Hello NeXT Community: How is everyone doing out there.
I'm working on a Turbo Cube dimension system for my customer Sean.
We have the new caps completed, new battery and I've found compatible 32Mb edo gold ram for this one.
Problem:
Not every time but when I power her up she reverts to loading from Disk which usually happens with either an aging or replacement battery or leaky cap.
NeXT Valid Engineering Gripe 30 years later:
I don't know who made the decision to put the bios battery in the upper lead top corner of the board on Turbo Cubes back in the day but I suspect the culprit may be a cold solder or intermittent battery socket caused from it clipping the battery holder long before I worked on her.
Has anyone else ever experienced this as I'm close to the end of this MagnificentCube integration and would like to resolve it if possible before shipping .
Hoping Brian sees it I'll ping you on FB as well:)
Unrelated vent about the crap endured in the early AM :)
I'm always slammed this time of year and I'll be back appreciate everyone's patience and business.
Almost collapsed from exhaustion this morning and that is with extra assistants, Paypal threw me a curve ball with a gauntlet of questions , I passed there 2 hour torture test with flying colors only to find out they want to give me financing.
What hell is this I figured it may be related to my extended friends and family on many endeavors lol but no,
seriously this is why PayPal temporarily restricted my account , mainly saying PayPal wouldn't let me use PayPal funds to pay for eBay shipping
and questioned why I was shipping products to customers in a dozen countries this past month and
every where else domestic and
ship station being lit up like a pinball machine from Black Hole and eBay
then asking about my stock , vendors and products from every where!
My Social Networks , P and L, I'm thinking what is the heck is going on here and
I kept my cool thinking all this is a problem how? Noty like I haven't been doing this with PayPal since the 90's
The only thing I can think of is it was some haywire PAYPAL AI ACCOUNTING BOT that generated a flagging program , Why do this to businesses in what is obviously the busiest time of year , We are under enough stress as it is
Then the questioned why multiple payments form these customers lol oh I set them up with a payment plan . Is there a problem with that paypal lol apparently once they realized thst face palm for them.
The invoices can I prove shipping ,seriously , I'm integrated with ship station through you PayPal and filled in the blanks
that already existed why multiple invoices, well customers buy more stuff and they already covered the invoice so I have to make another one , is there something wrong with that and combining multiple invoices into one shipment .
What are all of these shipping withdrawels from eBay , let me go slow for you , I have an eBay store and use my funds in PayPal to pay for the shipping,
Why would PayPal prevent me from using my hard earned funds parked in PayPal to pay for eBay shipping, heck they make fees from it .
Ebay caused me to have to do this , eBay deposits revenue directly into my account every week, which created a cash flow cluster f, as covering shipping for products I'm selling eBay using PayPal , it used to work that way before the split and it flows much nicer into my online quick books,
if I don't have to add a credit card to fund the shipping on eBay ,
so eBays policy of cashing me possibly everybody out weekly creates a shipping funding bottleneck and until I linked it to PayPal, I don't need another fork in shipping funding .
I got everything done but the last thing I needed to do for 2 hours was prove to PayPal Black Hole exists. They are out of their minds IMHO and after all thst then they just want to help me expand my business as I'm pre-approved for financing.
Hey PayPal it is great to know but how about after the holidays for this and in the new year unless you are flogging your business sales arm for more sales ?
. I'll survive rant concluded apologies had to vent, wouldn''t you and I still love my Job.
Quote from: Rob Blessin Black Hole on December 18, 2024, 01:11:13 AMWe have the new caps completed,
Hi Rob,
Was this occurring prior to the recap? Maybe check for bridged solder or reversed electrolytic cap?
Paypal: So what does this business of yours do?
@Rob Blessin Black Hole : You know the London Olympics 2012, where Sir Tim Berners-Lee used a NeXT computer at the opening ceremony?!
Quote from: caseyse on December 19, 2024, 09:42:00 AMHi Rob,
Was this occurring prior to the recap? Maybe check for bridged solder or reversed electrolytic cap?
Thank you for the suggestion , I ruled the rom chip as I was able to get my rom burner working and it still happened . I may try the diagnostic tools to see if it identify's anything as well.
