Here we go.... (48G fixing and MEMORY UPGRADE)

[lemontea]

Wish me luck :-)

[lemontea]

Something's still stuck.

I got all the rivets. Counted them. 10 bits dropped out.

The latches should have all been done but it won't open at the bottom.

[lemontea]

Crazy!!!

OK, now what. Suggestions?

[lemontea]

Is this part of the PCB trace the problem? Corroded and broken?

[lemontea]

Nope. Checked continuity with a multimeter. That part is fine.

[mfleming]

Time to remove the CPU board and check the keyboard connections. The CPU board can be removed by turning the six (three each in two rows) twist tabs at the top end of the PCB. You probably already have Han's description of opening the 48 (congrats btw on what looks like a good job of it).

You can do a quick check to see if the ON key comes through on the keyboard connector. Not sure if I've seen a pinout diagram, but I think it is like the Pioneer keyboard connector. There should be two contacts dedicated to the ON key to wake up the CPU and should measure several hundreds of Ohms when key contacts are closed. Check to see if it's working. If not, oh my! I can't suggest anything else that might help.

Hopefully you're dealing with the usual keyboard contact to CPU board problem, the fix for which is documented in several places.

Good luck!
~Mark

[lemontea]

(05-25-2017 02:19 AM)mfleming Wrote:  Time to remove the CPU board and check the keyboard connections. The CPU board can be removed by turning the six (three each in two rows) twist tabs at the top end of the PCB. You probably already have Han's description of opening the 48 (congrats btw on what looks like a good job of it).

You can do a quick check to see if the ON key comes through on the keyboard connector. Not sure if I've seen a pinout diagram, but I think it is like the Pioneer keyboard connector. There should be two contacts dedicated to the ON key to wake up the CPU and should measure several hundreds of Ohms when key contacts are closed. Check to see if it's working. If not, oh my! I can't suggest anything else that might help.

Hopefully you're dealing with the usual keyboard contact to CPU board problem, the fix for which is documented in several places.

Good luck!
~Mark

OK thanks!

I am a bit tired opening the thing (it's nuts! haha!) so will try this later.

While I'm at this, can anyone confirm if this is the 128K RAM chip to use? (If I can't fix this 48G it will be for my other one).

http://www.mouser.com/Semiconductors/Mem...LL&FS=True

UPDATE: I think I found the correct one here

http://www.mouser.com/ProductDetail/Cypr...j1wA%3D%3D

[lemontea]

The pinouts are a bit different. Will it be OK?

The original RAM chip has only 28 pins, so I am guessing the extra 4 pins on the 32 pin chip will be ignored?

I got the new part info from

http://www.finetune.co.jp/~lyuka/interes...hp48g.html

(Note that he has a typing mistake on the Web page for the Part # of the RAM chip. I checked the photograph of the chip itself for the Part #).

But the datasheet for the Cypress says it is NOT suitable for CMOS level, and only TTL level.

The existing RAM chip (on the 48G that I opened) is a Winbond CMOS chip.

UPDATE: The Datasheet on the Winbond says it is TTL compatible even though it is a CMOS chip. So, should be good.

[Paul Berger (Canada)]

(05-25-2017 06:44 AM)lemontea Wrote:  The pinouts are a bit different. Will it be OK?

The original RAM chip has only 28 pins, so I am guessing the extra 4 pins on the 32 pin chip will be ignored?

I got the new part info from

http://www.finetune.co.jp/~lyuka/interes...hp48g.html

(Note that he has a typing mistake on the Web page for the Part # of the RAM chip. I checked the photograph of the chip itself for the Part #).

But the datasheet for the Cypress says it is NOT suitable for CMOS level, and only TTL level.

The existing RAM chip (on the 48G that I opened) is a Winbond CMOS chip.

UPDATE: The Datasheet on the Winbond says it is TTL compatible even though it is a CMOS chip. So, should be good.

The extra 4 pins provide the extra address lines they are all pre-wired on the 48's circuit board, you will note that on the 32 pin 128K chip, the position that was Vcc for 28 pin chip is not an active high chip enable, which makes it very convenient to upgrade, no wiring changes required. Some care ie required when desoldering the old chip. I used a needle to lift away the pins as I heated them one by one.

I think that there are a few documents about upgrading memory in the HP48 file section on hpcalc.org, it seems to me that is where I found information when I upgraded some of them a long time ago, I have one here that I not only upgraded the base RAM to 128K I also added 128K port 1 and a further 512K for ports 2 to 5.

[TomC]

There are some pictures here:

https://sites.google.com/site/tchrapkiewicz2/hp48

In which the 'ON' signal is traced through the main board and up through the keyboard connector. This shows why pressure below the display helps the 'ON' key problem as it provides pressure to the 'ON' keyboard contact connection.

Hope this helps,
TomC

(05-24-2017 10:46 PM)lemontea Wrote:  Wish me luck :-)

[lemontea]

Thanks guys, all great stuff!

[matthiaspaul]

(05-25-2017 11:58 AM)Paul Berger (Canada) Wrote:  Some care ie required when desoldering the old chip. I used a needle to lift away the pins as I heated them one by one.

