#FujiNet Hardware Discussion

[Gavin1968]

10 minutes ago, mozzwald said:

part number in the BOM.

Ugh, you are correct sir, not sure what I was thinking.  lol  Sorry about that, I guess I need a little more coffee than I have had today. 

 

[Kyle22]

On 8/31/2020 at 10:44 AM, mozzwald said:

This was reported early by some users and I modified the SIO receptacle to be a little shorter which appears to correct the issue. In fact, yesterday someone had this problem and printed the newer port themselves and replaced it which fixed their issues. I can swap yours with another and hopefully that will work better for you. Also, I have the v1.1 boards with the fix for 1088XEL since I know you have one.

 

Anyone with a FujiNet made by me should let me know if they have this or any other problem and we'll try figure out a solution.

My FujiNet's SIO port is good and tight. Not too tight, but just right.

It fits in my 800 and 800XL nicely, as well.

 

[+skr]

Just received my FujiNet and I´m blown away. It perfectly fits and works out of the box. No installation, no hassle. Plug´n´Play at it´s best.

 

I only had to look for instructions about to do the Firmware upgrade. Didn´t find it (it was late already, working ~13 hrs/day currently) but it was easy enough for me to figure out that I just have to connect the FujiNet to my MacBook and the rest was just hit Update and watch it grabbing the current firmware automatically and flashing it.

[tschak909]

9 minutes ago, skr said:

Just received my FujiNet and I´m blown away. It perfectly fits and works out of the box. No installation, no hassle. Plug´n´Play at it´s best.

 

I only had to look for instructions about to do the Firmware upgrade. Didn´t find it (it was late already, working ~13 hrs/day currently) but it was easy enough for me to figure out that I just have to connect the FujiNet to my MacBook and the rest was just hit Update and watch it grabbing the current firmware automatically and flashing it.

Excellent!

-Thom

[BillC]

On 8/31/2020 at 7:27 AM, mozzwald said:

RE: uSD. I chose micro cuz they are available everywhere and are the standard (due to smart phones). I could have put the socket on the front side but wanted to keep everything on the back for ease of building which leaves limited space for everything.

A full-sized SD card socket can still be used with micro-SD cards in an adapter, many come with it. I use micro-SD cards more often than full-size with my SDrive NUXX.

[mozzwald]

1 hour ago, BillC said:

A full-sized SD card socket can still be used with micro-SD cards in an adapter, many come with it. I use micro-SD cards more often than full-size with my SDrive NUXX.

Yes, I have too many of those adapters  The biggest issue is size and simplicity. Keeping all the parts on one side of the board makes it easier to assemble and thus there is not enough room for a full size SD socket with everything else:

 

I don't even have enough room for a uSD push/push style socket in place of the push/pull that is currently there, let alone a full size socket

[a8isa1]

On 9/5/2020 at 4:18 PM, mozzwald said:

Yes, I have too many of those adapters  The biggest issue is size and simplicity. Keeping all the parts on one side of the board makes it easier to assemble and thus there is not enough room for a full size SD socket with everything else:

 

 

I don't even have enough room for a uSD push/push style socket in place of the push/pull that is currently there, let alone a full size socket

I've never encountered a push/push style microSD socket but I've had 2 full fized SD sockets of this type fail.  I greatly prefer push/pull style, having no mechanism to fail.

[Dropcheck]

On 6/7/2020 at 8:02 AM, mozzwald said:

Hardware is frozen at 1.0. I pushed the schematic to github. The only significant change is the power switch which now cuts any/all power to the FujiNet making it a proper on/off switch. Whichever input power has the highest voltage (SIO, USB or JTAG) will supply power to the FujiNet.

I got them from mouser for the prototypes. They are also available on Aliexpress a bit cheaper. 

