Acer X34P: Learning about DP / 120 Overdrive and Black Flicker

I bought one; Acer’s X34P. There are people that love them and others that have RMA’d them multiple times.

There are a few caveats; not only does mine work (not right away either) but its in a KVM config which really should be the kiss of death for this to work correctly.

Initially, the display was connected via DP to a KVM (GCS1934), to 4 machines (Win10/Nvidia670M, Win10/Nvidia980ti, Win10/Nvidia1060 and MacOSX). All of these connect at 60hz. All of these display full screen native resolution.

On the 980ti machine, I ramped the display up to 100hz and random intervals I’d get the DP port resetting (brief blackout screen and DP in the upper right corner). It was annoying. Running the overclock made it worse @ 120hz. Games or desktop. A lot of config fiddling led to slightly better results but still flickering. I turned off the audio (which was my plan from the start as I dislike monitor sound).

A lot of reading later, I unplugged the KVM from the mix. Surprise — worked. I need that KVM. So I pulled the segment of DP cables that that particular machine was using. I had the heavy duty (well made) and worked in my non KVM test as the monitor. This cable was really long. I had a few 3 footers that were decent quality. Once I made the run smaller, the KVM and the display worked at 120hz.

There are a lot of thoughts about cable quality — but sometimes length throws things out of whack. I tested this with 10 cables. Only the short ones (except for one) worked.

Now that it works with everything — I love this monitor. First thing I am doing is replacing the base stand with a desk arm.. man those feet are huge.

TROGUE64 – RGCD C64 16kb Cartridge Game Dev Compo 2019

I have been meaning to get involved with one of these compo’s for a while now — so here we go!

This is a rogue-like based game, where player starts at the bottom of mine (which was flooded) to trap all the zombies in the local town. For some reason the player was checking the containment and triggered the drain of the mine. Getting sucked to the bottom, they must make it back to the top before the denizens of the mine or the water flooding back into the mine drowns them…

Trogue is actually a word, it means the drains in a mine. Lucky 🙂

New Year 2019: It has been crazy

This previous year has been extremely busy — the birth of my son, William — has kept my days filled with fun and zero sleep.

I guess like most parents, I have this fantasy that being at home with him means I get some more time somehow for personal projects. Sleep deprivation seems to really mess with my concept of reality.

I woke up a year later blowing out his 1st birthday candle and realized it’s 2019 🙂

So, I suddenly have some time and I am getting back to projects.. no really.. I am

Cat Weasel v4 Running in an XP VM

I’ve managed connect my PCI Catweasel card to pass-though on my ESXi server. I took a bit of work and it’s not perfect (ie no Joysticks or Mice work) but the disk subsystem works great. I added to my server a Multi Function 3 1/2 drive with USB reader, which I thought was a stretch to get working — but in the end it was fine. I also added a TEAC 5 1/4 drive which reads and writes perfectly. The Disk Imaging tools work reasonable well but get stuck sometimes.

I tested some Amiga, Apple and C64 floppies and was surprised how fast they imaged or transferred down.

Stability was an issue initially, but after making some adjustments to processor and memory layout of the VM, XP and the Catweasel card are working great. The other subsystems on the card are not used, but what I really wanted was a multi format disk reader and writer.

Click on the thumbnail to see the desktop in action.

AL-1000: Saved from Oblivion.

The Commodore AL-1000, which is akin to a mini-PET (heavy and boxy). It uses tubes to display the numbers. There was some rough housing with this venerable calculator and it needed some TLC to get going. The calculator was abandoned a long while back and found after someone was cleaning out some buildings near by. A tip from a friend helped save this new addition.

On a Date with Ms. Pacman

I’ve been asked for about a year now to head out to my partners’ brothers studio, where an unhappy tabletop Ms Pacman has been sitting busted.

Three of the ROM’s on the board were loose, one had broken pins (3) and the processor board (z80 :)) cable had been stripped in some spots.

The real monster (after all the chip leg soldering – shaky hands) was the power switches. There are two; one for the main power and the another that shuts off parts of the game if you open the coin door. The coin door one was designed to prevent kids from sticking their hands in (if it was left open) and getting intimate with CRT discharging. This switch was so bad it was intermittent at best and kept the machine resetting or off (Not off, just the screen shut down).

Suggestions to others: remove the power switches, they are often crap and busted. then go to the fuses, then re-seat the chips. I should have imaged the ROM’s for a backup.

Oh, table tops are awesome — they are convertible. Flip the internal latches and it swings open.

