So I'm starting this thread after troubleshooting my Mimeo.
I had a problem where at first it couldn't load basic anymore, then I got it to load basic (bad ram chip) but after about an hour basic would not reload (E000R or E2B3R) unless I cycle the power and reload basic. It passed all memory tests for E000 and lower memory using Mike Willegal's memory test program for at least 100 cycles each.
So what was it...
Well after realizing my radioshack multimeter can't measure down to 47pf accurately and putting a scope and logic analyzer on the 74123 among other chips I saw some time drifting, I replaced the mylar cap and also decided to replace the B13 -5V decoupling cap (it measured spot on, but I replaced it with one that measured high)
It solved my problem and after about 8 hours of use, never once did I experience a problem.
So why is this thread called "replica problems vs the real thing"?
I never had this issue on the real thing, the mica caps are usually solid. I know I have been reading in the past that 2504V chips have been bad in Mimeo and Newton setups and that some people have had other problems. The only real Apple-1 I know that had a problem with 2504 chips was Wendell Sander's back in the 1977ish time frame when Steve Jobs replaced one under warranty for him.
There are other weak points in the real Apple-1 like board warping and the 7451 or 74123, but frankly too many potential issues to list here to send people with a Mimeo or Newton problem down a troubleshooting rat hole that don't apply to them, but I was wondering if people can post a brief description of the issues they have had with their replicas and what the solutions were. Please don't point to a thread, just summarize in this one. I'm hoping this new thread can be a quick cheat sheet to help people out in bringing up or troubleshooting their replicas.
Thanks,
Cheers,
Corey
Issues I have had with both Mimeo and Newton replicas include the following:
1) Flaky 2504 from Unicorn. Garbage characters being occasionally displayed.
Solution ==> replacement of flaky chip(s)
2) Spikes seen on negative power rails with my oscilloscope.
This caused memory reliability issues, data corruption occurring anywhere after a few minutes to a few hours.
Solution ==> I fixed this by tacking some tantalum capacitors under the board at every 3 or 4 DRAM memory chip
Are these issues unique to clones or have they been seen on real Apple I ?
I would guess the replica PCBs are close, but not 100% in terms of wiring routing, dielectric characteristics of PCB material, etc.
In my opinion, I don't think there is any significant difference in operation between the original and the reproductions. Areas where I've had problems with Mimeos, seem to be problematic with the original units. An example is the video dot problem, which shows up randomly on both originals and reproductions. This is a PCB layout issue, not so much a digital design issue. Regarding the -5volt/DRAM issues, LIza Loop was never able to get her Apple 1 to run properly. I suspect she was experiencing DRAM problems, but I have never been able to prove it. Here is a blog post describing a pretty good fix for that issue.
http://www.willegal.net/blog/?p=3668
The first large batch of 2504's that I bought all worked and I've never had a failure with any of them.
regards,
Mike Willegal
I think now what is happening is the demand for the old NOS parts has climbed and stuff is coming out of the woodwork that isn't up to snuff. This is why I started the thread, because Mimeo and the actual units are not the similar anymore because of this. The stuff that fails on the real thing is because the parts were used for a couple of years and then stored sometimes good (in a wine cellar like Wendell's) or bad in a humid garage.
BTW: I also have never had a bad 2504, but I bought bulk a few years back and have been using them up. I did get an accurate Capacitance Meter in the mail yesterday that can measure the 47pf cap. My bad cap was at 22pf cold (it's off the board now), so one can only imagine what happens as it is in use. This was a NOS 10% cap, so that's a bit more than 10% off. The new one measures 45pf, so my Mimeo is good now.
Cheers,
Corey