Hey mozzwald,

 

    Noticed that Mouser is listing these boards as NRND now.   Would this work instead ESP32 Wrover-IE

[mozzwald]

6 hours ago, Dropcheck said:

Hey mozzwald,

 

    Noticed that Mouser is listing these boards as NRND now.   Would this work instead ESP32 Wrover-IE

Yep, the WROVER-B has been replaced with the E which is a drop in replacement. I've tested it with the v1.1 board.

[Dropcheck]

9 minutes ago, mozzwald said:

Yep, the WROVER-B has been replaced with the E which is a drop in replacement. I've tested it with the v1.1 board.

?

[mozzwald]

There are 2 more hardware issues to deal with. Now that @jeffpiep has started working on cassette emulation we have figured out that there needs to be a 10k pull down resistor on the MOTOR line otherwise it's floating. The simple fix for 1.0 and 1.1 boards in the wild already is to add the resistor like so:

 

The other issue, found by @Mr Robot , is more involved. If FujiNet is turned off and connected to a powered up Atari, the inputs on the esp32 can be reverse biased and partially power up the board. This is noticeable as a buzzing or ticking noise thru the Atari SIO AUDIO_IN and then out to speakers/TV. I am looking at using a bus switch IC to cut off the inputs when FujiNet is powered off. Any other ideas are welcome

[Dropcheck]

13 minutes ago, mozzwald said:

There are 2 more hardware issues to deal with. Now that @jeffpiep has started working on cassette emulation we have figured out that there needs to be a 10k pull down resistor on the MOTOR line otherwise it's floating. The simple fix for 1.0 and 1.1 boards in the wild already is to add the resistor like so:

 

The other issue, found by @Mr Robot , is more involved. If FujiNet is turned off and connected to a powered up Atari, the inputs on the esp32 can be reverse biased and partially power up the board. This is noticeable as a buzzing or ticking noise thru the Atari SIO AUDIO_IN and then out to speakers/TV. I am looking at using a bus switch IC to cut off the inputs when FujiNet is powered off. Any other ideas are welcome

Couldn't you use a transistor rather than an IC?

[mozzwald]

24 minutes ago, Dropcheck said:

Couldn't you use a transistor rather than an IC?

I could, but it's 8 transistors vs 1 chip where space is already limited. I'm thinking about using SN74CBTD3384 which has 10 switches and does bi-directional level shifting. I will need another transistor to reverse the ENable pin logic

[jamm]

Time for a FujiNet XL (as in "extra large") board!  Or maybe a FujiNet expansion board...

 

[Dropcheck]

52 minutes ago, mozzwald said:

I could, but it's 8 transistors vs 1 chip where space is already limited. I'm thinking about using SN74CBTD3384 which has 10 switches and does bi-directional level shifting. I will need another transistor to reverse the ENable pin logic

I guess I need another can of DP.  I misread and thought you were only speaking of the one SIOAudio_IN signal.  Sorry.

[mozzwald]

57 minutes ago, Dropcheck said:

I guess I need another can of DP.  I misread and thought you were only speaking of the one SIOAudio_IN signal.  Sorry.

To be more clear, if the fujinet is turned off and Atari is on, any SIO pin that goes high will back feed power (I measure ~1-1.5V) to the esp32. This causes noise to be sent over AUDIO_IN back to the Atari.

 

[Kyle22]

On 9/10/2020 at 2:20 PM, mozzwald said:

To be more clear, if the fujinet is turned off and Atari is on, any SIO pin that goes high will back feed power (I measure ~1-1.5V) to the esp32. This causes noise to be sent over AUDIO_IN back to the Atari.

 

So, just don't turn FujiNet off?

[Dropcheck]

On 9/10/2020 at 1:20 PM, mozzwald said:

To be more clear, if the fujinet is turned off and Atari is on, any SIO pin that goes high will back feed power (I measure ~1-1.5V) to the esp32. This causes noise to be sent over AUDIO_IN back to the Atari.

 

Something I ran across while working and following another person's project might give you a start.  Not sure if it will work in this situation, but take a look.  SmartParallel curing the back-powering problem.

