NP-41 Emulator (may be)

[Massimo Gnerucci]

(03-16-2016 11:36 AM)Chris Chung Wrote:  Good catch on key legends. There are certain limitations w/ Fritzing, those characters are hi-asciis and I will make the adjustment for the new PCBs.

For the uppercase sigma maybe you could use an uppercase M rotated CCW?

[Chris Chung]

(03-15-2016 08:44 PM)jch Wrote:  1) White PCB, like the NP-41, but I prefer the lettering silkscreen placed on the left of the keys.
2) Battery on the bottom / optionally 4 pins header for external battery ?

Jean-Christophe.

I am actually working on the silkscreens pointed out by Helix, he had reminded me twice and I forgot, so it is wise to do this now. The "sum" sign, I found out, was actually a "M" sideway.

1) The current "lettering" prints are on the North-West corner of the button (upper left). Do we want to put in in line w/ the buttons (i.e. West). If so, I will have to print it w/o the round brackets. I don't really have a preference, if anyone has, please vote.

/EDIT for the NP41, there is no brackets on the letters. I want to make consistent between NP41 and NP41S, please look at the photos and suggest.

2) Power headers already on the side, the others are 2 pins for SBW programming (firmware), another 2 pins for UART. another 2 empty pins on open soldering pad for mods.

[Helix]

(03-16-2016 12:15 PM)Chris Chung Wrote:  I am actually working on the silkscreens pointed out by Helix, he had reminded me twice and I forgot, so it is wise to do this now. The "sum" sign, I found out, was actually a "M" sideway.

Thank you for your efforts.
If I can make some more nitpicking about the silkscreens, the comma seems to be replaced by a dot in the current PCB. Maybe you can change that too.

(03-16-2016 12:15 PM)Chris Chung Wrote:  1) The current "lettering" prints are on the North-West corner of the button (upper left). Do we want to put in in line w/ the buttons (i.e. West). If so, I will have to print it w/o the round brackets. I don't really have a preference, if anyone has, please vote.


/EDIT for the NP41, there is no brackets on the letters. I want to make consistent between NP41 and NP41S, please look at the photos and suggest.

I think that the brackets around lettering prints add too much complexity to the PCB, so they can be removed, like in the NP-41. (I have no preference about the location of the letters.)

[Helix]

(03-14-2016 12:42 PM)Chris Chung Wrote:  The NP41S looks thinner than the NP41, the battery is placed on top side. The LCD module do adds the thickness. I may adopt a non-back-lighted version to make it slimmer in the future. The current LED backlight is very weak and mostly useless when the coin battery is run down.

If the first batches use a back-lighted LCD, I will choose this version. I like the ability to use my NP-25 in low light conditions, and I would like the same feature with the NP-41S. It would be easy to reserve fresh batteries for the NP-41S, and put used batteries in the NP-25, or other devices.

[Sylvain Cote]

Hello Chris,

I have receive the 2016 NPC V1 prototype board yesterday. 8-)

Unfortunately, Canada Post was not very gentle with the package and things moved a lot inside.
The 3D printed case was a little damaged but easily repairable.
At first the board was not working but after replacing the battery everything started working perfectly.

First impressions ...
- I have spent about an 1 hour with the unit at this time
- I see that you have implemented a 41CV emulator.
- Very responsive keyboard, like the NP-25C v2 that I have
- The key labels placement on the PCB take some time to get used to, but after 15 minutes of usage I was fully running.
- Like the NP-25C, ENTER should be activated on both keys
- There is a bug on the power saving because a new battery is depleted within 12 hours (I am at my second new battery that I now remove when I am not testing)
- Is there a way to change the display from 4 stack display to a single line display like the real 41 ?
- When I insert the battery I am going into a some sort of test mode, could you give me more information as to how it work.
- I have not yet tested the launchpad
- Better indication as to how to insert the launchpad connector to the PCB is needed for the less technical people
- The 3D printed case is way too fragile around the launchpad connector and is breaking the case (I have removed the airline plastic bridge)

Keep up the good work!

Best regards,

Sylvain

[Harald]

(04-02-2016 02:53 PM)Sylvain Cote Wrote:  Hello Chris,

I have receive the 2016 NPC V1 prototype board yesterday. 8-)

Unfortunately, Canada Post was not very gentle with the package and things moved a lot inside.

So far I did not receive my NP 41. I recon it is stuck in customs.
Hope Canada Post wsd a bit more gentle with my parcel...

[Chris Chung]

Been busy for the last few days.

I had offered Sylvain a "refurbished" (salvaged LCD from the V0 board and combined w/ the V1 board) unit last week as my LCD modules are not arriving. The fact that both of us are in Canada allowed me to have a tester quick.

