Not working with some devices

[ffxx68]

Sometimes, board doesn't properly respond to commands.

[ffxx68]

584 = 0000 0010 0100 1000

Looks like Bit 0001 = SEL1 = R22 isn't working yet for you.

I have patched two different boards by bridging with 10K resistors (resulting in 5K ohm, after the patch) and both were fixed, testing them a PC1403 which DSKF gave 72 from DSKF, before the patch:

Note the stacked 10K resistors.

I had to redo the R22 patch twice for one of the boards, as after the fix it still gave me a 584.

Looks like reducing the resistors to 5K was enough to make the fix.

[RN-NOMATAN373]

On the same PCB, PC-E500 and PC-1475 return the correct value; only PC-1460 returns 584. Is it only PC-1460 that needs the SEL1 signal?

[ffxx68]

All models need the same signals. I really can't tell why your 1460 isn't working...

[RN-NOMATAN373]

Thanks! I'm just happy to be able to use LOAD and SAVE with PC-1475, PC-E500 and PC-1480U. I have another PC-1460, so I will try this PCB there as well.

[jbarbe]

Very good progress with the R22 R20 R16 hack,

Previously working and still working : 1360 & 1460

Other below was not working before hack : Now fully working : 1403H 1425 1280 (not Jap)

Working Save & Load (but with DSKF 1 with error or FILES not fully working (only first page, get error after)) E550 1490U E500S (1 unit) 1360K

Not working 1475 (DSKF 1 = 8) E500S (second unit) 1280K (create malfunction : keyboard not responsive properly making PC unusable. the PC works fine without the CE140F)

[RN-NOMATAN373]

Mr. FABIO! I prepared another PC-1460 and ran DSKF1 and got the correct numbers! The resistor is still 1kΩ.

Some older units may not work due to problems with the unit. Unfortunately, I don't have a CE-140F, so I can't confirm if that would work correctly with the PC-1460 that was giving me problems.

[ffxx68]

Working Save & Load (but with DSKF 1 with error or FILES not fully working

I think this is a problem of a different kind: not an hardware one, but board software not properly handling some timing in the initial device-code sequence, for these newer devices. For example, DSKF(1) including parenthesis, does work. I have to figure out yet the fix, but I'm pretty sure it's not the signal levels here... Not sure instead about your other devices, of the old series. This could still be the signal levels. Surely the level converter design should be revised, but that's the subject of another project entirely. I hope someone else could pick this up...

[jbarbe]

Not working 1475 (DSKF 1 = 8) E500S (second unit) 1280K (create malfunction : keyboard not responsive properly making PC unusable. the PC works fine without the CE140F)

After more test the e500s is okish. I'm able to save and load... Seems sometime need to reset the board.

The 1280k looks have some issue and sometime defective. So not including it on the testings list.

The only proven working unit still not compatible with the ce 140f is the 1475...

R16 20 22 have 5kohms value : 2x10k in //.

[ffxx68]

I have fixed the design on KiCad too https://github.com/ffxx68/Sharp_ce140f_emul/tree/main/KiCad_v1

[ht-deko]

PC-1470U (R16, R20, R22 = 5kΩ)

DSKF(1): 4616 FILES: ERROR 8 SAVE: OK LOAD: ERROR 8

DSKF(1) LOG:

s82665683 Device d 0x41 82771353 Device ID 0x41 82771394 CE140F 82787305 0:1D [1D] 82794667 1:01 [1E] 82801609 2:1E [3C] co 82851673 Processing... 82851726 in: 3 bytes (first 40 below) 82851776 <1D011E> 82851871 avg input timing (ms/byte): 10.07 82851953 checksum 0x1E vs 0x1E 82852025 command 0x1D 82852088 DSKF 82852140 diskspace 4680 82852208 out: 5 bytes (first 40 below) 82852287 <004812005A> 82873109 send complete 82873188 avg output timing (ms/byte): 4.11

[ffxx68]

@ht-deko may you please download and try this firmware? DSKF should return 65535: https://github.com/ffxx68/Sharp_ce140f_emul/releases/tag/v1.5

[jbarbe]

Looks like firmware 1.5 have fixed my 1475 issue :-) . (alongside the resistor hack) . And DSKF 1 return something now on E5Xx : 65535.

Conclusion : 1.5 is definitly a great positive move. Hope other people will see same improvements.

[ht-deko]

PC-1470U ・R16, R20, R22 = 5kΩ ・v1.5 firm ・Connect to PC

DSKF(1): ERROR 8 FILES: ERROR 8 SAVE: OK LOAD: ERROR 8

DSKF(1) no longer returns a value.

[ht-deko]

PC-1470U ・R16, R20, R22 = 3kΩ ・v1.5 firm ・Connect to PC

DSKF(1): 65535 FILES: No error SAVE: OK LOAD: OK

[ht-deko]

PC-1470U ・R16, R20, R22 = 3kΩ ・v1.5 firm ・Connect to USB charger

DSKF(1): ERROR 8 FILES: ERROR 8 SAVE: OK LOAD: ERROR 8

[jbarbe]

PC-1470U ・R16, R20, R22 = 3kΩ ・v1.5 firm ・Connect to PC

DSKF(1): 65535 FILES: No error SAVE: OK LOAD: OK

I have found similar behavior on some Pockets, looks like when connected to PC instead of USB charger. The card works better.

[YuuichiAkagawa]

FYI, I made new board.
https://github.com/YuuichiAkagawa/Sharp_ce140f_emul/tree/mod_74lv1t34/KiCad_v1

• Diode : CUHS20S30(SBD)
• Level converter : SN74LV1T34

This board works fine.

DSKF command on PC-1470U

[ffxx68]

Hi @YuuichiAkagawa - this is truly great advance. I wish I had your counselling during the development of my board. Now... time to order new pcbs ;) Joking. I had started actually a new design, without the Nucleo but the controller directly on the board, and battery power. But, the level converter was my doubt, and how to test a prototype ... I'm going to share what I've designed so far soon

[ffxx68]

Hi @YuuichiAkagawa - with respect to the converter you choose: SN74LV1T34, I looked for it, but there is no multiple-converter chip available (eg. 4 or 8 in-a-chip), as far as I can see. Do you know of any? That would help putting everything on a PCB, to make it more compact.

[YuuichiAkagawa]

I don't know if my choice is the best one.
Quad: SN74LV4T125
https://www.ti.com/product/en-us/SN74LV4T125 Octal: SN74LV8T245
https://www.ti.com/product/en-us/SN74LV8T245

[ffxx68]

All right - I actually started my v2 project using a different one (TXS0108EPW)

https://github.com/ffxx68/Sharp_ce140f_emul/tree/main/KiCad_v2

but please take this just as a DRAFT, or a starting point, mostly for the controller and surrounding power section, as it's NOT TESTED, and it would also need a software review.

[ffxx68]

Since the workaround of hacking the resistor values looks like solved most of the problem, while a more stable fix would need a major rework, I'm closing this issue for the time being. Thanks to all who contributed to the solution!