A $100 Lesson
Not too long ago I bought a Tektronix 496 Spectrum Analyzer off Ebay. The description said it came on but no further testing was done. It was a good price, $199 plus shipping and since I have wanted a decent spec-an for a long time now I jumped on it. It did look really nice. When it arrived I quickly pulled it out of the huge box and put it on the bench. Now I have bought a lot of stuff off Ebay that said it was not working or for parts and was able to get them to work, usually with not too much trouble or time or money. This thing was going to ruin my record.
It did indeed turn on. The screen lit up and all the little push buttons lit up and that was the end of it. Nothing worked. Nothing I did could change anything. Despair.
I found the service manual online for the thing and combed through its hundreds of pages for some clue as to why my spec-an would not work. One section of the manual described the boot up sequence and how to check it. Lights and everything. According to the manual an EPROM, a 2716, was the likely culprit. Well I had an EPROM programmer that would write 2716s and I found what was supposed to be updated code for it online. Great. I ordered a couple of ICs and when they got here I tried to program them. This is where the whole thing fell apart, not the spec-an, my plan.
My little TL866 programmer claimed it would write 25 volt EPROMS. It lied. I spent hours erasing and programming the 2716s to no avail. Finally I measured the program voltage on the TL866 and it did not even approach 25 volts, more like 18 volts. I even tried to update the firmware and software on the TL866. All that did was brick the thing so it is of no use at all now. Fail!
Undaunted I dug out my old Heathkit ID-4801 programmer, which I had to fix first since all the switches had corroded over and had to be cleaned. Another several hours building personality modules and writing EPROMS, or trying. I could make the thing write a 27128 which takes 12.5 volts VPP but no luck on the 2716s. I even took the time to hand enter the entire file into the 4801, 0000H to 07FFH. That took a little time. C0, Inc, FF, Inc, F0, Inc and so on for an hour or so. And the Heathkit was built when 2716s were a thing!
Ok, online searches brought me to a little Kree programmer that swore it would write 2716s and it was only $56. Great. I ordered it and a couple extra 2716s because I had erased the ones I had so many times. You only get so many chances with those. When it got here I found that I had to find an all wire 25 pin printer type cable and a computer with a printer port on it. OK. My old Lenovo with Vista Business that I use for digital ham radio would fill the bill. I tried and tried to make the new programmer work with no luck. It would write one byte and stop. No matter how I set the timing the same thing happened. Again, 27128s, no problem. Another fail.
In desperation I ordered a few CMOS 27C16s that required 12.75 volts VPP and after a lot of fooling around with the timing I was able to get one of those programmed. Success, at last.
Now I had previously taken out the EPROM in the 496 and read it. I never had any trouble reading the dang things on any of the programmers. The output looked nothing like the file I had downloaded from the web but the place I got the file said it was an updated version so I wasn’t too worried.
With great expectations I swapped out the EPROMs and turned on the 496. Nothing changed! ARGH!!!
A little farther down the page in the manual it said the boot up sequence was dependent on the 68000 microprocessor working so I took a look at the clock pulses, Good. Then I tried to look at the data coming out of the 68000. Nothing looked right. At this point I gave up. I could have gotten a processor board on Ebay for $75 but during the time I had been messing around with this thing I had acquired an IFR A-7550 with tracking generator and that is all I need and a bit better than the 496 since it had the tracking generator.
I offered the 496 for sale for $100 plus shipping and sold it almost immediately. A few days after shipping the 45lb beast I received a message from the buyer that he had replaced the processor board and all was well. I am so happy for him and disappointed in myself. Just a little bit more patience would have fixed it. I had followed the rabbit hole down to the very end and came nose to nose with the mole.
Lesson learned. Or Not.
So rare a processor goes down like that, but you got something better and didnt need the 496 anyhow.
That is true. The IFR is better in a lot of ways but it is just the principle of the thing. I don’t like failing.