I also ordered a V2 PCB design last week (white PCB, battery on underside, better key spacing).

(04-02-2016 06:34 PM)Harald Wrote:  

(04-02-2016 02:53 PM)Sylvain Cote Wrote:  Hello Chris,

I have receive the 2016 NPC V1 prototype board yesterday. 8-)

Unfortunately, Canada Post was not very gentle with the package and things moved a lot inside.

So far I did not receive my NP 41. I recon it is stuck in customs.
Hope Canada Post wsd a bit more gentle with my parcel...

I will try and find better packaging for future shipping.

Harald, not sure why it's taking so long. I would imagine about 10 days for Germany (NP25 experience). May be unlike the NP25 which is in an envelope, the NP41 is in a small box. Worse come to worse I still have parts for more units and if you don't see it in 2 weeks I can assemble any kit to replace it.

[Chris Chung]

(04-02-2016 02:53 PM)Sylvain Cote Wrote:  First impressions ...
- I have spent about an 1 hour with the unit at this time
- I see that you have implemented a 41CV emulator.
- Very responsive keyboard, like the NP-25C v2 that I have
- The key labels placement on the PCB take some time to get used to, but after 15 minutes of usage I was fully running.
- Like the NP-25C, ENTER should be activated on both keys
- There is a bug on the power saving because a new battery is depleted within 12 hours (I am at my second new battery that I now remove when I am not testing)
- Is there a way to change the display from 4 stack display to a single line display like the real 41 ?
- When I insert the battery I am going into a some sort of test mode, could you give me more information as to how it work.
- I have not yet tested the launchpad
- Better indication as to how to insert the launchpad connector to the PCB is needed for the less technical people
- The 3D printed case is way too fragile around the launchpad connector and is breaking the case (I have removed the airline plastic bridge)

Thanks for your quick review.

- Yes it's the CV ROM, but I don't know what is missing in my implementation.
- Next PCB will have the top 4 keys in normal position as we will move battery underside.
- I noted the "Enter" key issue. Should work on both.
- Have not noticed the sleep mode failure, will check on it.
- Single line display / stack display, setup options.

When you insert new battery, you will be at the setup screen. The following keys is used
. '7' turn on beep test.
. '4' turn off beep test.
. '-' and '+', toggle / cycle contrast level of LCD.
. 'CLX' exit setup, enter calculator emulation, may need to press twice.
. other keys will show key-code, also shows battery voltage on screen. (/EDIT)

When you are in calculator mode, and pressed 'ON' to turn unit off / timeout off.
. When you turn 'ON' again w/ the following keys (press together like NP25)
. 'LN' + 'ON', s/w reset, brings you to the setup screen like when you put new battery.
. 'TAN' + 'ON', turn on backlight for LCD, only on battery > 2.8V.
. 'SST' + 'ON', toggle "single line display" / "stack display".
. 'CLX' + 'ON', like HP-41, master clear, "MEMORY LOST".
* these 4 keys are the top right side keys. (/EDIT)

It will take me a least a few days to test a new firmware, as my own unit is showing problems. I was trying to add serial flash rom to it and messed up w/ my unit. There may be some hardware design flaws w/ the internal SPI design. For Sylvain's unit, I had pre-soldered a 1M byte SPI flash but we may not be able to use it because of this fault.

Sylvain's unit already has serial bootloader and ROM upload functions. I will write another reply for instructions.

[Chris Chung]

(04-04-2016 01:05 PM)Chris Chung Wrote:  

(04-02-2016 06:34 PM)Harald Wrote:  So far I did not receive my NP 41. I recon it is stuck in customs.
Hope Canada Post wsd a bit more gentle with my parcel...

I will try and find better packaging for future shipping.

Harald, not sure why it's taking so long. I would imagine about 10 days for Germany (NP25 experience). May be unlike the NP25 which is in an envelope, the NP41 is in a small box. Worse come to worse I still have parts for more units and if you don't see it in 2 weeks I can assemble any kit to replace it.

I found these shipping experiences from an ebay seller forum thread.


......

95% of my shipments go small packet international air so I can definitely shed some light on this for you

The delivery time varies by country, as you can imagine. DO NOT rely on Canadapost's or ebay's estimates

to

UK - 1 to 1.5 weeks (surprisingly fast every time)
france/scandinavia - 2 to 2.5 weeks
spain/austria/portugal - 2-3 weeks
germany - 2-4 weeks (i find german customs can be really slow sometimes)
romania/eastern europe - SLOW, up to 7 weeks (yes..this is for airmail...cant imagine how long surface mail will take)

australia - 2-3 weeks

China - 3-5 weeks
Japan/thailand/taiwan/korea/israel - 2-3 weeks


never shipped to central/south america

.....
It appears to be quite different from the "light packet" I used for the NP-25 where it's about 10-14 days everywhere.

