Discussion in '"Boat Anchor" & Classic Equipment' started by W7UUU, Jun 4, 2016.
That's veddy schmarrrt!
Well, now, there ya go - a "hybrid" solution!
73 de Jim, N2EY
Yep! Actually, I just built another - I'm now on the really awful sockets - not like the first three were much better, but they only needed small trace repairs. The last three are a different story.
But I just took my time and spent just over an hour building up another socket. I'm getting pretty good at trace repair on really really degraded, brittle, and decayed PC boards. It's ugly as hell, but the connections end up solid as a rock.
I installed another socket, and tested - 100% on the 4th socket. Before glueing, I'll test all 6 sockets one pin at a time. I did notice that on a couple of the ones I've finished that two end pins have a tendency to want to touch, just due to the delicate condition of the traces. So in those cases, to prevent shorting, I put in "cross braces" as well. I always check adjacent pins for shorts - and found these two pairs wanting to do it. I may just add a center insulator to all the pairs. What's another hour, after all I have into this?
For now, done for the day. It's back in its bin on the shelf and I'm off to do other things!
A good day's work IMO. Only two more sockets to go... Thanks for all the encouragement and tips, to Jim especially and everyone else who's been watching this progress and chimed in with suggestions and kind words all along the way
Anybody have a strip of PC Board repair "Lands" they could spare? Super hard to find unless you buy a $300+ "kit" for PCB repair... I've checked eBay for China suppliers, Amazon - there are a zillion "PC Board Repair Kits" out there - for stupid high prices. I miss the days you could just shoot over to Radio Shack and buy a strip for $2.99
I'm just making do with strips of copper tape, cut to size, poking a hole as best I can - but I have a bunch of missing/bad "lands" on the rest of the board.... I didn't bother with them for the Nixie Tube portion - too tight of a space - standard lands would be way too big. But for all the normal "through-hole" parts elsewhere, I could really use some.
I only need 15/20 or so.... happy to pay a fair price...
If you aren't sure what a "land" is, it's the little hole that a part goes through. Google it if still not sure.... I can cut path tape to fit, but it's really hard to make a good connection without a proper Land.... impossible to make a good looking connection without a proper land.
Please PM me if you have a small sheet of them... I also have LOTS of trading parts
Had a little time after dinner to play with my "cartouche"
It's pretty easy to fill in all the gaps with paper card stock - so why not? It's only 6 sockets - may as well do them all ways.
I ordered a $6 Amazon "micro squeeze bottle" - I will dilute Elmer's Glue to be thin enough to squeeze into all these tiny crevices - but that will make them all very solid when done.
Of course, will do a FULL 100% point by point test of every single socket connection before committing to glue in the pockets.
So far, I'm at 100% of the four sockets done. Will likely be another week before I finish the next 2 sockets - the traces / trace damage are really really bad - the worst of the lot - I've been honing my skills building up to the last two... and I have a boat load of work stuff in the coming week on top of everything else....
But I'm pretty happy with my results so far.
Photo below: added more insulators in the completed socket Cartouches....
Doing a great job. Gosh, looks labor intensive!!!
Yes - on one hand, it is pretty labor intensive. But what I like about this project is it's linear in layout (A goes to B goes to C etc.), and it's digital - just on or off. If this were a receiver or other fully analog circuit, I don't know that I'd go this deep. Not really sure I could. But with logic circuits, if you have proper continuity, and the original design is correct (which being a manufactured kit PCB, it is correct - proof being my "test 650" right on the bench working perfectly), it simply must work when you're done - provided all the chips are good. I've tested 100% of the replacement chips (thanks to Jim N2EY for the 7441 to replace my one defective 7441 of the 6 required - very hard chip to source!). I used the "good 650" as the test board to do that. Almost 100% of all the other parts are new as well, to limit failure points. The only reused parts are 2 seriously oddball resistors - one 6838 Ohm 1% (!!), and one 1900 Ohm 1% - and incidental stuff like transformer, PS RF chokes - stuff like that... stuff that doesn't generally fail or change value (transformer tests 100% and has been physically restored)
So I don't mind the labor to rebuild every Nixie socket... I know that if I maintain that 100% continuity, and keep that concept through the entire logic section of the board, it will work when I'm done.
The analog part of the board is a bit different.... I will just have to cross that bridge when I come to it. I'm not rebuilding it in "original Heathkit assembly" order - I'm rebuilding it logically. The very last part in fact will be the analog portions (3 HF amplifiers that feed waveforms to the counters). That way I can test 100% of the digital portion first.
