Microbotics M1230XA problem, possibly faulty?

[Jope]

Probably no harm, it all depends on how hard it is to get either type. I can almost guarantee that the original cap is not low esr :-)

[Amiga1992]

Well, I can't seem to find match-for-match on eBay, one of the specs usually difers.
Most 85C ones I find are 10V or more, or the right V value but different µF.
I can't seem to find any of this "brand" whatever that is.

What would you recommend as a most similar replacement, or should I just try those Panasonics?

Originally Posted by Akira

Well, I can't seem to find match-for-match on eBay, one of the specs usually difers.
Most 85C ones I find are 10V or more, or the right V value but different µF.
I can't seem to find any of this "brand" whatever that is.

What would you recommend as a most similar replacement, or should I just try those Panasonics?

Why are you going through ebay to get these caps when it's much easier to just order them from Digikey?

Either Panasonic, Nichicon, or Rubycon would be fine. The temp rating is just the max temp the caps are rated to withstand. If you get a 105c rated, assuming it's the same value, it should perform the same as an 85c rated cap (the difference being that the 105c rated one will last longer in high temp applications). Besides the temp rating, Digikey also lists the life expectancy @ a certain temp too.

Don't drive yourself nuts trying to get the lowest ESR caps you can find for this application. That cap is probably just being used as a filter so as long as the cap values are good the ESR rating shouldn't affect it. If it was a modern system it'd be an issue.

It doesn't look like it's an issue in this application, but one other thing to pay attention to is the height of the caps. This is more of an issue on the main board around the area of the IDE connector since taller SMD caps can interfere with certain CF card adapters (something that happened to me). If it's just on the accelerator card though, the height may not even be a factor in this case.

I didn't read the whole thread but if you list the value(s) of the caps you need i can try to find them on Digikey for you.

[Amiga1992]

Originally Posted by dethrattle

Why are you going through ebay to get these caps when it's much easier to just order them from Digikey?

Because I never used it ! Thanks for pointing that out, I'll check it out and order some caps there

[Daedalus]

Solid advice from dethrattle there. Don't get too stressed about finding an exact match when most of the time they're just used for preventing noise from being introduced to the power rails by digital chips. I'll add that the maximum temperature is related to the life expectancy, as in, the further you are from the limit, the longer the capacitor will last beyond the quoted life expectancy. It's a little simplified, but that's the idea, so all other things being equal, 105 degree capacitors will last longer than 85 degree capacitors. Both will last far longer than their stated life expectancy in an Amiga though, where it's unlikely to much exceed 30 degrees.

As for the voltage, higher voltage parts will generally work as well as the lower voltage ones, but for various reasons, don't go too high. For example I'd happily use 16V parts in the place of 10V parts, but would avoid using 35V parts in the place of 10V parts unless I couldn't get anything else.

[Amiga1992]

By the way, Digikey does not come up wth non-SMD caps with this rating:
https://www.digikey.com/products/en?...=100µF%206.3V

(I use eBay because it's easy and has everything I need, usually, so I don't run into this kinda situation)

Thanks for further tips Daedalus! I'll choose accordingly.
Also thanks dethrattle for pointing out capacitor HEIGHT. The ones on this card are actually lower height and fatter, so I will try to find something like that.

[Daedalus]

I buy lots of my parts from Farnell in the UK and Ireland, part of the Element 14 group and apparently called Newark over that side of the pond. They have a good range, and a good filter system for drilling down to what you need. For example, here are all the radial, through-hole, 6.3V 100uF capacitors, so you can add more filters for lead spacing, height, diameter, manufacturer, temperature etc. until you get down to what you need.

[Amiga1992]

I used Farnell a bunch when I was living in the UK, didn't know they had a US branch! Thanks for that Daedalus.

This is the only type these guys have that is the same height (but thinner):
http://www.newark.com/nichicon/urz0j...tor/dp/21M8183

Does size matter? Would this work? What about the capacitance?
HJmm.. lead spacing might be too small... Gotta measure.

[Daedalus]

Physical size doesn't matter except when it comes to mechanical compatibility with other nearby parts, so smaller height or diameter won't be a problem. Over the years capacitors have gotten smaller as the technology improves, so often replacement parts are a little smaller than their 25-year-old counterparts. It's particularly noticeable when replacing SMT caps in an A1200 for example, where the same capacitance and rating parts are often much smaller and end up looking lost on the spaces allocated for them. Still works fine though. The lead spacing for radial capacitors is not the end of the world if it doesn't match as they can be spread, but this will increase the height by a couple of mm to allow for the bent legs. It's also a little less mechanically stable, but is an option.

What do you mean about capacitance? The value is probably not that critical, but I wouldn't go for a different value unless you were sure of what you were filtering as changing the capacitance will change the characteristics of the filter it is a part of. Best to stick with the value that was designed.

[Amiga1992]

Sorry, I got confused with "capacitance tolerance".

[Daedalus]

Ah right. Don't worry about the tolerance - wide tolerances are pretty standard when it comes to capacitors, there are very few applications that require tight tolerance, and such capacitors are rare and expensive.

[kublai]

I just bought a 128MB RAm stick for my M1230XA board and I can't seem to test the memory. I keep getting 'unhibited mounting error' and would lock up. I'm using the v1.01 MBRTest-2 from their website and also the one that came with the board. I also tried my 32MB which was in there originally and got the same error. I tried booting from the MBRTest-2 floppy and it set my RAM to 80ns and only tested the 8MB and I saw my 128MB in the list but I can't choose it as it automatically test the 8MB.

