Help identify microcontroller in an XT keyboard

[soon668]

I have an XT keyboard (DataComp DFK-700FI) which the LED blink during boot, but after that does not response with any key press, not even cap lock. I have also ensured the keyboard lock jumper on the XT motherboard is set to unlock. I have tested the keyboard on 2 motherboards which was known to have worked before with this keyboard.

I suspect it could be the keyboard's micro controller that is bad. I would like to replace the IC but could not figure out what is the correct one to get. The IC is marked C35054E DATACOMP 8616. I have attached a photo for reference. The marking is a bit faded, but I am able to read under bright light.






Would appreciate any help to identify the IC and recommend a replacement if possible.

Thanks!

 

[modem7]

At [here], 'Datacomp' appears on a keyboard's controller chip that was made by Signetics. It is as though Datacomp paid Signetics to supply a Signetics chip preprogrammed with Datacomp supplied code.

As for your chip, presumably, the logo will reveal the chip maker and C35054E is the part number. If so, C35054E may be an in-house number, and may contain Datacomp supplied code.

 

[soon668]

At [here], 'Datacomp' appears on a keyboard's controller chip that was made by Signetics. It is as though Datacomp paid Signetics to supply a Signetics chip preprogrammed with Datacomp supplied code.

Hi modem7, thanks for finding the website. I found the Tatung_TCS-5000 resemble the keyboard that I have, except mine does not have the AT/XT switch. It is also have a DFK-700F2 model name, in which mine is DFK-700FI.

Another similar one is DFK-777F which have a INTEL P8049AH - 8-Bit Microcontroller.

If it is not possible to replace the microcontroller, do you have any suggestion to test out the keyboard? I don't have a logic analyzer that I can probe. Also, do you think XT bios make a different? as I recently replaced the Bios.

Thanks again!

 

[soon668]

Hi modem7,
I have just acquired a logic probe EM4610 , will you be able to guide me on what to check?

Thanks!

 

[modem7]

I have an XT keyboard (DataComp DFK-700FI) which the LED blink during boot, ...

Some keyboards (perhaps most) do some sort of self test as soon as they receive power, with that self test momentarily flashing on the LED's.

With the keyboard disconnected from the computer, supply +5V to the keyboard's DIN connector (ground to ground pin, +5V to +5V pin).
If the LED's momentarily flash on, then you can be confident that the controller in the keyboard is doing a self test on receipt of +5V.

... but after that does not response with any key press, not even cap lock.

Is the computer's power-on self test (POST) displaying a 301 error (keyboard error) or displaying an error that points to the keyboard ?

Also, do you think XT bios make a different? as I recently replaced the Bios.

No, because earlier your wrote, "I have tested the keyboard on 2 motherboards which was known to have worked before with this keyboard."
Although I am assuming that you only changed the BIOS on one of the motherboards.

But, if you changed the BIOS on both motherboards, and the keyboard problem started after that, then it is time to put back in the original BIOS', just in case.

I have just acquired a logic probe EM4610 ...

The first thing to do is to understand what your probe shows you in varying conditions:
* Measuring a LOW. Can be done by probing a ground pin.
* Measuring a HIGH. Can be done by probing a +5V pin.
* Measuring an alternating signal ('activity'). There will be plenty of those on a functional motherboard.

... , will you be able to guide me on what to check?

First, I am assuming that the motherboard's POST is displaying a 301 error, or displaying an error that points to the keyboard.
Refer to the diagram at [here], a diagram that shows what happens during the motherboard's POST.
That diagram is for the IBM PC, but it is sure to apply to other PC and XT class motherboards as well.

The LED's on your keyboard are momentarily flashing, so we know that the keyboard (particularly the keyboard's controller chip) is getting +5V.

Before jumping in with a logic probe, use a multimeter to verify that:
1. In the keyboard, there is continuity between the controller's clock pin/s and the clock pin of the keyboard's DIN connector.
2. In the keyboard, there is continuity between the controller's data pin/s and the data pin of the keyboard's DIN connector.

(Because the controller chip in your keyboard is unknown, you may need to examine the circuitry to work out the pins.)

If continuity, then use the logic probe to see if there is activity on the controller's clock pin/s and data pin/s during the POST.

