Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: darksun9210 on January 04, 2017, 11:50:15 PM
-
Hi all.
I think I've done something wrong/bad....
So, I thought I'd fire up the ol' A500 and see what state she was in.
current spec.
Rev5 board, 3.1rom, ECS Denise, MiniMegachip, 512k+RTC trapdoor, IndivisionECS, Jens 68k socket clockport board. GVP A530 8meg ram, 4GB CF via usual curmudgeon of adapters.
so. fired her up without the GVP, trap door ram or clock port adapter attached. flashing power light, and eventually get a green screen (chipram fault). reset, guru meditation, click left mouse, rom screen. boot off floppy, get the workbench, 512k ram. ok fine. power off. plug in the trapdoor slot expansion. nope. no screen. power light, no floppy click. reset seems normal as in the caps lock light comes on and goes off... flick the switch to disable the trapdoor expansion, all ok again. shes a bit wobbly from cold, but she gurus out eventually, and then left click resets to the insert floppy screen.
now. for the minimegachip, the JP2 lower solderpad has been cut, but the upper pad has not been joined. the side trace from near the trapdoor slot has been cut, but a quick poke with a multimeter tells me its still live? accidentally touched the metal shielding and got a connection sound. not sure that's right?
the fly lead from the megachip board is connected to pin 36 of gary, and that's a good connection. also get a good connection from pin 48 on the 68k without moving the lead.
i've never got the minimegachip working right. even with a jumper on the upper patches of JP2. had 1.5meg chipram at best once, and I know fat agnus is working.
seems something nasty is going on with the trapdoor slot but I don't know where to start looking.
-
pics for a bit of clarity on what's going on
-
Hi all.
I think I've done something wrong/bad....
So, I thought I'd fire up the ol' A500 and see what state she was in.
current spec.
Rev5 board, 3.1rom, ECS Denise, MiniMegachip, 512k+RTC trapdoor, IndivisionECS, Jens 68k socket clockport board. GVP A530 8meg ram, 4GB CF via usual curmudgeon of adapters.
so. fired her up without the GVP, trap door ram or clock port adapter attached. flashing power light, and eventually get a green screen (chipram fault). reset, guru meditation, click left mouse, rom screen. boot off floppy, get the workbench, 512k ram. ok fine. power off. plug in the trapdoor slot expansion. nope. no screen. power light, no floppy click. reset seems normal as in the caps lock light comes on and goes off... flick the switch to disable the trapdoor expansion, all ok again. shes a bit wobbly from cold, but she gurus out eventually, and then left click resets to the insert floppy screen.
now. for the minimegachip, the JP2 lower solderpad has been cut, but the upper pad has not been joined. the side trace from near the trapdoor slot has been cut, but a quick poke with a multimeter tells me its still live? accidentally touched the metal shielding and got a connection sound. not sure that's right?
the fly lead from the megachip board is connected to pin 36 of gary, and that's a good connection. also get a good connection from pin 48 on the 68k without moving the lead.
i've never got the minimegachip working right. even with a jumper on the upper patches of JP2. had 1.5meg chipram at best once, and I know fat agnus is working.
seems something nasty is going on with the trapdoor slot but I don't know where to start looking.
OK. Rev 5 boards shipped with either 8371 or 8372 Agnus. You don't have an 8371 Agnus with more than 512K of chip RAM ever, IIRC.
Now, broadly, you have an issue with chip RAM. It seems to be with the memory board bringing the system down, but it could also be a problem with the mod to give Rev 5 A500s more than 512K of memory too. ;)
A500 expansions come in all kinds of exotic variations, but they all do pretty much the same thing - add 16 bit memory. How that is configured depends partly on what Agnus revision you have fitted, as well as the card.
Post hires pics of the area around Agnus for any mods (bridges with solder between tracks cuts into tracks). Some of the expansion card to understand your issues with it would help.
