Falcon clock patch V4 series

[Steve]

I also agree that while troubleshooting you should be running at stock clock speeds. 66Mhz. If you think about it ... 100Mhz is the absolute maximum you can run an 060 CPU at, that's right at the brink!! I don't really understand why people seem to feel the need to run at such a speed, surely it could damage the CPU eventually. I run mine at 77Mhz and find the performance to be exceptional at such a speed.

I must also say, unscientifically, that I do not think FalcAMP responds well to a highly overclocked 060 system. I have worked on many Falcons over the years (I used to do Falcon services for people) and on multiple occasions I've seen that overclocking the 060 to very high speed (90Mhz+) puts a lot of strain on the Falcon bus. I've seen Falcamp start to fail at high speeds on more than one Falcon motherboard. You start to see errors like 'DSP timeout' as the system bus becomes swamped.

It may be prudent to expand your testing methodology. Falcamp is notorious for being very sensitive to these things. I personally think that you should run some synthetic tests, like Nembench, dml's FPU bench, Gembench etc. Also for super-high-stress testing, nothing is better than Badmood Doom engine .. make sure to get the latest versions.

(Also - bare in mind that when you have an 060 CPU, you wouldn't want to run Falcamp anyway. Falcamp is useful for 030 mode, as the 030 CPU isn't fast enough to decode MP3. But an 060 can easily decode MP3, so users running an 060 would run @mikro's mxPlay. I also think that the MP3 decoding of mxPlay is superior to Falcamp. I know this information is a bit ... besides the point ... but I'm trying to say theres no point testing some app that you wouldn't even realistically want to use in 060 mode)

[mikro]

Btw, wouldn't make sense to measure the expansion slot's clock before and after CT60 install? Just to make sure that the CT60 itself doesn't introduce some bad side effects.

[mikro]

overclocking the 060 to very high speed (90Mhz+) puts a lot of strain on the Falcon bus. I've seen Falcamp start to fail at high speeds on more than one Falcon motherboard. You start to see errors like 'DSP timeout' as the system bus becomes swamped.

I have a much dumber explanation for you: Falcamp can use non-handshaked CPU<->DSP communication at some points. Faster CPU => faster traffic out of sync.

[Steve]

That makes a lot of sense @mikro ! Thank you for some scientific explanation :)

[Swissfalcon]

I get your point. I will give a try and see if the Falcon behave more stable at 66 MHz and with others programs than Falcamp. In fact, I already did, I explained before that I saw the instability with others programs, like Mxplay.

But on the other side, the exact same system installation with the exact same CT60e and same 14 Mb ram expansion run Falcamp at 100 MHz on the mobo of my friend that have a basic Atari clock patch since 20 hours now:

IMG_3518.jpeg

Doesn't it validate the test as viable?

I was always told that Falcamp is very sensible and tend to crash. And it was in fact the software that crashed after a couple of minutes or when I was lucky after 30 minutes on my machine before I applied the V4 patch. This is why I saw it as the best test case, because it push the limits, exactly like we use Joy to test the clock patch itself, because it's push the limits.

The only thing I can tell now it that before we installed the V4, the machine was rock stable in 060 mode under MINT, except Falcamp. It was literally the only program that could hang my falcon (this, of course, without talking about well know buggy softwares that will crash because a bade code, like pixart or basilisk, that wit let the system in a instable state after quitting). Now, after the V4 fitting, the system is very instable in 060 mode under MINT, almost everything freeze after some minutes of using. Even doing nothing lead once on a total freeze.

[Swissfalcon]

Btw, wouldn't make sense to measure the expansion slot's clock before and after CT60 install? Just to make sure that the CT60 itself doesn't introduce some bad side effects.

This is exactly what I wanted. to do this evening when I come back with the oscilloscope, yes :-) I will measure on the pin without and then with the CT60e and will post the result.

My guess is that the signal will be good enough without the CT60e and as bad as Exxos one with the CT60e. this will explain why we both have crashes. And if I understand Steve correctly, he is running at a much lower speed, this can explain why it behave much more stable.

