WarpSE: 25 MHz 68HC000-based accelerator for Mac SE

[JTRetro]

@JDW -Thanks for the video-very informative! I haven't noticed any problems with the mouse pointer in my test, but then again, I've been testing item that I tend to use in my person setups to see how they perform....I actually never thought to try Mac Paint, although I have run Photoshop 1 with the WarpSE (Just a basic test, displaying a JPEG picture). I will try this test later today myself and see what happens.

I did manage to get all of my SCSI peripherals running at once, namely my 100Mb Iomega drive and my CD-ROM. Additionally, I had no problems accessing either drive. Again, this SE is running O.S. 7.1 and is running off of the original internal 20Mb SCSI drive, at least in this instance:

 

[JTRetro]

However, this morning when I went to fire up my SE with the same SCSI devices attached, I once again got a checkerboard:

So I shut the machine down, and then turned it back on.....and it fired right up. I'm still using the WarpSE.GW4410A.1.0 firmware that @Zane Kaminski sent me last week, and so far it seems to work well, other than a few minor hiccups.

I do also have to mention that I haven't quite had the uptime that @JDW has in one session with my SE- I think that the longest that I've kept it on in one shot with the WarpSE is a little over an hour. Guess I'll have to change that!

 

[phipli]

I found it!
@Zane Kaminski @JDW

It is a gestalt lookup.

Note "More Information" in the corner describes how it works. More information also mentions that depending on the age of the computer there are also a couple of different gestalt selectors that give similar results. But anyway, calling something like...

With the CodeWarrior headers I'm using, it seems to be pclk and bclk for processor and bus speed respectively.

 

[Zane Kaminski]

Hmm... testing has not been going as smoothly as I think we all have hoped. It's super frustrating that I cannot replicate the problem when starting up. At least 0.6b has been stable for @JDW. I am gonna have to clear off my workbench at home (as opposed to at the GW lab) and get all my Mac SE machines set up to try and replicate the issue. So far I have only been using one SE board on the lab bench at GW with no other original Apple stuff attached, only my power/video adapter and a SCSI2SD 4.2.

However, this morning when I went to fire up my SE with the same SCSI devices attached, I once again got a checkerboard:

@JTRetro Can you try updating to 0.6b and see if you can get it to happen again? I've attached the file since the old .zip archive with the 0.6-series firmwares has been lost up in the thread history by now.



@JDW When the Mac displays bars after boot, that means that it hasn't gotten to the screen clear routine where it puts up the 50% black/white dot pattern. This is either because the WarpSE is not releasing the system reset line or because of some other issue during executing sending the CPU off track before it can clear the screen. I have two things I'd like to ask of you, both of similar importance.

Firstly, can you confirm that 0.6b has continued to work? One weekend of testing should be plenty since the issue has manifested very quickly on the other versions.

Second, can you reflash 0.6c and check the reset line on the WarpSE with a multimeter next time you get the bars? The reset signal can easily be probed at any of these points shown here:

You can probe either side of the resistor or the pin of the little SOT-23 package chip. All of those should give you a good view of the reset signal voltage. If it's stuck low (under 0.8 volts) while the bars are displayed then the WarpSE is probably not releasing the reset line. Otherwise there must be some issue with the RAM/ROM/PDS logic. Also, once the bars are displayed and you probe the reset, if it's high (over 2.0 volts), does pressing the reset key on the programmer's switch do anything? What about pressing interrupt?

Once we determine whether or not it's because the machine isn't coming out of reset I can try and pinpoint the problem further.

I am also gonna be making new prototypes with a new board revision that will allow the WarpSE to be programmed outside of the Mac, plus a few other small fixes. Everyone is gonna get another one so that we can compare and see if the issues have something to do with the WarpSE hardware.

Thanks for your help @JDW @JTRetro @techknight!!

 

[JDW]

...I cannot replicate the problem when starting up...
...So far I have only been using one SE board on the lab bench at GW with no other original Apple stuff attached, only my power/video adapter and a SCSI2SD 4.2.

I've been testing exclusively in a real SE rig (SONY PSU with blue POT voltage adjustment turned up high but not max), both with the motherboard outside the chassis (using my extension cable) and while inside the chassis. I also am using a Main Wiring Harness that incorporates the 16AWG wire upgrade. All the tests I reported yesterday and all my tests done today were with the SE motherboard inside the metal chassis.

It very well could be that is making the key difference here.

@JTRetro Can you try updating to 0.6b and see if you can get it to happen again?

We absolutely need more of our WarpSE beta testers to do testing overlap using the latest firmware (a/b/c/d). Toward that end, I've sent a message to Petar Puskarich about an hour ago to ask if he might be able to make time to flash the recent firmware, starting with "d".

