Search the Community

On April 30, 2018, Allen Schweitzer sent Ken Lill, Michael White and myself a private email about a possibly undocumented Astrocade motherboard modification. I asked if I could toss this question out to the public. Allen said that that was fine with him. Here is what Allen said to us: Here are the eight pictures that he included in his email. After the pictures are some comments that Allen got from Michael. Allen followed up his original email to me with this telling me that he did get a reply from Michael about that board. Does anyone know anything about these modifications? Adam

Michael Matte, who is working on his Asrocade hi-res upgrade, recently purchased a dead Astrocade system and some cartridges on Ebay. He did manage to fix the console. Over the course of several emails, he described the process to me. With his permission, I've taken three emails, edited them a bit (mostly just the order) and am posting his process here. Hopefully some people might find this useful. Adam ---------------------------------------- Fixing My Astrocade Motherboard By Michael Matte September 20-26, 2017 [Part I] I've been testing my audio/composite video scheme using a motherboard I purchased from E-bay. This motherboard had evidence the custom data chip was running too hot. Now this motherboard has failed. Balcheck is indicating the Z80 and ROM are operating, but all 8 video data lines are in error. Yuck! I checked all the clocks and the power supply voltages to all the chips which are ok. I suspect the custom data chip, but it is outputting the 2 correct clocks. This data chip and custom address chip can be tested on another of my working motherboards. But, there a bunch of other chips that could also make Balcheck display the same error if one of those chips were bad. Don't want to use a "hit and miss" approach, ie, de-solder a chip and then test. I ordered a logic probe to help me troubleshoot this failure. I'm going to use this failure to see if I can develop and document, for future use, a logical and hopefully a quick approach to this kind of failure utilizing Balcheck, a frequency counter, logic probe and volt-ohmmeter. I have an idea that may help facilitate the use of a logic probe on the motherboard when both the Z80 and ROM chips are operating. Depending on how long it takes to fix this motherboard, I'm planning on documenting my "troubleshoot procedure" and may attach it to my Balcheck submittal to the Bally Alley. If it's just a bad custom data chip, then I will be done and will be back on my static RAM project. For me, this troubleshooting approach is long overdue and has higher priority over my static RAM project. I've been pretty lucky with other failures with quick fixes. This particular failure may take a while to resolve. I'm not crazy about having to de-solder a bunch of chips and installing low-profile IC sockets. This is not an easy task when conductor paths are on the top and bottom of the motherboard. You have to make sure the IC socket is completely connected using an ohmmeter. I prefer to de-solder once and then install an IC socket. Even with a 15W soldering iron you risk damaging conductor paths and then have to fix any damage. I'll let you know when I'm back on my static RAM project. [Part II] I bought this Astrocade on Ebay for the extra custom chips, motherboard parts, power transformer, RF modulator, console and game cartridges. I paid $76 for all that, a pretty good deal in my mind. This extra console would be for my possible 2nd hi-res Astrocade and the game cartridges for hi-res game conversions. The eventual Ebay motherboard failure turns out to be a blessing. In my preparation for troubleshooting this motherboard and an idea I had, I accidently created a new troubleshooting tool for a blank screen scenario when you turn on your Astrocade. This new tool alone has narrow-downed the possible bad chip count. I believe I know which 2 chips are possibly bad. I think I can test 2 outputs with my oscilloscope which I will try today. The custom data chip is working. Yippy! I'm expecting to receive the logic probe that I ordered tomorrow. This new tool that I created, which I am calling the "Set Screen" tool along with Balcheck are two very useful tools for troubleshooting a motherboard producing a blank TV screen. The good news is, I can add "Set Screen" to Balcheck II. I have hand written a lengthy rough draft "posting" intended for anyone with experience in troubleshooting a Bally/Astrocade motherboard. The posting discusses the blank screen scenario and details my "Set Screen" tool. I will key it in and send it to you after I finish troubleshooting this Ebay motherboard. SetScreen is just a 36 byte ML routine that resides at 2000H. All it does is set the screen colors and two other screen parameters, then forces the Z80 CPU to Halt. When the CPU halts, there's no write or read of RAM memory. The CPU attempts to refresh RAM and the custom data chip attempts to scan RAM for the video display. The CPU is forced into a known state and stays there. This just makes it a little easier to isolate a problem area. SetScreen takes over when the normal power up routine fails to initialize the TV screen parameters (screen goes blank) because of a related RAM failure. SetScreen will display at least 2 colors on the TV screen if the Z80 CPU, the ROM chip, the Microcyler data bus and the custom data chip are all operating properly. You have to program an EPROM/EEPROM and place it at address 2000H. My upcoming posting to you details SetScreen's purpose and limitation. I'm revising my already revised rough draft of SetScreen which is becoming lengthy. I'm doing some more investigating today related to the powering on of the motherboard. [Part III] My new SetScreen tool narrowed the problem area down to the custom address chip or the interface between the custom address/data chips and screen RAM. Because of what SetScreen was displaying on the TV screen, I suspected U23, the data read octal buffer, as the bad chip. I didn't have my logic probe, that I ordered, yet. So, I tried my 40 year old crude oscilloscope which is apparently only good up to 100,000 Hz. To my surprise, the scope showed enough on its CRT to tell me a signal line was operating, sort of like a logic probe with a pulsing LED. The CRT display was blurry, but informative. I checked the A0 thru A5 address lines, RAS and CAS lines which were operating. Then I checked U23's inputs which were present, but its outputs very little or nothing at all. Plus U23's outputs seem to follow what I was seeing on the TV screen which was colored pixels all fading out and in intermittently. SetScreen is really cool. So, I de-soldered the U23 chip and replaced it with a low profile machine tooled IC socket, which is not an easy task. You have to be careful when installing a socket so that it makes contact with the top motherboard solder pads. I had to check the socket contacts several times with an ohmmeter. During de-soldering, even with only a 15W soldering iron, I damaged one top soldering pad but still managed to make it contact the IC socket. I had just one DM81LS95 chip in my spares inventory. I pushed it in the socket, connected the motherboard to the TV and pushed on the power switch. Immediately, the Astrocade menu appeared. I ran Checkmate several times to test the motherboard which now runs great. Previously, I thought I might have a bad custom data chip because of what I saw at the bottom of the motherboard by the custom chip pins and a heat stain on the bottom console when I first received the EBay Astrocade that I purchased. What a relief to find this custom data chip running great. I can't express in words how valuable Balcheck and Setscreen are in troubleshooting a failed motherboard, especially one that produces a blank TV screen. They're like a one-two punch. Unfortunately, a chip that is not readily available went bad. But, I have an idea for a IC adapter that can be built and then plugged into an IC socket on the motherboard. This idea should work allowing one to run a 74LS244 chip via the adapter in a U23 IC socket and also a 74LS245 chip in a U10 IC socket. U10, DP8304 is also a 20 pin chip that is not readily available. I may build and test the two adapters after my static screen RAM project is finished. [End of Michael's description of his Astrocade fix.]