PGram+ Card --- Loading up the DSR

[humeur]

I made the modification I have the assembly diagram the DSR was version 2. I will take a photo.

For your part, have you just tested P-Gram?

 

J'ai réaliser la modif j'ai le schéma de montage la DSR était la version  2.

Je ferais une photo.

 

 

 

 

Edited August 13 by humeur

[humeur]

Photography

 

Changing the regulator, removing CR6 diode and fixing the jp1 jp2 bridge temporarily.

The card is far from the flexcable

 

Use PGDSR Version 2                 version 2.1 no work

 

 

Changement du régulateur,  suppression diode CR6  et bidouille pour le pont jp1 jp2  en provisoire

La carte est loin du flexcable

 

 

[Gary from OPA]

4 minutes ago, humeur said:

Photography

 

Changing the regulator, removing CR6 diode and fixing the jp1 jp2 bridge temporarily.

The card is far from the flexcable

 

Use PGDSR Version 2                 version 2.1 no work

 

 

Changement du régulateur,  suppression diode CR6  et bidouille pour le pont jp1 jp2  en provisoire

La carte est loin du flexcable

 

 

Looks good.. what part is that for the voltage regulator replacement? Does Mouser sell it?

 

You only have RAM for PGRAM mode not the Plus version.

 

Battery should be a little bigger one to hold the data better if off for more than a few days.

[humeur]

The purchase regulation system on ebay LM2596 DC-DC

For my test just the classic P-Gram config, I will now test for the P-GRAM+ config the dsr 2.1 does not work at all in my case, ok with 1.0 and 2.0

 

The battery is a 2032 for testing this is enough.

 

Have you tried with DSR version 2.0

[humeur]

Changed the SRAM 128k to a slower one of 100 ns instead of the current one of 80 ns and it works without problems.

test with dsr 2.0; position 6 in the box.

No work with 2.1

 

 

Changement de la SRAM 128k par une plus lente de 100 ns au lieu de celle présente de 80 ns et la fonctionnement sans soucis.

Edited August 13 by humeur

[Gary from OPA]

7 minutes ago, humeur said:

Changed the SRAM 128k to a slower one of 100 ns instead of the current one of 80 ns and it works without problems.

test with dsr 2.0; position 6 in the box.

 

 

Changement de la SRAM 128k par une plus lente de 100 ns au lieu de celle présente de 80 ns et la fonctionnement sans soucis.

As long as you not running two PEBs together or having a fully packed PEB with external sidecars as well you should have any timing issues with loading up all 4 module banks.

 

I think I am only using pgram2k DSR, not sure what this 2.1 is, is that the Y2K fixed version? Give me a link to it and I will test it out.

[humeur]

i use this DSR

 

PG-RAM-Card-Hoddie.dsk

 

 

00 00 00 00 00 00 00 00 00 00 00 00 00 00 54 6F ..............To
00001F70 6E 79 20 4B 6E 65 72 72 20 20 20 20 20 20 20 20 ny Knerr        
00001F80 20 20 20 20 20 20 13 73 02 69 85 C9 E6 D0 61 6C       .s.i....al
00001F90 36 10 50 DA 6E 4C 56 C6 01 01 80 80 40 40 20 20 6.P.nLV.....@@  
00001FA0 10 10 08 08 04 04 02 02 01 01 80 80 40 40 20 20 ............@@  
00001FB0 10 10 08 08 04 04 F4 46 73 BB B9 A5 E8 CE 77 6F .......Fs.....wo
00001FC0 16 37 14 08 24 08 02 02 01 01 80 80 40 40 36 31 .7..$.......@@61
00001FD0 98 16 8C 8D A7 25 43 73 B9 A1 80 80 40 40 20 20 .....%Cs....@@  
00001FE0 10 10 08 08 04 04 50 47 20 42 61 6E 6B 20 32 20 ......PG Bank 2
00001FF0 76 65 72 20 32 2E 31 20 20 34 2F 30 31 2F 30 31 ver 2.1  4/01/01

 

 

 

.Ro
00001F70 62 65 72 74 20 4C 2E 20 4A 6F 6E 65 73 20 20 20 bert L. Jones   
00001F80 20 20 20 20 20 20 8C 0C 04 06 52 E4 02 11 85 AD       ....R.....
00001F90 CA E4 20 53 3A 17 08 08 04 04 02 02 01 01 80 80 .. S:...........
00001FA0 40 40 20 20 10 10 08 08 04 04 02 02 01 01 80 80 @@  ............
00001FB0 40 40 20 20 10 10 91 5B 2D CD C7 32 A1 02 81 E0 @@  ...[-..2....
00001FC0 6A 6E 31 31 90 1C 08 08 04 04 02 02 01 01 85 91 jn11............
00001FD0 C8 40 70 68 37 B7 19 48 04 64 66 96 21 03 CD 95 .@ph7..H.df.!...
00001FE0 E4 D2 64 65 10 10 50 47 20 42 61 6E 6B 20 32 20 ..de..PG Bank 2
00001FF0 76 65 72 20 32 2E 30 20 20 31 2F 32 38 2F 38 39 ver 2.0  1/28/89

[humeur]

Be careful, take the 2.0 but not the Tony Knerr one for me, none of it works

 

 

 

Attention prends bien la 2.0 mais pas celle  de Tony Knerr  chez moi aucune ne fonctionne

[+dhe]

Hi @humeur 
  I'm needy today.

