On a mobo I used to have there was a non functional 83 mhz option that was asynchronous I believe the reason it didn’t work was the chipset or mobo wasn’t able to go async. I suspect your no boot option may be that. Mine also would not boot. My understanding was it was because of the rest of the mobo, not the chip.

Sphere478 wrote on 2023-09-01, 03:53:

On a mobo I used to have there was a non functional 83 mhz option that was asynchronous I believe the reason it didn’t work was the chipset or mobo wasn’t able to go async. I suspect your no boot option may be that. Mine also would not boot. My understanding was it was because of the rest of the mobo, not the chip.

You might be right but... according to my research, "HiZ" means "floating", so I'm inclined to believe that it actually doesn't produce any signal (I may be wrong).

Hi-Z should mean "high-impedance", while all the outputs of the clock generator need to be low impedance during regular operation.

Keep in mind this circuit was used by "PC-Chips" for their mainboards, so you´re supposed to expect some fake information... 😉

majestyk wrote on 2023-09-01, 05:13:

Hi-Z should mean "high-impedance", while all the outputs of the clock generator need to be low impedance during regular operation.

So if I understand correctly, this would mean that it effectively doesn't produce any useful signal (no 83MHz).

majestyk wrote on 2023-09-01, 05:13:

Keep in mind this circuit was used by "PC-Chips" for their mainboards, so you´re supposed to expect some fake information... 😉

Yes, but... is CYPRESS in any way related to PC-Chips and its shenanigans? I wasn't aware of that.

To be fair PC-Chips/Hsing Tech Enterprise made some good boards. That doesn't compensate all the damage they did with their defective and fake cache boards...

FWIW this board is working rather well so far. The only one with the TX chipset that is fully compatible with the R9250 DVI output (that I could find) and.. I appreciate the 33MHz bus option 🐢

Just joking here - Cypress is / was a renowned company.

majestyk wrote on 2023-09-01, 05:45:

Just joking here - Cypress is / was a renowned company.

Oh, ok, that makes sense 😄

Hopefully I'm not bringing this too far out of the grave...
I have a Socket 1 SBC here from Teknor, a VIPer809 with a Cirrus 7548 VGA chipset. PLL is a Cypress CY2291F. Datasheet and manuals attached.
I've figured out that pins 14/15/17 are S0/S1/S2 respectively and coincide with header W24 pins 2/4/6 on this board.
I've laid out the various frequencies I've been able to get out of it in the table below.

There's also a header W3 which is mentioned in the quick reference and in the manual, but it only ever appears to be left open. This is rather weird. if I apply a jumper to it I get a blank screen if it makes it far enough through to initialize the graphics chip, but no POST. All of these tests were done with an AMD 5x86 X5-133ADZ which I have successfully tested in the past (like 15 years in the past) to do the sweet spot of 160MHz. I was hoping this board would allow me to try and go further and see if I have a golden chip capable of 180-200MHz, but I guess not at this moment.

The PLL datasheet says there's 8 possible combinations for frequencies which I do not understand if there's only three possible pins to jumper, and so far I can only get two.
Maximum that the PLL can do is either 66.6MHz or 90MHz depending on if it's provided with 3.3v or 5V in respectively, and if an external crystal is used for the reference clock, which in this case it is, so this is likely being fed 5v.

Pin 6 of W24 is an unknown frequency going to/from the PLL. I don't have a scope handy to run the board externally and check what it's outputting, either, which is a bit of an issue. I'm planning on getting my hands on a cheap scope later on down the line but for now I wanted to drop this information in here lest it help someone else down the line with one of these PLLs on this board or any other.

lazibayer wrote on 2017-09-21, 19:20:

Stolen from

Kamerat wrote:

The Gigabyte GA-6EM have a clock generator capable of 33MHz FSB, the PLL52C66-05. You might have to modify the board to get access to the lowest clocks.

It goes from 33MHz all the way up to 100MHz.

Hello! Could you please attach pdf datasheet for this clocker ? I have super socket 7 GA-5AA motherboard and want to change clocker for more lower. For me it would be great from 33 to 100. But i cannot find what i can use. pinouts all of clocker are not comparable 🙁

I'd like to bump the thread with the question; can there be no universal guide on translating these tables into the jumpers on the board you happen to have in your collection?

Because from what I can see there are no universal rules to doing this other than "try it and see" but I'm afraid of damaging some hardware I own by moving some jumpers around that I shouldn't be moving.

