Multicomp MAX705 Reset Circuit - Going for the burn!

[Richard Deane]

I recently built a Multicomp board from the PCB design of Kurt Mueller, which is a tidy up from Multicomp Cyclone IV iib.

It has added a MAX705 reset circuit, a 5v PS2 KBD supply, and supports 3 to 4 serial ports.

I carefully sourced components and soldered it up and gave it a run. At first I didn't notice a problem, except that the reset switch on the PCB did nothing (the one on Cyclone IV piggy-back board works ok).

A while later I noticed that R9 (10 ohm) was burning hot, and the connected C13 10uF Ceramic looked as though it had burnt out. I removed C13 , R9 and MAX705 and the rest of the board continues to work with no hot components. KBD, VGA and Serial 1,2 (RS232) verified.

I have faith that Kurt's design is OK but wonder whether a typo crept into the schematic, or I messed up on decoding the circuit diagram to create a BOM . For example the MAX705 label does not identify all the part number so perhaps I bought the wrong component (Mouser 700-MAX705EPA , MFG Part No: MAX705EPA+ Analog Devices / Maxim )

I attach the diagram for the reset circuit, in the hope that someone with max705 experience might be able to do a sanity check.

If R9 is 10 ohm then I comprehend that it can get  a high current and get too hot, but that would imply that the voltage on pins 6,2,4 of max705 is held low.

Should R9 be a higher value?

Why would C19 burn out? Surely no current is expected to flow through that. Perhaps excessive heat from R9 could have damaged it. C19 was a 10uF MLCC capacitor (schematic says Ceramic)

The full details for the board are at https://www.retrobrewcomputers.org/doku.php?id=builderpages:muellerk:multicomp_july_2019, though the circuit and files are the April21 version.

I tried to contact Kurt through the retrobrewcomputers forum  with no success hence reaching out here for inspiration.

Cheers

Richard

[Tadeusz Pycio]

The symptoms indicate that C13 was short-circuited, which is why R9 heated up. I am assuming that the MAX705 is operational and this was not the cause of this failure.

[Richard Deane]

I have to assume the max705 was good, bought from Mouser so quality should be ok.  

I don't know that the max705 is still good, but in a week or two (to give time for more feedback)  I can try a rebuild with a new capacitor and resistor and see how it goes.  Hopefully the PCB tracks aren't damaged. The cost of a rebuild on a fresh PCB (I have them) is not warranted as I can at present live without the PCB-mounted reset button. It will take a while to sort the Multicomp FGA firmware and software, configured with CP/M for all the serial ports. Currently the CP/M is supposed to be sorted with  the  ROMMJS software which I have working on Cyclone II, but should work on Cyclone IV where there are more serial ports, however at present I only have the MC2G1024 software working on Cyclone IV. I have a vague hope that the actual CP/M implementation is compatible across both .

At present I am distracted (in the UK) by my PT Cruiser ECU dying, forcing the car to be immobilised. Due to Chrysler pulling out of the market a number of years ago, all the former dealers are getting selective amnesia about how to reprogram a replacement ECU for the correct VIN and to match the SKIM module (key security)

Cheers

Richard

On Wed, 6 Sept 2023 at 09:52, Tadeusz Pycio <ta...@wp.pl> wrote:

The symptoms indicate that C13 was short-circuited, which is why R9 heated up. I am assuming that the MAX705 is operational and this was not the cause of this failure.

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/9a3081c2-d61b-47b3-9c67-acbbca07ac47n%40googlegroups.com.

[Tom Storey]

The WDI input should be left floating if you're not using that function. Otherwise if it is pulled high or low for more than 1.6 seconds and doesn't change then it will initiate a reset.

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/CAG2GuRbshnUg2TnkQgRr66xp0oJKGpXaQonJYkHrCrPPMc5ePw%40mail.gmail.com.

[Tom Storey]

As for the myriad of part numbers, as long as you've bought one for the correct Vcc then you should be fine. Many of the part numbers are just variations of packaging, temperature range (commercial/industrial/etc), and how they get supplied from the factory (tape/reel/tube/loose/etc), so the same physical component may have several different part numbers but are all otherwise identical. The datasheet should have a section that lists all of these variations.

On Wed, 6 Sept 2023, 10:36 Richard Deane, <rwd...@gmail.com> wrote:

[Tom Storey]

On Wed, 13 Sept 2023, 19:55 Tom Storey, <t...@snnap.net> wrote:

The WDI input should be left floating if you're not using that function. Otherwise if it is pulled high or low for more than 1.6 seconds and doesn't change then it will initiate a reset.

Ah actually, never mind. Watchdog resets are generated by the WDO output which is unconnected in this case.

[Richard Deane]

Thanks for the comments. Over the cold days of winter I think I will breadboard just this sub-circuit, put a scope on, or logic analyser and examine a bit more. I have one of the salea clones which I've used successfully under Windows, but not yet under Linux Mint. I've only used the logic analyser for auto-baudrate serial decoding so far.

Also buy a spare max705 next time I place an order (and excuse / reason to build something else)

Cheers

Richard