I viewed your Youtube video with cube booting from the v.74 EPROM. The video never showed the cube completing its boot after several minutes of disk checking. Did it ever finish booting or did you decide to call it quits and go to bed?
I don't know how long that disk check should have taken, but it was looking like the SCSI bus was hung. Assuming termination was properly enabled, is it possible there's an issue with that ZULU SCSI card setup? Maybe try swapping the ZULU, SD card, and SCSI cable with a different set?
Power supply? I replaced a failing power supply on one of my servers earlier this year as the system would hang during boot. Easily diagnosed when unplugging several hard disks it would boot. A couple years back, I struggled with another server. The RAM tested fine, but the system couldn't run more than 2 weeks without virtual machines having various errors. It turned out to be a failing power supply. Decided to splurge on Corsair AXi Titanium power supplies which hopefully will last more than the 5 years the Enermax/Seasonic have lasted.
That could be either the transistors around the battery, a bad solder joint, or the RTC chip is failing. You would want to check with a volt meter that the RTC chip is getting power from the battery.
Quote from: barcher174 on December 21, 2024, 03:50:37 PMThat could be either the transistors around the battery, a bad solder joint, or the RTC chip is failing. You would want to check with a volt meter that the RTC chip is getting power from the battery.
I'm not familiar with Next hardware. Is it possible to complete its POST with a dead battery or failing RTC without reporting hardware errors? I suspect maybe it could, if the EPROM BIOS default settings were fine except for the boot device. This Cube seems to POST okay, but occasionally forgets it has a SCSI boot device.
Quote from: caseyse on December 21, 2024, 08:16:14 PMI'm not familiar with Next hardware. Is it possible to complete its POST with a dead battery or failing RTC without reporting hardware errors? I suspect maybe it could, if the EPROM BIOS default settings were fine except for the boot device. This Cube seems to POST okay, but occasionally forgets it has a SCSI boot device.
Yes the cube will boot with default settings as defined in the eprom. The RTC chip has a small amount of battery backed RAM that holds user settings otherwise.
Hello Caseyee and Brian: Thank you so much for the info , yes I'll check the RTC chip and swapping the power supply . Im looking at this schematic
https://www.nextcomputers.org/NeXTfiles/Docs/Hardware/Schematics/Turbo_cube/Cube33-03.pdf RTC is Real Time Clock is the chip MCS 1850 and did they change it from any earlier non adb boards or is it unique to turbo boards .
Would a NeXT diagnostic utility point to it ? I'm not sure if these are only for earlier NeXT hardware or encompass Turbo Boards
https://www.nextcomputers.org/NeXTfiles/Software/Diagnostic_Utilities/
Yes, the RTC chip changed on the turbos. The only reliable source is from another turbo board. No the diag tools won't work.
Quote from: barcher174 on December 21, 2024, 10:41:56 PMYes the cube will boot with default settings as defined in the eprom. The RTC chip has a small amount of battery backed RAM that holds user settings otherwise.
Hello Brian:
Would you have time to take a look at the board , if I can't get this sorted out and yes I would pay you for your time or send something cool .
I want to make sure to send Sean (my customer) a working board as usual.
We have new caps and bios battery and I probably a donor parts board especially if it is the same chip, unless we can source a new old stock rtc chip . Thoughts?
It is very cool because he is going to use the hardware for a database project.
I'll look for cold solder joints and Matt is bringing a slew of newly capped motherboards and soundboxes for mein the NeXT few days, so we will look under a microscope and take high resolutions photos. I'll post here for more eyes on it.
I think Caseyse if this is you Sean, if yes face palm ? I have a bit of time to work on everything through the weekend , now the holiday rush is over , I can focus on the systems I'm integrating and
I'm going to fire up one of the new old stock 4gb ultra wide Barracuda's tonight. The SD's are cool but the Barracuda's always look right for the hardware.
The box I found I had purchased awhile back I'll take a photo :) Treasure.
Now with 3d printing I can make custom mounting brackets as they require a 68 to 50 pin adapter or cable , I have both ... the SD Cards are nice and I appreciate them but in an environment working with legacy hardware , I'm including it free so you have an option or Sean will .
