Introducing the Mikey tester

oski_98 · January 1, 2022

Hello all, this is a continuation of sorts from my other thread where I've been trying to troubleshoot a C104129-001 REV. 2.6 Lynx II. Given that I was sort of stuck at what could be wrong with that Lynx, I decided to build a simple circuit to test Mikey. The circuit is very simple and consists of PLCC68 socket for Mikey, a crystal oscillator, and a power-on reset circuit built with components matching the schematics I've found in the wild. I am powering the circuit using 4.5V power supply. I used this to test both the Mikey in my broke Lynx and a spanking new Mikey I ordered from the good folk at Best Electronics. Here's a picture of the circuit:

I also tested all pins and compared the Mikey from my broke Lynx II and the new Mikey:

And finally some pictures of the only waveforms I observed. This first one from X16M has a sharp spike on the rise:

Here's XTAL0:

Here's AUDOUTHI, a clear square wave:

And AUDOUTLO, another square wave but different:

And that's it. Which was somewhat disappointing because I was expending something on the /RAS and /CAS probes, D0-D7, A0-A7, DL0-3, or CLK1-3, but they are just fixed at static values. On the other hand the behavior of both the Mikey from the busted Lynx II and the one from Best Electronics seems exactly the same, so there's some evidence to suggest that Mikey is not the issue with this Lynx. The next steps I'm contemplating:

* Modify the Mikey Tester circuit to include the pin 31 to pin 33 short that should produce "INSERT GAME", and/or build out other parts of the circuit that might be necessary to have Mikey do more interesting things (though I'd only be best-guessing what these might be). I'd appreciate any suggestions from folk on this forum on how to modify the circuit to provide a more conclusive test that Mikey is WAI?

* My other best guess is that the issue is the DRAM. The chips seem to start with a request to write to DRAM (/WE high) and perhaps if the DRAM chips are dead Mikey freezes? I could just replace those and see what happens. I've been able to source a modest number of D41464L-12, these are not the same as the D41464L-80 the Lynx II came with. From the datasheet, it seems like the D41464L-80 is providing 80ns row access time whereas D41464L-12 is providing 120ns row access time. So I guess the DRAM I got is just slower, but other posts on the forum suggest that it's still compatible.

If anyone on this forum has suggestions or ideas, I'd love to hear them. Thanks in advance.

oski_98 · January 2, 2022

Update: I was finally able to get some waveforms on the RAM access pins from Mikey! Yesterday, I spent an hour carefully studying the schematic as well as this documentation from 42Bastian's Lynx page. I came up with two theories on why Mikey was frozen:

The /POWERON pin needs to set or pulsed
The /BUSGNT pin needs to be set, apparently this is the pin that Mikey/Suzy use to coordinate/share access to the shared bus

So today I modified my Mikey Tester circuit to set both these pins high through a 1K resistor and also added a push-button switch connected to ground to set them low again momentarily. At first I had no luck, nothing going on. Even the static high pins I was seeing before were low. I thought I fried the chip, so I decided to undo the changes I made one at a time to get back to the previous state. First I removed the resistor connecting /POWERON to VCC then the resistor connecting the BUSGNT, each time checking on the oscilloscope. The leads were still attached to the breadboard though and then I saw waveforms! So I started playing around with the connections a bit and realized that it was BUSGNT that has to be set low, so with BUSGNT connected to ground via a 1K resistor (the resistor is probably not necessary but I left it just in case to limit current) I was seeing logic waveforms on the pins. So I'm guessing that if BUSGNT is floating, or set high, then Mikey sleeps. That means SUZY is a more likely the damaged chip in my Lynx. I'll have to think about it some more to know what to try next but it seems that Mikey is alive for sure. Here's a pic of the waveform on the RAS pin:

I hope this information helps someone else trying to troubleshoot their broken Lynx. Happy new year.

oski_98 · January 2, 2022

Here's the final version of the Mikey Tester circuit. The operation is as follows:

The right switch when pressed sets the /RESET pin to low, when released the power-on reset circuit slowly raises the voltage of the reset pin to high
The left switch when pressed sets the /BUSGNT to low, when released it sets the pin to high

Here is the progression of states after power on.

Press and hold the left switch and all pins should read the static states captured in my notes in the first post of this thread, while the switch is held (e.g. /RAS is held high).
Press the left switch a second time and you should now see logic wave forms (e.g. see pic in the previous post for a sample of /RAS waveforms).

Some final pictures:

Close-up of the circuit.

I'm now reasonably confident the problem with my broken Lynx II is not the Mikey chip and my current suspect is Suzy. So next steps for me are to replace Suzy, put back the original Mikey chip, and see what happens. I think one useful discovery from this experiment is that, unlike suggested in other threads, no waveforms from Mikey on the signal buses does not necessarily mean it's dead. There is documentation, and now experimental evidence, that other components of the Lynx (such as Suzy) can "silence" Mikey by setting /BUSGNT high (though I haven't tested this with a Hayato). It was also interesting to discover that /BUSGNT had to go HI-LO twice to start seeing waveforms.

And that's that for now on this particular thread, unless someone wants me to post a schematic of the circuit.