1088XEL Atari ITX Motherboard DIY Builders Thread

[flashjazzcat]

12 minutes ago, MacRorie said:

FJC, I think MyTek is referring to the fact that since they *are* shrouded, it is easy to mis-align the U1M when attaching it to the motherboard.

I fully appreciate that and understand exactly what he means.  I'm saying it's easy to align it correctly, as well, and that doing so is a skill worth having when your U1MB turns up with shrouded headers (as they all now do, thankfully). Just in case anyone thinks it's a good idea to risk modifying a perfectly good board with shrouded IDC headers just for the sake of having sight of the open pins.

 

[+DrVenkman]

I agree with all of the above - the shrouded headers on the U1MB I put into my XLD are great, as they saved me over an hour of work changing them from right-angle pins as I had to do with my XEL build. And buried back in this very thread is my own tale of woe and a perfectly innocent U1MB board rendered useless by my first mistakes removing the pins (and yes, I did watch Jon's video first!). Bah. I did things just enough differently the second time that it went completely perfectly. Live and learn, and about $80 down the drain replacing a U1MB. 

But with the shrouded headers, Michael is also right - if you remove and reinsert the board, it's easy to shift the board by a row in either direction with no indication you've done so except a non-booting computer. Fortunately, misalignment doesn't seem to damage anything - it just doesn't work. So that's a good thing to remind people to check if their machines don't boot on first attempt after assembly, or if they've been working on the system with the U1MB removed and replaced in the process.

 

On the whole, however, I'll take the IDC box headers over the surgery otherwise required any day!

[+mytek]

On 6/30/2019 at 9:09 AM, tf_hh said:

Hi,

 

I found a little "issue" with the 1088XEL V1.1 - only with PAL chips and crystal equipped (I think so).

 

After building 7 pieces of the 1088XEL V1.1 for different people, I have every time the opinion, that the screen is too greenish. Look here, one example:

 

 

Of course I know, effects like camera setup, color settings at your computer or mobile device is always falsifying what I want to show, but I think most people trust me to judge this behavior the right way ?

 

One of 7 boards I have assembled didn´t make any right color, it´s shows totally wrong colors even when I set the color pot to the maximum. So I start to examine this issue and found out, that the maximum voltage which can be produced by the loading pump at the 1088XEL circuit is approx  6.5 volts. Some of the 7 pieces only reach 6.0 volts, one only 5.9 volts.

 

As every hardware guy knows, the color settings depends on GTIA, CPU (which PH1 signal is used for the loading pump) and in part also ANTIC. So I try to change them with several other, best known working parts and the issue changes slightly, but never the wrong color behavior is solved.

 

After changing several GTIAs and CPUs, I look about the differences between Mytek´s solution and the one found in the 800XL five-chip standard version. The one and only difference is the transistor Q9. Atari used the 2N3904 transistor nearly everywhere, Mytek uses the 2N2222. So I change the following transistor to a 2N3904:

 

 

This fixes the behavior perfect. Now the range of voltage setting with the color pot is up to 8.5 volts, which matches my reference 800XL. I didn´t look closer into the datasheets of the transistors, but the 2N3904 works better here. Maybe it´s also my 2N2222 series I purchased, but generally all parts of the 1088XEL I have assembled work fine.

 

@Michael: Would suggest you to mention that in the BOM. I can´t test now, if NTSC is also affected, but I think not, as most systems are build from NTSC using guys IMHO.

 

Jurgen

 

Jurgen I checked out two of my XELs (both in PAL mode) and one had just barely over 4V going to the color trim pot, and yet the other was a very healthy 9V. Hard to say if both have a PN2222 from the same manufacturer or not, but I was ordering some parts for another project from Jameco.com and added some PN3904s to the cart. When those arrive I'll swap out the poor performing PN2222 and see if that fixes it.

 

I'll report back when I know more.

 

[flashjazzcat]

12 hours ago, DrVenkman said:

if you remove and reinsert the board, it's easy to shift the board by a row in either direction with no indication you've done so except a non-booting computer.

The open headers do not absolutely prohibit the possibility of misalignment, and I fancy with shrouded headers, it's only possible to misalign on one axis, and with open headers, both axes.

 

 

[+mytek]

9 hours ago, flashjazzcat said:

The open headers do not absolutely prohibit the possibility of misalignment, and I fancy with shrouded headers, it's only possible to misalign on one axis, and with open headers, both axes.

