Getting Compaq Portable 1 running

[Bill-kun]

Okay, Christmas break with family is now over, so getting back to the computers.

Although this is the 286 Technical Reference, the Video Display should be effectively the same between the Portable and Deskpros: https://onedrive.live.com/?authkey=!AJOmtRmVhlhL6_o&cid=7D567B5161FE0EB7&id=7D567B5161FE0EB7!47144&parId=7D567B5161FE0EB7!37014&o=OneUp

Thank you for the link. The schematic pack has schematics for 3 different versions of the VDU, on PDF pages 63-81. I looked through all of them but I don't see anything marked "video RAM" or similar or "U62" through "U69." Again, this is the first time I've ever tried to read a schematic, so please tell me if I missed it in there.

So it looks like there are 3 paths I can take to trace the VDU circuit board.

Path 1: start at edge connector.

With the equipment you have the easiest way is start at the edge connector. If you look online at the pinout data 7 is B2. Follow the trace from that pin to a chip or it may go to multiple chips. If you can't follow the trace you will need to start checking with your meter on beep to see where it goes. If you get a beep check with ohms and verify its really close to 0 ohms.

Path 2: start at SRAM chip. I assume that means U62-U69.

Also can start from the SRAM chip data pins Check the chip datasheet and find the same bit on the "other side" of the chip. D7->Q7, A7->B7 etc. Follow that signal until you get path from SRAM to connector. The SRAM chip uses separate pins for data output and input on 8 and 12. You want to trace to one of those pins.

Path 3: start at the MC6845 chip.

It may be simpler to trace to the MC6845 data pins since we know which D7 (pin 26) must be for the board to function. Likely less layers of chips to trace through than to edge connector. This will be time consuming. I can look some at my board if you get stuck.

I chose path 3 first, starting at the MC6845 chip since it gave me a single starting point, specifically pin 26, and sounded like it would fork the least.


Here are good photographs of the front and back of the VDU board. I started by printing these out on paper and then using a red ink marker to literally trace the traces I was finding. I should have continued doing that, but I stopped it at some point and will have to catch up on it, or redo it since it the marker writing got too hard to read. This allows me to write down what physically leads to what.




In parallel to the red ink tracing, I started writing a "trace flowchart" on a separate piece of paper. This allows me to write down in words what leads to what. Here is my paper on that so far, aided critically by looking up data sheets on the chip numbers.

The network forks out several times as I follow is from MC6845 pin 26.

1. Does this trace flowchart look sound so far?
2. Are there forks where I should cease tracing, and instead go down others? If so, which to not bother with and which to definitely continue?
3. Is I0 the correct pin I want to go out from at SN74LS151NDS? Here is the pinout.
   1. 

 

[Bill-kun]

I've now also done some more tracing, this time on path 1: start at edge connector.

With the equipment you have the easiest way is start at the edge connector. If you look online at the pinout data 7 is B2. Follow the trace from that pin to a chip or it may go to multiple chips. If you can't follow the trace you will need to start checking with your meter on beep to see where it goes. If you get a beep check with ohms and verify its really close to 0 ohms.

Here are the marked up photographs with the physical tracings and the tracing flowchart.




I went as far as I thought I could before showing this work here and asking the same type of questions:

1. Does this trace flowchart look sound so far?
2. Are there forks where I should cease tracing, and instead go down others? If so, which to not bother with and which to definitely continue?
3. What chip is M3864-04? I could not find a datasheet for it. Or am I searching for the wrong number from the chip?

Other questions in general, now that I've done this this far.

1. I haven't gotten anywhere close to U62-U69 on either of these paths so far. How long is this supposed to go before I connect with them?
2. How can I read the schematics linked in post #39 for the 3 versions of the 286 VDU to help me?

 

[djg]

looked through all of them but I don't see anything marked "video RAM" or similar or "U62" through "U69." Again, this is the first time I've ever tried to read a schematic, so please tell me if I missed it in there.

No time to read all of your writeup this morning. We know the U#'s on the schematic won't match your board so they won't be U62-U69. In the previous photo posted of the board that went with this schematic you can see the memory is U22-U29. Took a quick look and video ram is on page 65 of PDF in the upper right. They used one symbol to represent all 8 parts. Note the U22-U29 for the reference designator. Below that is LSB->MSB so MSB is U29. Part # is 6167 and I found a UM6167 datasheet which is a 16kx1 SRAM. Think the not very readable text next to the part number is 16kx1 also.

 

[mykrowyre]

I'm very interested in your diagrams, I am troubleshooting the same card. I have the SAMS photofacts, from 1987, but the board it shows is slightly different than mine (which is the same as yours). My card worked fine and then I lost horizontal and vertical sync. One is stuck at 5V and the other is showing a 164khz square wave. ??

