New Project wp34s micro usb flash cable revisited

[MarkHaysHarris777]

Greetings,

All due respects, Harold Pott's micro usb board idea is bass-ackwards. We don't want to mount the chips in the wp34s, we only want to mount the micro usb female plug in the wp34s!

Why mount a usb board in every device in your collection? Why not make one 3.3v usb cable containing the crossover resistors, along with 3.3v power source, which you may plug into any of your 3Xs devices? All you need mount in the wp34s (in the end of the unit, not side) is the female micro usb jack and associated fine bell wiring, and voila! (sorry Walter, no violets for you this time :-p )

So, the ultimate simple and reliable wp34s flash|update cable is going to start here:



This photo shows the power line moved from the 5.1v to the 3.4v source, for powering the wp34s. The next step is to install the crossover resistors (miniature 1k), then seal the plug. On the other end will go the male micro usb jack (Tx Rx) already crossed, 3.3v power source, and ground.

The fine bell wiring will be according to Pott's idea (thanks Harold) and, again, voila!

We start tomorrow, photos coming!

Cheers,

[brouhaha]

It's possible I'm confused about what you're doing, but doesn't that result in a calculator that has a micro-USB connector that isn't electrically USB? that doesn't seem like a good thing to me.

I'm also unclear on the motivation; USB-to-serial chips are fairly inexpensive. I'm fine with putting one in each calculator.

[MarkHaysHarris777]

(02-04-2015 11:09 AM)brouhaha Wrote:  (1) It's possible I'm confused about what you're doing, but doesn't that result in a calculator that has a micro-USB connector that isn't electrically USB? that doesn't seem like a good thing to me.

(2) I'm also unclear on the motivation; USB-to-serial chips are fairly inexpensive. I'm fine with putting one in each calculator.

(1) That is correct. It is RS232 (sort-of) made available on the micro usb connector which eliminates the need to use the pogo pads for update, provides 3.4v power for the unit, provides a rock solid connection in a tiny space, and serves the specific purpose of updating a functional wp34s.

(2) I'm ok with mounting the FTDI PL2303HX in the wp34s... and I may do that on one of my units... but I'm still going to start with the FTDI connector. One of the things I don't like about Harold's board is the paradigm of mounting the micro jack out-the-side, and the need to modify the case to get the board and jack into place due to size. The FTDI plug can be disassembled so that only the board needs to be mounted (small drop of epoxy, the board is tiny) and then fine bell wiring runs to the pcb AND to the female micro jack (which instead of being mounted through the side, which is ugly, it is mounted in the back plate under the battery cover; when not in use you don't even see it).

I'm still mulling it around on the drawing board...


cheers

[walter b]

(02-04-2015 06:00 PM)MarkHaysHarris777 Wrote:  

(02-04-2015 11:09 AM)brouhaha Wrote:  (1) It's possible I'm confused about what you're doing, but doesn't that result in a calculator that has a micro-USB connector that isn't electrically USB? that doesn't seem like a good thing to me.

(1) That is correct. It is RS232 (sort-of) made available on the micro usb connector which eliminates the need to use the pogo pads for update, provides 3.4v power for the unit, provides a rock solid connection in a tiny space, and serves the specific purpose of updating a functional wp34s.

One thing I don't like about this concept is the paradigm of using a standardized data jack for deviating purposes. I concur with Eric in that matter. Barry's approach (cf. here) looks better to me.

(02-04-2015 06:00 PM)MarkHaysHarris777 Wrote:  One of the things I don't like about Harold's board is the paradigm of mounting the micro jack out-the-side, and the need to modify the case to get the board and jack into place due to size. The FTDI plug can be disassembled so that only the board needs to be mounted (small drop of epoxy, the board is tiny) and then fine bell wiring runs to the pcb AND to the female micro jack (which instead of being mounted through the side, which is ugly, it is mounted in the back plate under the battery cover; when not in use you don't even see it).

Beauty is in the eye of the beholder. Personally, I prefer an easy access to that jack over the need to open that overbuilt battery door every time.

d:-/

[Sanjeev Visvanatha]

(02-04-2015 08:40 AM)MarkHaysHarris777 Wrote:  All due respects, Harold Pott's micro usb board idea is bass-ackwards.

This is a matter of reference frames. It is a well thought out addition to the WP 34S.

It has the following advantages IMO:

1. Calculator lies flat on your desk while updating or transferring data with the computer;

2. I'd rather have some cheap micro USB cable in my bag or at work, than a custom cable being transported around.

[BarryMead]

(02-04-2015 07:00 PM)Sanjeev Visvanatha Wrote:  

(02-04-2015 08:40 AM)MarkHaysHarris777 Wrote:  All due respects, Harold Pott's micro usb board idea is bass-ackwards.