    1) What did the heat go to with DC-DC Converter?

    2) Can you take a picture of what the change looks like on the board?

    3) Can you send a link with the LM2596 you purchased?

 

Thanks in advance.

[+Ksarul]

I don't see the wires or the resistor required to make the 128K chip work properly. Here is a picture of Jon Guion's modifications for the 128K chip. Note also that the middle 8K chip must be 32K

[Gary from OPA]

15 minutes ago, Ksarul said:

I don't see the wires or the resistor required to make the 128K chip work properly. Here is a picture of Jon Guion's modifications for the 128K chip. Note also that the middle 8K chip must be 32K

This is basically the same as my version of recreating the plus version on a normal PCboard. Adding the two address lines needed. But I didn't add the pull up resistor on the middle 32k chip, but not sure if that is really needed it used to be common on horizon Ramdisk when using 32k chips.

 

What resistors do you think are missing. My recreating the plus on the normal board is working fine.

[+Ksarul]

3 hours ago, Gary from OPA said:

This is basically the same as my version of recreating the plus version on a normal PCboard. Adding the two address lines needed. But I didn't add the pull up resistor on the middle 32k chip, but not sure if that is really needed it used to be common on horizon Ramdisk when using 32k chips.

 

What resistors do you think are missing. My recreating the plus on the normal board is working fine.

I was looking at the picture of @humeur's board. That one seems to be missing the changes in the RAM area. On the resistors, the only one you seem to be missing is the one on the 32K RAM chip, @Gary from OPA. You note that the mod works fine without it, so it may be useful but not necessary. The wires to utilize the larger chips seem to be doing the trick for you--and if @humeur is using the same or similar wiring changes on his board that I don't see here (possibly due to reverting it to standard P-GRAM mode), his "should" work as well.

[Gary from OPA]

14 minutes ago, Ksarul said:

I was looking at the picture of @humeur's board. That one seems to be missing the changes in the RAM area. On the resistors, the only one you seem to be missing is the one on the 32K RAM chip, @Gary from OPA. You note that the mod works fine without it, so it may be useful but not necessary. The wires to utilize the larger chips seem to be doing the trick for you--and if @humeur is using the same or similar wiring changes on his board that I don't see here (possibly due to reverting it to standard P-GRAM mode), his "should" work as well.

Humeur has a actual pgram+ PCboard so no extra wires, majority of units sold were the + version. Very few original ones out there with the added wires. But the plus design also had some minor changes over on the right side were all the resistors are stacked as well.

[humeur]

10 hours ago, Ksarul said:

I don't see the wires or the resistor required to make the 128K chip work properly. Here is a picture of Jon Guion's modifications for the 128K chip. Note also that the middle 8K chip must be 32K

Jim c'est une P-GRAM+ la mienne donc c'est déjà câblé pour la 128k

 

[humeur]

14 hours ago, dhe said:

Hi @humeur 
  I'm needy today.

    1) What did the heat go to with DC-DC Converter?

    2) Can you take a picture of what the change looks like on the board?

    3) Can you send a link with the LM2596 you purchased?

 

Thanks in advance.

 

 

Hi @dhe

 

For the converter buy on ebay  

 

https://www.ebay.fr/itm/204809996288?_nkw=convertisseur+lm&itmmeta=01J58BN1339K6H6CXPEX1TNA3Y&hash=item2fafa08800:g:6ZoAAOSwtu9mWDGq&itmprp=enc%3AAQAJAAAA8HoV3kP08IDx%2BKZ9MfhVJKndKd7B7Q62O0OJwLrTDCf1VT9ql%2FtZE0ZtPR0SeWE0iqBCgmWlZSf6f8yaW%2FkQOr7jvQzSHXg%2BDCIDXkmLljiXnCV1Hs1E5%2BV59fYeqhe00JxyxRCdI9g8dQv09CsoCaBjrboHy%2B%2BmIZ2UVLBsF1GeYX7ioPeUf5etR6Nr0UZefnKBn7Qzgwa0ucwGokkyXVk00wWFHmYpArW41SRtY2fhtoyelQOx6Jk1aP%2BWG6qTKwUYqEqS%2Fg2vuOb6CU%2BiWKciO1wcRgu7JZan2rrWmziS5TMQr%2B4Q2LvcsqV2YiaRLg%3D%3D|tkp%3ABk9SR9SR1IuqZA

 

For heat dissipation as the consumption is less than 1 A no need for a cooler, moreover the technology used has a higher efficiency than the classic 7805 less heat in the end.

