386P CPU Upgrade

  • As many know, the current post processors create bazillion lines of code for contouring as they are unable to use any of the Maho's in built geometry programming, E-parameters etc. even in 2.5D strategies.

    The common problem is data exhaustion and movement studder when higher feed rates are being used. But also the data transfer bit rate reduction while BTR transfer, especially if there are a ton of lines with only small movement lines. The latter causes the BTR buffer to get depleted, machine to halt with spindle running and to wait while filling up the BTR buffer. If the program is being run from the on board memory, the operation seems to be a bit more effortless.

    I think that the bit rate reduction is mainly caused because the CPU struggles with data operations, even there are separate DMA controller units.

    I have found some clues that the 532 control would have a 386 CPU that is running at 20MHz as the basic 432 386 unit runs at 16MHZ.

    There are three crystals on the CPU. One seems to be clocking the CPU and the 802380 DMA processor (double the clock rate as the CPU / 32MHZ for 16MHz and 40 for 20MHz with the 532), the one upwards from that is running at 24MHz and the crystal next to the N80C452 microcontroller is running at ~11MHZ. The difference seems to be that the RAM and EEPROM's are rated faster on the 532: 80ns vs 120ns and 120ns vs 200ns acess times. Not quite sure if that is significant, but the chips would be easy to replace.

    I noticed that the two crystals that are not clocking main CPU and 802380 are both running at same speed in 432 and 532 boards (11MHz and 24MHz). This rises a question that can the CPU speed be increased to better deal with the incoming bit flow and avoid the transfer rate reduction?

    As a test, I installed TX486DLC CPU and ULSI DX co-processor (originally they are separate items / 386SX and 387 unit) to see if there would be some proprietary issues to run a different CPU than intel. So far no problems, but haven't had the time to make any tests if there is any difference.

    Last week I acquired a complete 432/10 control with 386 CPU as spare parts and will be getting it next week. I will then make some more advanced tests after I have made backups of the EEPROM's, getting my hands on a faster rated 802380 to replace the current that is rated to 16MHz.

    Florian added a bus description for 432 control. It can be seen that there are two clocked pins: 1107 and 1114, 1,5MHz and 3.0MHz accordingly. Does anyone know what are the clock sources for there as neither seems not to be directly related crystals on the CPU module => times ten would be 1,5->15MHz and 3->30MHz (well, da'h)?
    There seems to be a 6000kHz crystal on Contr.Telet, would that be it (divided by 2 and 4)? And on Graph Mod there is a 24000kHz crystal hence the 24M oscillator on the CPM?

    Has anyone attempted anything in similar?

  • Dieses Thema enthält 52 weitere Beiträge, die nur für registrierte Benutzer sichtbar sind, bitte registrieren Sie sich oder melden Sie sich an um diese lesen zu können.

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!