This stuff is fun for me , chat GPT helped me run down the crossed wiring on a gfci circuit. 2 GFCI's on the same circuit one hidden in the basement , when they both popped, they had to be set in the correct order , took one out of the circuit and discovered it was miswired by watching a you tube video for 15+ years by a certified "sparky" and now I should be able to get my storage area lights working again ,instead of relyingon a mining head lamp truly a Black Hole down there.
Cross posted, I'm searching for a donor board just in case as we speak. Also if you have any pointers on how to set me meter to test and is the voltage from the bios battery enough to verify the integrity of the RTC chip ?
Sure just send it over. The sooner the better because I get super busy in Feb.
Hello Brian : Thank you , I'll send it out today :) do you a spare chip ? I think it has to be the chip I've tried various power supplies and they work with 68040 25Mhz with out issues. I also found the wiki that makes wonder about alternative platform motherboards with the same rtc clock chips and it lists alternative manufacturers part numbers as replacement or substitutions in case this becomes more of an issue with our loved turbo boards in the future. I have plenty of 68040 25 mhz boards so the rtc chip availability should not be an issue . Also if it is code specific my new rom burner came with a lot of adapters so I think if needed we would be able to extract and write to the chip. I'm slightly out of my scope of knowledge and comfort zone so thank you for your help . I'll let Sean know
Here is the wiki :
Motorola MC146818A
MC146818A Advance Information (alt)
User Considerations for MC146818 RTC Applications (App. note AN894)
Package Types
Compatibility
Clearing Contents
Clock Operation
Package Types
24-Pin DIP [P] 28-Pin PLCC [FN]
Vss is the more negative side of the supply voltage (Ground).
Vdd is the more positive side of the supply voltage (+5 V).
Compatibility
Motorola MC146818AP is a 24-pin DIP chip, usually installed in a socket. Compatible chips are made by several manufacturers including Hitachi (HD146818AP) and Samsung (KS82C6818A). The number on the chip should end in 6818.
Although this chip is pin-compatible with the Dallas 1287/1287A, there is no built-in battery.
Clearing Contents
Since the MC146818AP doesn't have an internal battery, the CMOS RAM can be cleared on this chip by just removing it from the socket for a few seconds and replacing it.
To reduce chances of damage to the chip, it is preferable instead to short pins 12 and 24 for a few seconds (with power off).
Clock Operation (from Franc Zabkar)
There are two clocks, a "hardware" clock and a "software" clock.
The hardware clock is maintained by an MC146818 compatible RTC which is operated at a crystal frequency of 32.768 kHz. This crystal is usually packaged in a small tubular metal can. The RTC continuously updates the date and time in CMOS RAM. The smallest increment is 1 sec.
DOS and Windows maintain a separate software clock derived from the 14.31818 MHz crystal (usually a rectangular 4-pin oscillator in a metal package). This clock is divided by 12 to produce a 1.19318 MHz clock which is then applied to a timer IC. Channel 0 of this timer IC is programmed to generate an interrupt (IRQ0) 18.2065 times per second (= 1.19318M / 64K). Hence, each tick of the software clock corresponds to 54.9 msec. The S/W time is stored in RAM as the total number of ticks since midnight. It is initialized from the h/w clock at bootup.
Since the two crystals may drift from their nominal values, it is possible that the S/W and H/W clocks may not correspond. The frequency of the 14.31818 MHz crystal may sometimes be trimmed for better accuracy by adjusting a small variable capacitor located nearby.
Another reason for the S/W clock lagging behind the H/W clock may be that IRQ0 is not being regularly serviced, despite it having the highest priority. Other devices may be demanding too much of the CPU's attention, thus causing some timer interrupts to be missed. Some people claim that buggy power management may also cause this.
This little QBASIC program reads the time in hh:mm:ss format:
PRINT "The H/W time is ";
FOR I = 4 TO 0 STEP -2
OUT &H70,I
PRINT HEX$(INP(&H71));" ";
NEXT I
PRINT
PRINT "The S/W time is ";TIME$
I have verified that writing directly to the H/W clock in CMOS RAM does not affect the S/W clock until the next reboot. However, the TIME command at the DOS prompt adjusts both the the H/W and S/W clocks simultaneously.