This is a matter of reference frames. It is a well thought out addition to the WP 34S.

It has the following advantages IMO:

1. Calculator lies flat on your desk while updating or transferring data with the computer;

2. I'd rather have some cheap micro USB cable in my bag or at work, than a custom cable being transported around.

The really great thing about "Open Source" hardware and DIY electronics, is that EVERYONE can have it their own way. There is never a right or wrong way to do things, and differing opinions are always welcome. Creativity is enhanced when people can "Do their own thing" and share with others. While I personally LOVE Harald's USB board, and thank him for providing it at such a reasonable price to the community, I also respect Mark's desire to try something new.

[Sanjeev Visvanatha]

(02-04-2015 08:43 PM)BarryMead Wrote:  The really great thing about "Open Source" hardware and DIY electronics, is that EVERYONE can have it their own way. There is never a right or wrong way to do things, and differing opinions are always welcome. Creativity is enhanced when people can "Do their own thing" and share with others. While I personally LOVE Harald's USB board, and thank him for providing it at such a reasonable price to the community, I also respect Mark's desire to try something new.

I also respect Mark's plan. I was just providing some advantages to one method that he may not have thought of.

[BarryMead]

(02-04-2015 09:25 PM)Sanjeev Visvanatha Wrote:  

(02-04-2015 08:43 PM)BarryMead Wrote:  The really great thing about "Open Source" hardware and DIY electronics, is that EVERYONE can have it their own way. There is never a right or wrong way to do things, and differing opinions are always welcome. Creativity is enhanced when people can "Do their own thing" and share with others. While I personally LOVE Harald's USB board, and thank him for providing it at such a reasonable price to the community, I also respect Mark's desire to try something new.

I also respect Mark's plan. I was just providing some advantages to one method that he may not have thought of.

I personally would not use a micro-USB connector for a NON-USB purpose, because some bone-headed friend might plug a standard USB cable into the thing and damage the calculator. But if Mark is CERTAIN that he will be the only one handling his special calculator with the non-standard USB connector, then he should be alright.

[MarkHaysHarris777]

Hey, thanks everyone for your great ideas... lots to think about.

I really like Barry's setup with the phone jack (very nice). The TI 84 Plus Silver edition uses a phone jack (tip, ring, and gnd) for I/O along with their micro usb port; and it works great. Its also easy to wire, is easy to come by, and doesn't take up much space.

I see everyone's point about deviating from a standard interface.

I want the micro jack behind the battery cover to keep it clean. One of the things I hate about my Android Samsung is that the micro usb port is exposed out one end, and its always filling up with naval lint... its disgusting, and it makes for bad connection.

Well, Barry's right, we can all have it our way with these projects... I'll share what I come up with and see what y'all think about it.

Thanks again for the ideas to think about!

Cheers

[matthiaspaul]

I was thinking about designing a USB to TTL serial converter to be build into the calculator myself, but with the Micro-B USB receptable (and IR LED) facing to the back of the calculator, not its side. This way, the cable gets into the way less frequently. (Compare this with the cableing of a computer mouse.)

While USB is convenient, I still have alot of RS232-based equipment as well, and therefore would have to fit in a serial port as well. In order to save space I was tinkering with the idea of multiplexing the 3.3V TTL RX/TX signals with the USB D+/D- lines, so that, by default, the calculator would route the RX/TX signals to the connector and reassign them to D+/D- only if the USB host provides 5V via VBUS. While this would continue to allow to power the calculator via USB it would rule out the possibility to power the calculator via an external RS-232 level translator. The difficult part would be to multiplex the signals while drawing zero power from the calculator itself (which is no option as it is powered by coin cells only).

The alternative to build in a "smart" interface would be to just add a connector providing RX, TX, GND and VCC (3.3V), but in this case I would rather use a female 2.5 mm 4-conductor TRRS connector instead of a USB connector. The drawback would be that a multitude of different pinouts exist for such connectors, so some protection would have to be included in order to avoid the calculator being bricked when plugging in the wrong cable by accident.

Does someone know if the calculator's RX and TX signals are ESD protected already?

Greetings,

Matthias

[BarryMead]

(02-04-2015 11:03 PM)matthiaspaul Wrote:  Does someone know if the calculator's RX and TX signals are ESD protected already?

The HP-20b/HP-30b schematic shows that the serial inputs are connected with jumpers (JP5, JP6) directly into inputs/outputs on the processor. So the only ESD protection would be the standard I/O input protection diodes on an ARM micro-controller input/output. (Not much)

[MarkHaysHarris777]

(02-04-2015 07:00 PM)Sanjeev Visvanatha Wrote:  This is a matter of reference frames. It is a well thought out addition to the WP 34S.

I agree, its a very well thought-out solution for a combination IR transmitter | update | usb cable; and he should be proud of it!

