Pentagon 128k debugging

[Shockwav3]

Hi there,
I'm currently in the progress of building a Pentagon 128 (Revision 2011 produced by zorel). Good things first, it already boots, accepts keyboard input and even lists betadisk directories. Now comes the bad part, after a few seconds I get junk characters on the screen, but the machine remains operational for a few more seconds til it crashes (mostly after 30-45 seconds). Sometimes it reboots, but most of the time it just hangs requiring a hard reset. Any ideas what could be wrong?

Further info:
- IC type is all 74lsxxx except for dd02, dd11 & dd12 (used 74hct93 there). I used the translation table from here: http://sblive.narod.ru/ZX-Spectrum/P...ntagon128k.htm (except for dd100 which is wrong there and should be cd4069).
- RAM chips are 41256 and have decoupling capacitors
- random decoupling capacitors have been placed on the board
- resistor and capacitor values have been used according to pentagon.pdf from here http://zxspectrum.it.omegahg.com/doc/Pentagon.pdf

Here's also a picture:
http://img855.imageshack_.us/img855/...ntagon128k.jpg

So any ideas what to try next?

[Advertiser]

[null_device]

Are all the errors of the board have been eliminated?

http://zx.pk.ru/showpost.php?p=535392&postcount=371

Is the board AY \ YM, it also is an error.

[Shockwav3]

Thanks for letting me know about that one - didn't know about that one before, fixed that, didn't help much
I've also fixed the missing track on dd31 and added R50 (56kohm)

[null_device]

Sorry if this mу advice does not help.
In the next branch, discussed the assembly and adjustment of the board, try to ask a question there.

http://zx.pk.ru/showpost.php?p=574471&postcount=983

[Shockwav3]

Just to let you guys know: both of my Pentagons are up and running
Appears the first one was suffering from a flakey PSU (I was using my PC's) and C7 being soldered in the wrong place and the second one must have had a cold solderjoint, as it magically started working once I re-soldered every chip.

So ... what are good measures for testing that everything is working alright?
I've tried loading and running a demo from disk (Alienate by Hooy-Program), loading and saving a program from/to tape, testing the INT and RAM using Test 4.30 (which also confirms the tact at 171680 cycles).
One more thing that comes to my mind would be testing a demo that uses covox sound and that would be pretty much all.

[solegstar]

testing the INT and RAM using Test 4.30 (which also confirms the tact at 171680 cycles).

How many INT cycles? "Too short", "Too long" or..?

One more thing that comes to my mind would be testing a demo that uses covox sound and that would be pretty much all.

If you connect DAC (R-2R or 8bit chip) on "printer port"... only after that you can hear covox sound.

[Shockwav3]

Test 4.30 gives me 32-28
TestInt reports 40-36
Hidden part of the Insult Megademo shows just fine.

[solegstar]

Test 4.30 gives me 32-28
TestInt reports 40-36
Hidden part of the Insult Megademo shows just fine.

there are no more questions) pentagon work properly. my congratulations!!!

[solegstar]

try to replace silicon diodes on germanium diodes or Schottky diodes with small voltage drop on P-N transition. The R11 resistor shall be 10 kOhms. On scheme (http://sblive.narod.ru/ZX-Spectrum/P...28VMGPage3.jpg) used /OIRQ signal and germanium diodes Д9.

[Shockwav3]

try to replace silicon diodes on germanium diodes or Schottky diodes with small voltage drop on P-N transition. The R11 resistor shall be 10 kOhms. On scheme (http://sblive.narod.ru/ZX-Spectrum/P...28VMGPage3.jpg) used /OIRQ signal and germanium diodes Д9.

I'll try that with adjusting R11 to 10kOhms, in case that alone won't help, I'm going to drive /RD, /OIRQ and A5 through OR-gates as shown here: http://8bit.yarek.pl/hardware/zx.joystick/kjschem.gif