[rmzalbar]

Checking in on this thread with my own MBX 1230 XA board. Just finished resurrecting my A1200 from long-term storage, only to find out that it locks up whenever the data cache is enabled (instruction cache is OK.) So.. can't use setpatch, can't use autoXA, etc.

I'm going to try replacing the 68030, but that would be a weird CPU failure.

UPDATE: CPU was found to have become temperature sensitive, stable with a fan blowing on it. Replaced with new CPU, a C mask which runs cooler. All "good" now. "Good" being that I'm running into the incompatibility problems experienced by the other M1230XA owners above. No Indy 500 for me.

[rmzalbar]

I discovered something today. About 35% of WHDLoad wouldn't start at all. Access faults, 8000 0004 gurus, 0100 000C recoverable alerts, line $1111 emulation errors, etc.. sometimes a game needed NOMMU tooltype to run, but sometimes this would break a running but flaky game.

I discovered the problem was with the 128MB SIMM I had installed. I switched back to my old 16MB SIMM and these issues largely went away. I'm still testing, but I estimate about 95% of the games are working now.

I don't think the problem is with the SIMM itself - it works fine in normal use, including rendering, etc, and passes every memtest I could throw at it. I think the issue is with the MMU handling a memory of that size. I don't know if the CPU or the Microbotics/Paravision ASIC is at fault, but I think the MMU is either not working properly or not quickly enough. So, if you're using a large SIMM, try going to a smaller one and test again - remember to set the speed and size in SetXA before swapping, especially if it is a slower SIMM.

EDIT: OK.. a lot more games work now that didn't work at all - these were mainly games that used a kick image. However, I still get access faults and illegal instructions randomly in many games. Man this M1230XA is a POS, isn't it?

[rmzalbar]

UPDATE: I found the cause of the access faults and illegal instructions on my setup. NoVBRMove tooltype (on the offending games) was the last thing which fixed that. So let me summarize:

Hardware configuration
Computer:
A1200 NTSC

Storage:
32GB CompactFlash card on A1200 IDE controller
HDToolBox MAXTRANSFER=0x1FE00 <- Despite Hyperion saying no need to adjust, I ended up having to change it on my partitions. Was getting corrupt disk image errors on loads.

M1230XA:
MC68030RC50C @ 50MHz
MC68882RC50 @ 50MHz
16MB 60ns non-parity non-EDO FPM SIMM <- NOTE: a major contributor to crashes and exceptions was a 128MB EDO SIMM I had been using. Errors were red guru meditations (8000.0004) and orange recoverable alerts. Replacing it with a standard FPM (non-EDO) 16MB module fixed this problem. I have another couple of FPM (non-EDO) simms on the way, including a 128MB simm, so will test to see if the SIZE of the memory is the issue, or if EDO is a contributor (or I had a one-off wonky SIMM)

Software

OS:
Kickstart 3.1.4 | Workbench 3.1.4

WHDLoad:
Version 18.5
Global settings S:WHDLoad.prefs
MMU <- The MMU traps exceptions, prevents games from crashing the computer - and I've found a few that wouldn't run at all without MMU enabled.
NOAUTOVEC <- AUTOVEC exceptions fixed by this.
NOMEMREVERSE <- not sure if this is really needed. This was something I added while tracking down problems ultimately resolved by NoVBRMove. Will remove and test, update here when done

Per-game tooltypes used:
PAL [or NTSC] <- some games won't run in the other timing standard, so try this first if it won't go.
NOVBRMOVE <- for games that give exceptions or access error notifications. If you don't have an MMU enabled, they may simply crash instead of giving an error notification. Only add this on a game-by-game basis, because most don't need it and you lose quitkey functionality on some of them with NOVBRMOVE.

Now, all of the few dozen games I'm testing with are working, with the exception of the original Gunship that scrambles up after taking off. I will test other games that users have reported as being especially problematic (Turrican series?) and report back. I'm clicking my heels with joy. I really hope this helps others with the M1230XA!

Here's a few off the top of my head that needed NOVBRMOVE:
Mortal Kombat 2 AGA
Seek & Destroy ECS and AGA
Saint Dragon
Double Dribble
Blades of Steel (Konami games in general, seems like)
Second Samurai
Nitro NTSC
Turrican 3

[rmzalbar]

Turrican I: OK
Turrican II: no joy, error is consistently repeatable (illegal instruction at same address)
Turrican 3: OK with NOVBRMOVE

Speedball 2: Crashes the machine typically within seconds to a minute. So does the floppy version - have to pull the accelerator, then the Crystal-cracked AGA-fixed version works.
Speedball 2 CD32: OK with PAL (waits endlessly at "loading game" screen in NTSC, though doesn't crash)

[rmzalbar]

8MB and 128MB FPM simms tested: No difference.

The new 128MB FPM SIMM runs no better than the EDO one, guruing on load of larger games.

The 8MB SIMM works OK but did not add any additional compatibility to my test cases.

This card seems to struggle with larger memory sizes. It also seems to have (what I suspect is) a timing issue when communicating on the 14mhz bus. This shows as as graphic glitching, and can be most easily seen when running Lionheart. More with the 128MB simm, less with the 16MB and 8MB size.

I'm going to look into Jens Shoenfeld's clock signal filter cleanups, and see if any make sense to apply it to my board and test again.