C64: First 8K Generic Cart

As of today, I made my first C64 cartridge. Most of my programming experience is either through floppy or tape based storage. Making a ROM/BIN file and getting it to work was a tougher than I thought. It was more problems with getting a CRT/BIN combo into the tool chain for testing in an emulator, than to actually make a cart and burn an EEPROM. My first cart is nothing amazing, but it pumps out the INC $d021’s.

C64 NIC: Networking on the Commodore 64

I have a few of these guys in a box, mostly for a long term project that I have been working on. I haven’t had much need for them in the last while as VICE has great networking support.

Lately, it has been crazy fun doing TCP/IP communications with the IP65 networking stack. It’s definitely the way to go for making games that need updates from time to time. Write a bootloader that fits in ROM, place it in the socket and off you go. Yes there are issues with a ROM update later on and that you need to have internet access to enjoy the software. Oh well, its a connected world.

One of these guys needs to get crossed with a 16 MB REU. It would be pretty awesome combo. But for now; this thing is still great.

Amiga 2000 / PC Side Upgrades and Repairs

YURIPCMK2x300Yuri’s Goldfinger card, was having some really strange issues — the PAS AUDIO 16 ISA card that was handling audio for the PC side of the computer was also throwing up in certain DOS configurations. The PC on a card PEAK650VLB2 ended up having BIOS corruption after finding a surface mounted resistor had popped off the card. I gave the card a hard look to see where the resistor was from, but it was next to impossible (the card has hundreds of them). If I had a card to compare against maybe, but its easier to pickup a new one. I managed to locate a replacement far far away for super cheap, which was a surprise when it actually powered up.

The card itself is a newer revision, PEAK650D4 type. The headsink for the chipset required some ‘Mazzolla’ to let the processor heatsink slide into its spot. Whomever manufactured the card put the chipset heatsink on the wrong way. Oh well, I got it in (the processor) — with a little rough housing.

I found a local source for a Soundblaster  AWE64 ISA card. All my starting computers that were PC’s had a variety of soundcards, but never a real Soundblaster. I have a collection of the best soundcards of that era (PAS16, GavisUltraSoundPro, TurtleBeachOriginal) as I did audio work at that time. Soundblaster, though is the best card for video game support — which is why this machine is setup. Other than some funny hardware warnings from Creatives’ plug and Play loader — the card works perfectly. Doom2, Dune2 and CNC all love the config now. The PAS16 would get one or crash another.

Getting the right memory configs in DOS6.22 is such a pain. I conveniently forget the massive pain it was in that era to actually get different games to work.. let alone mixing in mouse and a LAN stack for multiplayer action. Memories of NE2000 and IPX, smells like burning toast.


C64: (teaching old dogs new tricks)


I will say three things about Michal’s c64 programming series: his video’s are fun, fast and packed with information.

I am not a bad assembly programmer, -not the best- but not bad. After watching a few seasons of his videos I started to notice subtle ways to improve my style.

There is a common problem with programmers in general. We need a problem to learn something and anything else in the way gets filtered out. So when Michal’s videos had Basic language sections, I was going to skip them for more interesting topics. I’m glad I didn’t. In fact, a few assumptions (over the years) that I’ve made were proven wrong (about the DATA statement in Basic). Suddenly, I wanted to know if there were other things that I was mistaken about.

Before you complain about it being a paid service; its worth it. Its a good refresher and its interesting to see how he works and his logic. He even manages to get automated testing into the mix.. on an 8bit 64k machine. It’s worth the money and I’m happy when a new episode appears in my inbox.

Amiga + Gotek Floppy Emulation

GotekFloppyEmuWell, the Gotek Floppy has been running relatively well. I dragged a bunch of of ADF’s over to a USB stick (with the menu ADF) and it was pretty much working.

I wish the drive face was at least the right size, to fit flush with the enclosure.

I’ve noticed a few problems here and there with some images; also when there isn’t a USB mounted you get a dead icon on the screen for DF0:. I’m testing a variety of floppy images to put it through the paces. I’ve even made a 3.x Emergency ADF and mounted it.

Amiga 1020 5 1/4 Floppy

Amiga1020I’ve had this one in the basement for a long time, never got it working correctly — until now.

Device = trackdisk.device
 Unit = 2 /* first external unit */
 Flags = 1 /* important ! */
 Surfaces = 2
 BlocksPerTrack = 11
 Reserved = 2
 Interleave = 0
 LowCyl = 0 ; HighCyl = 39
 Buffers = 20
 BufMemType = 1 /* or 3 if you run OS 1.x */

Drive is formatting 🙂

A3000D: Random problems finally sorted.