If you can't photograph, examine and describe. It's probably a faulty component on the expansion board, or it could be something else too. Most all expansions have at least one strange behaviour.
It could also be the power supply is dying, or you have a local brown out for power. A530s can be greedy beasts. Disconnect it.
Examine the expansion and connector THROROUGHLY to see how well they connect, and if any bits of metal are touching where they should not.
-
Rev5 board, 3.1rom, ECS Denise, MiniMegachip, 512k+RTC trapdoor, IndivisionECS, Jens 68k socket clockport board... 4GB CF via usual curmudgeon of adapters.
That's a fair bit of hardware for many Amiga standard power supplies. Is there a floppy connected too? And how is the CF wired, to the SCSI bus, or to the floppy, or some other way?
-
right bear with me a bit, and i'll get you some photos.
i'm ruling the A530 from the start. she boots fine without the A530 or the trapdoor expansion, and boots fine with only the A530 - so i'll leave that out as not to confuse the matter. but just for interest, off of the A530's internal scsi bus, there is an Acard scsi2ide adapter, an IDE2CF, and a Kingston 133x 4Gb CF card in that.
also removed the clockport adapter card.
have connector blocked the floppy power line to add an additional externally routed male molex connector that would usually plug into a female molex soldered to the back of the A530 PSU plug. its not pretty but its safe, and works.
the PSU does "sing" with a tiny bit of capacitor squeal, but it drives the A1200 I have loaded with a bliz040+scsi just fine (how long for remains to be seen ;)
so. as we currently are.
rev5 A500 board
3.1 rom + fly lead
minimegachip with 8375 agnus + fly lead to pin 36 of gary
IndivisionECS with ECS Denise socketed
512k+RTC trapdoor expansion set to disabled. (was known good till "mod")
amigakit floppy drive
and NOTHING else.
-
trapdoor track "cut" but still seems connected when tested with buzzer function of voltmeter
https://1drv.ms/i/s!ArUfT_xm4N1vguUKScgGkivu8cDN8w
JP2 solder pad, bottom pad cut, but top pad not linked
https://1drv.ms/i/s!ArUfT_xm4N1vguULsd7cMzlSUAGOfQ
trapdoor expansion (set to disabled)
https://1drv.ms/i/s!ArUfT_xm4N1vguUNO51FZsLxh2CJtg
trapdoor tracks again, but with flash
https://1drv.ms/i/s!ArUfT_xm4N1vguUOueLyxgG1RjsJdw
JP2 Pad, CPU, KS3.1rom with fly lead
https://1drv.ms/i/s!ArUfT_xm4N1vguUPSQlBPGPEuN3QqQ
MiniMegaChip "Agnus". flylead solder pad looks a little lifted
https://1drv.ms/i/s!ArUfT_xm4N1vguUQnPE6uZ_zbynyBQ
Indivision + ECS Denise
https://1drv.ms/i/s!ArUfT_xm4N1vguUSC1U1JUo13yPnmA
Gary with flylead to minimegachip - connection fine, also tests fine to pin 48 on CPU
https://1drv.ms/i/s!ArUfT_xm4N1vguUTw14fXuBYqWbuyg
board overview
https://1drv.ms/i/s!ArUfT_xm4N1vguUUaiD1nCJFwEWYXA
proof that it all works :D
https://1drv.ms/i/s!ArUfT_xm4N1vguUWkkJtmuS-uVIIXw
-
ok, utter hackiness going on here.
in my annoyance at not being able to patch JP2 properly, I figured i'd trace what the solder pads connect to. took a bit of googling and poking with a sharp stick, but found that the bottom pad is pin 52 (address 23) of the 68k. the middle pad is pin 59 of agnus (Address 19). the top pad is pin 47 (address 19) of the 68k.
for reference, pin36 of gary/pin48 of 68k is Address20 that the megachip flylead clips to.