[Dropcheck]

Mozzwald,

 

  Is it possible to use a different USB Interface chip such as a CH340G/C?  I'm looking desperately for a hand solderable chip. 

[tschak909]

deleted

Edited September 14, 2020 by tschak909

[mozzwald]

11 minutes ago, Dropcheck said:

Mozzwald,

 

  Is it possible to use a different USB Interface chip such as a CH340G/C?  I'm looking desperately for a hand solderable chip. 

Yes, you can use any USB to serial chip you like. There is no 'code to write for it'. The chip interfaces with the PC so you just need the right driver installed for it to show up as a serial port.

On 9/12/2020 at 6:21 AM, Dropcheck said:

Something I ran across while working and following another person's project might give you a start.  Not sure if it will work in this situation, but take a look.  SmartParallel curing the back-powering problem.

Thanks for the tip, I'll check it out

On 9/11/2020 at 7:57 PM, Kyle22 said:

So, just don't turn FujiNet off?

Yes, that is the fix for all boards in the wild. FujiNet is awesome, I dunno why you'd want to turn it off 

 

Edited September 14, 2020 by mozzwald
clarify usb serial

[+DrVenkman]

31 minutes ago, mozzwald said:

Yes, that is the fix for all boards in the wild. FujiNet is awesome, I dunno why you'd want to turn it off 

I generally agree with that, but my version 1.1 board has a specific reason for me to leave it powered: the power switch doesn't toggle cleanly between states. It moves down (to Off) easily, but you have to tug/force it to Up (On) - so much so that the switch stem is bent and flexible now. The only way to definitively power on the board is to literally use a fine flatblade screwdriver pressed up against the base of the switch and forcing the toggle mechanism to move. My earlier version 1.0 board didn't have this issue. I think the switch must have gotten crushed or the housing squeezed or something at some point.  

 

So I just leave it powered on. I figure I can always just unplug it. If worse comes to worse, I can buy or get someone to send me a new toggle switch and use my hot air station to remove the old one and replace it.

[mozzwald]

2 minutes ago, DrVenkman said:

I generally agree with that, but my version 1.1 board has a specific reason for me to leave it powered: the power switch doesn't toggle cleanly between states. It moves down (to Off) easily, but you have to tug/force it to Up (On) - so much so that the switch stem is bent and flexible now. The only way to definitively power on the board is to literally use a fine flatblade screwdriver pressed up against the base of the switch and forcing the toggle mechanism to move. My earlier version 1.0 board didn't have this issue. I think the switch must have gotten crushed or the housing squeezed or something at some point.  

 

So I just leave it powered on. I figure I can always just unplug it. If worse comes to worse, I can buy or get someone to send me a new toggle switch and use my hot air station to remove the old one and replace it.

I can send you a new switch or swap units. You can remove the old switch with hot air, but do not use hot air to put a new one on unless you want a pile of plastic goo.

[+DrVenkman]

13 minutes ago, mozzwald said:

I can send you a new switch or swap units. You can remove the old switch with hot air, but do not use hot air to put a new one on unless you want a pile of plastic goo.

No need for a full swap - everything else works great.  Just mail me a switch - I'll use Kapton tape and light heat to remove the old one. I expect my fine solder point and magnification will be enough to tack the new switch into place.

 

And if not ... I'll just solder-bridge the pads for an always-on configuration and be no worse off.  

[Mr Robot]

On 9/14/2020 at 7:11 PM, DrVenkman said:

the power switch doesn't toggle cleanly between states. It moves down (to Off) easily, but you have to tug/force it to Up (On) - so much so that the switch stem is bent and flexible now. The only way to definitively power on the board is to literally use a fine flatblade screwdriver pressed up against the base of the switch and forcing the toggle mechanism to move.

Just wanted to mention that one of my Fujinets now has this problem too, it's the primary one I've been using for testing so it's been getting the bulk of the abuse. I have a bunch of these switches so I can just change mine but it looks like maybe a more robust switch might be a good idea for the next revision of the board.