Jon I think you are missing the point, or perhaps you are just jesting with us.

 

Yes agreed, it doesn't prevent misalignment one way or the other, but having better visibility without the shroud in the way certainly helps us old folks with bad eyesight to get it right . I myself have accidentally installed both the U1MB and the XEL CF3 off by one pin position several times. You'd think I'd know better, but then again I'm doing a lot of experimenting with constant unplugging and re-plugging so sh_ts bound to happen.

 

[+DrVenkman]

2 hours ago, mytek said:

Jon I think you are missing the point, or perhaps you are just jesting with us.

 

Yes agreed, it doesn't prevent misalignment one way or the other, but having better visibility without the shroud in the way certainly helps us old folks with bad eyesight to get it right . I myself have accidentally installed both the U1MB and the XEL CF3 off by one pin position several times. You'd think I'd know better, but then again I'm doing a lot of experimenting with constant unplugging and re-plugging so sh_ts bound to happen.

 

I can take Jon's point as well, though. The first time I powered up my XEL it failed to boot - I spent over an hour going over all the solder joints (and admittedly did clean up a few that I wasn't happy with) but that didn't allow the board to boot. Ultimately it was because I had shifted the U1MB laterally by 1 pin. As soon as I moved it over, success! Since then I've been super-careful but it's especially difficult now with the shrouded headers on current U1MB. However, the outline on the board silkscreen is very helpful here.

 

But again, on the balance of things, I'll take the inconvenience of not being able to see the headers over the hassle and risk - however small for a very experienced tech - of removing and replacing header pins through such tiny vias on modern PCBs and that awful lead-free solder.

[flashjazzcat]

2 hours ago, mytek said:

Jon I think you are missing the point, or perhaps you are just jesting with us

I'm suggesting shrouded headers completely prevent the possibility of misalignment on one axis, and that open headers allow misalignment on both axes. Since the discussion concerns accidental misalignment, I'm saying the shrouded header is arguably safer in one sense. I understand that your argument is that open headers ease correct insertion by eye. But it's not so hard to align the shrouded header by eye, and it's only possible to misalign the shrouded header on one axis. If I am wrong (if the open headers cannot be misaligned by one row), please correct me, but if I am right, I see no reason for argument.

[+mytek]

You'll get no further argument from me .

 

[Bikerbob]

ok.. another boo-boo

 

The header on the U1MB  that would solder to the PBI on the computer..   P4 S0.. does it need to be connected to the 1088xel???  I cut the header to an 8pin block not a 10pin block. 

 

Thanks

 

James

Edited August 1, 2019 by Bikerbob

[+mytek]

Yes all of the signals provided on the U1MB small pin header P4 will need to mate up with the XEL via a DIL 5x5 tall female header. This allows the XEL to supply the additional signals required by the U1MB, as well as provide access to the digital I/O pins that control various functions on the XEL side (Stereo, Vgate, 2nd IRQ, and one accessory).

 

So when you say you cut the header to an 8 pin block instead of a 10 pin block, I assume you are referring to the tall female header on the XEL board. Is it soldered in already? If not, please wait until the proper header is available.

 

[Bikerbob]

Well I ended up with 2 x 4x4 headers after making the other two. So I then took them and made a 2x2 and 3x3 used the dremel to clean them up so they will mate and align.. and have a 5x5 out of them.. more hacking and dremel than I wanted but p4 s0 is now included in the party.

[+MacRorie]

6 hours ago, Bikerbob said:

Well I ended up with 2 x 4x4 headers after making the other two. So I then took them and made a 2x2 and 3x3 used the dremel to clean them up so they will mate and align.. and have a 5x5 out of them.. more hacking and dremel than I wanted but p4 s0 is now included in the party.

If you haven't soldered them in yet, I could send you a 2x5.  AT the very least it would be easier to solder in.

[Bikerbob]

6 hours ago, MacRorie said:

If you haven't soldered them in yet, I could send you a 2x5.  AT the very least it would be easier to solder in.

Thanks.. but I am fine with it.. Now all I have to worry about is if I torched any traces on my original attempt to desolder the one header.. before I saw Jon's video and pulled the pins individually on the 2nd one.

