PDA

View Full Version : Incompatability of AST Sixpackplus and NEC V20 Experienced with Clone XT



SailorEd
March 2nd, 2015, 02:41 PM
As noted in previous posts, I have also experienced the incompatibility of the SIxPackPlus board with the NEC V20 CPU chip.
Booted OK until installed board. When installed, NOTHING; only a few rotations of fan in switching power supply. Remove board and boots OK.
Wanted to install card to expand memory from main board of 256K up to 640K. I have now ordered Intel 8088 on e-bay to see if the SixPackPlus will boot on it.
It is a clone system with two floppy, HDD and no math co-processor.

modem7
March 3rd, 2015, 11:22 PM
When installed, NOTHING; only a few rotations of fan in switching power supply. Remove board and boots OK.
Classic symptoms of a short circuit in the card (which overloads the power supply, stopping it from working).
The short circuit is possibly in one of the tantalum capacitors.

NeXT
March 7th, 2015, 07:30 AM
Yeah, that screams of a rail shorting out on the card and less of a CPU failure.

Stone
March 7th, 2015, 07:53 AM
Very easy to test for this with an ohmmeter.

SailorEd
March 8th, 2015, 08:43 AM
Thanks for the suggestion. I have the ohn meter but not the knowledge to do the test. Would appreciate your advise on how to do it or where to go for the test procedure. Thanks Ed

modem7
March 8th, 2015, 09:35 PM
Thanks for the suggestion. I have the ohn meter but not the knowledge to do the test. Would appreciate your advise on how to do it or where to go for the test procedure. Thanks Ed
1. Remove card from computer.
2. Turn card over so that the solder side of the card is showing.
3. Put mutimeter into resistance mode.
4. Refer to the diagram at [here (http://www.minuszerodegrees.net/images4/card_edge_connector_power.jpg)].
5. Measure the resistance of the +12V line by placing one of the meter probes onto the +12V pin, and the other probe onto either of the two ground lines. Note the resistance.
6. Repeat step 5, except use the -12V pin. Note the resistance.
7. Repeat step 5, except use the first +5V pin. Note the resistance.
8. Repeat step 5, except use the second +5V pin. Note the resistance.
9. Repeat step 5, except use the -5V pin. Note the resistance.

If any of the resistance measurements are zero or a few ohms, then there is definitely a short citcuit (on the measured line).
Perhaps report back your measurements.


Note: Both +5V lines are measured, just in case a card maker decided to use only one (i.e. because of very low +5V consumption).

Note: There is an assumption that the card maker has used both grounds on the solder side of the card, which is what I would expect. That can be proven by verifying that there is zero ohms between the two grounds.

Stone
March 9th, 2015, 04:32 AM
I prefer to use a slightly simpler method. This involves testing for shorts where the PSU connects to the motherboard with the power off.

First, remove all the cards from their slots. Then check the +5V, -5V, +12V and -12V connections for shorts to ground. If all is good there, insert the questionable card into a slot and repeat the test for shorts at the same four locations as just done. If there is now a short on one (or more) of the power rails it is due to the card you just inserted.

SailorEd
March 12th, 2015, 07:16 PM
Thanks for the instructions. Will get it done tomorrow. I physically examined all the capacitors and they appear to be without physical distress.
Ed

SailorEd
March 12th, 2015, 07:18 PM
Thanks for the additional tip. Will try when have time to pull all cards from machine. Now it is running (knock on wood) with another card that I populated with chips.
Ed

SailorEd
March 13th, 2015, 11:18 AM
I have checked the resisance and got these results 20K scale Pin to Grd: 3-.29 5-open 7-0 9-0 29-.29
When using 20M scale appears that capacitors where being charged as readings on on 9 and seemed to start about 1.9 and go toward 0
Does this tells us anything about the condition of the board?
thanks
Ed

modem7
March 16th, 2015, 02:35 AM
I have checked the resisance and got these results 20K scale Pin to Grd: 3-.29 5-open 7-0 9-0 29-.29
When using 20M scale appears that capacitors where being charged as readings on on 9 and seemed to start about 1.9 and go toward 0
Does this tells us anything about the condition of the board?
Well, the +5V and -5V appear to be okay.

The +12V and -12V show "0", suggesting a short circuit on both lines, however, it is highly unlikely that that is the case (both short).
I see that you are using a high scale. What do you read when you use the 20 ohm scale ?

SailorEd
March 17th, 2015, 09:31 AM
Checked using 200 ohm scale (no 20 scale on meter) with following results: 3-Open; 5-Open; 7-7.4; 9-Open; 29-Open.
Thanks for you input on what to do next?
Regards
Ed

modem7
March 18th, 2015, 03:13 AM
Checked using 200 ohm scale (no 20 scale on meter) with following results: 3-Open; 5-Open; 7-7.4; 9-Open; 29-Open.
Thanks for you input on what to do next?
Partial short on the -12V line. Could be capacitor; could be something else. But forum history weighs heavily towards capacitor.

You can prove the 'short circuit on -12V line' hypothesis for yourself. Put a piece of sticky tape over the B7 contact, and then carefully (so as not to affect the tape) insert the card back into your computer. If the computer starts, it proves the hypothesis.

Provide us with a photo of your SixPakPlus card, or, point us to a photo of same on the Internet. If your SixPakPlus matches mine, I am in a better position to advise further.

