Damned Audio Glitches !

Badstarr · 01:15, 02 May 13

Well, my GX4000 mod is all (apart from the HxC) "test installed" inside the 464 case and its running sweet apart from one thing... Audio Glitches!

It's not consistent and I have had one example of this happening on my first 464 Plus...Basically when I load up FLUFF the theme tune is a little glitchy, there are weird whistles and pops, it plays pretty much fine the first time the tune plays through but when it begins to play again its like the "squeaky"(if you know what I mean) part of the tune gets interrupted, and its just the bouncy riff and drum sound that continues.

The other one is once on Orion Prime the sampled music at the end of the intro was glitchy like the sample data was corrupted. I power cycled and the problem has not returned at any point but the problem with Fluff remains.

I have experienced the Orion Prime glitch once on my 464 plus but I chalked it up to the rather poorly desoldered DRAM ICs that had been installed by a previous owner, I replaced them and the problem has never returned but that could just be coincidence I suppose.

I have run a full RAM test and all is fine so it doesn't appear to be a RAM issue at least! Also I have checked over the AY and its solder joints and all seems fine there, its really just Fluff that has the problem by the looks of things. I will post a video when its a more convenient time.

Has anyone experienced audio glitches on their plus machine like the ones I have described?

Could this be a GX4000 hardware difference? Some special audio effect that requires a very specific timing or something? I'm inclined to believe its maybe something that has occurred while I've been working on the machine but as its never been able to load disc software until now I have no basis for comparison.

Any theories greatly appreciated!

Bryce · 08:41, 02 May 13

The AY may be damaged, do you have a spare to swap it out?

Bryce.

Badstarr · 14:00, 02 May 13

Drat! I was hoping this would be a known "plus glitch" lol! I do have at least one spare AY now so I guess I will have to roll my sleeves up and switch it round. I will report back when finished, cheers Bryce!

Badstarr · 15:23, 02 May 13

Ok I have swapped in a new AY and Fluff still has the same glitches, Orion Prime has never had the same glitch again, all other in game music seems perfect too BFE Demo plays fine too, intro music for Prehistorik is playing through fine also "groovy samples!" and all.

Maybe this is the lower clock speed somehow, it seems counter intuitive as it just should mean everything happens a minute fraction slower than on a normal CPC (?).

Just wondering if there are any diagnostics for the sound ? I know there is a plus diagnostic cart but right now I don't really have the time to modify a cartridge :-(

Bryce · 15:28, 02 May 13

Have you time to up the clock frequency to the standard CPC speed?

Bryce.

ralferoo · 16:40, 02 May 13

Doesn't the ASIC on the plus support DMA of data into the AY registers? Maybe this part of the chip is damaged somehow...

redbox · 17:05, 02 May 13

Yeah, sounds like a DMA issue rather than an AY one.

Could be bad routines in Fluff - has this ever beeb tested on a GX4000?

Bryce · 20:07, 02 May 13

Is Fluff CPC+ only? Or will it run on a Classic CPC? Does it even use DMA?

Bryce.

redbox · 21:38, 02 May 13

Quote from: Bryce on 20:07, 02 May 13
Is Fluff CPC+ only? Or will it run on a Classic CPC? Does it even use DMA?

Yes, it's Plus only.

Can't really tell if it uses the DMA from the video, but it wouldn't surprise me as the game is running at 50hz so he's squeezing a lot into each frame.

Badstarr · 02:19, 03 May 13

Hi guys been out most of the day so haven't had much time in front of the machine. I did leave the cpc running the Orion Prime screen for most of the day and no glitches at all after many hours. I then switched to Prehistorik 2 and let its sampled music run for a good few more hours (pretty sure it's using DMA but correct me if I'm wrong) all is fine and dandy. Well at least as far as that goes, I power cycled the machine and wouldn't you know it, the MegaFlash went undetected bah blast I thought and decided to spend 10 minutes before I went to sleep finding the cause. Well as you can see its rather late and that ten minutes turned into a couple of hours lol! Found the cause eventually a loose wire going to ASIC pin 133 MegaFlash reinstated, machine in bits again! Won't take long to reassemble it though!