And after the POST, I would expect that whenever you pressed a key, that you would see activity on the controller's clock pin/s and data pin/s (all the way out to the clock and data pins on the keyboard's DIN connector). Try many keys in case some are working and some are not.

 

[soon668]

Hi modem7,

First, I am assuming that the motherboard's POST is displaying a 301 error, or displaying an error that points to the keyboard.

No, the POST does not display an error. When boot, the LED blink for 3 times, then stop, POST continue as normal, and I am able to get to DOS prompt. Keyboard does not response to any key pressed.

1. In the keyboard, there is continuity between the controller's clock pin/s and the clock pin of the keyboard's DIN connector.
2. In the keyboard, there is continuity between the controller's data pin/s and the data pin of the keyboard's DIN connector.

My best guess is the controller chip used should be similar to the one used in DFK-777F, which is P8049AH, datasheet here. On the keyboard circuitry, I traced the DATA pin from the DIN connector to PIN 39 of the controller, and I got continuity. However, I can't get continuity of the CLK from DIN to any of the PIN of the controller. According to the datasheet, PIN 1 is suppose to be the CLK output. When I trace the CLK connection from DIN, I see there is a 4.7 K Ohm resistor on the way to PIN 1 of the controller.

 

[modem7]

No, the POST does not display an error. When boot, the LED blink for 3 times, then stop, POST continue as normal, and I am able to get to DOS prompt.

That is odd. I would have thought that at least one the two motherboard you have used would have displayed a keyboard related error.

If you remove the keyboard, does the POST display a keyboard related error ?

According to the datasheet, PIN 1 is suppose to be the CLK output. When I trace the CLK connection from DIN, I see there is a 4.7 K Ohm resistor on the way to PIN 1 of the controller.

A P8049AH is a generic device. If I was to look, there may even be one in my old microwave oven. That even applies if the microcontroller is an 8048, or many other microcontrollers produced for different applications. The engineer decides which pins on the microcontroller to use, and what other circuitry is required.

But if you are sure about the circuit you have determined, then maybe the engineer put the 4.7 K Ohm resistor into the clock line for a purpose.

 

[soon668]

If you remove the keyboard, does the POST display a keyboard related error ?

Yes, it display "Keyboard bad", then proceed with memory count, and booted successfully to DOS again.

 

[soon668]

But if you are sure about the circuit you have determined, then maybe the engineer put the 4.7 K Ohm resistor into the clock line for a purpose.

Here is the snapshot of the circuit.

The DATA line goes to Pin 39 of the microcontroller, passing by HD7407P.
The CLK line goes to a 4.7kohm resistor, and further go through a jumper before ended in Pin 40 of the microcontroller.

Here is the snapshot of the front of the board, flipped so that it is easier to match with the circuit

 

[modem7]

If you remove the keyboard, does the POST display a keyboard related error ?

Yes, it display "Keyboard bad", ...

Suggesting that when the keyboard is attached, the motherboard is seeing an AA byte from the keyboard.

Here is the snapshot of the circuit.

First, look at [here], the circuit diagram for the keyboard that IBM supplied with the IBM PC and IBM XT. Note that between the microcontroller and keyboard is U2, an open-collector driver.

With the photos that you provided, I started to trace out some of the circuit, which I show at [here]. R5 is a pull-up resistor.

Using your logic probe, when you press a key on the keyboard, do you see activity on the clock and data lines ?

 

[soon668]

Thanks modem7 for the diagram. I was also trying to do something similar, but got confused :-(

Using your logic probe, when you press a key on the keyboard, do you see activity on the clock and data lines ?

No, the Data and Clock lines remain HIGH when keys are pressed. I have updated your diagram with the result.

 

[modem7]

Thanks modem7 for the diagram. I was also trying to do something similar, but got confused :-(

I am very experienced in 'recreating' circuit diagrams. But it is still easy to get confused. For example, this morning, I realised that I had resistors R4 and R6 swapped. I updated the diagram.

No, the Data and Clock lines remain HIGH when keys are pressed.