[Harald]

(04-04-2016 01:41 PM)Chris Chung Wrote:  

(04-04-2016 01:05 PM)Chris Chung Wrote:  I will try and find better packaging for future shipping.

Harald, not sure why it's taking so long. I would imagine about 10 days for Germany (NP25 experience). May be unlike the NP25 which is in an envelope, the NP41 is in a small box. Worse come to worse I still have parts for more units and if you don't see it in 2 weeks I can assemble any kit to replace it.

I found these shipping experiences from an ebay seller forum thread.


......

95% of my shipments go small packet international air so I can definitely shed some light on this for you

The delivery time varies by country, as you can imagine. DO NOT rely on Canadapost's or ebay's estimates

to

UK - 1 to 1.5 weeks (surprisingly fast every time)
france/scandinavia - 2 to 2.5 weeks
spain/austria/portugal - 2-3 weeks
germany - 2-4 weeks (i find german customs can be really slow sometimes)
romania/eastern europe - SLOW, up to 7 weeks (yes..this is for airmail...cant imagine how long surface mail will take)

australia - 2-3 weeks

China - 3-5 weeks
Japan/thailand/taiwan/korea/israel - 2-3 weeks


never shipped to central/south america

.....
It appears to be quite different from the "light packet" I used for the NP-25 where it's about 10-14 days everywhere.

Thanks for the offer, I'll keep you update.

Cheers,
Harald

[emece67]

(03-09-2016 06:21 PM)Chris Chung Wrote:  

(03-09-2016 03:35 PM)Harald Wrote:  I think the build will be easy with those instructions.
However, I didn't quite understand how your key scanning works


Do you have a schematic for the keyboard? That might help me understand what you are doing.

Here is a simplified schematic for a 5 pin key multiplexer to illustrate the connection and scanning, it yields 4 + 3 + 2 + 1 keys.



In the firmware I start by placing pin 0 (say from top) as output set to high, pin 1, 2, 3, 4 as input w/ pull-down to low. And I would scan for the 4 input pins to see any of the 1st row keys has been hit.

That's interesting. I've searched a bit and have found another approach, named "charlieplexing" that allows scanning a 42 switch keypad with only 7 pins. It also uses 7 diodes. In general, such approach, when using n pins, can scan up to \(n*(n-1)\) switches, requiring n diodes.

I'm now studying it to see if it can handle (with the aid of software) the simultaneous activation of 2 switches.

[Chris Chung]

(04-06-2016 11:43 AM)emece67 Wrote:  

(03-09-2016 06:21 PM)Chris Chung Wrote:  Here is a simplified schematic for a 5 pin key multiplexer to illustrate the connection and scanning, it yields 4 + 3 + 2 + 1 keys.



In the firmware I start by placing pin 0 (say from top) as output set to high, pin 1, 2, 3, 4 as input w/ pull-down to low. And I would scan for the 4 input pins to see any of the 1st row keys has been hit.

That's interesting. I've searched a bit and have found another approach, named "charlieplexing" that allows scanning a 42 switch keypad with only 7 pins. It also uses 7 diodes. In general, such approach, when using n pins, can scan up to \(n*(n-1)\) switches, requiring n diodes.

I'm now studying it to see if it can handle (with the aid of software) the simultaneous activation of 2 switches.

It's interesting to introduce diodes expand keys. There are also implementations that uses resistor series w/ switches and use ADC to "read" which switch has been depressed based on how it alter a voltage divider via shorting out resistors.
In my case, there are enough I/O lines and I would avoid additional components.

[Chris Chung]

Instructions for load new firmware, NP41S

Mainly for Sylvain. There are a few different ways to do this. I will explain the most simple one.

1. Download mspdebug, an open source tool to flash msp430 devices. Choose download archive and choose version 0.23 (the latest)

2. unzip the package into a folder.

3. download the latest firmware "np41.hex" and place it into the same folder.

4. run a command shell and navigate to the folder.

5. attach launchpad via usb cable, also connect 6 wires to the NP41S unit. I would usually remove the battery and have the NP41S powered via launchpad. Please follow photo, can also checked and pair labels for Reset (r), Test (t), V+ (+), Gnd (G), Tx (T) and Rx (R)

   

6. run the following command

mspdebug rf2500 "prog np41.hex"

   

7. when done, detach unit and use.

[Chris Chung]

About the newer firmware and another evaluation unit.