 

James

[+Stephen]

20 minutes ago, Bikerbob said:

Thanks.. but I am fine with it.. Now all I have to worry about is if I torched any traces on my original attempt to desolder the one header.. before I saw Jon's video and pulled the pins individually on the 2nd one.

 

James

If ya make it down to VCFMW this year, I can take a look at the board.  goven a few people I've spoken with, I am pretty sure I'll be bringing my desoldering pump, soldering station, etc.

[Bikerbob]

2 minutes ago, Stephen said:

If ya make it down to VCFMW this year, I can take a look at the board.  goven a few people I've spoken with, I am pretty sure I'll be bringing my desoldering pump, soldering station, etc.

I have all that.. guess I am just not very good at it yet.

 

James

[Bikerbob]

LOL.. sorry guys.. I know all this is two years old.. but hope this is bringing back memories of you first doing the board.. 

 

That being said this is a technique question. I assume on this board with pads on the top and bottom that even for resistors etc.. I should be making sure I flow enough solder to get all the way through??

 

So is there a way to know that I have done this without just flipping the board over and seeing that I have flowed enough through?

 

James

Edited August 2, 2019 by Bikerbob

[_The Doctor__]

depend on your solder and if it's solid or flux core... generally you get used to feeding the solder according to hole size and pcb thickness. The axial lead size in the hole. I normally push about the same amount of solder as the pcb is thick or a little more. so it's not very much. that usually fills the through hole and pads very nicely. Your solder diameter and, as stated solid or flux core will determine the exact amount but what I've said is a good place to start. Once you test out how much a couple of times on an old useless pcb from lets say a junk old alarm clock etc. you will zero in on the amount that generally works for you. After that you will get a sense of how much by eye and feel. Even the best of skilled soldering maniacs occasionally get an underfill or even sometimes an overfill. So sweat the joints but don't sweat yourself, it's all going to be fine.

Edited August 2, 2019 by _The Doctor__

[+DrVenkman]

You should - theoretically - but for plated through holes it’s not “necessary” but it is good technique. 
 

The best way I found to get solder all the way through this thick 4-layer board was to apply a bit more solder than I would on a normal thinner 2-layer board. Basically flow solder around the component for just a second or so longer, and then after removing the solder, keep the iron in the leg a bit longer to allow the solder to stay molten and flow all the way down and through the thickness of the board. 
 

Now having said that, this shitty factory soldering on one my C64’s has kept the thing running for 35 years, lol. 

[Bikerbob]

Ok so on MYTEKs board here we have contact all the way through the hole?... but a trace on the top or bottom of the board would need contact with that pad correct?

 

James

[_The Doctor__]

From both an electrical and mechanical standpoint it is best to fill all the way through from pad to pad. This is true even with the best boards.

[Bikerbob]

OK, so the follow up.. I have installed most of the resistors now.. Some i see the solder at the component side.. most I do not.. so should I reflow some? ... if so.. do I touch the lead on the compent side.. to draw the solder up to the heat? or add more solder on the solder side.

 

James

[+DrVenkman]

15 minutes ago, Bikerbob said:

OK, so the follow up.. I have installed most of the resistors now.. Some i see the solder at the component side.. most I do not.. so should I reflow some? ... if so.. do I touch the lead on the compent side.. to draw the solder up to the heat? or add more solder on the solder side.

 

James

Just heat from the solder side until the solder turns glisten-y and has melted all the way through. You might even notice it “sag” a bit as it flows deeper into the via. Add a bit at that point once it’s good and hot.

 

[+Stephen]

1 hour ago, Bikerbob said:

OK, so the follow up.. I have installed most of the resistors now.. Some i see the solder at the component side.. most I do not.. so should I reflow some? ... if so.. do I touch the lead on the compent side.. to draw the solder up to the heat? or add more solder on the solder side.

 

James

Make sure to use flux on both the top and bottom of the board.  This will also help "draw" the solder all the way through the board.

[flashjazzcat]

Use good quality rosin core 60/40 solder and make sure the iron is hot enough and that the tip is large enough to conduct heat well. If all the prerequisites are satisfied, solder should have no trouble travelling to the other side of the PCB. I'm not sure what kind of solder is in use here, but lead-free alloys are not pleasant to work with in my experience.

 

[Bikerbob]

I am using Silver Solder 60/40 and did not use flux on both sides.. hmm never thought of that. OK.. thanks guys.

 

James