I'm hoping I can glean enough from the schematic if the newer version of the board to troubleshoot the old board. Reverse engineering the schematic from the PCB has been pretty difficult. I just got the printed photofacts a few days go.

My schematics are probably not helpful to you, but I'd be happy to share any info from my 4/1987 printing.

If you are booting to MS-DOS with no working display you can connect a printer and hit CTRL+PRT SCR to send all output to the printer.

 

[mykrowyre]

Ok I looked at your original post to me and I agree it looks like bad ram. I'd feel the ram chips along the bottom of the board to see if any of them are warmer than others.

I looked over the schematics and they are still just too different to be of much use for the Rev 1 board that we both have.

 

[djg]

Version 3 schematic is DRAM so not relevant. Version 1 and 2 are closest to your card. Version 2 has the SRAM chips separately drawn but pin numbers hard to read so I'll use version 1. U10 D7 pin 26 is net BD7. Following the net the SRAM U29 pin 12 should be connected to it. Pulling out my board. My test doesn't show any connection from U10 pin 26 to any of the SRAM pin 12.

Lets look at the data in. That goes to DO bus. That goes to the new big chip U4 thats not present on your board. Also goes to U13 pin 13 through chip output pin 12 to U14 pin 23 ROM. Buzzing out my board looks like ROM U41 pin 23 goes to I42 pin 9 output. U42 pin 8 input goes to U69 pin 8 so looks like U69 is the bad chip unless they addressed the ROM differently.

Checking out your first signal trace my board matches your U39 pin 26 to U44 pin 9. Just buzzing with meter I find U44 pin 8 does go to U69 pin 8 so that agrees with previous the U69 is the bad chip. Looks like you didn't find that connection in your tracing. I think your approach was fine but sometimes just brute force checking with the meter on beep dragging the lead down rows of pins finds connections better than trying to follow trace on a populated board.

What chip is M3864-04?

What U number is it? Didn't see that number looking at my board. If other questions are still relevant ask again.

 

[Bill-kun]

What U number is it? Didn't see that number looking at my board. If other questions are still relevant ask again.

M3864-04 is U41, the second-largest chip on the board.

I think your approach was fine but sometimes just brute force checking with the meter on beep dragging the lead down rows of pins finds connections better than trying to follow trace on a populated board.

I was doing a lot of running the probe over chip legs and tapping VIAs and such. I apparently have to do it more slowly, though; it took me at least 2 or 3 sweeps to find one of the beeps on a chip leg while drawing my trace flowchart. Or maybe higher-quality multimeters are faster at getting a beep?

The two documenting methods I invented here were more for creating a record for myself to come back to, or at least for myself in the far future if I wanted to look back at past work. The act of tracing has even just so far taken so long that it is a multi-session task, and I'd forget everything if I didn't write it down, and I'd be unable to follow the writing unless I gave it an orderly fashion like I did. Those two documents are really neat tools, and in doing future circuit board tracing, I plan to use them more, improving as I go, like using graph paper instead of blank white. And familiarity with chip numbers and what they contain and what those contents do, would help me direct my tracing efforts faster down more probable paths.

I haven't gotten anywhere close to U62-U69 on either of these paths so far. How long is this supposed to go before I connect with them?

Growing up, I never imagined that all these green boards made of wire that I saw inside VCRs, stereos, and computers had taken so many people so many months just to plan their layout. As a guess from you, to help me understand how big the board network is, and thus how wide and deep the tracing flowchart would be, how many node points do you think it would take to go:

• from B2 at the edge connector to U69?
• from MC6845 (U39) pin 26 to U69?

Ok I looked at your original post to me and I agree it looks like bad ram. I'd feel the ram chips along the bottom of the board to see if any of them are warmer than others.

I put the VDU board back in, turned on the computer for about 5 minutes, and felt the chips on the board. U62-U69 did not feel any warmer than a few other chips on the same board.


But I am ready now to order a new U69 chip. U69 is marked "AAN / INMOS8417-D / IMS1400P-70". Googling "INMOS8417-D" gives nothing. Googling "IMS1400P-70 data sheet" gives a few results, and a few of them lead to data sheets for just "IMS1400."

• What does "AAN" mean?
• What does "INMOS8417-D" mean?
• What does "IMS1400" mean?
  • There are more possible suffixes than "P." What do the suffixes "P" and "-70" mean?
• What is an acceptable chip number for me to order? How close do I need to be to exactly the number "IMS1400P-70"? (I'd like to get an exact replacement.)
• What are good reputable online places to buy it from? Can you provide links?