I had received my 4 lcd modules and I had Etienne's unit built but want to hold it for a few days before shipping it out. Mainly the sleep power consumption issue is not completely resolved and if it is hardware related, it's better to fix it before shipping it out.

I checked the sleep mode power consumption and found that I had intermittent power problems. Usually I would get about 7uA during sleep, so for a 200mAh CR2032, we can sleep for about 1200 days. But sometimes when I put the unit to sleep it ends up at 230-240uA, which is 30+ times the norm.

I suspected it's the LCD module failed to get into sleep mode, as per specification it's active power is about 200uA and when sleep the datasheet said 4uA max.

I altered the code to send "sleep" command sequence twice, plus add some timing to allow the SPI bus ample time to finish communicating. Which improved the odds and I thought I had solved it. Still after the fix I got again some instances that the sleep power is excessive at 200+uA. So more debugging on this.

The main addition in this new firmware is that a bootloader is available so that future firmwares can be loaded via a USB-Serial cable. I will describe to procedure in a separate post. The USB-Serial cable is also used to load ROM files of up to 16K.

[Sylvain Cote]

(04-08-2016 02:39 PM)Chris Chung Wrote:  Mainly for Sylvain. ...

Hello Chris,
Thank you for the additional informations.
I will update the unit this weekend and I will report the result here when its done.
Best regards,
Sylvain

[Chris Chung]

Power Consumption
I had nailed the power leak during sleep mode. It was from the ADC peripheral that I used to detect supply voltage. It was not turned off after use. Now the unit measures 3.6mA on active use and 6.1uA during sleep. This is not ideal as typically the CPU should only consume no more than 1uA during deep sleep. It would be the LCD module which could take up to 4uA during sleep. And for now I won't pursue to lower it again, there is not much room.

Printed Case
I had added support to the design and printed a solid back on the next one, which Etienne will receive. Now it is solid and good to hold on. I will print a replacement for Sylvian in a near future.

BootLoader
The bootloader is ready in the latest firmware. To use the bootloader you need.
1. install a terminal program that can download binaries, I used "realterm".
2. have a usb serial adapter ready, I tried w/ a CP2102 type w/ 3.3V supply ($2).
3. attached Ground, 3.3V, Tx and Rx between unit and usb adaptor. Tx/Rx are crossed joined between usb adaptor and unit.
   
4. attach usb adaptor while press and holding "user" key. This will place the unit into bootloader mode.
5. run realterm and identify serial port (double click on "port" in the port menu tab). Set protocol to 9600, N81. Connect. If already connected, be sure to disconnect and connect again by pressing the "open" button twice.
   
6. send the '@' key, you should get a '?>' prompt to indicate bootloader is ready. any other key will fail.
8. go to the "Send" menu tab, select a new firmware file (supplied by me) np41.bin in the "Dump File to Port" entry. Click "Send File" to start loading firmware. This will take about one minute.
   
9. exit realterm, remove usb adaptor, done.

* for Sylvain's setup, there is a usb-serial adaptor included in the launchpad system via the 6-pin cable, you can query windows "Devices and Printers" to identify and use it w/ realterm.
/EDIT ** the bootloader is minimal / not informative and only response to the '@' command, this is typical as it need to conserve space, do only one thing fast. I had designed the flashing sequence to avoid potential "bricking". The critical section (loss power and it will turn into a brick) is only 10-20 cycles so this should be safe.

.... will continue in next reply....

[Chris Chung]

ROM Loading

There are 20K bytes flash reserved for 4 banks of 4K (10bits) "pages".
The loading hardware and tools is the same as bootloader usage. I.e. binary transfer capable terminal program and a usb-serial adaptor. I had only tried w/ "realterm".
I am loading *.rom raw binary rom files. NOT *.mod.
I had only tried a MATH-1C and a PHYSICS ROM.

1. connect unit to PC (similar to bootloading procedure) but without press-and-hold any key.
2. you would see the normal setup menu on the unit.
3. start realterm and connect to the appropriate serial port, again 9600,N,8,1.
4. press any key to get the menu. the pages 1, 2, 3, 4 in the menu corresponds to page 8, 9, a, b in the memory system (I think).
5. erase the appropriate page(s) by pressing the capital A, B, C or D.
6. load rom to a page by pressing the keys 1, 2, 3, or 4.
   
* 4k ROM files are of 8K byte size, NP41S will compress it into 5K bytes during transfer.
7. remove connection after done.

I am still not very familiar w/ the ROMs. I can load my 2 ROMs and run the functions, also they show up on 'CAT 2'. I do not know the relationship / importance to load to specific pages, etc. So this is still experimental.

