Threaded Mode | Linear Mode

Claudio L. · 09-02-2016, 02:11 AM

(09-02-2016 12:08 AM)JDW Wrote: Really? More economical than opening the 50g and attempting a hardware repair, which is technically an "engineering" solution? :-)

Clearly not.

So I will bite the bullet and yank off the LCD screen cover, unscrew the 50g, and attempt the repair, which will cost me nothing and perhaps take less than 1 hour in total.

Attempting a repair is one way to go, but it only becomes a solution if you can successfully fix it, so I guess we'll have to wait and see. If you fix it then I'll stand corrected, as that's the most economical solution. You didn't seem so confident based on previous posts.
Please make a post explaining what you see and how you did it, and take pictures of every step for the rest of us.

JDW · (This post was last modified: 09-06-2016 10:28 PM by JDW.)

Here is "part 1" of a 1080p 60fps video I made of OPENING and REPAIRING the 50g:

Sept.7 Update: I replaced the above YouTube video with the full version showing the jumper wire repair.

There is no Part II yet because as you see at the end of the video, I really need either (a) a schematic of the calculator or (b) a diagram of that exact same SD Card Slot Mechanism in order to complete repair. If a kind soul among you can provide me with such information (PM me if you like), I will complete attempt to complete the repair in a second video, then merge the two and upload the unified video on YouTube. That is why I left the current video unlisted for now, since I may delete it if I ever make a unified video showing that last part on how to fix it.

Be sure to watch the video in full screen mode, especially on a larger screen device, to see the PCB in its full glory, seeing more clearly the points I desoldered and tested.

Once again, the repair objective is to fix whatever problem exist in the 50g that is making it falsely believe SD cards are Locked (when in fact they are not). Whatever kind assistance any of you could provide would be greatly appreciated.

Thank you.

UPDATE: I've been spending the last couple hours viewing diagram after diagram on SD Card Sockets (connectors). There is not standard for the CD (Card Detect) and WP (Write Protect) pin-outs from what I can see. But most tend to put the WP output on the far edge of the card, which if you watch my video above would appear to be the line connected to resistor R66. But as you see in my video, my test seems to indicate that the R66 line is actually CD. The reason I think that is because when I removed R66 and tested, you can see in my video that when I push the FORMA button nothing happens and when I refresh the card (Slot 3) vanishes in the File Manager. However, when I subsequently removed resistor R68, with R66 attached, the card was detected (although I still got a write protect error).

I really need a datasheets on that exact SD Card Socket to know for sure, but as you can see in my video, it certainly seems that the WP output is the PCB trace attached to resistor R68.

The thing is, I have been reading a lot of SD Card Socket datasheets and all of them say the same thing regarding how WP is output. In every case, all the datasheets say that the WP output is Grounded when the card is NOT protected (not Locked). And they say the WP output is Open Circuit when the card is protected.

But if you watch my video, you will see that I used a Fluke meter set to Continuity Check and then touched one probe to the pad associated with R68 and then touched the other probe to the metal frame of the SD card socket. I must assume that the metal frame is always grounded. If so, and if R68 truly is WP, then such would seem to show that WP is being grounded properly. And if "WP OFF" truly is defined by "WP = GND" then why would my 50g complain about the card being write protected all the time?

Also note in my video that even after I remove resistor R68 (which again, I assume is the WP line, although I could be wrong), the card was detected by the 50g but I got the same Write Protect error as always. Any thoughts on this?

Again, if anyone can direct me to a schematic on the exact SD Card Socket used in the 50g, and if any beautiful soul out there has a full schematic on the 50g, please let me know. Regardless, I would appreciate hearing your thoughts. Thanks.

Vtile · (This post was last modified: 09-03-2016 09:26 AM by Vtile.)

You did check if the resistors aren't foult? Especially the last one you took? I personally would try to trace the route to IC from the last resistor R68? you desoldered to see if there is foulty connections / bad resistors.