Also, if you have not already, do a detailed visual inspection of the card, because the faulty component may be visually showing as damaged. With the capacitors, you are looking for signs like:
http://www.minuszerodegrees.net/images/bad_tantulum_1.jpg
http://www.minuszerodegrees.net/failure/faulty_decoupling_cap.jpg

SailorEd
March 19th, 2015, 12:36 PM
I physically examined both sides of card and see no physical deterioration of any components or leads touching.
I placed tape over the B7 contact and installed installed card. Computer started OK!! Of course did not recognize card had been installed.
I have attached pic of board.
Thanks again for all the help.
2329023291

modem7
March 20th, 2015, 03:48 PM
I have attached pic of board.
Your SixPakPlus card matches mine. Down near the edge connector are three blue tantalum capacitors, C44, C45, and C43. The middle one, C45, sits on the -12V line. Remove it (unsolder, or clip off) to see if it is the source of the partial short.

By the way, when you remove the sticky tape, it is a good idea to clean the contact in order to remove any 'sticky' residue.


Computer started OK!! Of course did not recognize card had been installed.
I expect that the -12V is only being used to support the serial port. I covered up the B7 contact on my card, and as I expected, my card's RAM was still seen at power-on. That suggests that your card has additional issues. Switch/jumper configuration? Bad RAM in bank 0.

SailorEd
March 20th, 2015, 06:30 PM
Thank you for checking with your card. I will continue forward with looking into ram and switches/jumpers. Seller of card told me it was removed from machine with 256K on the main board the same as my machine so I expected switches to be set correctly.

I noticed that the sixth row of ram chips (counting from the end of the card inward) is different from the other five rows. The sixth row of chips are MCM6665BP20 all of the other chips are TMM4164AP-15. This is correct?

Before getting your comments, today I checked the resistance of the three capacitors and the C45 is the only one that shows resistance. Tomorrow I will try to remove it from the card and see what happens. Will let you know what happens.

Thanks again,

Stone
March 20th, 2015, 06:38 PM
I don't think you can accurately check the resistance of a capacitor that is still in the circuit.

modem7
March 20th, 2015, 08:46 PM
Before getting your comments, today I checked the resistance of the three capacitors and the C45 is the only one that shows resistance.
Stone is correct. Making a measurement of C45 with it in-circuit, results in a measurement of the combined resistance of all devices on the -12V line, not just C45.


I noticed that the sixth row of ram chips (counting from the end of the card inward) is different from the other five rows. The sixth row of chips are MCM6665BP20 all of the other chips are TMM4164AP-15. This is correct?
Not a problem. My card uses a mixture of both MCM6665BP20 and TMS4164-15NL.


Thank you for checking with your card. I will continue forward with looking into ram and switches/jumpers. Seller of card told me it was removed from machine with 256K on the main board the same as my machine so I expected switches to be set correctly.
On my card, switch 7 on the switch bank is off. Looking at the documentation for the early versions of the card, switch 7 does not do anything, and there is nothing to indicate if switch 7 needs to be left off, or left on. You could try switch 7 off and see if it makes a difference, but I don't think it will.

SailorEd
March 21st, 2015, 09:00 AM
Making Progress. Removed C45, checked with Ohm meter and short was gone. Installed in machine and it booted through the ram check and reported error as follows:
512K OK
08 RAM Bad 80004

Does this indicate location of bad chip? How are the banks of ram and chips identified on the AST card?

Thanks again for all the help getting to this point. I have some additional ram chips that I expect are OK; but not 100% sure.

Stone
March 21st, 2015, 09:32 AM
It looks like the error is in the 5th bank. You can verify this by switching the 5th bank chips with the sixth bank chips. If the error moves accordingly you will know that there is a bad chip in the (now) 6th bank. Replacing those chips one by one with a known good chip will locate the bad one.

SailorEd
March 21st, 2015, 12:48 PM
How are the banks numbered? From the BUS End (inside row out) or end of the card? Is the parity chip the bus plug edge or top edge of the card? Thanks Ed

modem7
March 21st, 2015, 01:40 PM
512K OK
08 RAM Bad 80004
So, bit 3 in the bank starting at 512K.


How are the banks numbered? From the BUS End (inside row out) or end of the card? Is the parity chip the bus plug edge or top edge of the card? Thanks Ed
A quick experiment shows that it is the same as used in the first version of the card:
Banks run from center of card to edge of card.
Bits: From top to bottom are P/7/6/5/4/3/2/1/0

modem7
March 21st, 2015, 02:14 PM
Making Progress. Removed C45, checked with Ohm meter and short was gone.
I am sorry if any of the following is 'suck eggs'.

For a replacement capacitor, you have the option of using an aluminium electrolytic type. In fact, that is what AST used on my card. Photo at [here (http://www.minuszerodegrees.net/temp/1/h63ndxcgd90345.jpg)].
Height must not be such that the capacitor can make contact with an adjacent card.

Positive leg of capacitor into positive hole. For C45, the positive hole is the one closest to the DB25 connector.

In this application, a larger capacitance value is okay, e.g. 33uF

You can also use a higher voltage rating, e.g. 25V

SailorEd
March 21st, 2015, 03:49 PM
SUCCESS!! Rotated memory bank 4 & 5 and that got RAM OK up to 576. Then played swap and test until found two bad chips. Everything is now working except serial port given removal of C45. Really do not need it now and will leave it for a future activity as I will have to buy solder sucker, iron and capacitor(s).

Actually, the card does not have all chips populated, so I expect the serial chip(s) are what is missing.

Thank you so much again for all your guidance and coaching. I would not have been able to get it running without your and the forum support.

Buying from e-bay is a risk; sometimes you win and other times you are challenged to find value in what you buy.

We can now consider this thread closed and successful. Have a good weekend.
Best regards
Ed

schreiben
March 1st, 2017, 06:18 AM
I placed tape over the B7 contact and installed installed card. Computer started OK!! Of course did not recognize card had been installed.