the agnus socket has rows of tiny holes all around the socket, stipped the ends of a piece of wire wrap with my teeth (don't do this at home kids), stuff one end in the hole next to pin 59 of agnus, success! connection from the other end of the wire wrap to the middle pad on jp2.
now wrapped the other end of the wire round pin47 of the 68k and fired her up.
no green chip ram error or flashy power light. booted straight up!
https://1drv.ms/i/s!ArUfT_xm4N1vguUXauc1EHAZH5aVvA
1.5meg chip ram! BOOM!
https://1drv.ms/i/s!ArUfT_xm4N1vguUYlXBp_1yCM8_bgQ
ok, so seeing as 1.5meg is the most I've ever had, and the trapdoor must have been faulty for longer than I understood.
-
Could just be the trapdoor RAM doesn't like working as chip RAM. A501s don't, the first version with 16 chips. Chips are too slow.
EDIT: BS, you just fixed it by bypassing the usual chip RAM mod. But it can be an issue with upgrading A500 chip RAM.
Great stuff! :)
-
quick update. turned the trapdoor expansion on to see if the magic is still happening.
its not.
blackscreen. on power on, the floppy sounds like it does a quarter spin but nothing else. power light is steady. no capslock flash code. no screen output at all. no floppy clicks, no nothing.
powered off, turned off the trapdoor ram module, turned back on, and all is well... either the ram expansion is faulty, or I've screwed up cutting the traces to the trapdoor slot I guess.
-
Could just be the wire to Agnus flapping about. Try shaking it when working properly.
Wire wrap mods like that usually go on the back of a board, to keep them out of harms way.
If it just won't start with the expansion, check the chip temps. It's probably gone short somewhere and gets hot really quick.
-
yeah, I remember the A501 wasn't the most... efficient...
the board I have is a four chip board with the chips rated at 80ns (I think?)
anyway, looks like i'm going to ami/fleabay to find another trapdoor ram card.
doesn't solve my ultimate issue though. a vampire is no going to fit with the megachip board :~( ....
good call on the chip temps though. going to check that now
-
If you want to build a new one from a blank board, it can be done.
http://eab.abime.net/showthread.php?t=85395&highlight=A500+RAM
EDIT: Hmmm, extra 512K of chip RAM or a Vampire... no contest. You might lose a litltle compatibility, but you get huge CPU power, and the HDMI output can give you way better resolutions than the native A500 ones.
-
nope. nothing on the chip temps.
got the kids IR temp monitor. so much better than trying to stick a piece of plastic in a wriggling toddlers ear. does surface temps and calcs body temps. so, in surface mode:-
agnus hovers at around 44 deg C when all is working fine.
power off,
switch trapdoor exp on.
power on,
trapdoor ram chips don't warm up. hover at around 19.5 - 19.7 deg C
Agnus however gets up to about 50 deg C?
power off.
switch off trap door ram,
power on.
all ok...
as for the vamp... well.... yeah there is that, but the minimegachip takes me from 0.5 meg chip to 1.5 atm... and that's got to be helpful with WHDload stuff. plus I won't be worried about running out of chip ram if I have to do stuff in native resolutions... maybe if I can get the 1meg chip thing working that'll ease my stress in that regard...
-
doesn't solve my ultimate issue though. a vampire is no going to fit with the megachip board :~( ....
Not sure how it compares size-wise, but I know a Megachip board fits inside an A500 with a VXL*30 accelerator (which is a pretty honkin' big card), just have to use an extra set of riser pins on the 68000 socket.
Might be more tricky with the Vampire because of that IDE header sticking out of the top. Maybe in conjunction with one of the adapters that brings the 68000 socket further back inside the case?
Have you checked on any of the dev forums? I'm sure someone who owns this combination can offer some suggestions for getting it to fit. Good luck! :)
-
maybe if I can get the 1meg chip thing working that'll ease my stress in that regard...