Also are you positively sure you actually were testing WP as write protect detection and not CD card detection? As from NXP aplication note (you have propably already read it) they say as follows "Additionally card holder can provide mechanical switches for card detection (CD) and write protection (WP) mechanisms. "

That said, I'm disappointed I always thought that the switch is actually mechanical device to cut the trace inside the card. *Meh* it is only a wimpy software lock.

Just shooting blind, I even have not opened my device.

matthiaspaul · 09-03-2016, 10:24 AM

(09-02-2016 11:27 PM)JDW Wrote: But if you watch my video, you will see that I used a Fluke meter set to Continuity Check and then touched one probe to the pad associated with R68 and then touched the other probe to the metal frame of the SD card socket. I must assume that the metal frame is always grounded.

The metal frame of SD card readers is grounded in some cases, but definetely not in all cases. It all depends on the specific SD card reader.

When not connected to GND, it may be floating, connected to the card's supply pin or connected to one of the switches (in this case, the frame is used to save the space necessary to route that switch signal to a pad located elsewhere).

Greetings,

Matthias

Didier Lachieze · 09-03-2016, 11:15 AM

(09-02-2016 11:27 PM)JDW Wrote: I really need a datasheets on that exact SD Card Socket to know for sure, but as you can see in my video, it certainly seems that the WP output is the PCB trace attached to resistor R68.

Have you checked if the state of the PCB trace connected to R68 change when you change the WP lock switch on the SD card ? Grounded if unlock, not grounded if locked ?

matthiaspaul · 09-03-2016, 11:39 AM

(09-03-2016 08:59 AM)Vtile Wrote: You did check if the resistors aren't foult? Especially the last one you took? I personally would try to trace the route to IC from the last resistor R68? you desoldered

Tracing the WP signal to the controller would also help Claudio to implement the WP feature in newRPL.

Quote:That said, I'm disappointed [...] it is only a wimpy software lock.

I agree.

Greetings,

Matthias

Vtile · (This post was last modified: 09-03-2016 05:49 PM by Vtile.)

I did find an easy method to open the calculator... No need to remove the screen shield.

My trusty veroboard strip cutting drill bit is ultimate hacktool. Tongue

PS. The mechanism is really interesting.. It seems that it is "single pole double throw" switch or how one should describe it.. where is my multimeter and magnifying glass. R67 I assume is the WP line ??

Edit. Yes, I would look at R67. my quick proping indicates that it is in not connected with the contactor springs at the bottom part of the connector near the R68 when the card is WP. While those springs are really painfully breathles to probe as they are so sensitive to move around. The mechanism in my eyes do CD task in the bottom part and the WP part at the top part (bottom is the R68 end and TOP is where you insert the card and where the R67 is.)

Vtile · (This post was last modified: 09-03-2016 09:11 PM by Vtile.)

I have interesting evening probing the 50g and using my toys.

Here is the connections between testpoints between R66, R67 and R68. Since this forum doesn't allow tables to be used in BBCode I post it as CSV to open in spreadsheet program. The measurement should be valid and my unit is working properly.

R67 is connected directly to ground and I did measure it as 17k resistor, but incircuit with normal ohms range so it might not be accurate. The continuity should be accurate, since that old measuring device is using only 25mV test voltage, so it doesn't turn silicon nor germanium devices to conducting state.

Quote:HP50g SD Card CD/WP;;;;;;
;;;;;;
No Card;False;F;F;F;F;F
Card+Lock;False;F;F;F;F;T
Card+NoLock;False;F;F;T;T;T
;TPG to R66;TPG to R67;TPG to R68;R66&R67;R67-R68;R66-R68
;;;;;;
;False = NO connection;;;;;
;True = Connection;;;;;
TPG = TPGND found near battery wiring point
;;;;;;
Measurement done without batteries (button cell included);;;;;;
Measuring device Meratronic Meratester/Conway model 639, FET Multitester ;;;;;;
;Continuity / Ohms range, test voltage 25mV / 30uW;;;;;

There is also really interesting text in that PCB seen also in your video at 17:50..

Hope this get you further.

JDW · 09-03-2016, 10:08 PM

Gentlemen, thank you for your replies.