 

Can you take a picture of what the change looks like on the board?   What you want is the connection of the module?

 

it is necessary to adjust the voltage with the potentiometer before mounting on the card, make an empty setting around 5.4 and on the card remove diode CR6.

the TTL will have a voltage lower than 5.4 volts because of the CR5 diode which will lower the voltage.

 

 

JL

[humeur]

10 hours ago, Gary from OPA said:

This is basically the same as my version of recreating the plus version on a normal PCboard. Adding the two address lines needed. But I didn't add the pull up resistor on the middle 32k chip, but not sure if that is really needed it used to be common on horizon Ramdisk when using 32k chips.

 

What resistors do you think are missing. My recreating the plus on the normal board is working fine.

@Gary from OPA  advice change your 128k 55ns too fast much too fast for our old TI finds more around 100ns.

 

The 100ns are for reading but the control signals are much faster see datasheet https://www.alldatasheet.fr/html-pdf/82433/HITACHI/HM628128/1759/7/HM628128.html

 

JL

[Gary from OPA]

59 minutes ago, humeur said:

@Gary from OPA  advice change your 128k 55ns too fast much too fast for our old TI finds more around 100ns.

 

The 100ns are for reading but the control signals are much faster see datasheet https://www.alldatasheet.fr/html-pdf/82433/HITACHI/HM628128/1759/7/HM628128.html

 

JL

The weird thing is for me with dual PEB setup, changing to 55ns help to reduce data errors, compared to 100ns.

 

I am still waiting on getting the 74F logic chips to even reduce the amount of time even more.

 

As my problem still is the RDBENA signal takes too long to go active because of the amount of complex address decoding logic the P-GRAM card does.

[humeur]

I'm not sure that this will play a lot, look here it's in French, in addition the F have greater consumption, the regulator is already very hot.

 

http://meteosat.pessac.free.fr/Cd_elect/perso.wanadoo.fr/f6crp/elec/sc/famtec.htm

 

To be sure that this is the RDBENA you would need a logic analyzer.

 

If you look at the data we fall to 10ns depending on the signal.

JL

 

[Gary from OPA]

11 minutes ago, humeur said:

I'm not sure that this will play a lot, look here it's in French, in addition the F have greater consumption, the regulator is already very hot.

 

http://meteosat.pessac.free.fr/Cd_elect/perso.wanadoo.fr/f6crp/elec/sc/famtec.htm

 

To be sure that this is the RDBENA you would need a logic analyzer.

 

If you look at the data we fall to 10ns depending on the signal.

JL

 

Yes. That is another problem switching to F chips.

 

Using ACT would be best just as Fast and not much power usage. But it's hard to find supply of the needed ACT chips, places like Mouser don't stock them.

 

Even better would be changing the regulator to output 3.3volts and changing the logic to LVT those are even faster in timing!

Edited August 14 by Gary from OPA

[humeur]

Your DSR is not loading at all or is it crashing while loading?

 

Ta DSR ne charge pas du tout ou ça plante en cour de chargement ?

[Gary from OPA]

Just now, humeur said:

Your DSR is not loading at all or is it crashing while loading?

 

Ta DSR ne charge pas du tout ou ça plante en cour de chargement ?

My card as I stated before works perfectly when in a setup with a single PEB , the DSR and all 4 grom banks load up and run great.

 

But when in dual PEB setup, the timing is off, DSR and the Cartridge 6000 space is perfect, but GROM read/write is too slow, data errors occur if multiple bytes are read or written in a sequence.

[humeur]

The distance is long between your two boxes

your other cards in this case work without problem

 

 

Edited August 14 by humeur

[+dhe]

@mizapf - This weekend, I was going to play with loading the PGRAM. I wanted to run PGTEST1,2.

 

The way the test are ran, is to use E/A to load PGTEST1, then you are asked to remove the module.

 

 How can I simulate doing that in MAME?

[Gary from OPA]

1 minute ago, dhe said:

@mizapf - This weekend, I was going to play with loading the PGRAM. I wanted to run PGTEST1,2.

 

The way the test are ran, is to use E/A to load PGTEST1, then you are asked to remove the module.

 

 How can I simulate doing that in MAME?

You would power down everything, reset the ti99 and then exit mame correctly.

 

Then restart mame and run pgtest2 and it should say ok. On real hardware it is testing the battery backup. On mame it should remember the data, like it was an horizon Ramdisk.

[+Ksarul]

16 hours ago, Gary from OPA said:

Using ACT would be best just as Fast and not much power usage. But it's hard to find supply of the needed ACT chips, places like Mouser don't stock them.

Digi-Key stocks various ACT chips. I just looked up the 74ACT244 and they have several hundred of the DIP version in stock.