That's the way I'd run it. But then again, I don't have much need for more than 1Mb chip, and am probably very untypical in that respect.
-
Oldsmobile_Mike
that's pretty much what I'm after! with just with the clockport socket adapter on the 68k socket puts the 68k up against the bottom of the keyboard, so need a different shape to move it back, and maybe then over the indiECS.
sheesh, i'm not after much am I ;)
clock port is handy as there is no Ethernet (clockport/usb to Ethernet dongle) solution for the 500 bar a zorro adapter for the side car slot. slow but it works. plus usb sticks are hany for moving large amounts of files around between machines.
used to had SMBFS setup to map network shares as logical volumes on the miggy.
anyway. final concern. I think the trapdoor ram is failing because the trap door trace I've cut as per minimegachip (dis/in)structions is shorting to ground, so I think I may have cut through to the ground plane if there is one on the 500.
or is it actually a ground line? I suspect it should be an address line that switches it from the slowram address range to the 2meg chipram address area, but I cant find any info on what this trace is to fix/repair/circumvent my butchery.
I'm going to give it some pokes with my circuit tester with the trapdoor ram on/off switch to see if I can find what changes between being enable and disabled.
-
quick update. turned the trapdoor expansion on to see if the magic is still happening.
its not.
blackscreen. on power on, the floppy sounds like it does a quarter spin but nothing else. power light is steady. no capslock flash code. no screen output at all. no floppy clicks, no nothing.
powered off, turned off the trapdoor ram module, turned back on, and all is well... either the ram expansion is faulty, or I've screwed up cutting the traces to the trapdoor slot I guess.
Hi,
It sounds like you have JP2 set correctly. For your memory expansion board, I'm not sure you even need to cut the trace on the motherboard. I think having the enable switch set to *disabled* will be equivalent.
The trace that gets cut for the Rev 5 board connects Gary pin 32 (_EXRAM) to pin 32 on the memory expansion header. You could test continuity between these two points to check whether the trace is truly cut.
_EXRAM is connected to ground to indicate that the memory expansion is present (as non-chip ram). If this pin is disconnected, it gets pulled up to 5V by the Amiga. For a 1MB chip configuration, this line must be disconnected from the expansion card, either by cutting the trace (rev5) or by cutting JP7A (rev6).
The Commodore A501 simply grounds pin 32 to indicate is presence to the Amiga. I suspect the enable/disable switch on your card grounds pin 32 when in the Enable position. You can verify this by checking continuity between pin 32 on the card and ground (pins 3/4 on the expansion, or any point along the thick trace that connects to these pins). If your switch does ground pin 32, then putting the switching in the disable position is equivalent to cutting the trace on the motherboard.
You can also check the voltage on Gary pin 32 to see that it has the correct state for 1MB chip. With the expansion installed, this pin should be a constant 5V in the 1MB chip configuration. If this pin is 0V, the expansion is present in the 512k configuration.
I don't see evidence from your picture that the enable/disable switch does anything more than connect pin 32 to ground, but I can't say completely from the picture. It looks like ground connects to the lower pin on the switch, and some signal line to the center pin. That probably goes to pin 32 on the connector, but that trace is on the bottom side of the board. Take a look at both sides of the board and see if there are any more connections to the switch besides the two traces visible on the top side.
Assuming the switch does nothing more than connect pin 32 to ground, I would leave the switch in the disable position and check voltages on the expansion board. You should see 5V between pins 10 and 20 on the M514256 RAM chips.
Robert
-
Could just be the trapdoor RAM doesn't like working as chip RAM. A501s don't, the first version with 16 chips. Chips are too slow.
That is weird, as slow ram and chip ram are the same speed.
Converting from slow to chip ram just changes the address that it appears to the CPU.
-
aha! that's great info, many thanks rkidd7952! i'll have a poke and a prod this evening and see what is what. either ground, nothing, or 5V - and if connected to pin 32 of gary.