 

[mykrowyre]

M3864-04 is U41, the second-largest chip on the board.


I was doing a lot of running the probe over chip legs and tapping VIAs and such. I apparently have to do it more slowly, though; it took me at least 2 or 3 sweeps to find one of the beeps on a chip leg while drawing my trace flowchart. Or maybe higher-quality multimeters are faster at getting a beep?

The two documenting methods I invented here were more for creating a record for myself to come back to, or at least for myself in the far future if I wanted to look back at past work. The act of tracing has even just so far taken so long that it is a multi-session task, and I'd forget everything if I didn't write it down, and I'd be unable to follow the writing unless I gave it an orderly fashion like I did. Those two documents are really neat tools, and in doing future circuit board tracing, I plan to use them more, improving as I go, like using graph paper instead of blank white. And familiarity with chip numbers and what they contain and what those contents do, would help me direct my tracing efforts faster down more probable paths.


Growing up, I never imagined that all these green boards made of wire that I saw inside VCRs, stereos, and computers had taken so many people so many months just to plan their layout. As a guess from you, to help me understand how big the board network is, and thus how wide and deep the tracing flowchart would be, how many node points do you think it would take to go:

• from B2 at the edge connector to U69?
• from MC6845 (U39) pin 26 to U69?

I put the VDU board back in, turned on the computer for about 5 minutes, and felt the chips on the board. U62-U69 did not feel any warmer than a few other chips on the same board.
View attachment 1237443

But I am ready now to order a new U69 chip. U69 is marked "AAN / INMOS8417-D / IMS1400P-70". Googling "INMOS8417-D" gives nothing. Googling "IMS1400P-70 data sheet" gives a few results, and a few of them lead to data sheets for just "IMS1400."

• What does "AAN" mean?
• What does "INMOS8417-D" mean?
• What does "IMS1400" mean?
  • There are more possible suffixes than "P." What do the suffixes "P" and "-70" mean?
• What is an acceptable chip number for me to order? How close do I need to be to exactly the number "IMS1400P-70"? (I'd like to get an exact replacement.)
• What are good reputable online places to buy it from? Can you provide links?

16K Static ram chip, 70 is the speed, 70ns

I see some on ebay here:

IMS1400P-70 INMOS 20 Pin Dip Chips (2 Parts) | eBay

Find many great new & used options and get the best deals for IMS1400P-70 INMOS 20 Pin Dip Chips (2 Parts) at the best online prices at eBay! Free shipping for many products!

www.ebay.com

Here's the datasheet:

https://4donline.ihs.com/images/VipMasterIC/IC/IMOS/IMOSS00045/IMOSS00045-1.pdf?hkey=D9A213CC6FEE7D103EF6B88F2AEB20B8

 

[djg]

• from B2 at the edge connector to U69?

For edge connector to U69 I would expect it wouldn't go through more than 2 chips.

from MC6845 (U39) pin 26 to U69?

Checking out your first signal trace my board matches your U39 pin 26 to U44 pin 9. Just buzzing with meter I find U44 pin 8 does go to U69 pin 8 so that agrees with previous the U69 is the bad chip. Looks like you didn't find that connection in your tracing.

You did see my reply which answered that question? There is one chip U44 between U39 and U69.

M3864-04 is U41,

That is the character ROM which maps the ASCII code to dot pattern to display. Its a custom ROM so you won't find the part number in any standard datasheet.

What does "AAN" mean?

Don't know. There are frequently extra marks on chips which indicate where manufactured or other information.

What does "INMOS8417-D" mean?

Chip was made by INMOS company in 17 week of 1984. No idea what D indicates

What does "IMS1400" mean?

Its the part number. IMS would likely be a chip family designation such as for SRAMS and 1400 the particular part.

There are more possible suffixes than "P." What do the suffixes "P" and "-70" mean?

P indicates packaged in plastic DIP. These are manufacturer specific so you need to check the data sheet. -70 is the time from chip enable to data valid in nanoseconds. The datasheet I linked to doesn't list chips that slow. Didn't find older datasheet in a quick look.

What is an acceptable chip number for me to order? How close do I need to be to exactly the number "IMS1400P-70"? (I'd like to get an exact replacement.)

Matching the IMS1400P-70 is safest but I suspect the board isn't going to be too picky about the chip. I normally see if I can get the same part at a reasonable cost. If not I either see if I can find cross reference information, also in databooks. Also google the generic 16kx1 SRAM to try to find other chips. You then need to compare the datasheets to make sure the pinout is the same. Then you need to check the speed. Normally manufacturers will use the cheapest chip that works which is normally the slowest so using a slower chip is risky. I try to use the same speed but will use a somewhat faster chip if that's what is available at a reasonable cost. There is some risk of not working properly if the chip is too much faster.