But it's only suppositions for now, let's wait the real measure :-)

[Swissfalcon]

overclocking the 060 to very high speed (90Mhz+) puts a lot of strain on the Falcon bus. I've seen Falcamp start to fail at high speeds on more than one Falcon motherboard. You start to see errors like 'DSP timeout' as the system bus becomes swamped.

I have a much dumber explanation for you: Falcamp can use non-handshaked CPU<->DSP communication at some points. Faster CPU => faster traffic out of sync.

It's certainly on the edge, yes. But it's in some cases still ok (like on my friend's falcon that stream at 100 MHz since 20 hours).

But I didn't use only Falcamp to validate the fact that my mobo isn't stable anymore after the V4 path fitting. I used mxplayer too. In fact, mxplayer is my acid test for stability since I discovered it (thanks to you). I have a big playlist for more than 700 MP3 and I let it run overnight. Before V4 fitting: always ok, never crashed. After V4 fitting : it crash after maximum 30 minutes (tried 4 times).

So, it's definitively less stable with the V4 fitted than without in 060 mode, using the same long test method.

[Swissfalcon]

(Also - bare in mind that when you have an 060 CPU, you wouldn't want to run Falcamp anyway. Falcamp is useful for 030 mode, as the 030 CPU isn't fast enough to decode MP3. But an 060 can easily decode MP3, so users running an 060 would run @mikro's mxPlay. I also think that the MP3 decoding of mxPlay is superior to Falcamp. I know this information is a bit ... besides the point ... but I'm trying to say theres no point testing some app that you wouldn't even realistically want to use in 060 mode)

Yes, I agree... Except that Falcamp is the only one program that can stream webradios. I really would like a webradio plugin for MXplay :-)

[exxos]

It's certainly on the edge, yes. But it's in some cases still ok (like on my friend's falcon that stream at 100 MHz since 20 hours).

It is not just the accelerator itself which is important here, it is how it interacts with the entire system. Completely different motherboard is a completely different set of tolerances placed upon the accelerator. While it is an indication that the accelerator might be okay, it does not mean it could necessarily run exactly the same speed in a different Falcon.

It would be like driving a car up a hill and comparing it to the same car driving down the hill. One person can say the car goes very fast, the other person can say the car goes very slow. Even though it is the exact same car ! This is due to "external factors".

You also said earlier.

Ok. I tried Falcamp under TOS 404 (with local MP3) too, in 030 mode ok, in 060 mode crash too,

So please confirm that you can play Falcamp in 030 mode (remove the CT60) try Falcamp and joy demo to see if its stable.

I don't think it has been mentioned the power supply you are using either. But indeed trying to debug multiple things at once isn't going to work. So removing the CT60 and doing all tests should be the first step. Of course, your original Falcon power supply should also be recapped to be stable.

I also patched a better 4700uF low ESR on my falcon motherboard as the original is likely bad these days.

Also when you measure the SDMA clock (across the 47pF) please put minimum and maximum voltage reading on your scope!

Also verify all the soldering is good on the V4, in particular where it rests on top of the GAL chip as this gives the board power.


I tried the 68R. Signal is now a lot better. But I cannot really say made any difference to my machine as it was working fine before anyway.

68R.JPG

68R2.JPG

I also tried with and without the 47pF and there was really not that much difference.

47pf.JPG

no_47pf.JPG

Maybe 22pF would be slightly better, but I went around in circles with all these things previously anyway. There's never going to be a perfect solution as every falcon is different.

Currently I have had Falcamp running on the CT60e for a hour & half on loop without problem. Its running over 90Mhz. I will let it run for a little longer. I only have the PSU fan blowing on the 060 at the moment, so its probably getting pretty hot by now anyway.

[exxos]

I looked back at some of your images and the soldering on the centre pad on the V4 in particular does not look good. So please check all the soldering is good.

measure V4 R216.jpeg