I also warned Petar about what happened to @techknight and the tiny USB connector breaking off, advising him to use an extension cable to move his SE motherboard outside the chassis when flashing new firmware. I pointed him to the right post in this thread with links to all 4 firmware versions, so he won't need to waste time reading the entire thread to be brought up to speed. I was sure to tell him to watch keep a close eye out for (1) Vertical Bars at Power-ON, and (2) Arrow Pointer oddness. I linked my video yesterday for him so he can see all the details.

USER=5]@JDW[/USER] I have two things I'd like to ask of you, both of similar importance.
Firstly, can you confirm that 0.6b has continued to work?

Zane, I've spend many hours testing 0.6b firmware, including today. It has never shown the bars at power on. Not even once. It's rock solid in terms of NOT showing those bars. (And by bars I mean the bars I normally see when the USB cable connects WarpSE to a PC.)

The ONLY time the bars appear (outside a flashing session) when using 0.6b firmware is when I deliberately invoke them by way of pressing-and-holding INTERRUPT and then switching on power. And that is correct behavior. I then release INTERRUPT, the bong sounds, and the SE boots in the stock 8MHz configuration.

The one and only problem I have had with 0.6b is what I showed in my video yesterday; namely, while moving the mouse, the arrow pointer (or paint brush or whatever is moved by the mouse) will suddenly FREEZE and then a fraction of a second later it will jump to another, distant location on the CRT. It only does that every several seconds. In other words, once that freeze-and-jump happens, it doesn't happen again until several seconds later (assuming I am still moving the mouse). If I am not moving the mouse, it will not do anything bad or unusual, of course.

Tomorrow on my lunch break, before I do any flashing, I will move WarpSE to my stock SE motherboard in order to reconfirm the Mouse Movement problem remains. That test will be proof that my motherboard is not at issue here.


I spent more time testing many things today, including running Apple Personal Diagnostics:

Like I said, everything except the mouse movement is working perfectly.

SCSI performance is for the most part not enhanced by WarpSE when using 0.6b that I can see. To help show that I have prepared SCSI Director Pro 4.0 benchmark results showing that firmware 0.6b slows down SCSI to such an extent that WarpSE's SCSI performance matches the stock 8MHz SE (numbers below within margin of error):

=============================
WHILE BOOTED INTO SYSTEM 7.1
=============================

WarpSE ENABLED:

WarpSE DISABLED (stock 8MHz SE):

===============================
WHILE BOOTED INTO SYSTEM 6.0.8
===============================

WarpSE ENABLED:

WarpSE DISABLED (stock 8MHz SE):

Below are some past test results I made (for sake of comparison), all done while booted into System 6.0.8...

Levco SuperMac SPEEDCARD (16MHz 68000 & 68881 FPU):

SpeedCard with MacSD (instead of BlueSCSIv1):

SpeedCard with BlueSCSIv2:

Stock 8MHz SE with MacSD:

Stock 8MHz SE with BlueSCSI v2:

If my fellow beta testers post any such benchmark results, please make mention of your OS, because it matters. You get higher scores when booted into System 6.0.8, as compared with 7.1. (And I can tell you that S7.5.5 yields even lower numbers than S7.1.)

Second, can you reflash 0.6c and check the reset line on the WarpSE with a multimeter next time you get the bars?

Because the only PC I have access to is at the office and tomorrow is Monday, I will test that tomorrow on my lunch break and report back at that time.

But just to confirm, you want me to test "c" and not "d", correct?

 

[techknight]

I've been testing exclusively in a real SE rig (SONY PSU with blue POT voltage adjustment turned up high but not max), both with the motherboard outside the chassis (using my extension cable) and while inside the chassis. I also am using a Main Wiring Harness that incorporates the 16AWG wire upgrade. All the tests I reported yesterday and all my tests done today were with the SE motherboard inside the metal chassis.

It very well could be that is making the key difference here.


We absolutely need more of our WarpSE beta testers to do testing overlap using the latest firmware (a/b/c/d). Toward that end, I've sent a message to Petar Puskarich about an hour ago to ask if he might be able to make time to flash the recent firmware, starting with "d".

I also warned Petar about what happened to @techknight and the tiny USB connector breaking off, advising him to use an extension cable to move his SE motherboard outside the chassis when flashing new firmware. I pointed him to the right post in this thread with links to all 4 firmware versions, so he won't need to waste time reading the entire thread to be brought up to speed. I was sure to tell him to watch keep a close eye out for (1) Vertical Bars at Power-ON, and (2) Arrow Pointer oddness. I linked my video yesterday for him so he can see all the details.



Zane, I've spend many hours testing 0.6b firmware, including today. It has never shown the bars at power on. Not even once. It's rock solid in terms of NOT showing those bars. (And by bars I mean the bars I normally see when the USB cable connects WarpSE to a PC.)

