exxos's Atari, Amiga & retro forum

Bought all the necessary parts to build and program a TF 536 but I've run into problems. I can't get the damn thing to boot..
Tried in 2 Amiga 500's, a 500++ built recently and an older Rev 6a 500, both machines just give a dim power LED showing that the system is just not initialising..

The CPU I'm using is a M68030RC33C which works quite happily in one of Matze's TK68030 boards at 50 Mhz but I quite fancied the smaller board size and extra RAM of the 536.

I have checked for bridged connectors and loose legs (did find one on the CPLD but this didn't help) and buzzed out all the lines from the CPLD to their destinations and all seems fine. The only thing that confused me using the board layout diagram is that A15 shows as connected to GND and A3 to 3.3V.? Is this right?
When I buzzed mine out, A15 was present on X1 pin 43, IC1 Pin 11 and IC3 Pin 125 and A3 was present on X1 pin 31, IC6 Pin 7, IC3 Pin 138 and IC1 Pin 47

The CPLD has been flashed with the latest RC3 candidate, flash and verify were successful.
I can see 100Mhz coming form the clock chip but only 7Mhz going to the CPU

High res pics here:

You will see 7Mhz to the CPU unless its running in fastram. This is expected. What does diagrom do on the serialport.

We've got zilch on serial output from DiagROM, The system doesn't come out of reset, it basically acts like there's no proc installed..
I'm wondering if the CPLD is knackered. I had to rework it as my first attempt to get it soldered on was a failure. One of the sides must have slipped while I was tacking it down. I had one side fully soldered before I found it had shifted enough that another side had all the pins almost exactly between the pads. had to use hot air to remove it and may have cooked the chip. Thought I might have gotten away with it as the chip did program succesfully.

not coming out of reset means reset is held low. Is it low or?

It's very unlikely that the CPLD is the issue. If you see 7M.. that indicates its working perfectly. Its more likely a soldering fault. Does AS/DTACK wiggle at the 68000 socket? It's also absolutely nothing to do with the SDRAM or the buffers that side of the board if it's not booting diagrom.

Did you remove the 0hm resistor nets since you have installed the bus buffers?

Time to break out a microscope and start looking for bridges. Don't start just replacing parts without figuring out what's wrong.

Did some more work. Noticed that 50% of the time in startup I was only seeing 1.5V off the LM1117, likely due to a bad solder join on the capacitor. Resoldered that down and now seeing a steady 3.3V. Still no joy on starting it up though.
Checked signals, I did see a steady square wave on DTACK on one boot but have never seen it since now just getting noise.
Reset line was showing a noise at roughly the same as the clock speed, scope showed 7Mhz wave but very unsteady.
Checked the IDE buffer chips, Pins 13-18 all showed a short on IC5 but nothing visible on the microscope. Tried an ultrasonic cleaner in case some detritus had got stuck under the chip but made no difference so had to remove the IDE header and remove, clean up the pads and resolder the chip. Need to get a new IDE header as it didn't survive the removal process.
Unfortunately it appears the ultrasonic bath has killed the Oscillator chip as it no longer outputs a signal

Are you sure the osc simply doesn't need resoldering ?

My first ever built TF card sparkled to life after replacing oscillator. And it is the only time I have never seen any output from diagrom?

If you haven't spread, give your pcb a good spray with isopropyl alcohol and go over it with a tooth brush. You may have some dodgy pins that look like they are soldered when they are only glued down with flux.

Also, round that oscillator looks pretty crusty, so it may be shorting out?

Also I've seen what may look like some bridged pins on the ram.

You really do need to use a microscope to check your work when soldering chips of that complexity as its very easy to miss solder paste balls under pins that are shorting them out.

Good luck.

Vypr wrote: 26 May 2021 16:03 Did some more work. Noticed that 50% f the time n startup I was only seeing 1.5V off the LM1117, likely due to a bad solder join on the capacitor. Resoldered that down and now seeing a steady 3.3V. Still no joy on startiung it up though.
Checked signals, I did see a steady square wave on DTACK on one boot but have never seen it since now just getting noise.
Reset line was showing a noise at roughly the same as the clock speed, scope showed 7Mhz wave but very unsteady.
Checked the IDE buffer chips, Pins 13-18 all showed a short on IC5 but nothing visible on the microscope. Tried an ultrasonic cleaner in case some detritus had got stuck under the chip but made no difference so had to remove the IDE header and remove, clean up the pads and resolder the chip. Need to get a new IDE header as it didn't survive the removal process.
Unfortunately it appears the ultrasonic bath has killed the Oscillator chip as it no longer outputs a signal

Have you removed the RNs? You need to remove the resistor networks under the Buffers if you fit buffers.

I've replaced the oscillator now and got my 100Mhz clock back.
The resistor networks R5,R6,R7 and R8 have been removed before I started work. I may put them back and remove the IC4,5 & 6 as a last resort.

After replacing the oscillator again I have found something new. I can boot to DiagROM if, and only if, I put pressure on the 68000 socket area.
Once it has started booting I can let go and DiagROM boots fine and all tests pass. CIA & IRQ tests OK, keyboar dis fine, 64Mb FAST RAM checks out OK, I can even CTRL-A-A and it will reboot fine but if I power off and back on again without pushing down on the board, it won;t boot.

I'm very confused now.