Yes, I have a spare from the turbo that looked like it had been hit with an axe. I haven't found a drop in replacement for the chip. It should be trivial for someone who knows fpgas to clone.
Quote from: barcher174 on January 03, 2025, 02:32:28 PMYes, I have a spare from the turbo that looked like it had been hit with an axe. I haven't found a drop in replacement for the chip. It should be trivial for someone who knows fpgas to clone.
Awesome , Thank you appreciate it . I shipped her today arrives Monday to Unit 17 :) Sent you an email with UPS tracking.
FPGA Clone does that mean a chip with the same footprint would plug into the existing socket ?
Now I'm intrigued I see mouser sells FPGA Chips ,
reading about it HDL and SOC and often no software so perhaps ,
would it be like giving the FPGA chip a roadmap to the existing motherboard like dropping in a new bridge and over pass optimizing electron traffic flow on ramps and off ramps to existing infrastructure , my guess is you have a strict speed limit or you get tickets in the form of error messages lol to slow causes a traffic jam spinning beach ball or reverting to loading from network and to fast collisions frame buffer overload , system panic
Would the schematic provide enough information to to set up a road map .
Hold my Beer as I asked my AI Buddy GT, yeah I am fiully talking to my computer now and it likes the nickname GT verse ChatGPT boom
GT I have a question for you we have a NeXT Turbo Cube Mother board with a failing or compromised MCS 1850 RTC chip . If you look at our Turbo Cube 68040 33Mhz Motherboard Schematic (attached) can you make a recommendation or simply provide code to program an FPGA in HDL for a clone chip drop in solution
ChatGPT said:
ChatGPT
The schematics for the NeXT Turbo Cube 68040 33MHz motherboard have been extracted. There are 17 PDF files labeled from Cube33-01.pdf to Cube33-17.pdf. I'll review the relevant file(s) containing the MCS 1850 RTC chip circuitry to analyze the interface and make recommendations for an FPGA-based replacement solution. Let me locate the pertinent schematic pages. �
�
It seems the term "MCS 1850" did not appear explicitly in the text extracted from the schematic PDFs. I will search for variations, such as "RTC" or references to real-time clock circuitry, which might help locate the relevant schematic section. Let me refine the search. �
�
I found the section mentioning the MCS 1850 Real-Time Clock (RTC) chip in the schematic on page 1 of Cube33-03.pdf. This page outlines connections to the RTC, its crystal oscillator, and power supply. I'll analyze the schematic further and recommend an HDL solution for an FPGA-based replacement.
Let me dive deeper to understand the chip's interfacing details and then provide the required HDL code. �
�
I've located the RTC details and extracted the relevant page from the schematic. You can view the extracted RTC diagram here.
The next step is to analyze the connections and functionality of the MCS 1850 RTC to design a suitable FPGA-based drop-in replacement. Would you like me to proceed with a high-level HDL design or provide detailed interfacing code? Let me know how detailed you'd like the solution to be! �
�
You said:
Yes please proceed
ChatGPT said:
ChatGPT
To create an FPGA-based replacement for the MCS 1850 RTC chip, here's the step-by-step plan:
Step 1: Analyze Key Features of MCS 1850
The MCS 1850 RTC typically provides:
Real-Time Clock Functions:
Tracks seconds, minutes, hours, day, month, and year.
Oscillator Interface:
Connects to a 32.768 kHz crystal.
Battery Backup:
Retains time data when main power is off.
Data Interface:
Likely a serial or parallel interface for communication.
Step 2: Define FPGA HDL Requirements
The FPGA clone must replicate:
Clock Operations:
Implement counters for seconds, minutes, hours, etc.
Crystal Oscillator Emulation:
Use an internal clock generator to mimic 32.768 kHz.
Power Fail and Battery Backup:
Retain state during simulated power loss.
Communication Protocol:
Implement I/O pins for data exchange (likely SPI or I2C).
