2012
Location: Fyn / Denmark
Posts: 1,244
Quote:
Originally Posted by Sim085 View Post
I have read this elsewhere. Can the CIA chips not get the E clock signal from the CPU and instead have this provided from the original 7MHz - slowed to whatever value required by the CIA chips?
Yes - check the schematics for Amigas with later/bigger processors (they don't have the 'E' output).
hooverphonique is offline  
 


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Similar Threads
Thread Thread Starter Forum Replies Last Post
14Mhz Accelerator Kai Hardware mods 111 06 May 2021 14:45
A2091 14Mhz SCSI mod SpeedGeek Hardware mods 1 25 December 2012 19:55
AGA on 14mhz 68000 little request.UAE Wishlist 6 03 May 2012 23:09
A500 MTEC 68030/14Mhz bebek Hardware mods 9 20 January 2010 22:30

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +2. The time now is 06:02.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, vBulletin Solutions Inc.
Page generated in 0.08718 seconds with 16 queries