Ide card build issue.

[RickyDean]

Since this is unrelated to the ide cards I got from Shift838 and I can't quite find a similar topic, here it is.

A few years ago I bought two ide cards from a gentleman, one partially populated with some of the smd chip and so forth. Now I being unfamiliar at the time with reflow workstations and such, tried to YouTube and do some toaster oven work with solder paste. It seemed to work though there was some overage of paste in some areas, but looked fairly good. But the Benchmark bq4847ymt's that were provided bubbled at the top some(yes I know , why didn't I remove them first). I bought some from a China source on ebay and I have never gotten either card to work. I get a light on the front when inputting my cru, but when I go to card test and it ask if I see the back light lit. I have to say no, then I get the error code in the picture. When I try a sram test, I get 40000 errors. So do I have bad sram chips, bad bq4847ymt chips, or where should I start. I do have new sram chips on order from digikey. The ones Shift838 uses on his Geneve 384 board and maybe his built Ide cards.

Edited May 8, 2021 by RickyDean
spelling, 1am in the morning.

[F.G. Kaal]

There is a idetestman.txt file in the download with clues.

idetestman.txt

 

Where to start: There is no use in doing further tests if a previous test fails. First fix that.

[+FarmerPotato]

Have you inspected the SMD pins with a magnifying glass? I hear a 10x handheld loupe is the handiest and beats a USB microscope overall. You’re looking for solder bridges. 
 

if you have a logic probe ($20) a quick test for SRAM is to check the CE, WE, and OE pins. The logic probe should indicate pulsing on these pins during the ram test. 

but like Fred said you should figure out why the LED is not getting what it needs. Maybe it’s on backwards, maybe there is a short somewhere. you’d have to trace it backward to its driver. 
 

Edited May 9, 2021 by FarmerPotato
Short, not shirt

[RickyDean]

1 hour ago, FarmerPotato said:

Have you inspected the SMD pins with a magnifying glass? I hear a 10x handheld loupe is the handiest and beats a USB microscope overall. You’re looking for solder bridges. 
 

if you have a logic probe ($20) a quick test for SRAM is to check the CE, WE, and OE pins. The logic probe should indicate pulsing on these pins during the ram test. 

but like Fred said you should figure out why the LED is not getting what it needs. Maybe it’s on backwards, maybe there is a shirt somewhere. you’d have to trace it backward to its driver. 
 

No I used the microscope, it is a new addition and I can see very clearly. No bridges, that I can see. And,I did use flux, desoldering wick and a slightly square nosed iron to suck up any excess solder, tricks I've recently learned for smd. I looked with the new microscope very carefully, but when I first did the job 2 years ago, I did check the pins for bridges. The bq4847ymt's were already soldered to the board, but the previous owner, and I left them thinking it wouldn't take long to get the solder paste hot enough to melt and set the chips, but I think I may have left it a little too long and cooked the sram and bq4847ymt. Never had any real instructions to assemble and used Fred's picture to orient the chips correctly, so I don't know if I'm supposed to do something if I use the  bq4847ymt. Not sure what the picture is really telling me. With the sram and the bq4847ymt I do have a cap at c23. Is that supposed to not be there?

Edited May 8, 2021 by RickyDean
add content.

[RickyDean]

Well, I removed the cap and my sram errors went down from 40000 to 30000, per the picture. For the moment I am assuming bad sram, then I will go from there. The other board had a couple pads torn off, by me, and I may remove that sram and place it on this board to see if it's doing the same thing.

[F.G. Kaal]

7 hours ago, RickyDean said:

Well, I removed the cap and my sram errors went down from 40000 to 30000, per the picture. For the moment I am assuming bad sram, then I will go from there. The other board had a couple pads torn off, by me, and I may remove that sram and place it on this board to see if it's doing the same thing.

Ther LED at the back of the IDE card:

 

3) Switch the IDE-card back led on.
   This is done by continiously reading address >404E (IDE status register).
   If this test fails check the orientation of the LED at the back, U8, U9
   and all the other gates used.
 

 

 

Again you are allready in the total SRAM test. What is the result of the simple read write test of the IDE test program:

 

4) Simple read/write and read only test
   Writes a 16 bit pattern at address >4000 and >5000 and tries to read
   it back. When this fails, there is something wrong with the SRAM
   selection or SRAM is always read/only. Check circuitry for CRU bit 0 & 5
   on IC16, and IC8, IC10, IC15, IC23 and all the gates used.
   