I am also not sure if the unit should accept ROM or MOD files. It seems MOD files have additional information that may benefit.

Next

I am preparing the next prototype unit to go to Etienne. The new all white PCB are done and I hope to received them in 20 days.

My next priority is to develop a way to save and load the RAM space, i.e. user program steps and register contents.

There are still odds and ends like the "TONE" command frequency and behaviour, CV ROM time functions and such and the longer term.

[Harald]

(04-08-2016 02:39 PM)Chris Chung Wrote:  Instructions for load new firmware, NP41S

Mainly for Sylvain. There are a few different ways to do this. I will explain the most simple one.

1. Download mspdebug, an open source tool to flash msp430 devices. Choose download archive and choose version 0.23 (the latest)

2. unzip the package into a folder.

3. download the latest firmware "np41.hex" and place it into the same folder.

4. run a command shell and navigate to the folder.

5. attach launchpad via usb cable, also connect 6 wires to the NP41S unit. I would usually remove the battery and have the NP41S powered via launchpad. Please follow photo, can also checked and pair labels for Reset (r), Test (t), V+ (+), Gnd (G), Tx (T) and Rx (R)



6. run the following command

mspdebug rf2500 "prog np41.hex"



7. when done, detach unit and use.

Hi Chris,

I have built my NP41 and was going to flash it using a launchpad. Then I discovered that I only had GND, Vcc, TxD and RxD on the pinheader. So two pins are missing compared to the NP41S. Is the procedure different in that case? Or do I just connect those four lines? Do I have to cross RxD and Txd between launchpad and NP41?

Cheers,
Harald

Edit: And even more questions:

1) Which file do I have to use? I can only find NP41S and NP42 files. Am I correct in assuming it is the NP42 one and it is supposed to be called NP41?

2) I have an MSP-FET430UIF here, but that would appear to be JTAG only - so no use for me

3) Are we using the UART Bootstrap loader? I also have an MSP-EXP430GS launchpad (which appears to be the one you use for flashing the NP42S). Will that work? Or do I have to get the MSP-EXP430FR4133?

4) Will the mspdebug tool work?

Cheers,
Harald

[Chris Chung]

(04-13-2016 02:06 PM)Harald Wrote:  Hi Chris,

I have built my NP41 and was going to flash it using a launchpad. Then I discovered that I only had GND, Vcc, TxD and RxD on the pinheader. So two pins are missing compared to the NP41S. Is the procedure different in that case? Or do I just connect those four lines? Do I have to cross RxD and Txd between launchpad and NP41?

Cheers,
Harald

Edit: And even more questions:

1) Which file do I have to use? I can only find NP41S and NP42 files. Am I correct in assuming it is the NP42 one and it is supposed to be called NP41?

2) I have an MSP-FET430UIF here, but that would appear to be JTAG only - so no use for me

3) Are we using the UART Bootstrap loader? I also have an MSP-EXP430GS launchpad (which appears to be the one you use for flashing the NP42S). Will that work? Or do I have to get the MSP-EXP430FR4133?

4) Will the mspdebug tool work?

Cheers,
Harald

Harald,

The 4 pin connector on the left are marked G, +, R, T. They represent Ground, V+, Reset and Test. Sorry for the confusion, the R and T are not Rx and Tx.

1) my bad, I had upload the latest firmware. NP41.hex is the hex file you should use via a SBW programmer.

2) I have never used the MSP-FET430UIF (expensive), if it supports SBW (spy-bi-wire) then it could be used.

3) no, the stock BSL is difficult to use as it requires extra lines like DTR etc. Common usb-serial adaptors will not work.

4) yes, I am using the same one (mspdebug). I do use it via a launchpad as programmer. I can use the FR4133, FR5969, FR6989 and F5529 launchpads. You just need to redirect the jumpers (remove jumper blocks and use female to female jumper wires) for the dual line header. Here is a photo of my current setup (using the F5529 as it is smaller).

   

Also the new firmware I just uploaded has the uart bootloader (my own creation, not the stock one). You can use the same procedure to try it out file is NP41.bin. The ROM download also works. But everything is kind of not mature (I just finished them this morning). You can see I temporarily attached (fiction fitted) a 4 pin right angle header beside the "ON" key for uart testing.
If I am making new PCBs, I will find a permanent place to put the uart Tx / Rx pins.

[Chris Chung]

Hi Harald,

The flashing command is different from the NP41s, it is because the programming device is different.

mspdebug tilib "prog np41.hex"

If you are using a new launchpad, mspdebug may complain and ask you to re-flash new firmware for the program. You can follow the instruction to re-flash the Launchpad and continue w/ the programming to the unit.