-
ok. update time.
pin32 on Gary seems to be grounded.
pin32 on the trapdoor slot is NOT connected to pin 32 on Gary. so i guess the trace IS cut?
playing with the switch on the trapdoor expansion doesn't seem to effect this trace. no change from looking like a ground line. which would tally up with it being cut?
attached photo of the back of the expansion board.
pin 32 from what I can count is the fourth pin down on the right side from the top of the expansion connector.
this only seems to be connected to the 2nd down on the left side of the DIP chip on the - as you look at it, left side of the board. it's highlighted basicly by the letter "A" in the word RAM on the back of the board.
none of the pins on the switch seem to be connected to that?
secondly, I can't get 5V across pins ten and twenty. but let me double check I've got the right pins, so i'll update in a second when I find that info
(edit - rotate that picture 90 degrees clockwise and what I said makes sense)
-
ok I was being a bit of a bell and measuring between 10 and 11 on the ram chips. now I have the right pins, i'm getting 4.77Volts between 10 and 20. had to confirm that DIPS follow a standard pin pattern.
Marstons EPA is only 3.6% but it's obviously enough to need me to double check my work at the very least.
-
ok I was being a bit of a bell and measuring between 10 and 11 on the ram chips. now I have the right pins, i'm getting 4.77Volts between 10 and 20. had to confirm that DIPS follow a standard pin pattern.
Marstons EPA is only 3.6% but it's obviously enough to need me to double check my work at the very least.
Haha! It sounds like the voltage is OK then.
Thanks for the picture of the bottom side of the board. There's more going on with the switch than I thought at first. The traces on the top side do indeed connect pin 32 to ground, but there are also traces on the bottom to enable the memory and clock chips. The switch will have to be in the enable position to enable the chips. That means the trace will need to be cut to use the board as chip RAM.
As for locating pin 32 on the expansion card, I don't think you've found the right one. Starting from pin 1, count across the narrow (2 pin) width, then proceed down the length of the connector. Page 4 in the A500 R6 schematic at http://amiga.serveftp.net/schematics.html#A500 (http://amiga.serveftp.net/schematics.html#A500) has a good illustration of the numbering scheme.
Looking at the top side of the card, the odd pins are the upper/outer set of pins, and the even ones are the lower/inner set. (Upper meaning the layer of pins closer to the keyboard when the card is installed.) It looks like pin 32 connects to a short trace that connects to a via just to the left of the center resistor pack.
If you found that Gary pin 32 is at 0V with the expansion installed and enabled, it sure sounds like the trace is not completely cut. Double check that you've found the correct pin on the connector and see if you have continuity between Gary and the connector.
Also try checking the voltage on Gary pin 32 with the expansion installed, but with the switch set to disable.
Robert
-
good stuff! I've been after a pin out of the trapdoor slot to be sure
yes I had the wrong pin. I was following the dip pin numbering scheme. I didn't twig it would be the same as an edge connector/card slot.
will test this when I get home tonight
-
Right! Good News!
so, as per your advice, I counted alternating pins till I found pin 32 on the trapdoor expansion. as a bonus, there was a visable trace going from that pin on the trapdoor expansion card to a nice fat solder pad.
checked the continuity between that and pin 32 on Gary.
got a beep for both the ram being enabled, and the ram being disabled on the switch.
crap. ok. kitchen pairing knife, pressed and cut another point on the trace as carefully as I could manage.
no beep this time for either enabling or disabling the ram
test power on.
boots fine. still 1.5meg chip
power off.
enable trapdoor ram module
power on.
hard drive light pulses as the GVP scans the scsi bus. - excitement! never got this far before!
guru, left click, reboot - worry.
drive light pulses again. - excitement!
boots! oh my god.
take picture as attached.
JOB DONE!
thanks everyone for all your help :) learnt a lot. I think the main thing was finding what to test to find the problems :D
-
That's great! Glad to hear you got it working.
Robert