The ONLY time the bars appear (outside a flashing session) when using 0.6b firmware is when I deliberately invoke them by way of pressing-and-holding INTERRUPT and then switching on power. And that is correct behavior. I then release INTERRUPT, the bong sounds, and the SE boots in the stock 8MHz configuration.

The one and only problem I have had with 0.6b is what I showed in my video yesterday; namely, while moving the mouse, the arrow pointer (or paint brush or whatever is moved by the mouse) will suddenly FREEZE and then a fraction of a second later it will jump to another, distant location on the CRT. It only does that every several seconds. In other words, once that freeze-and-jump happens, it doesn't happen again until several seconds later (assuming I am still moving the mouse). If I am not moving the mouse, it will not do anything bad or unusual, of course.

Tomorrow on my lunch break, before I do any flashing, I will move WarpSE to my stock SE motherboard in order to reconfirm the Mouse Movement problem remains. That test will be proof that my motherboard is not at issue here.


I spent more time testing many things today, including running Apple Personal Diagnostics:

View attachment 18131

Like I said, everything except the mouse movement is working perfectly.

SCSI performance is for the most part not enhanced by WarpSE when using 0.6b that I can see. To help show that I have prepared SCSI Director Pro 4.0 benchmark results showing that firmware 0.6b slows down SCSI to such an extent that WarpSE's SCSI performance matches the stock 8MHz SE (numbers below within margin of error):

=============================
WHILE BOOTED INTO SYSTEM 7.1
=============================

WarpSE ENABLED:

View attachment 18122

WarpSE DISABLED (stock 8MHz SE):

View attachment 18123


===============================
WHILE BOOTED INTO SYSTEM 6.0.8
===============================

WarpSE ENABLED:

View attachment 18124

WarpSE DISABLED (stock 8MHz SE):

View attachment 18125


Below are some past test results I made (for sake of comparison), all done while booted into System 6.0.8...

Levco SuperMac SPEEDCARD (16MHz 68000 & 68881 FPU):

View attachment 18126

SpeedCard with MacSD (instead of BlueSCSIv1):

View attachment 18128

SpeedCard with BlueSCSIv2:

View attachment 18129

Stock 8MHz SE with MacSD:

View attachment 18127

Stock 8MHz SE with BlueSCSI v2:

View attachment 18130

If my fellow beta testers post any such benchmark results, please make mention of your OS, because it matters. You get higher scores when booted into System 6.0.8, as compared with 7.1. (And I can tell you that S7.5.5 yields even lower numbers than S7.1.)



Because the only PC I have access to is at the office and tomorrow is Monday, I will test that tomorrow on my lunch break and report back at that time.

But just to confirm, you want me to test "c" and not "d", correct?

how are you even getting color pictures/screenshots from an SE?

 

[JDW]

how are you even getting color pictures/screenshots from an SE?

I save the test results in SCSI Director Pro 4.0, then Shutdown my SE and remove the SD card from my BlueSCSI. I put the SD card into my modern Mac and launch Basilisk II. I then launch SCSI Director Pro and open my saved results files. I then use my modern Mac to make a screenshot of that and Paste into my reply here on TinkerDifferent.

 

[JTRetro]

Soooo......I installed firmware .6b onto my WarpSE accelerator and then fired up the machine with, once again, the caddy-loading CD-ROM (Apple CD300) and an 100Mb Iomega SCSI drive attached to the external SCSI port, and the machine fired right up!

I also did a test similar to what @JDW did with Macpaint; i did a quick and dirty perspective drawing, all while having the SE control my CD300 to play music in the background. Here is a video of the results:

As you can see, I did not have the issue of the mouse jumping around. So I think we are on the right track, at least in my case!

 

[JDW]

@Zane Kaminski
You want me to test 0.6 "c" and not "d", correct?

 

[Zane Kaminski]

@Zane Kaminski
You want me to test 0.6 "c" and not "d", correct?

Yes, since this is the only one you could get to boot at all. Then if the screen is showing bars, check the reset signal voltage as I described in my earlier post. I may however have a new firmware update within the next 1-2 hours with a possible fix for this issue. Since the new initialization sequence in 0.6b is working well for you, I focused my attention on the fast-to-PDS bus bridge. I discovered a timing problem in the PDS master controller affecting communication with the VIA chip. I also increased the data setup time at the fast processor by an average of 20 nanoseconds. This may affect the ability to do back-to-back PDS operations and make the WarpSE even slower than stock during slowdown, so it could break floppy compatibility, but it will ensure that data coming from the PDS has sufficient setup time at the fast CPU. If this fixes the problem then we can tune it further to solve both sides of the problem.

 

[Zane Kaminski]

@JDW Hmm, well finally I am able to replicate the problem and this made it much easier to find! New firmware version 0.6e is attached which should fix the problem. If you have time, try this instead of 0.6c or 0.6d.

Thanks again for all your help testing!