I have no idea if my very delicate 40673 dual-channel FETs will work - Jim donated one, another kind ham donated a pair - 2 of the originals had rusted and their leads were gone - and on top of that, if you just look cross-eyed at a 40673 you'll kill it. So I have 4 of the 40673, and 2 3N211 dual-channel FETs that could substitute for the BFO and LMO channels - meaning I may in fact only need ONE of the 40673 to be good. Having 3 or 4 candidates means I at least have a good chance of getting one to work for the critical HFO input... the other two channels would probably do just fine with the 3N211's
But even if the three amplifiers have issues initially, I can still confirm digital operations using a signal generator to Q2 Q4 and Q6, the NPN high-speed switching transistors that follow the 40673's and produce the actual "pulses to be counted" by the TTL sections.... just feed sine-waves to the bases of those transistors and all three counts should appear, if all the TTL is in fact correct. Then I can work out the input side as the very last step.
So in my opinion, it's worth the work. I feel very challenged by this project, but really clear in purpose: I can make this work again. I actually find working on this silly frequency display cathartic in context to the craziness of the business world where I spend my weekdays.
Over Sunday morning coffee... NIXIE SOCKET CARTOUCHE ART!
So finally a couple hours available back in the shop.... biz travel, kitchen remodel, it's nuts.
But - SOCKET FIVE (not in proper order mind you - the fifth of the six - I started with the easiest/cleanest of course) is now in and tested to the 7441 socket. As all the rest, not exactly pretty - but 100% connection to the 7441 chip socket.
ONE MORE TO GO!!! Yay!! Hoping to get that one built up tomorrow evening (these hard ones take about 2 hours EACH to build up - the three easy ones that didn't have such crappy traces I built up in about an hour and a half).
The rest of the board? Ha! Should be a piece of cake by comparison. Sure, a few trace repairs and whatnot - but nothing so miserable as these sockets.
But before I move on to the rest of the board, I plan to do one last "pin by pin check" from socket to 7441 (and the 100k B+ resistors from each tube of course) - if all passes 100%, I'll then pot each cavity with slightly-diluted Elmer's Glue (cuz it can be easily dissolved later), injected with a small needle. That will firm up all the pins in each socket - because of the very compromised and weak / repaired traces, some of the socket pins are like loose teeth - the are 100% in continuity, but only 80% physically attached!! Crazy, right? But I think the "potting" with the water-soluble glue will give them just about as much strength as the originals.
Anyway, so here's the last couple hours' work building one lousy 12-pin Nixie socket
Did I already mention ... ONE MORE TO GO?!!
[Just noticed the one filler paper moved between the bottom pins right when I moved the board for photos LOL! It's just little pieces of card stock - before potting, ALL the little gaps will be filled with their proper paper dividers]
Check out this eBay listing: http://www.ebay.com/itm/RARE-Vintag...749459?hash=item33be392a13:g:cnIAAOSwOdpXxdzn
I see a fair number of SB-650 boxes on eBay over time, and I do wonder if any of them ever actually work! But it's kinda fun to view them as they come up and be able to pretty much know at least from what I see what will need to be fixed.
In the link above, in one photo it displays 99999 - according to the manual, "all 9s" is either a flaky IC25 (SN74H103N dual flip-flop) or the circuit around that chip. Note in that photo the leading zero is off - as it must be, because it's not connected to anything!... but in the shot where all the zeros are as they should be (flip-flop flip-flopping as it should for that photo), that first digit is a blur. That means most likely that the 7441 decoder/driver is failing, and lighting multiple segments, but sometimes settles down and goes blank like in the 99999 photo. Could be just the chips need to be reseated - but it's fun to know enough to at least speculate.
Lastly, note that there is no decimal point - V8 is most likely burned out (like mine was) or has a broken lead or bad dropping resistor. It's not easy to fix, as the entire circuit board has to come out - another aspect of poor design IMO... something as simple as the two lamps (DP and KHz) are soldered in, not in sockets.
Either way, despite the "Great Overall Condition - no other tests done", it's done broke! And my bet is the next buyer will just shelve it and sell it off next year as "great overall condition - no further testing done"
At least this seller is honest enough to say, essentially, "I have no idea if it actually works"
But still fun for me to "armchair diagnose"
Photos below are of the eBay unit for after that link is dead