What are good reputable online places to buy it from?

I've managed to get a reasonable stock of parts so haven't been buying that much lately. Other may have better recommendations. I try to stay with US suppliers that have been around for a while and try others when necessary. Looking at my bookmarks one is Unicorn electronics. They have some 16kx1 in 20 pin but I didn't check the datasheets. Minimum order may give you more spares that you want unless you need other parts. I keep lists of parts low on or for future projects to fill minimum orders.
I've used Quest components. They say they have the IMS1400 in stock. They are also a parts broker reselling other peoples parts. In stock with price may indicate they aren't reselling. Also in my bookmarks mistersprockets has somewhat fast 16kx1. Should check datasheet. Have others but couldn't quickly see if they had useful parts.

I've gotten parts OK on eBay from US suppliers that don't look like fronts for overseas sellers but more risk here.

 

[Bill-kun]

16K Static ram chip, 70 is the speed, 70ns

I see some on ebay here:

IMS1400P-70 INMOS 20 Pin Dip Chips (2 Parts) | eBay

Find many great new & used options and get the best deals for IMS1400P-70 INMOS 20 Pin Dip Chips (2 Parts) at the best online prices at eBay! Free shipping for many products!

www.ebay.com

Here's the datasheet:

https://4donline.ihs.com/images/VipMasterIC/IC/IMOS/IMOSS00045/IMOSS00045-1.pdf?hkey=D9A213CC6FEE7D103EF6B88F2AEB20B8

Thank you. I ordered one pair from that seller. It should arrive by January 20. I wish I knew what the "L" suffix means (the datasheet doesn't say) but other than that the part number is identical.

Checking out your first signal trace my board matches your U39 pin 26 to U44 pin 9. Just buzzing with meter I find U44 pin 8 does go to U69 pin 8 so that agrees with previous the U69 is the bad chip. Looks like you didn't find that connection in your tracing.

You did see my reply which answered that question? There is one chip U44 between U39 and U69.

I suppose that was one detail I forgot to circle back on. Sorry. I have now verified that continuity. That was definitely one I missed myself in tracing the board. Hopefully the next time I will trace a circuit board like this I will have its schematic to help me.

 

[djg]

I wish I knew what the "L" suffix means

For SRAM normally indicates low power. Should be fine.

 

[Bill-kun]

The foam and foil keypad replacements arrived today from Texelec. Fortunately, the 8-Bit Guy did a video of the foam and foil keypad replacement process for the Compaq Portable 1. I will start on this in the next day or two. It will be tedious, so I will have audiobooks or podcasts ready to keep me company.

And Tech Tangents too.

And this guy too.

 

[mykrowyre]

I found the foil tends to rub on the edge of the key socket making a little squeaking noise which is like nails on a chalk board.

I used my side cutters to snip off the edge of the foil anywhere that it rubbed the key socket, testing all keys over and over until it sounded normal.

Really the foil needs to be slightly smaller in diameter for the compaq keyboard.

 

[Bill-kun]

I've still been trying to determine how to best remove the bad microchip from the VDU board. I thought it would be a piece of cake: melt the solder surrounding the legs and take the chip out. The problems I have encountered so far have been:

1. All legs have to have liquid solder simultaneously. That is easy to make happen with a soldering iron on a resistor. That is impossible to make happen with a soldering iron on a 20-leg microchip.
2. Another option for having the solder be liquid is a heat gun. But since I am not skilled with a heat gun and I think me using one would be dangerous, that is not a path.
3. I tried using copper solder wick. After working for a couple hours getting the solder wicked from the bottom of the board, the chip will still not lift off. I can't see any more solder that is holding it on, but apparently I am wrong. I can't just yank it off despite solder still being there or I risk destroying the traces on the board.
4. Another option is to cut each of the individual microchip legs and then the body of the microchip will be free. Then I would just unsolder each loose leg. The problem with this method is that the microchip is so close to a neighboring microchip that I have yet to find an instrument that can reach in and snip the individual legs. What I really need is needlenose pliers, except the nose has to be bladed for cutting, not knurled for gripping. I just bought a set of small cutting pliers, but they still won't get in there. This was the smallest tool I could find after going to 2 hardware stores.

Here is a picture with a ruler for scale. Can anyone please give a hyperlink for buying a tool for this work on Amazon or anywhere else?