The point of this thread is to provide info on how to interprete the various ways how motherboard manufacturers implemented multi clock jumper configurations. There was no standard for that. I don't think that you can damage hardware by trying out different multipliers or, for that matter, different fsb settings. Just do not experiment with voltage settings, those can really hurt.
Here you find an article that gives some insight in the basis of multi clock implementation:
http://pchardwarelinks.com/cpuspeed.htm

Best way to do it is to look up datasheet for the clock gen and cpu then follow traces

TMC/MyComp/MyNix/Megastar - TI5TT (VER. 1.1)
https://theretroweb.com/motherboards/s/tmc-my … -ti5tt-ver.-1.1

ICS9148F-25

Interestingly, my board only has half the DIP switches populated

I've traced the switches with a meter and 2 is connected to FS0. Unpopulated 6 is connected to FS1 and unpopulated 7 is connected to FS2. This agrees with the truth table above, FS1 and FS2 will always be high as they have no switch to ground them. FS0 is used to toggle between 60 and 66.8 MHz clock.

I've ordered an 8 switch DIP to see if I can open up the other clock options on the table above. 75MHz operation is quoted in the board manual but it will be interesting to see if 83.3MHz and 55MHz also work.

I assume this means the rest of the switches (1, 2, 4, 5 and 8.) are setting the multiplier? Is the multiplier also on this ICS IC? Or do those switches head elsewhere?

Sudos wrote on 2024-01-11, 08:04:

The PLL datasheet says there's 8 possible combinations for frequencies which I do not understand if there's only three possible pins to jumper, and so far I can only get two.

000
001
010
011
100
101
110
111

Welcome to binary 😀

The multipler goes to one to three pins on the cpu which will be one to three of four possible pins.

Bf0,bf1,bf2 (amd/cyrix),bf2 (intel)

These pull low at about 25-500 ohm
And pull high at about 300-2100 ohm depending on mfg

Your resistor combo should see the pull high resistance much higher than the pull low resistance

It’s a voltage divider meant to over power the pull high resistor

Lostdotfish wrote on 2025-03-05, 15:32:

TMC/MyComp/MyNix/Megastar - TI5TT (VER. 1.1) https://theretroweb.com/motherboards/s/tmc-my … -ti5tt-ver.-1.1 […]
Show full quote

TMC/MyComp/MyNix/Megastar - TI5TT (VER. 1.1)
https://theretroweb.com/motherboards/s/tmc-my … -ti5tt-ver.-1.1

ICS9148F-25

The attachment Screenshot 2025-03-05 152247.png is no longer available

Interestingly, my board only has half the DIP switches populated

The attachment Screenshot 2025-03-05 152520.png is no longer available

I've traced the switches with a meter and 2 is connected to FS0. Unpopulated 6 is connected to FS1 and unpopulated 7 is connected to FS2. This agrees with the truth table above, FS1 and FS2 will always be high as they have no switch to ground them. FS0 is used to toggle between 60 and 66.8 MHz clock.

I've ordered an 8 switch DIP to see if I can open up the other clock options on the table above. 75MHz operation is quoted in the board manual but it will be interesting to see if 83.3MHz and 55MHz also work.

I assume this means the rest of the switches (1, 2, 4, 5 and 8.) are setting the multiplier? Is the multiplier also on this ICS IC? Or do those switches head elsewhere?

75MHz bus unlocked

83.3MHz bus unlocked

K6-2 400 happily chugging along at 6x75

Not too long, and you'll be looking through alldatasheet for pin compatible replacements.

It's p2+ era, but here's a bunch I can add to the thread.
Should paste into excel and make sense

1ICS						hublink					
2xx	Chip format	CPU#	CPU volts	SDRAM#	PCI#	66mhz?	Jumpers	minMhz	Maxmhz	Serial VID max clocks	Other/notes
3ICS9250-08 	56ssop	3	2.5	17	7	0	4	66	133	150	
4ICS9250-10 	56ssop	3	2.5	9	8	2	2	66	100	133	
5ICS9250-12	56ssop	4	2.5	0	8	4	2	100	133	NA	2xcpu/2. For rambus systems
6ICS9250-18 	56ssop	3	2.5	17	7	0	4	66	150	166	
7ICS9250-23 	56ssop	2	2.5	13	8	2	5	66	166	same as jumpers	
8ICS9250-27	56ssop	3	2.5	9	7	3	3	66	133	same as jumpers	
9ICS9250-50 	56ssop	2	2.5	13	8	3	5	66	133	undocumented	
10ICS9250-16 	56ssop	3	2.5	9	8	0	3	66	133	150	
11ICS9212-13 	-	-	-	--	-	-	-	-		-	rambus
12ICS9250-28 	56ssop	2	2.5	13	2	3	3	66	133	same as jumpers	
13ICS9150F-04	56ssop	5		16

55MHz bus is also working.

My K6-2 400 AFK craps out if I set 83.3MHz bus but my 450AFX is happy

Who needs Super Socket 7...