   Next sets CRU bit 5 so that SRAMS becomes read only and tries to write
   to >4000 and >5000.

 

(I suppose that IC16 etc means U16 etc here).

 

It can be the SRAM or any other chip in the circuitry to select it.

 

Fred

[RickyDean]

@F.G. KaalWill perform these steps outlined hopefully this night, but due to Mothers Day here in America, it may be tomorrow night. Thanks for your advise.

[Flottmann1]

32 minutes ago, RickyDean said:

but due to Mothers Day here in America

I have to put my IDE card from Shift838 into operation

 

 

ps. here in Germany it is also Mother's Day   ??

Edited May 9, 2021 by Flottmann1

[RickyDean]

3 hours ago, F.G. Kaal said:

Ther LED at the back of the IDE card:

 

3) Switch the IDE-card back led on.
   This is done by continiously reading address >404E (IDE status register).
   If this test fails check the orientation of the LED at the back, U8, U9
   and all the other gates used.
 

 

 

Again you are allready in the total SRAM test. What is the result of the simple read write test of the IDE test program:

 

4) Simple read/write and read only test
   Writes a 16 bit pattern at address >4000 and >5000 and tries to read
   it back. When this fails, there is something wrong with the SRAM
   selection or SRAM is always read/only. Check circuitry for CRU bit 0 & 5
   on IC16, and IC8, IC10, IC15, IC23 and all the gates used.
   
   Next sets CRU bit 5 so that SRAMS becomes read only and tries to write
   to >4000 and >5000.

 

(I suppose that IC16 etc means U16 etc here).

 

It can be the SRAM or any other chip in the circuitry to select it.

 

Fred

Also, Fred I'm using idediag, so I will look on your site and see if I can find this idetest. Didn't realize, that there was another program. Had these cards for about 5 years, but just did the soldering attempt, that I described two years ago, with no success, so put them on the back burner. Didn't keep up with the programs that were created for it. Just downloaded. Now that I have more experience and my reflow workstation, I can get'em going.

[F.G. Kaal]

2 hours ago, RickyDean said:

Also, Fred I'm using idediag, so I will look on your site and see if I can find this idetest. Didn't realize, that there was another program. Had these cards for about 5 years, but just did the soldering attempt, that I described two years ago, with no success, so put them on the back burner. Didn't keep up with the programs that were created for it. Just downloaded. Now that I have more experience and my reflow workstation, I can get'em going.

THe IDETEST program and manual is in the same ZIP download as the IDE DSR!

[RickyDean]

43 minutes ago, F.G. Kaal said:

THe IDETEST program and manual is in the same ZIP download as the IDE DSR!

Yes got it, never updated anything as the card has never worked. But after a careful examination, I found the 10k resistor next to the heat sink for the power, is missing. Looks like it's there in the pics but it's gone, this will probably cause the issue, right?

[RickyDean]

I replaced the 10k resistor, still not lighting the back light up. Then I ran bodge wires to the two pulled pad pins on the 512k memory, and 3 previous pad damaged pins on tge ls138 on the other board mentioned earlier. Then replaced the back led. But it doesn't even show the front light when running idetest. Says to check the rotary switch and cru circuitry. So I need to go to Thierry's site and read up what is the cru circuitry here, on this one.

I am using my microscope and a set of China made glasses, that are basically two jewelers loops together. I can see close up, no bridges on both boards. Waiting on the new srams to rule those out. Then I will start, as money permits, purchasing the boms to fill the two shift838 boards, I bought, and get extra to replace the smd on these boards, if the simple things don't fix them.

I might be able to use the good parts on one board to get the other rolling, then buy new parts for the bad board.

I asked the question earlier, on the board that the front led lights up on, do I need to remove the c23 capacitor when adding a bq4847ymt, as it seems to indicate on the silk screen on the IC25 through holes. But the board on Fed's site, showing the pin one locations is using a bq4847ymt and had the c23 cap in place. So I guess that is answered. Hope to give good reports soon.

Edited May 10, 2021 by RickyDean
spelling

[Shift838]

15 hours ago, RickyDean said:

I replaced the 10k resistor, still not lighting the back light up. Then I ran bodge wires to the two pulled pad pins on the 512k memory, and 3 previous pad damaged pins on tge ls138 on the other board mentioned earlier. Then replaced the back led. But it doesn't even show the front light when running idetest. Says to check the rotary switch and cru circuitry. So I need to go to Thierry's site and read up what is the cru circuitry here, on this one.