I'm really trying to decide whether I really want the IR transmitter. If I put it in I want the driver circuit described by HP in their IR manual (looks like Harald's board has that) and even though I don't like the micro usb hangin out the side, you're right about the calc being able to lay flat.

Can anyone who has Harald's board installed tell me whether there is a crystal on the back side of the board? One thing I'm noticing about the boards that work with MySamba and ones that do not doesn't seem to have anything to do with the PL2303HX chip (they're all made by FTDI); it has to do with whether the board is crystal controlled. The board I have has a crystal mounted on the back-side. Harald's pcb 2 the same?

Another thing I'd like to know is whether he gets usb transmit distances of more than 50 cm (like across the room) or is that not possible?

I guess at this point if I decide to place the IR transmitter in this thing, Harald's board is the way to go... otherwise, I am really liking Barry's idea of the tip-ring-gnd solution with a miniature gold phone jack (just for solid connection for flash|update). I am concerned about cutting a whole in the side of the calc with a hand-held file... afraid it would look like a fifth-grader did it... no offence to any of you fifth-graders. Just saying...

Cheers
marcus

[rprosperi]

(02-07-2015 11:16 AM)MarkHaysHarris777 Wrote:  I am concerned about cutting a whole in the side of the calc with a hand-held file... afraid it would look like a fifth-grader did it... no offence to any of you fifth-graders. Just saying...

I agree, I'm afraid to permanently scar my 34S and forever brand it with blatant proof of my cutting skills, though I wonder if a tiny bit on a dremel may do the trick. Any tips from anyone who has succefully done this?

[walter b]

(02-07-2015 03:27 PM)rprosperi Wrote:  

(02-07-2015 11:16 AM)MarkHaysHarris777 Wrote:  I am concerned about cutting a whole in the side of the calc with a hand-held file... afraid it would look like a fifth-grader did it... no offence to any of you fifth-graders. Just saying...

I agree, I'm afraid to permanently scar my 34S and forever brand it with blatant proof of my cutting skills, though I wonder if a tiny bit on a dremel may do the trick. Any tips from anyone who has succefully done this?

Don't be afraid. Use a small, flat, hand-held file and take your time. Look from the side to check parallelity of the whole hole. DON'T use a Dremel.

d:-)

[BarryMead]

(02-07-2015 04:13 PM)walter b Wrote:  Don't be afraid. Use a small, flat, hand-held file and take your time. Look from the side to check parallelity of the whole hole. DON'T use a Dremel.
d:-)

Walter is right. When I modified my latest calculator to accommodate a 2.5mm headphone jack, I used a small round jewelers file, and it looks great. For the micro-USB, of course, a small flat file is needed. A Dremel tool, can cut away plastic so fast that a tiny slip, can make a big ugly scar. I have such a scar on my first calculator modified for a 2.5mm jack. (I learned the hard way, but you don't have to. Thanks to the experience shared by those like Walter)

[Marcus von Cube]

(02-07-2015 11:16 AM)MarkHaysHarris777 Wrote:  Can anyone who has Harald's board installed tell me whether there is a crystal on the back side of the board? One thing I'm noticing about the boards that work with MySamba and ones that do not doesn't seem to have anything to do with the PL2303HX chip (they're all made by FTDI); it has to do with whether the board is crystal controlled. The board I have has a crystal mounted on the back-side. Harald's pcb 2 the same?

Harald's board doesn't have the crystal mounted. The crystal needs to go to pads on the calculator to work reliably. If you want the IR led you will have to add the crystal because the IR software requires accurate timing.

The flash process has nothing to do with the crystal. Since the boot loader is part of the Atmel chip it cannot rely on the crystal to be present but only on the less accurate RC oscillator. So even with the crystal present it is not used during flashing.

[BarryMead]

(02-07-2015 11:16 AM)MarkHaysHarris777 Wrote:  Another thing I'd like to know is whether he gets usb transmit distances of more than 50 cm (like across the room) or is that not possible?

If you have 1K resistors in series on your TX and RX lines cumulative cable capacitance becomes more of an issue. My cables don't have the 1K resistors and they seem to work just fine. I think the PULL HIGH voltages on these chips operating at 3.3 Volts probably isn't much above 3 Volts anyway. To cause problems the voltage would have to go .6 to .7 volts above 3V which would make the input protection diodes turn on like SCR's. The data transfer speeds for programs and data (Not flashing) is at 9600 baud, and for flashing is 115200 baud, so one would expect that the emulator/calculator cable distances could be slightly longer than the flashing cable distances, but I have never tried to test these limits.
Hope this helps, Barry

[matthiaspaul]

(02-05-2015 12:14 AM)BarryMead Wrote:  

(02-04-2015 11:03 PM)matthiaspaul Wrote:  Does someone know if the calculator's RX and TX signals are ESD protected already?

