VOGONS

386DX board always stays in DEturbo mode

Topic actions

First post, by analog_programmer

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

Some time ago I got motherboard with AMD Am80386DX-40 CPU soldered on it. The only text on the mobo related to its model "says" TK-82C491/386-4N-D02C MADE IN TAIWAN R.O.C., so I think it is Magitronics board.

The board came to me without BIOS chip and with missing cache chips. I installed an EEPROM Winbond W27C512 BIOS chip and started to test the motherboard with some BIOS dumps for compatible mobos or clones of the very same board, which I took from TRW site: https://theretroweb.com/motherboards/s/tk-tk- … 86-4n-d02c#bios, https://theretroweb.com/motherboards/s/abit-ab-ak3#bios, https://theretroweb.com/motherboards/s/auva-c … -tam-40-u2#bios, https://theretroweb.com/motherboards/s/gemlig … gmb-386umc#bios, https://theretroweb.com/motherboards/s/qdi-386-4n-d04a#bios and https://theretroweb.com/motherboards/s/unknow … -386-uct-q#bios.

It turned out that this TK-82C491/386-4N-D02C motherboard can't POST without cache chips installed (I don't have any suitable chip for TAG-RAM) and Award BIOSes, but it boots to DOS with most of the AMI BIOSes I tried.

And here's the problem: I can not switch the CPU in "Turbo mode", no matter if the "TB SW" (TURBO switch connector J5) is shorted or open and thus with 40 MHz setting for the onboard AMD 386DX-40 CPU according to NSSI and Norton SysInfo 8 always runs slightly faster than 386SX-16 🙁

Here is the picture what I managed to trace for the "TB SW" circuit:

The attachment TB-SW_traces.jpg is no longer available

When I short with jumper J5 "TB SW", the "green"connection/trace on the picture gets "low" i.e. no +5 V (coming through the resistor R18), but nothing changes with the CPU speed. I can't trace where is going the inverted signal from pin 8 of the HEX Inverter Motorola MC74F04N. The non-inverted signal from "TB SW" and MC74F04N's pin 9 goes directly to UMC UM82C491F cipset's pin 5 which according to datasheet is "TURBO".

I don't have an oscilloscope, but now I have one cheap chineese signal-tester. Unfortunately it came to me without any user manual, so I don't know how it can be used. This is how it looks:

The attachment signal_tester.jpg is no longer available

Any advises on how to troubleshoot the permanent "DEturbo mode" on this board will be appreciated!

P.S. I've desoldered the old Ni-Cd CMOS battery from the mobo and currently there's no new battery installed. I checked all the pins of the chipset and CPU for loose/bad solder joints to the board's pads and I found none. There are no battery or capacitor leakage, visible scratches or broken traces on the motherboard.

The word Idiot refers to a person with many ideas, especially stupid and harmful ideas.
This world goes south since everything's run by financiers and economists.
This isn't voice chat, yet some people overusing online communications talk and hear voices.

Reply 1 of 4, by Deunan

Posted on 
User metadata
Rank l33t
Rank
l33t

Some BIOSes have an option to enable/disable the turbo switch. The older (286, early 386) mobos would change the CPU clock for (de)turbo, newer ones inject wait cycles on RAM access. If your programs detect 40MHz 386 (usually by timing the mul or div instructions) but you get SX-16 like performance then it's either some very odd cache issue or injected waitstates. Perhaps with cache chips missing entirely the chipset/BIOS enters some fail-safe mode with max relaxed timings and thus the performance is badly affected.

Reply 2 of 4, by analog_programmer

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

That's a good pint, Denuan. But with the AMI BIOSes I set cache to disabled and still this changes nothing 🙁 There are some waitstates settings for the memory in the AMI BIOS menus, but I'm not thinkering with them (just using BIOS defaults).

If the problems come from the missing cacehe chips, it will be very odd that this board becomes virtually unusable without them.

Maybe I really have to find some cache chips for this mobo, but I have no idea what make and model SRAM chips will be fully compatible with it.

The word Idiot refers to a person with many ideas, especially stupid and harmful ideas.
This world goes south since everything's run by financiers and economists.
This isn't voice chat, yet some people overusing online communications talk and hear voices.

Reply 3 of 4, by st31276a

Posted on 
User metadata
Rank Member
Rank
Member

Some 386es disabled cache for deturbo, perhaps that is why it stays in deturbo because there is nothing to enable. The cache makes a significant difference in the performance of a fast 386.

Reply 4 of 4, by analog_programmer

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

I don't think AMD 386DX-40 (without onboard cache) will be slower than AMD 386SX-40 (also without cache), not to mention 386SX-20 as in my case. But who knows what all those BIOSes are capable of...

The word Idiot refers to a person with many ideas, especially stupid and harmful ideas.
This world goes south since everything's run by financiers and economists.
This isn't voice chat, yet some people overusing online communications talk and hear voices.

Go to top of page Go to top of page
Back to General Old Hardware
Close window
Close window