The signals on the clock and data lines are of very short duration. For example, at [here], is an oscilloscope capture of an AA byte being sent to the motherboard by an IBM 83-key keyboard (a keyboard that normally holds the data line low). Top waveform is the clock, and bottom waveform is the data.

Perhaps you should use your logic probe to monitor the clock and data wires on a good keyboard, in order for you to be confident of what your logic probe is expected to show on the faulty keyboard (when it sends a byte).

We know that the microcontroller is at least partially working, because it is what is flashing the LED's when it receives +5V. For your faulty keyboard, when the microcontroller generates a byte and corresponding clock, the clock signal is leaving pin 38 of the microcontroller and the data signal is leaving pin 37 of the microcontroller. If you do not see any signals on either of those pins at power-on and later, when you press a key, then it suggests that the microcontroller is at fault.

 

[soon668]

Hi modem7, thanks for the help and patience in explaining to me.

Perhaps you should use your logic probe to monitor the clock and data wires on a good keyboard, in order for you to be confident of what your logic probe is expected to show on the faulty keyboard (when it sends a byte).fault.

I will try to source for another keyboard locally or at least from China, it is not easy to find an XT/AT keyboard anymore.

If you do not see any signals on either of those pins at power-on and later, when you press a key, then it suggests that the microcontroller is at fault.

On another thread that I posted about problem with my XT motherboard, I purchased an exact same motherboard from eBay and have waited 2 months to get it, only to realised the keyboard is also not working :-( I was thinking to further troubleshoot the faulty motherboard after I am able to at least try out the new motherboard. But seem luck is not with me...

I will most likely report back 2 months later after I get another keyboard from ebay :D

 

[modem7]

I will try to source for another keyboard locally or at least from China, it is not easy to find an XT/AT keyboard anymore.

Consider too the option of an AT class keyboard together with an adapter like the one at [here].

BTW. I updated the circuit diagram even more.

 

[Chuck(G)]

Based on behavior, I suspect that the keyboard is operational, but for some minor issues. But it would take a 'scope to verify.

 

[soon668]

Consider too the option of an AT class keyboard together with an adapter like the one at [here].

BTW. I updated the circuit diagram even more.

Thanks for the update modem7. One thing I found which I am not sure it raise any concern to you is, when I plug in the DIN connector to the motherboard (power off), and check continuity using multimeter on C5 or R9, both ends of this components actually reports continuity (beep on multimeter) to the ground. From your diagram, I can understand if one of the end is connected to ground. However, the other end although report continuity, the meter show a small resistance of about 57ohm. If I unplug the connector, it is not behaving this way.

 

[modem7]

Thanks for the update modem7. One thing I found which I am not sure it raise any concern to you is, when I plug in the DIN connector to the motherboard (power off), and check continuity using multimeter on C5 or R9, both ends of this components actually reports continuity (beep on multimeter) to the ground. From your diagram, I can understand if one of the end is connected to ground. However, the other end although report continuity, the meter show a small resistance of about 57ohm. If I unplug the connector, it is not behaving this way.

C5 and R9 run from ground to +5V. When you have the keyboard plugged into the (unpowered) motherboard, measuring the resistance across C5 (or R9) is also measuring the resistance of the +5V on the motherboard, a measurement that is expected to be quite low.

 

[soon668]

C5 and R9 run from ground to +5V. When you have the keyboard plugged into the (unpowered) motherboard, measuring the resistance across C5 (or R9) is also measuring the resistance of the +5V on the motherboard, a measurement that is expected to be quite low.

Thanks modem7 for explaining, not something to be concern of then

Based on behavior, I suspect that the keyboard is operational, but for some minor issues. But it would take a 'scope to verify.

Thanks Chuck for your suggestion, will be difficult to justify a scope. Will a "Virtual Oscilloscope Logic Analyzer" like LHT00SU1 do the job?

 

[Chuck(G)]

How handy are you? There are basic oscilloscope/logic analyzers based on the "Blue Pill" microcontroller boards.
But yes, a USB "virtual scope" would do for this purpose.

 

[soon668]

Consider too the option of an AT class keyboard together with an adapter like the one at [here].

Hi modem7, I should be able to purchase a Fujitsu FKB8720 keyboard locally. How do I go about building the AT2XT convertor?