Fluff still has the glitch after the AY swap all other music seems good no glitches, the Zynaps title music is different in that you don't hear a ear obliterating sqeel when it starts like on my 6128, just like on my other plus machines.

As for clocking the machine up, I think I have the right crystal somewhere was wondering if I should maybe dig it out but was worried that the GX main board might throw a spazz?

I was thinking maybe I could make a Fluff GX cart and try it on one of my other GX4000s and see if the glitch appears there too may be worth a shot?

Bryce · 08:10, 03 May 13

If you can try Fluff on a different GX4000 it would definitely help narrow down the cause.

Changing the crystal should have any negative effect on anything else on the board accept SCART output, but I don't think that will be an issue either.

Bryce.

redbox · 09:51, 03 May 13

Hmmm, I just checked Fluff and it's not using any of the DMA registers (&6C00/1, &6C04/5, &6C08/9, &6C0F) that I can see...

Quote from: Badstarr on 02:19, 03 May 13
I was thinking maybe I could make a Fluff GX cart and try it on one of my other GX4000s and see if the glitch appears there too may be worth a shot?

Yeah, I will try this too and see if I get any glitches.

I can only test it with the monitor output on the GX4000, not the SCART though - does this make a difference?

Bryce · 11:51, 03 May 13

Hmmm, that's a posibility too. Maybe the modulator is filtering out higher frequencies. If this is the case, then no GX will play those sounds.

Maybe you should try connecting a speaker directly to the AY outputs to see if the sound can be heard thorugh that.

Bryce.

Badstarr · 14:10, 03 May 13

Interesting! I hadn't considered filtering as a possibility, however, since I can hear the audio the first time the tune runs through it would suggest that the filtering is inconsistent. The only other thing I can think of is that I have a MegaFlash installed, perhaps its some kind of memory overlap?

Bryce · 14:31, 03 May 13

Pop the Flash chip out of the MegaFlash, then you'll know. without the Flash IC it uses no memory. But that also wouldn't explain why it works the first time.

Bryce.

Badstarr · 19:07, 03 May 13

Ok I will try removing the IC, however the expansion bus gremlin has reasserted itself so I guess the problem lies elsewhere in the wiring. So I think now (well next time I have some spare time) is as good a time as any to replace the wiring in the machine with something a little more tidy hopefully I will fix the error during that process. Basically the machine stalls at the Amstrad PLC part of the boot when initialising the MegaFlash ROMs its got to be some loose connection somewhere I will track it down yet! I checked the ROMEN etc etc continuity at the ASIC and checked for shorts, even decided to tidy up the soldering there and no joy so I suspect its a connection on the address lines or data lines but if you or anyone else have a more likely theory then as usual I'm all ears! :-)

Bryce · 20:13, 03 May 13

Sounds like the data lines. Probably D2 or D3.

Bryce.

Badstarr · 04:49, 04 May 13

Thanks Bryce, I will concentrate on those first when I next work on the machine. I should have a couple of hours free after work with a bit of luck :-)

Badstarr · 01:55, 05 May 13

Okaydokay... I had a poke around inside the machine and found everything to be in order, but still the MegaFlash would not initialise so I swapped in my DIY version and it worked! Progress at least, put the proper MF back in and back to the same problem so I had a closer look at the MF pcb and found a single strand of wire had become stuck to it god knows how or what it was shorting out after removing it I tried again and got a screen full of gibberish, I swapped out the Flash IC and this time it worked. So it appears the contents of the ic are corrupt, no amount of hot swapping will allow me to reinstate any ROMS on that IC so I'm waiting on an adapter to arrive so I can check it on my EPROM Programmer.

The other curious thing is that when I use the MegaFlash ROM Manager software I seem to have CPM installed at slot 7(on the good IC). I attempted to change it to PARADOS but it stays stuck as CPM, perhaps IM being a bit thick and missing something simple?

I should also probably mention that running the machine with no Flash IC installed in the MegaFlash did not change the audio glitch at the Fluff title screen so that rules out the MegaFlash as the cause at least, it was a long shot I suppose.

News:

Damned Audio Glitches !