The HP-20b/HP-30b schematic shows that the serial inputs are connected with jumpers (JP5, JP6) directly into inputs/outputs on the processor. So the only ESD protection would be the standard I/O input protection diodes on an ARM micro-controller input/output. (Not much)

Thank you, Barry. I have meanwhile found the HP 30b schematic and SDK as well.

From personal experience (with other designs), I would not want to rely on the CPU's internal clamping diodes alone when directly exposing I/Os to a connector. I've seen too many chips partially or fully blown out due to ESD outside the lab environment. Therefore, I'd like to recommend to add protection diodes at least to the two data lines.

Greetings,

Matthias

[matthiaspaul]

(02-07-2015 11:16 AM)MarkHaysHarris777 Wrote:  Can anyone who has Harald's board installed tell me whether there is a crystal on the back side of the board?
[...]
The board I have has a crystal mounted on the back-side. Harald's pcb 2 the same?

Marcus has already answered this question, but I'd like to add, that, according to a photo of Harald's board (in another thread), he's using the 16-pin FTDI FT230X rather than the Prolific PL2303HX (which is for the most part pin-compatible with the 28-pin FTDI FT232RL).

Neither the PL2303HX nor the FT230X has a provision for an external crystal, whereas the FT232RL has (the PL2303HX has the pins, but they are documented as reserved and not connected). However, over the years I have never experienced problems omitting the crystal on the FT232RL when used in bus-powered or dynamically powered configurations (that is, when VBUS > 4V was available).

Quote:Another thing I'd like to know is whether he gets usb transmit distances of more than 50 cm (like across the room) or is that not possible?

USB 1.x/2.x is specified for up to 5m. Provided that proper USB cables (meeting the USB specs) are used, this will work with the FTDI as well as the Prolific chips, otherwise they weren't USB-compliant.

Actually, the TTL serial side between the USB chip and the ARM CPU is more critical than the USB side, since it is an unsymmetrical signal. While it is a very good idea to insert some series resistors into the lines for a number of reasons (to limit the current in case of external shortcuts or if someone would accidently connect two TX drivers, to help deal with the 3.0V/3.3V level transition, to reduce the current in case of an ESD event and generally to reduce ringing), I think, 1k is too much and may introduce problems if longer cables are used. I would recommend to not use more than 810R instead. Most probably, lower values would be sufficient as well, even if they would temporarily overload the CPU's output capabilities. In either case, the resistors should be located inside the calculator, not your external USB converter.

Quote:I guess at this point if I decide to place the IR transmitter in this thing, Harald's board is the way to go... otherwise, I am really liking Barry's idea of the tip-ring-gnd solution with a miniature gold phone jack (just for solid connection for flash|update).

Hm, the IR transmitter is just the IR LED and a resistor. It is controlled by a separate I/O port pin of the CPU, not by the TTL serial or USB interface. Instead of using the resistor on Harald's board, you could just as well solder the resistor onto the calculator's PCB or use a threaded resistor...

I too like Marcus' and Barry's 2.5mm 3-pin phone jack solution very much, but in order to play it safe, I would add some ESD protection diodes. Also, I would use a (3-pin compatible) 4-pin jack (female side: Tip = 3.0-3.3V RX, Ring 1 = 3.0V-3.3V TX, Ring 2 = 3.0-5.0V VCC, Shaft = GND) for an optional external power supply. (In case of 3-pin 3.5mm phone plugs, assembled cables with built-in USB converters are readily available on the market - I haven't seen them for 3-pin 2.5mm plugs or for 3.5mm or 2.5mm 4-pin plugs, though.)

Greetings,

Matthias

[matthiaspaul]

(02-04-2015 08:40 AM)MarkHaysHarris777 Wrote:  So, the ultimate simple and reliable wp34s flash|update cable is going to start here:



This photo shows the power line moved from the 5.1v to the 3.4v source, for powering the wp34s. The next step is to install the crossover resistors (miniature 1k), then seal the plug.

If you have opened the converter already, you might check if the RTS/CTS pins (and in case of the FT232RL or PT2303HX also the DTR/DSR/DCD pins) are connected with each other or not. If they are still open (that is, not connected to anything else), you could short them. With this patch in place, it is no longer necessary to disable hardware handshake when opening serial connections through USB - it will work both ways. While this isn't normally a problem when using dedicated software it'll make things easier when setting up a connection in a terminal program.

BTW. Based on a photo of Harald's USB board, he didn't connect RTS and CTS as well, so this patch also applies to his board.

Greetings,

Matthias

PS. Where do the 5.1V and 3.4V figures come from? Shouldn't this read 5.0V and 3.3V?