My 50g is at the office, where I have my tools, so I cannot touch it until Monday, Japan time. For now, I would like to reply to Vtile regarding the "CSV" data. I copied that text and pasted it in TextWrangler, then saved it as a *.csv file, but when I open that file in Excel, it appears simply as imported text that looks exactly the same as the mess we see in your post here. So could you please open Excel on your computer, then take a screen shot, then post that screen shot? I ask this because I cannot make sense of the data you presented, nor can I tell if in your test you discovered that R67 is truly the WP line or not. (I did not test R67 when I made my video.)

You mention 17:50 in my video, but that is the point at which I showed my continuity check between R68 and the SD card metal frame. Please explain what specific "interesting text on the PCB" that you were talking about.

As to whether the resistors are faulty, I need to of course verify that on Monday, but in my years of experience, I would say "it's unlikely." I see no burn marks nor smell any evidence of component burning, which is usually what causes resistors to fail. (Another common reason for non-burned resistors to fail are solder balls or other soldering defects at the time of manufacture, which are not present in my calculator from what I can see.) Moreover, those resistors are there as simple protection mechanisms since the SD card interfaces with the outside world. You want to use resistors like that to prevent a dead short which could otherwise harm the MCU.

So, Vtile, if you answer the questions above, I will be well equipped to perform further checks on my 50g this Monday, at which time I will report back here.

Vtile · 09-03-2016, 10:33 PM

(09-03-2016 10:08 PM)JDW Wrote: Gentlemen, thank you for your replies.

My 50g is at the office, where I have my tools, so I cannot touch it until Monday, Japan time. For now, I would like to reply to Vtile regarding the "CSV" data. I copied that text and pasted it in TextWrangler, then saved it as a *.csv file, but when I open that file in Excel, it appears simply as imported text that looks exactly the same as the mess we see in your post here. So could you please open Excel on your computer, then take a screen shot, then post that screen shot? I ask this because I cannot make sense of the data you presented, nor can I tell if in your test you discovered that R67 is truly the WP line or not. (I did not test R67 when I made my video.)

Here are my findings as an image.

I did though that you might need some reference for your measurements. I hope I didn't make any fools mistake on this. The R67 did to me seem to be somesort of pulldown resistor, since the trace branches to throught the board connecting hole and I have not taken the board off. I'll leave the further measurement to you.

(09-03-2016 10:08 PM)JDW Wrote: You mention 17:50 in my video, but that is the point at which I showed my continuity check between R68 and the SD card metal frame. Please explain what specific "interesting text on the PCB" that you were talking about.

That didn't actually relate to the current problem at all. The board have markings that indicates that same PCB is used for 50g, 39g and 40g from "current" production.

(09-03-2016 10:08 PM)JDW Wrote: As to whether the resistors are faulty, I need to of course verify that on Monday, but in my years of experience, I would say "it's unlikely." I see no burn marks nor smell any evidence of component burning, which is usually what causes resistors to fail. (Another common reason for non-burned resistors to fail are solder balls or other soldering defects at the time of manufacture, which are not present in my calculator from what I can see.) Moreover, those resistors are there as simple protection mechanisms since the SD card interfaces with the outside world. You want to use resistors like that to prevent a dead short which could otherwise harm the MCU.

So, Vtile, if you answer the questions above, I will be well equipped to perform further checks on my 50g this Monday, at which time I will report back here.

Hopefully you find the reason for the problem.

JDW · 09-03-2016, 10:48 PM

Thank you for the screen shot.

And just to clarify the screenshot...

TPG = Test Point GROUND

False = (no continuity, no connection)

True = Connection

So from examining your screen shot data, it appears that you found R67 to possibly be the WP line. I will therefore focus on that when I test tomorrow.

Thank you.

Vtile · 09-03-2016, 10:54 PM

Those are correct.

JDW · (This post was last modified: 09-05-2016 01:33 AM by JDW.)