I am using my microscope and a set of China made glasses, that are basically two jewelers loops together. I can see close up, no bridges on both boards. Waiting on the new srams to rule those out. Then I will start, as money permits, purchasing the boms to fill the two shift838 boards, I bought, and get extra to replace the smd on these boards, if the simple things don't fix them.

I might be able to use the good parts on one board to get the other rolling, then buy new parts for the bad board.

I asked the question earlier, on the board that the front led lights up on, do I need to remove the c23 capacitor when adding a bq4847ymt, as it seems to indicate on the silk screen on the IC25 through holes. But the board on Fed's site, showing the pin one locations is using a bq4847ymt and had the c23 cap in place. So I guess that is answered. Hope to give good reports soon.

I can't tell from the photo of the card but ensure your dip switch has all 4 position open (off).  Also if your photo originally had the 10k resistor and now it does not, chances are the PEB cage hit it and ripped it off.  I always will put a piece of electrical tape over this resistor as it's just too close to the cage.

 

One thing to note.  the clock chips control writing and accessing of the SRAM.  If your battery is dead in the clock chip you will see errors when testing.

 

i've gone through a Frankenstein battery replacement on these HERE

 

[F.G. Kaal]

16 hours ago, RickyDean said:

I replaced the 10k resistor, still not lighting the back light up. Then I ran bodge wires to the two pulled pad pins on the 512k memory, and 3 previous pad damaged pins on tge ls138 on the other board mentioned earlier. Then replaced the back led. But it doesn't even show the front light when running idetest. Says to check the rotary switch and cru circuitry. So I need to go to Thierry's site and read up what is the cru circuitry here, on this one.

I am using my microscope and a set of China made glasses, that are basically two jewelers loops together. I can see close up, no bridges on both boards. Waiting on the new srams to rule those out. Then I will start, as money permits, purchasing the boms to fill the two shift838 boards, I bought, and get extra to replace the smd on these boards, if the simple things don't fix them.

I might be able to use the good parts on one board to get the other rolling, then buy new parts for the bad board.

I asked the question earlier, on the board that the front led lights up on, do I need to remove the c23 capacitor when adding a bq4847ymt, as it seems to indicate on the silk screen on the IC25 through holes. But the board on Fed's site, showing the pin one locations is using a bq4847ymt and had the c23 cap in place. So I guess that is answered. Hope to give good reports soon.

You don't need C23 because the SRAM is powered by the battery of the clock chip.

Drawings (no ASCII art) of the circuitry are available on my website.

And also what Chris said ... if battery is dead then no SRAM.

 

https://www.ti99-geek.nl/Projects/idecard/idecard.html

https://www.ti99-geek.nl/Projects/idecard/Files/ide_sch1.gif

https://www.ti99-geek.nl/Projects/idecard/Files/ide_sch2.gif

https://www.ti99-geek.nl/Projects/idecard/Files/ide_sch3.gif

 

Edited May 10, 2021 by F.G. Kaal

[RickyDean]

Okay @F.G. Kaal
I was wondering then, did you pull the c23 off the example picture on your website, that showed all the pin one location for the ic's. It does show it as having the same clock chip and the c23 is installed too. Does it matter? @Shift838, yes it could be dead, or the Chinese versions are no good, but I have done the Frankenstein process on dead Dallas clock chips, so do understand the principal. Thanks for both of you and your replies.

[RickyDean]

Well, I pulled half the top off one of the Texas marked bq847ymt's. I have the whole battery down to the sides of the battery exposed, when I test the voltage, I'm getting a reading of 3.080 volts, so I assume that it is good, and would power the sram. But I am going to remove the battery tomorrow and place in a new in a holder, for quick changes later. If I understand the data sheet correctly the bq847ymt designed to power the sram, if the power falls below a certain level. Is the ide card not designed to work this way, maybe I'm misreading the datasheet.

Anyway, if it's anything like some Dallas chips, it was added and sealed on top of an older model chip to upgrade that maybe a bq4845? Maybe I'll find out tomorrow, if the China version uses that or another chip.

Edited May 11, 2021 by RickyDean
add content.

[+Schmitzi]

 

Hi, can you make one more pic later, from the "empty" chip inside, means without the battery ?