A3000DI upgraded the A3000D with as much stuff I had in storage and few things off of Amibay. After the RTC fixes, HD LED, misc capacitors and replacement floppy I started adding:

GVP IOEtxender (high speed serial an LPT)
BigRam+ (256MB fast ram)
X-surf-100 (network adapter)
Picasso IV with Audio add-on.

The mainboard initially had standard original custom chips (Ramsey 4, Super DMAC 2, Super Buster 7), which got upgraded to Ramsey 7, Super Buster 11. The WDC SCSI chip on the board was a type 04, which its now a type 08. The fast ram (8MB) was a mixed bag of chips and speeds, but they were all static column type which gave the 10% boost. A newish SCSI drive, 8.5GB which replaced a 200MB drive I found in it.

After installing the base OS, and the networking adapters — problems started to come out. Random checksum errors, enough to corrupt the drive a few times. After a few Disk Salv 4 sessions, I kept checking all the parts. I removed the all the cards but the NIC and still got the errors. I had already upgraded the SCSI chip, which should have stabilized the SCSI chain. I checked the diode at D800, its direction and the voltage inside, on the port and through the cable. I double checked the drive to supply power and termination already and everything was terminated correctly.

Removing the BigRam+ for the remaining tests yielded a slightly more stable system, but the errors started coming back. I think I formatted the drive and installed the OS  too many times, I was clicking and editing things while doing other things without thinking about it. The system would last a little longer with each tweak, then it would just get worse. It’s really strange, my Amiga 2000 with a GVP Accelerator installed without a problem, ever!. For some reason I thought the A3000D was going to be open shut case once the repairs were done. This was not the case.

I had a batch of 514402AZ-60 ZIP modules, which are supposed to be static column RAM, but are not. I thought maybe a homogeneous block of chips might help the stability, also I had 16MB’s of it. A 10% fast ram speed loss for system stability I can live with.

I was reading though other people’s issues, when an article about problems with Ramsey/Super DMAC revisions. They are paired chips, according to many. Others state that sometimes the mixed bag works. I pulled my Ramsey and returned the original Ramsey. Well +1 to the Paired Chips theory, system is rock solid.

A3000: Amber’s Variable Resistor

AmberSyncThe A3000 was randomly deciding to output a proper signal to VGA. I know the monitor I was using (or LCD) was capable of displaying in the right frequency range — but only 1 out 5 power ups would yield an active VGA display.

It was easy to tell when a good powerup would occur, as the VGA display would have some very dark gray vertical lines in the black screen.

Putting a scope on the VGA port showed signals on the RGB and H/V sync pins. Each signal looked okay from a casual inspection (comparing a good powerup versus bad). I checked all the pins around the board to see if there was anything missing on the way in to the VGA… but if the VGA signals looked ok either way (on or off) then it must be a very slight signal variance.

The variable resistor on Amber was very touchy. I had a spare that would work, but I re-soldered the old part just to be sure.. it looked like there had been some work previously done on it. Using a metal screwdriver, I noticed the signals jumping around on the V/H pins. Removing the screwdriver would change the signal as well. I had a plastic screwdriver for altering TBC signals — TBC’s did the same thing with metal tools.

Without worrying about over turning the resistor, I swung it pretty far. I then noticed it was loose. After screwing it down for a while, I felt a bump and then the signal seemed to tune easier. I found the sweet spot, rebooted — VGA stable. I waited 30 minutes with the machine off tried again, still good. Off for two days, powered it up — still working fine. Nailed it 🙂

A3000 Clock Spagetti

clockspagettiThe A3000 I am working on has a few issues, the first was NO CLOCK reported by SYSINFO and other tools. This is of course true, as the battery had been removed after I unboxed this machine. After the battery was removed, the gross damage to the motherboard by the leaking battery is obvious in the picture to the left.

I cleaned the board as best I could, noticed that the back of the board was OK… somehow. I didn’t want to mess with the area around where the battery was, as there was a lot of damage there and I didn’t want to add to it.

I temporarily jumped a CR2032 in a holder (the BLACK and RED leads),  to get get the clock powered. I then went into the Prefs and set the time. After a reboot, the system recognized the clock — but the time had not incremented while the machine was off.

The clock crystal was not outputting anything which ended up being a connection from C192 to R193 was lost. I jumpered them with the BLUE lead. After checking all the connections out I found C190 was not connected either, so i jumpered it with the YELLOW lead.

clocksignalAs opposed to a flatlined output from the Oscillator, I got a signal.

I loaded up the system, reset the time powered it off and waited a few minutes. Turned it back on and the time was in sync. I went out for dinner, had a few drinks and came back — powered it up, clock still working.

