PDA

View Full Version : IMSAI 8080 - Tantalum Capacitors



psteichen
December 13th, 2014, 07:27 AM
Thanks everyone for your helpful comments on my previous post. I have another question regarding this machine. To summarize, I have an IMSAI 8080 that has been in storage for 20+ years, and probably more like 30+ years since it's been hooked up to a terminal for any meaningful work. Not knowing about the tantalum capacitor issue, I pulled it out of storage two weeks ago and fired it up. Everything seems to work fine. The two drives both spin up, all the LEDs (save one on the PIO card) light up. The processor seems to be running as I get some rapid flickering of the address bus LEDs when I select RUN.

So my next step is to toggle in some simple programs to test the processor further. Possibly even a program that will output something to the terminal. I don't have the original terminal used with the machine, but I have several old serial terminals that I will try. I'm also waiting on delivery of some 10 sector, hard sector 5.25" diskettes from a generous anonymous source who was willing to part with some of his personal supply.

But first I wonder if I need to take a step back. Like I said I already ran it for over an hour testing it a few weeks ago. Now I read on various forums that I'm risking a catastrophe by powering this thing up. Is this true? Or have I already verified that the caps are good by running for an hour without issue? Also, are there any issues with the large caps in the power supply? I can't imagine a very destructive fire being caused by a little cap the size of my pinky nail....but those bad boys on the power supply could certainly do some lasting damage.

Any thoughts are appreciated.

1980s_john
December 13th, 2014, 10:54 AM
Hi,

I think you have tested your computer the quick way. On some of my old 1980s computers I have had tantalum capacitors fail after many hours, and yes they do catch fire so you need to have easy access to switch the power off if this happens. Tants might be on the S100 cards, or on the floppy disk electronics. When replacing the advice is to use tants with about double the voltage of the cap used in the original circuit.

The big power supply capacitors are electrolytic, these can dry out and over heat when old. However, the manufacturers tended to use really good quality ones (ie expensive) that can last several decades.

The cautious approach (which is a but late in your case - you were lucky!) would be to remove all the S100 boards, then use a variac or series light bulb (or both) to gradually build up the mains input voltage from 20V to whatever your max line voltage is (120V/240V). This allows the electrolytic caps to reform. If you apply full voltage to a cap that had lost some of its dielectric then it may short or overheat. What tends to happen is that the capacitor will draw a small current whilst it is reforming, this drops to under a milliamp when reformed.

http://www.vcomp.co.uk/tech_tips/reform_caps/reform_caps.htm

Having applied a low voltage you would use a DC volt meter across to smoothing caps to check all was well.

Having proven the PSU you should then plug one S100 board in at a time and check if any cause a short circuit. I have a mains power meter (about 15), this displays watts drawn, and a sudden jump is a sign a board has short, typically due to a tant capacitor. On a PC this would cause the PSU to shut down gracefully, but with a powerful linear supply found in an IMSAI there is little to limit the current a board can draw (hence a variac and/or series light bulb helps prevent a high current being drawn through a shorted board).

Hope your IMSAI works, some photos of what you have would be most interesting.

Regards,
John

PS You don't have any switch mode power supplies to worry about, these have other issues capacitors eg X2 suppression caps that burn out (eg those made by Rifa in the 70s/80s), and DC smoothing caps that go high ESR.

psteichen
December 13th, 2014, 01:08 PM
Thanks for your advice John. So it sounds like I lucked out, but at this point should be safe going forward. Is it a good idea to install inline fuses between the PSU and the rest of the unit to shut down the supply should a future short occure?

I'll post some photos below. Sorry about the quality, they were taken in a dimly lit basement.
21788
21789
BTW...the two missing blue paddles will soon be restored, I found a few on eBay.

1980s_john
December 30th, 2014, 04:17 PM
Is it a good idea to install inline fuses between the PSU and the rest of the unit to shut down the supply should a future short occure?

I am not familiar with the IMSAI, but I would expect there must be fuses on the three outputs of the PSU (+8V, +16V -16V) , for example:

http://www.s100computers.com/My%20System%20Pages/The%20Box/The%20Box.htm

It might be a good idea to fit smaller fuses if you are only running with a few S100 boards. The PSU fuses are there to protect the PSU against damage, not to protect an individual S100 board against damage. For example the +8V line might typically have a 25A fuse. A 25A fuse will carry 50A for a few seconds before blowing too, for info about fuses see:

http://www.tlc-direct.co.uk/Book/3.6.3.htm

The PSU output fuses are likely to be fast acting. The mains input fuse is likely to be a time delay fuse (typically these will have a small coil of fuse wire visible), due to inrush at switch on whilst the big smoothing capacitors charge up. See:

http://www.schurter.co.uk/content/download/194051/5552460/file/Guide_to_Fuse_Selection.pdf

Regards,
John

Chuck(G)
December 30th, 2014, 07:28 PM
You don't want to fuse individual power supply lines--the reason is that the old 3-rail PMOS and NMOS chips (such as your CPU and many memory chips) require the power supplies all to be present and in the proper order. Drop out a key one and you can ruin a chip. Use a current-sensing circuit on each line and have a fault on any line trip the primary supply. There are plenty of ideas on the web, but basically you sense current by examining the voltage across a shunt.