So that I can see what is the lowest level to saw away the "cap",i.e. with a Dremel or so, from the outside.

 

 

This is the line that I am looking for:

 

 

 

[RickyDean]

6 hours ago, Schmitzi said:

 

Hi, can you make one more pic later, from the "empty" chip inside, means without the battery ?

So that I can see what is the lowest level to saw away the "cap",i.e. with a Dremel or so, from the outside.

 

 

This is the line that I am looking for:

 

 

 

Will do @Schmitzi.

[RickyDean]

I'm also going to investigate the possibility of using a Dallas DS1251Y-120 rtc on the card. I have about 30 of those in tubes in my chip box. It has onboard 512k memory.

Edited May 11, 2021 by RickyDean
spelling, stupid auto correct.

[F.G. Kaal]

21 hours ago, RickyDean said:

Okay @F.G. Kaal
I was wondering then, did you pull the c23 off the example picture on your website, that showed all the pin one location for the ic's. It does show it as having the same clock chip and the c23 is installed too. Does it matter? @Shift838, yes it could be dead, or the Chinese versions are no good, but I have done the Frankenstein process on dead Dallas clock chips, so do understand the principal. Thanks for both of you and your replies.

A missing capacitor or a capacitor to much over the VCC and GND of a TTL chip doesn't matter to much.

[F.G. Kaal]

12 hours ago, RickyDean said:

Well, I pulled half the top off one of the Texas marked bq847ymt's. I have the whole battery down to the sides of the battery exposed, when I test the voltage, I'm getting a reading of 3.080 volts, so I assume that it is good, and would power the sram. But I am going to remove the battery tomorrow and place in a new in a holder, for quick changes later. If I understand the data sheet correctly the bq847ymt designed to power the sram, if the power falls below a certain level. Is the ide card not designed to work this way, maybe I'm misreading the datasheet.

Anyway, if it's anything like some Dallas chips, it was added and sealed on top of an older model chip to upgrade that maybe a bq4845? Maybe I'll find out tomorrow, if the China version uses that or another chip.

If you place the clock chip back on the card then you must be able to measure some voltage on the power pins of the SRAM (between 2.0 and 2.5V or so). If the IDE card is placed in the PEB and the PEB is switched on then you must be able to measure a higher voltage on the SRAM (3V or so).

If it doesn't then there is no correct 5V reaching the clock chip. If the clock chip has no 5V then it keeps the SRAM in a disabled state.

 

[RickyDean]

Okay, will check that tonight, busy welding on a truck frame right now. I have never had any real experience with clock chips other than the aforementioned Dallas battery replacements. Thanks for your help.

[RickyDean]

11 hours ago, Schmitzi said:

 

Hi, can you make one more pic later, from the "empty" chip inside, means without the battery ?

So that I can see what is the lowest level to saw away the "cap",i.e. with a Dremel or so, from the outside.

 

 

This is the line that I am looking for:

 

 

 

@Schmitzi, here is one with the battery lifted up. Keep in mind this is a Chinese chip marked with Texas  Instrument's logo, and it is a rework of a Benchmarq bq4845yp chip. From the bottom of structure of the chip to the top of the internal benchmarq chip is about 6mm, or a quarter inch. If you look at the right exposed corner of the chip right behind the battery, in the second photo, there is a bit of exposed metal. I believe this is the added oscillator. I will next sometime this afternoon, open the top of the bubbled up Benchmarq, to see if it built similarly. Trying to finish my frame welding.

[+Schmitzi]

 

@RickyDean thanks a lot so this is like massive plastic all over...    ?

I will try it with some old chips maybe, to see how it works, but this looks very difficult

 

 

[RickyDean]

1 hour ago, Schmitzi said:

 

@RickyDean thanks a lot so this is like massive plastic all over...    ?

I will try it with some old chips maybe, to see how it works, but this looks very difficult

 

 

Really, more like an epoxy encapsulation, with a plastic cover. Careful grinding with the Dremel, small srewdriver and box knife can reveal a lot.

If you use @Shift838's method above, it's not difficult at all. but I like to get the old battery out of the picture entirely.

I did go ahead and open up the Benchmark bq847ymt and found the battery was mounted negative side up, unlike the other bq847ymt  and I do not see an older chip under the battery which seems to indicate that this was made new as a complete unit. The battery is d.e.a.d. on it as well as the it was puffed up.

Edited May 11, 2021 by RickyDean
add content.