I know the clock is not so important in this day and age, as NTP can handle the time sync. But who wants a broken machine?

Next on the list.. Display warm up… sometimes the display shows something, sometimes it takes a few power cycles to get it going…

A3000T: New friend :) Mouse and Keyboard :(

Well, I’ve got 2 Amiga 3000’s going, which means they are in 30 pieces on my workbench.

One is a A3000D, with 2MB/16MB/256MB of various RAM types. I’ve managed to get the latest custom chips on the mother board, all but a Super DMAC04. This machine has been in the basement for a while, with battery damage and a few blown caps. It now boots with 3.1 KS and 3.9 OS. The VGA port takes a little while to warm up, the machine needs to be on for about 2 minutes and then power cycled to get VGA back up. Working out the VGA is the next major task…

Number Two, is an A3000T. I can’t think of a heavier computer. This one booted right off the bat, but for some reason the mouse and keyboard wasn’t working. Scouring the motherboard for bad components/traces, reading forums and voltmeter came up empty. I could understand why one port would go or just the keyboard… but all three ports? The system would boot to a temp OS disk I mounted, but as soon as the OS was up (or even the KS screen) not keyboard or mouse action. Weird. Even the 3 finger salute on the KS screen..

There is something to be said for computers with a keyhole embedded in its face plate. Yes, when in the locked position — no keyboard or mouse. Funny… I have never had this happen in the last 30 years. The trick that works on old bike locks (Plastic Bik pens, cut in particular way) works on the lock in A3000T. I can see that the lock has been man handled before. When I opened the lock, everything started working.

This machine needs a lot of TLC. The case wasn’t intact, missing the front and top sections. It also has older chips than my A3000D. I was kind of hoping this one was going to have a late revision Super DMAC… oh well 🙂




A2000: Business up front, party out back..

YuriFrontPanelWell it might be the other way around. Recently I picked up a really thin/small USB 4 port hub, which would get in the way (the way it was wired). But I found a use for him. The PC side of Amiga Yuri needed USB ports and other interface options.

So we begin with the left side front panel, which are two toggle switches (with LED’s :)), one momentary switch and a USB hub. The first toggle is a red LED, which is tied to the PSU for the system. When the system is powered down, the red LED lights up. The second toggle is for the Amiga’s 030 Accelerator, which can now be disabled (off) or active (on), the blue LED in this toggle is for the PC side’s HD light. The small momentary switch is to reset the PC Card, as this doesn’t have a power off/on so it would be handy to reboot it without having shutdown both sides. The USB is wired into the PC side, to allow me to mount CD-ROM, flash drives etc for the Goldfinger card.

YuriRearPanelIn the rear of the machine, I mounted all the ports that would be needed (but not all, there is not enough backplane for it :)).  From left to right:

1. Amiga SCSI2SD card drive (swap out cads for different setups DH0:)
2. Amiga Rapidroad/Xsurf combo (USB and Ethernet)
3. Ventilator Fan (strong, lots of hot stuff in there)
4. PC Audiotrix Pro sound card
5. PC IDE2CFCARD interface (different PC configs Drive C:)
6. PC Serial Ports
7. PC Goldfinger Card (PS2, 2xEthernet and VGA)


YuriFrontCaseI left the machine running doing some speed tests to generate heat and everything worked out great. I was monitoring the BIOS for the PC side, thermal sensors didn’t rise as much as I expected. I put thermal sensor on the Amiga MB and Accelerator and it’s temperatures only went a few degrees higher than normal readings. Amazing!

The case needed some cutting and with a dremel, which turned out pretty clean. I’ve always wanted to use that space for something, it always seems like a waste of front panel. Just have front panel power is great, as when I was done I put it in my rack, along with a cheap VGA KVM switcher. So both machines use the same monitor 🙂






A2000: Goldfinger – Xsurf/Rapidroad – SCSI2SD

Hardware Changes and Adjustments (Part 1 of 3)

  1. OS3.9 on DH0: and OS1.3 on DH1:  — Easy booting to each OS.
  2. Replaced the AMIGANET 1.1 / Hydra with and XSurf100 using AmiTCP. Networking is going great in OS3.9
  3. Combined with XSurf100, is the RapidRoad USB addon card. I’ve mounted a variety of devices — setup was a little weird but once it was working, it worked well. Was impressed by the device support.
  4. Got 2 more GVP 4MB ram modules, Chip: 1MB, Extended: 16Mb on the 030
  5. With the 030 Processor Upgrades to 50MHrz and an MMU, FASTROM Enabled.