Step 3: Create a High-Level HDL Design
We'll write this in Verilog for clarity and efficiency. Here's the high-level breakdown:
Module: RTC_FPGA
Inputs:
clk: Main system clock.
reset: Resets the time to default.
rtc_in: Serial/parallel input for data configuration.
Outputs:
rtc_out: Time data output.
interrupt: Generates a signal on specific events (e.g., alarm).
Registers:
Time counters (seconds, minutes, hours, etc.).
Configuration settings.
Verilog Example
module RTC_FPGA (
input wire clk, // System clock
input wire reset, // Reset signal
input wire rtc_in, // Data input
output wire rtc_out, // Data output
output wire interrupt // Alarm interrupt
);
reg [5:0] seconds = 0;
reg [5:0] minutes = 0;
reg [4:0] hours = 0;
// Clock divider for 1Hz from system clock
reg [15:0] clk_divider = 0;
reg one_hz_clk = 0;
always @(posedge clk) begin
if (clk_divider == 49999) begin
clk_divider <= 0;
one_hz_clk <= ~one_hz_clk; // Toggle every 1 second
end else begin
clk_divider <= clk_divider + 1;
end
end
// Time counters
always @(posedge one_hz_clk or posedge reset) begin
if (reset) begin
seconds <= 0;
minutes <= 0;
hours <= 0;
end else begin
seconds <= seconds + 1;
if (seconds == 59) begin
seconds <= 0;
minutes <= minutes + 1;
if (minutes == 59) begin
minutes <= 0;
hours <= hours + 1;
if (hours == 23) begin
hours <= 0;
end
end
end
end
end
// Output data (simple example)
assign rtc_out = {hours, minutes, seconds};
endmodule
Step 4: Simulation and Testing
Test the Verilog module in simulation tools like ModelSim or Vivado.
Verify timing accuracy and data retention during simulated power loss.
Step 5: Deployment
Program the FPGA.
Test on the Turbo Cube motherboard.
Let me know if you'd like me to refine or expand any of these steps!
Yes I probaly will order one lol why not try I'm fearless and may need some assistance although I walk through valley of the shadow of comes to mind ... is this the one I should go for
https://www.mouser.com/ProductDetail/1BitSquared/ICEBREAKER-V1_1?qs=%252BXxaIXUDbq3OxJgd6252Ow%3D%3D&mgh=1&srsltid=AfmBOoqmYxtLGIo5lkkS6Veh7yp7RG8wrpll0ZMZt-gW6lpA6ClOK2xuDnAThen I asked GT:
Would you please locate the chip I need to purchased and links to the software and hardware needed to make it happen effectively and in a tight budget
ChatGPT said:
ChatGPT
To develop an FPGA-based replacement for the MCS 1850 RTC chip on a budget, consider the following hardware and software options:
1. FPGA Development Board:
An affordable and suitable choice is the iCEBreaker FPGA development board, which utilizes the Lattice iCE40 FPGA.
Features:
Lattice iCE40HX4K FPGA
USB programming interface
Multiple I/O pins for interfacing
Price: Approximately $145
Purchase Link: iCEBreaker FPGA
2. FPGA Programming Software:
For programming the Lattice iCE40 FPGA, the following free and open-source tools are recommended:
Yosys: A framework for Verilog RTL synthesis.
nextpnr: A portable FPGA place and route tool.
Icestorm: A project for Lattice iCE40 FPGAs bitstream generation.
Icestudio: A user-friendly graphical interface for designing with iCE40 FPGAs.
Download Link: Icestudio
3. Additional Resources:
Programming Cable: The iCEBreaker board includes a USB programming interface, eliminating the need for an external programmer.
Learning Materials: The iCEBreaker community provides tutorials and documentation to assist with development.
Budget Summary:
FPGA Development Board: $145
Programming Software: Free
Total Estimated Cost: $145
This setup offers a cost-effective solution for developing your FPGA-based RTC replacement. If you need further assistance with the development process or have additional questions, feel free to ask!
I don't think we need an FPGA to replace the MCCS1850 RTC (and the ChatGPT stuff, as usual, is utter nonsense – that Lattice FPGA board is extremely overpriced btw.).