It's seems I lost R68 sometime after I removed it a few days ago. And since my video isn't quite high enough resolution to see the silkscreen on it, and since I don't have a schematic, I don't remember what value it is. It might be 104 (100kΩ) from what I see in my video, but I can't be sure. Vtile, if your 50g is still open, can you check the value for me, please?

Anyway, I took a high-rez 50g PCB photo just now (with R68 removed).

I put my Fluke meter back into Continuity Check mode (which beeps when both probes are connected to each other) and put one probe on the round pad near R67 and put the other probe on the SD cart slot itself (metal) and got a beep. So putting the probes between that R67 pad and the R68 pad beeps regardless of whether the card is Locked or Unlocked on my 50g. That is different from what Vtile tested on his 50g. I'm pretty sure that R67 pad is ground though because it connects to the foot (mount) of the SD card slot itself, nearest to where you insert the SD card. You can see all of that in my video, especially if you view it at full 1080p fullscreen. And yes, I probed both sides of R67 and my meter beeps both times when I touch the other probe to the SD card slot metal body.

I then probed the opposite side of R68, the side of that resistor which is furthest from the SD card slot, and found my Fluke meter beeped when I touched SD card pin-2, which is CMD (a data line).

After that I probed R66 (marked 102, which is a 1kΩ resistor), the side of the resistor facing opposite the SD card slot and traced it to R63, the side of R63 facing opposite of the SD card slot.

Still not sure what to think.

I wish a had a 50g schematic.

Vtile · (This post was last modified: 09-05-2016 08:51 AM by Vtile.)

There is difference if you probe the system batteries in or without. If you probe it batteries in it will give different results with different meters, depending on the internal measuring setup of the meter. That is what have tought to me and that is what I noticed. The truth table I gave you is done without any own current flowing inside of the DUT.

In my unit the card case is not connected to the Testpads found on the R66 .. R68 while measuring the device without batteries and while my meter is old it propably beats average handheld Flukes pretty much in any respect input impedance included.

The resistors are following.
R66 - 102
R67,68 - 104

JDW · 09-05-2016, 09:28 AM

Vtile,

Thank you for the resistor size confirmation.

I removed all batteries, including the button cell, and then I waited a few minutes for the capacitor to fully discharge. I then probed the PCB again.

When I put one probe on Ground (where the black wires solder to the PCB, coming from the batteries), and put the other probe on the SD card slot metal body, I get a beep, confirming that the metal frame of the SD card slot is definitely Grounded. So without any batteries, that point is absolutely confirmed.

R67<->SC Card Slot Metal

With the SD card REMOVED, I put a probe on the round pad on the R67 trace, and put the other probe on the SD card slot metal, and I do NOT get a beep. (I also probed between Ground and the R67 pad, and no beep either -- no surprises there since the SD card body is grounded.)

With the SD card INSERTED and UNLOCKED, I put a probe on the round pad on the R67 trace, and put the other probe on the SD card slot metal, and I do NOT get a beep.

With the SD card INSERTED and LOCKED, I put a probe on the round pad on the R67 trace, and put the other probe on the SD card slot metal, and I do NOT get a beep.

R66<->R68:

With the SD card REMOVED, probing between the R66 pad and R68 pad does NOT give me a beep.

With the SD card INSERTED and UNLOCKED, probing between the R66 pad and R68 pad gives me a beep.

With the SD card INSERTED and LOCKED, probing between the R66 pad and R68 pad gives me a beep.

R66<->R67:

With the SD card REMOVED, probing between the R66 pad and R67 pad does NOT give me a beep.

With the SD card INSERTED and UNLOCKED, probing between the R66 pad and R67 pad does NOT give me a beep.

With the SD card INSERTED and LOCKED, probing between the R66 pad and R67 pad does NOT give me a beep.

R68<->R67:

With the SD card REMOVED, probing between the R68 pad and R67 pad does NOT give me a beep.

With the SD card INSERTED and UNLOCKED, probing between the R68 pad and R67 pad does NOT give me a beep.

With the SD card INSERTED and LOCKED, probing between the R68 pad and R67 pad does NOT give me a beep.