 

[jafir]

Something like this might be smaller to assist in cutting out a chip: https://www.amazon.com/dp/B000GTMZHG

I ended up splurging on a desoldering gun. It works very well. Sometimes you’ll still leave some solder behind, like you do with the wick. You can try adding fresh solder and then desoldering. Also many times you just have to carefully wiggle the legs and you’ll get a little pop when they free from the through holes. But there is always the risk that you do some damage.

Another option is low melt solder like chip quik. I’ve not used it, but it looks like it should work well.

 

[Bill-kun]

Something like this might be smaller to assist in cutting out a chip: https://www.amazon.com/dp/B000GTMZHG

I got an identically shaped similar tool one and it finally worked. I was able to get in and snip the individual legs of the bad chip. The bad chip is finally now out for good, and destroyed, even.

I soldered the new-used chip in and the screen is now far worse than it was when I started. Here is a still picture. Many of the characters that you are seeing flicker--not blink, but flicker.


What I believe is the cause of this is solder flux still being on the board. I tried to get off as much as I could before soldering on the new-used chip, but this stuff is awful to try to clean off. I followed these instructions that I found at pcbway.com.

1. Dip a clean toothbrush in isopropyl alcohol or acetone. Flick away excess by flipping your wrist two or three times.
2. Use the toothbrush to gently brush the solder flux on the PCB, applying the alcohol or acetone. Do not use too much pressure, as this may cause the solder point to break. Repeat as necessary, to remove remaining flux residue.
3. Wipe down the cleaned area with a clean rag. Rinse the toothbrush in distilled water and allow the PCB to dry. Remove any flux dust with canned air.

In my use of solder flux a few weeks ago to try to unsolder the bad chip, the flux I used spread to the area of about half of a dollar bill.
1. I used an old toothbrush with isopropyl alcohol. I was liberal in the use of it since it evaporates easily, and there was a lot of area to cover.
2. It didn't really do much. It spreads it around. Because of the texture of the flux, it doesn't just rinse off or get scrubbed off, like dried coffee drops in a cup.
3. Wiping down the circuit board is also nearly impossible, since the flux doesn't get wicked into the cloth, and especially on the back there are thousands of legs that will infinitely snag any cloth, any cotton swab, and any paper towel.

First, questions about solder flux in general.

1. Is the flux the likely cause of this problem? I piggybacked the new-used chip on the bad one before I tried anything to get the bad one off, and the screen worked perfectly then.
2. Is solder flux electrically conductive? If yes, why is it made that way? It sure seems like being conductive makes using it far more of a problem than a benefit.
3. What is flux made of? It's got the texture of Neosporin ointment, and is just as frustrating to clean off of something as Neosporin. Doubly so, since this is a circuit board. Triply so, since this is a circuit board with chips that you cannot get underneath to clean, and fragile-legged capacitors and things on it that can break off.

So, how do I really get solder flux off this circuit board? Here are some ideas I have. I have no idea if they are smart or not.

1. If I use acetone, will that damage the circuit board if I use too much of it? Will that run off all of the solder flux, or will I have to scrubs and somehow get underneath chips?
2. Can I just pour really hot water (steaming or nearly boiling) over the circuit board? That would melt the solder flux and run it off, and be in a large enough quantity to be the volume that I need.
3. Can I use a heat gun to melt all the solder flux at once? I really dislike the idea of using a heat gun because it's so powerful and dangerous to inexperienced me. How careful do I have to be? And even if I do melt all the flux at once, what then? I still have to scrub or rinse it off somehow since there would still be residue even if most of it dripped off.

 

[tradde]

Is it possible you used the wrong type of solder? I have never seen solder flux spread out that large. Or are you just applying flux itself? I have never had a problem so have never had to try to remove it. So I can't make a suggestion there? I would think a heat gun if too hot could unsolder other components nearby? Someone more knowledgeable will come forth and answer all your questions.

 

[SomeGuy]

Flux would have zilch to do with that problem.

It looks like you either damaged a trace or made a bad solder joint.

Use an ohm meter to check each trace to each leg of the chip.

Just alcohol and a toothbrush should clean enough of the flux off. You can rinse this card with water as long as you make sure to let it dry completely first.

 

[djg]

Is solder flux electrically conductive?

Depends on the flux. I like the water soluble flux which is somewhat conductive but easy to remove. Some flux is designed to not need cleaning and other must be cleaned off.

For us to answer your other questions you need to say what flux you used. There are many types of flux.

 

[Bill-kun]

For us to answer your other questions you need to say what flux you used. There are many types of flux.

The flux I used is Oatey #5 lead-free solder paste.

https://www.amazon.com/Oatea-Oatey-53017-Paste-Flux/dp/B000UTSSQO/