 

[Zane Kaminski]

Hmm wait a minute! Now it's not working for me again... Let me look into this further.

 

[Zane Kaminski]

@JDW Never mind lol I think I just forgot to unplug the USB cable, which forces the Mac into reset hahahah. 0.6e I just posted should be good!

 

[JDW]

@Zane Kaminski
Somewhat bad news.

First, I moved WarpSE to my stock Apple motherboard (I had been using my SE Reloaded motherboard all weekend), and removed BlueSCSI entirely so I could boot of a MacPaint v1.5 floppy. I repeated my test that I showed in the video. No problem with the mouse. But I lacked time to test more extensively than that. (On my lunch break.)

I then flashed 0.6c (prior to reading your recent posts), and at first power-on there were no bars. So my oscilloscope measurement at the pads you specified has no meaning. But here it is anyway (ignore the spikes because I'm using a very long probe GND wire):

I then flashed 0.6e, and I got the bars while probing with my scope! Goodie! But as you can see, the test point stayed high. (The first part going from GND to HI is when I flipped on power to the SE.)



While everything was connected and running, I then pressed the RESET button and got a temporary GND...



I then pressed INTERRUPT but the waveform remained HI.

Lunch break over. :-(

 

[ppuskari]

I just did rev D firmware and it was FAST! when it worked. Had to reboot about 20 times to get one good boot

C firmware did the same thing but I didn't test it much. BOTH gave the checkerboard pattern on the screen most boots.

I am just now running B and it boots properly to 7.1 on a external scsi Zip 100 (Epson drive)

Will get the sound apps and speedometer apps on it tomorrow.

For now B seems to be pretty good. I had been running "fastscsi" firmware and it was doing quite nicely actually. What seemed to matter for me especially since I"m using external Zip 100 removable drives here is that I use a 68000 compatible removable hard drive driver. I'm currently running Scsi Director Pro 4.0 drivers on all of my test carts. Working great and supports removable inserts and ejects of media when using 1 or 2 zip drives on the chain.

I have noticed in both D when it booted and now as well in B that Localtalk is broken over the serial ports... With "FastScsi" that worked REALLY well.

 

[JDW]

@ppuskari & @JTRetro
Both of you seem to be getting a checkerboard pattern at times. That is something I have never seen, with either of my SE motherboards. Instead, the "artifacts" I see when flashing are thick vertical bars when using my SE Reloaded motherboard, and mostly garbage when using my stock motherboard. On both boards, I have the OLD ROMS and IWM chip (not SWIM).

Also, whenever I use firmware that won't bong at power-on, what I see on the CRT matches what I see when flashing (with USB cable connecting WarpSE to a Windows PC).

So both of you see that checkerboard pattern when flashing and also whenever a given firmware has power-on issues, correct?

I'm guessing the answer to that question is YES. And if so, it then makes me wonder why we don't all see the same thing on the CRT. Is it ROM? The IWM or SWIM? Or something else?

 

[ppuskari]

@ppuskari & @JTRetro
Both of you seem to be getting a checkerboard pattern at times. That is something I have never seen, with either of my SE motherboards. Instead, the "artifacts" I see when flashing are thick vertical bars when using my SE Reloaded motherboard, and mostly garbage when using my stock motherboard. On both boards, I have the OLD ROMS and IWM chip (not SWIM).

Also, whenever I use firmware that won't bong at power-on, what I see on the CRT matches what I see when flashing (with USB cable connecting WarpSE to a Windows PC).

So both of you see that checkerboard pattern when flashing and also whenever a given firmware has power-on issues, correct?

I'm guessing the answer to that question is YES. And if so, it then makes me wonder why we don't all see the same thing on the CRT. Is it ROM? The IWM or SWIM? Or something else?

Well, Not sure if it matters but I have the internal SCSI HD removed entirely in my SE, and only running term power enabled and termination set on the external scsi buss side.

Past that I'm running 4meg of 70ns 1meg sticks if that matters....

Yes checkerboard screen at startup and sometimes it will make it to boot for me on D but that was 1 in 20 boots.

I'm going to try E next...

 

[JDW]

Yes checkerboard screen at startup...

And you see the same exact checkerboard when flashing WarpSE too, correct?

 

[ppuskari]

Before bed finally. The latest E boots properly and seems nice and fast and the mouse even seems to be faster? At least it feels faster in responsiveness. BUT Finder gives and Address Error on shutdown or restart like several of the other firmware did. Also I have noticed that the address error on restart usually happens when I also hear a single click on exiting or launching random acts. It's usually different. BUT if I hear that click, from the built in speaker then it seems the address error on shutdown or restart will also be there. Weird. Also Localtalk is still borked.

 

[ppuskari]

And you see the same exact checkerboard when flashing WarpSE too, correct?

that is correct it is the same during flashing and boot when it's the bad boot.