NOTE: Beep = True = Connection, No Beep = False = No Connection

Thoughts?

Vtile · (This post was last modified: 09-05-2016 04:26 PM by Vtile.)

I have not much time atm. to delve into your finds, but as a quick scan of data it seems that if connecting testpads R66 and R67 should fix electrical HW problems of the WP issue.

I read your post, now with a tiny bit more time and tough. My first intuition seems to be correct. For me ir seems like ( said above) you can permanently disable the writelock, by jumpwiring the pads R66 & R67 to fool the calculator hardware that the card is inserted, but not locked. This is the easy fix if it works, but only partial solution since it disables write protection permanently.

Claudio L. · 09-05-2016, 01:54 PM

(09-05-2016 09:28 AM)JDW Wrote: Thoughts?

The SD card socket in the 50g looks like an older version of this Tyco component:

http://www.digikey.com/product-detail/en...ND/4731771

While this one has the detect pads on the side, perhaps the company kept the same pattern:

* Card detect pin
* Common wire
* Write protect

Look at their datasheet, there's a schematic of the electrical connection between these 3 pins. It's a long shot, but since it looks so similar mechanically, it's likely connected the same way just the pins were moved to the sides.

I hope it helps.

Claudio L. · 09-05-2016, 08:15 PM

(09-05-2016 01:25 PM)Vtile Wrote: I have not much time atm. to delve into your finds, but as a quick scan of data it seems that if connecting testpads R66 and R67 should fix electrical HW problems of the WP issue.

I read your post, now with a tiny bit more time and tough. My first intuition seems to be correct. For me ir seems like ( said above) you can permanently disable the writelock, by jumpwiring the pads R66 & R67 to fool the calculator hardware that the card is inserted, but not locked. This is the easy fix if it works, but only partial solution since it disables write protection permanently.

Based on your tests and the datasheet I found, it's clear that either R66 or R68 is the common wire, the other one is card detect. That leaves R67 for WP.
Write protect goes active when it's unlocked per your probing, and because when the card gets inserted R66 is always in contact with R68, doesn't matter whether you connect R67 to R66 or R68, either should get the job done.

Vtile · 09-05-2016, 08:22 PM

(09-05-2016 08:15 PM)Claudio L. Wrote:
(09-05-2016 01:25 PM)Vtile Wrote: I have not much time atm. to delve into your finds, but as a quick scan of data it seems that if connecting testpads R66 and R67 should fix electrical HW problems of the WP issue.

I read your post, now with a tiny bit more time and tough. My first intuition seems to be correct. For me ir seems like ( said above) you can permanently disable the writelock, by jumpwiring the pads R66 & R67 to fool the calculator hardware that the card is inserted, but not locked. This is the easy fix if it works, but only partial solution since it disables write protection permanently.

Based on your tests and the datasheet I found, it's clear that either R66 or R68 is the common wire, the other one is card detect. That leaves R67 for WP.
Write protect goes active when it's unlocked per your probing, and because when the card gets inserted R66 is always in contact with R68, doesn't matter whether you connect R67 to R66 or R68, either should get the job done.

Oh, nice that you can confirm it from another route. R66 is better it saves copper!

JDW · 09-06-2016, 12:18 AM

Vtile and Claudio,

Thank you. I connected R67 to R66 and now the SD Write Protect problem is solved. And thanks to Vtile's R68 confirmation, I was able to replace that resistor with ease.

Here is a high resolution 50g PCB Photo showing SD Card Unlock Wire Fix.

I don't mind that the 50g now thinks all SD cards are unlocked. Honestly, I never would put a locked SD card in the 50g anyway.

Lastly, I know some of you spoke about "tracing to the GPIO pin" but as you can see the SAMSUNG chip is potted all around, making that job rather difficult. Since this calculator is now defunct anyway, and since HP has released other documents to the community in the past, it would certainly would be nice of them and appreciated if they would one day release the full 50g schematic for us, which obviously would be used for repair purposes, or for hobbyist use such as Claudio's newRPL. I'll send Tim Wessman a PM about it and see where it leads.

Thanks again.