Since this method carries some risk to rip pads off the PCB, I'd like to suggest a different desoldering method, which can be applied whenever the old chip does not need to be salvaged (for it is cheap to replace or known to be broken anyway):

You can use a scalpel or *sharp* cutter knife and simply cut off the pins next to where they come out of the plastic case (using the chip's case as a guide). Afterwards, the pins can be desoldered from the pads one after another without any risk to apply too much heat or force to the pads. This may sound "brutal" at first, but it is actually quite easy to do and very reliable.

Greetings,

Matthias

[Dave Frederickson]

(05-25-2017 08:55 PM)matthiaspaul Wrote:  You can use a scalpel or *sharp* cutter knife and simply cut off the pins next to where they come out of the plastic case (using the chip's case as a guide). Afterwards, the pins can be desoldered from the pads one after another without any risk to apply too much heat or force to the pads. This may sound "brutal" at first, but it is actually quite easy to do and very reliable.

Take care to not cut through to the PCB and through traces. Otherwise I've also found this to be an acceptable method of desoldering if you don't have the fancy tools.

[lemontea]

Ok before I do any of that, a couple of questions:

1) Would just heating and pressing solder wick against the pins work?

2) What "fancy tools" are we talking about? You mean dedicated desoldering gear? (I actually have a solder sucker gun but wonder if it will "suck" too much in this case).

Again, big thanks guys! This is all a lot of work but somehow very rewarding, like working on old cars or guitars I suppose (never did those; I'm sure some here do!).

[Paul Berger (Canada)]

(05-25-2017 09:22 PM)lemontea Wrote:  Ok before I do any of that, a couple of questions:

1) Would just heating and pressing solder wick against the pins work?

2) What "fancy tools" are we talking about? You mean dedicated desoldering gear? (I actually have a solder sucker gun but wonder if it will "suck" too much in this case).

Again, big thanks guys! This is all a lot of work but somehow very rewarding, like working on old cars or guitars I suppose (never did those; I'm sure some here do!).

1. No not for surface mount there is just a thin layer of solder under the pins I don't think wick would be able to clean that out.

2. In the case of smt parts probably a hot air station with a tip for SOIC packages so you can heat all the pins at once and lift off the chip, some even have a vacuum probe to aid with removing the part. A solder sucker will likely not be much use in removing the chip but if it has a good strong vacuum it can help with cleaning up the pads after, but again care is needed as too much heat may break the bond between the pad and the board and you could suck the pad right off the board.

Paul.

[lemontea]

Ok thanks, excellent points.

Also, I remember seeing a YouTube of someone rigging up a paperclip for heating all the pins at once.

[Dave Frederickson]

(05-25-2017 09:51 PM)lemontea Wrote:  Ok thanks, excellent points.

Also, I remember seeing a YouTube of someone rigging up a paperclip for heating all the pins at once.

The trace amount of solder left after desoldering is sometimes called a "sweat joint". It can be broken by rotating the pin, which works fine for round through-hole leads but not so much for flat SMT leads. Paul is right, the fancy tool would be either a hot air tool with an appropriate nozzle or a U-shaped soldering iron tip for heating up all the pins at once. Again, the tip is sized for the chip.

Ah, yes, making a soldering iron from a paperclip taped to a lighter.

The desire here is to minimize the amount of heat injected into the PCB. In this case, since you're not trying to salvage the chip, the X-Acto method is an acceptable choice.

Dave

[matthiaspaul]

(05-25-2017 09:22 PM)lemontea Wrote:  1) Would just heating and pressing solder wick against the pins work?

While this may work in some cases, I would recommend against it. Sometimes applying significant amounts of new tin before using soldering wick will do the trick, but in general soldering wick is best used to clean the pads once the part has already been removed or to remove short circuits in difficult to access areas.

In general, it is important to keep the heating duration as short as possible and to try not to apply much force to the pads, otherwise they may come off (the longer the soldering process the higher that risk). A rather common mistake is to apply a lifting force to a pin before the tin has become fluid already; this may easily rip the pad off the PCB.

Quote:2) What "fancy tools" are we talking about? You mean dedicated desoldering gear? (I actually have a solder sucker gun but wonder if it will "suck" too much in this case).

Professional desoldering tools include (de)soldering irons with dedicated heads to heat all part pins at once, however, they are expensive and of limited use without a head exactly matching the part to be desoldered. Hotair stations can be very useful as well, but if not used properly they can also cause damage to surrounding parts (they may start to "swim" or get overheated). While it is nice to have such equipment for more difficult problems, I would not use them for easy stuff like having to desolder a chip like this one.

Greetings,

Matthias

[lemontea]

Thanks, will take caution applying the heat.

[lemontea]

Thanks again for everyone who offered their advice. For those who are curious what has since happened, I have some interesting updates on this but I am a bit swamped at the moment. So, stay tuned! I will try to take good photos as I progress.

(Just a recap, the main issue to first fix is that the unit does not turn on. Then the next stage is the RAM upgrade).