I had two Primary machines in Artschool, one was Yuri and the other was a Pentium 133 in the school lab. Years later, after working at the same school and passing the baton on to another — someone showed up to a house party with my Old computer in tow. The machine had been in storage and I had a disk backup of the machine. The IDE drive was cooked and there was some damage to the machine’s PSU. It was a gift I wanted to get going.

In order to get the rest of the changes completed, I needed to pull out the motherboard from the case.

PEAK650VLBI found a Goldfinger card on Ebay, which if you want to know more about them: Here and Here. The card is a PEAK650VLB, which is basically everything you need for a computer (minus a soundcard) on one ISA card. It’s like the bridgeboard for the Amiga, but a million times better. It has a 1ghrz Pentium 4, with 64 megs of ram, EIDE controllers, USB and two NIC’s (100mb). It’s a 16bit ISA Card, with a PCI edge near the front of the card. So it can be placed in either BUS system type.

The Amiga has ISA slots — 2 8bit and 2 16bit, which are dumb slots. And the PEAK650VLB is a card that takes over the ISA BUS, so I could add a few ISA cards into the mix. Last time I checked, the only 8bit ISA cards I had were a modem and PS2 port. Everything else (including my P133) was 16bit.

The Amiga 2000’s 8bit ISA slots can be upgraded to 16bit, if you solder on the extra connector. The PEAK650VLB is a monster full sized card (the heat sinks add to the girth), but because of the PCI connector it will not fit in the stock 16bit card slots.. only the outer slots (8bit) have the space to take the card. So it looks like some soldering is needed.

*I desoldered two of the 16bit ISA card slots from my old P133*.

ISAI did not know what I was thinking when I did that. It was such a brutal exercise initially. A solder sucker and iron took forever to get the first slot, which I will never get that time back again. Even after the solder comes off, trying to find the pins that had a minuscule amount of solder on them (preventing the slot from coming out) was impossible. Solder wick was faster on the second attempt, but still the connectors were impossible to pull out from the sheer amount of pins.

If you ever have to pull out a lot of stuff from a board, something like a rework station is needed. Sparkfun has one: 303D. There are moments in life where you think: Where was this 10 years ago? The third slot and the remaining two came out in a fraction of the time. I wish I had a 303D from the beginning. Awesome! Chopped off the 16bit part of the slot and solder to the Amiga.

So with the two slot’s upgraded, we get the next round of upgrades:

  1. PEAK650VLB with Pentium4 1Ghrz, 64MB ram.
  2. AudioTrix Pro Sound card (ISA) with Soundblaster Emulation
  3. Compactflash Drive to EIDE Converter. 4GB storage on CF.
  4. Custom made bracket for the CF Card, to load out the back of the Amiga
  5. DOS 6.22, Windows 98 and Debian Linux installed.

DUNE2DOS 6.22, was a pain to get working with a more modern system. PNP and BIOS configs to get the sound card to work in DOS was something I pleasantly forgot.. but after a little brain massaging it worked. I installed DOOM3D and DUNE2, both of which I still have the original install disks. I copied over my original school/work backups from an ISO I made years ago. FRACTINT still looks great 🙂

I still need to make a backplane for the USB and Serial Ports for the PC side.


SCSI2SDThe last of the upgrades is SCSI2SD. This device basically is a hard disk replacement for computers with SCSI-2 interfaces. Insert an SD card, connect the device to a USB computer and configure how many SCSI devices and their sizes. Connect the device to the computer’s SCSI chain, boot and partition. It was pretty easy.

I did some speed tests before and after. I was a little sad with the results:

Seagate HAWK: 2.1MB/sec
SCSI2SD: 1.6MB/sec

Maybe a different card choice or configuration might improve the stats. But the real amazing thing about this is that you can take the SD card out, back it up as an image and put it back. Multiple cards, multiple setups. I’m going to try and workout why the transfer loss, but at the moment I am happy with what I couldn’t do that easily before the change.


GVP SCSI 4.15 Update for Yuri

GVPROMIt took a little while to locate EPROMS that would work for a ROM replacement/upgrade, but success! I’ve made a backup of my original ROM (ver 4.13) and burned a new one.  I picked up the ROM file from which is a great Amiga resource. Ralph Babel, thanks for the update 🙂

After Brutalizing myself with the PCB45, I found a Canadian company selling UV Eraser and GQ-4X combo. Wow, the difference in setup is night and day… I know that this model doesn’t have clock support for the weird Commodore ROM’s, but as a general purpose ROM burner, this one was pretty easy to use.