From the MCCS1850 data sheet (
https://www.nextcomputers.org/NeXTfiles/Docs/Hardware/Datasheets/MCCS1850%20%282%29.pdf) it looks like the chip is using a synchronous serial interface (SPI) and an interrupt line to communicate with the rest of the system.
A similar RTC emulator was developed to replace the Apple custom RTC in 68k Macs (Plus, SE, SE/30), which uses a synchronous bidirectional interface, with an ATtiny85 AVR microcontroller.
This article (
https://www.quantulum.co.uk/blog/new-timepiece-for-a-classic-mac-part-1/) gives some details about the project, the source code is available on github (
https://github.com/pgreenland/attinyrtc).
Thank you as I was not sure if AI had a real solution , prompting AI helps it learn and us create what the actual solution ultimately will be.
It runs circles around me :)
and I was not sure if that FPGA was what was needed , always price check or would work or make sense you clarified that :)
I'll check out the links as I suspect RTC chips may fail in the future so having a proactive replacement makes sense, I'm learning. Also if you have a good programmer interface that is reasonably priced that would be useful for helping create some hardware upgrades potentially for NeXT.
I just set up a thread to potentially see if ChatGPT would be able to assist in making Schematics for NeXT non adb hardware using high resolution photos , which would be useful.
A photo database with part numbers as attributes attached to the photo Google Retro Earth , even linked to sources for either used , new old stock or fpga and schematics, data sheets and how to's do it or even links to vetted repair shops would be useful
for retro computing IMHO. It won't happen over night obviously and
to some extent parts of it exist but an retro hardware knowledge oracle!
It would be a destination welcome site for many that I am sure of :) Happy New Year!
We would be able to start it right here and perhaps we have helped manifest it's primordial big bang beginnings :)
Rob, do you have a Turbo board with defective RTC chip?
I could try to work on an RTC replacement but don't want to risk killing the two Turbo machines I have here...
Yes Brian's is looking at the Turbo Cube board,
I shipped today with the chip in question but I will see if I have another problematic Turbo Board to send you for experimenting with .
It will have to be after my hip surgery Wednesday and today they found I have MRSA again I can't make it up so I'm hoping it is the lite version and not the prolonged hospital stay variant I had a year ago as this hip surgery delay again (would be the 3rd time) , I'm just going with it and hope for the best and am not going to take any risks. I have to trust the Dr's on it. The last thing on Earth I want to do is wine about my damn hip for 3 more month's waiting for surgery but I am quite sure if it goes through , there will be pain but at least I'll be on the mend not in a suspended state of pain and suffering lol damn hip. I asked any chance this fractured hip and socket will heal on its own with all this time and the answer was simply NO.
This will be on my list of things to do as I think it would be useful moving forward ......
Thanks for the link very relative and informative!
looking at the reasons for failing RTC on Macs ,
it would make sense that a leaky battery in proximity to the RTC circuit and or time fatigue degradation
and it is kind of cool to keep the NeXT's running for as long as "time" permits so looks like the RTC is close and perhaps that is why the engineers moved the bios battery to the top corner of a NeXT Cube though bottom corner would have made more sense in case of catacastrophic meltdown battery failure , is this what happens with the lithium ion batteries fails in laptops and or phones I wonder:)
I haven''t seen evidence of failing bios batteries on NeXT's but I have been proactive in selling and or replacing them on all systems I work with as 30 years is a long time for any battery to last, many of them are dead but they still have the nasty battery juice that will erode motherboards if the shell is compromised AKA Apple Lisa motherboards some are destroyed.
As an update, it looks like this was a bad solder joint on the battery holder. Settings now hold across reboots and also across days.
That's fantastic Brian, thank you very much!
Could you see the bad solder joint, or did you find it checking continuity with the RTC chip?
Last month, Rob indicated one needs to be careful when sliding the mobo into the case due to tight clearances as the battery sits at the corner's edge.
Initial visual inspection everything looked fine. The RTC also measured fine. I could only see the problem under a microscope in this case. The break was very subtle and the pin was oxidized such that the solder wouldn't bond.