Frequency counter won't count.

[neo]

Well at least that is working. Just as an FYI when you switch between resolutions(with no input) you can expect all of the digits to light up briefly before the blanking circuit disables all but the right most 3 digits.

It does exactly that.

[neo]

Forgive me if this is impetuous but i am a bit too stoked to care.

I WON!!! I WON!!!!! FOR ONCE IN MY LIFE I'VE WON!!!!

Sorry i don't win a lot, now that i have got THAT out of my system. The high frequency counter works! It even gave me a big clue as to the problem elsewhere in the unit! Any number above the first 3 is a garbage number, in GHz mode they will light but are useless because they don't display correctly. In the 1M range they will not light at all past the first 3.

Also Cub, would you mind telling me how to read the serial # so i can know the year for sure?

[neo]

Pictures.

[Cubdriver]

The instruments made after 1960 have a two part serial number - 3 or 4 digits forming a date code followed by a dash (early units) or a letter (that seems to indicate country of manufacture for later units) and then another 5 digits for the unit's serial number.

My understanding, which may not be 100% correct, is that the first digit (or first two, for four digit prefixes), when added to 1960, will give the year that production run started, and that the last two give the week it began.  For example, the 1328A beginning on my 5340A would indicate mid 1973 production - 1960 + 13 gives 1973, and 28 for the 28th week.  'A' designates US made. 

A lot of the LCR meters were made in Japan; my 4262A starts with 2022J which would indicate 22nd week of 1980, made by Yokogawa in Tokyo.

Some of the signal generators were made in Böblingen, Germany; they have a 'G' in the serial number.

The thing I'm not certain of is the production week numbers, as I've seen instruments with late production week numbers (40+) that have semiconductors with early date codes from the following year.  That's why I think that has to do with when a run started, but that the instruments may have been fabricated over a longer period.

-Pat

[neo]

The instruments made after 1960 have a two part serial number - 3 or 4 digits forming a date code followed by a dash (early units) or a letter (that seems to indicate country of manufacture for later units) and then another 5 digits for the unit's serial number.

My understanding, which may not be 100% correct, is that the first digit (or first two, for four digit prefixes), when added to 1960, will give the year that production run started, and that the last two give the week it began.  For example, the 1328A beginning on my 5340A would indicate mid 1973 production - 1960 + 13 gives 1973, and 28 for the 28th week.  'A' designates US made. 

A lot of the LCR meters were made in Japan; my 4262A starts with 2022J which would indicate 22nd week of 1980, made by Yokogawa in Tokyo.

Some of the signal generators were made in Böblingen, Germany; they have a 'G' in the serial number.

The thing I'm not certain of is the production week numbers, as I've seen instruments with late production week numbers (40+) that have semiconductors with early date codes from the following year.  That's why I think that has to do with when a run started, but that the instruments may have been fabricated over a longer period.

-Pat

Well with that convention mine was minted in 1978 BUT there a was a sticker on the front that said "M1973" when i first got it. So really it is at best a guessing game on this one. I thought they had all but completely switched over to leds by 78 though.

Also the more i go into depth with the display issue the weirder it gets. Is it just my perception of it that makes it that way or does it seem fairly weird to you?

[Cubdriver]

Well with that convention mine was minted in 1978 BUT there a was a sticker on the front that said "M1973" when i first got it. So really it is at best a guessing game on this one. I thought they had all but completely switched over to leds by 78 though.

Also the more i go into depth with the display issue the weirder it gets. Is it just my perception of it that makes it that way or does it seem fairly weird to you?

I just looked through my catalogs.  Surprisingly, they show and call out the nixies as the readout up to the 1982 catalog.  The seven segment LEDs appear in 1983.  Perhaps the switchover began during '82.

I think your counting issue will wind up being simple once it's figured out.  As is said in the other thread, I suspect the carry signal is getting hosed somewhere along the line, but I haven't dug in enough to how it counts to be sure. 

The older counters were pretty much literally daisy chained decade counters that counted input pulses for as long as the gate timer held them on.  The timer would say 'start', and the conditioned input signal would clock the first digit, which would count from zero.  When it reached ten, it would send one pulse to the second digit, which would increment by one.  The second digit would pulse the third every ten counts, and so on, until the count time expired and the gate closed.

Newer ones (and this one, I'm assuming) used more sophisticated methods that I'm not yet 'up' on, so I'm of limited help until I do some more digging on the operation.

-Pat

[Cubdriver]

And on that note, I must cease my nocturnal ways and try to get some sleep, as it's after four am in these parts.   

-Pat

[neo]

And on that note, I must cease my nocturnal ways and try to get some sleep, as it's after four am in these parts.   

-Pat

Hope you get some good  . Also yeah i think it will be quite simple once i "get it" but for now it is a very strange well defined phenomena.

[bitseeker]

I though the first set of digits was the date code of the design on which that particular unit was based. Not a big deal as neither interpretation nails down the manufacture date. You'd still have to find some component date codes inside in order to approximate that.

[neo]

Interesting update, in the 100k range it used to count 3 digits of the frequency for example (4)61.5 MHz the 4 not being visible as the left 5 digits were blanked but now it does the same thing with the junk numbers.

[Cubdriver]

I though the first set of digits was the date code of the design on which that particular unit was based. Not a big deal as neither interpretation nails down the manufacture date. You'd still have to find some component date codes inside in order to approximate that.

That sounds like a plausible explanation as well.  Thus far I've only seen early following year date code components in units that had high 'week' codes, but I admittedly haven't dug too deeply into that.  I'll try to pay more attention from now on.  It would be telling to find one with components that were a year or more  'younger' than the date code would indicate (eg 1975 parts in a 1973 coded instrument, rather than just early 1974 parts in a late 1973 unit) - I can't imagine that every instrument had a new design iteration every year.

On another note, my extender card is on the move and has allegedly made it to Springfield, MA.  At this rate perhaps I'll see it by Tuesday.  <fingers crossed>

-Pat

[neo]

I though the first set of digits was the date code of the design on which that particular unit was based. Not a big deal as neither interpretation nails down the manufacture date. You'd still have to find some component date codes inside in order to approximate that.

That sounds like a plausible explanation as well.  Thus far I've only seen early following year date code components in units that had high 'week' codes, but I admittedly haven't dug too deeply into that.  I'll try to pay more attention from now on.  It would be telling to find one with components that were a year or more  'younger' than the date code would indicate (eg 1975 parts in a 1973 coded instrument, rather than just early 1974 parts in a late 1973 unit) - I can't imagine that every instrument had a new design iteration every year.

On another note, my extender card is on the move and has allegedly made it to Springfield, MA.  At this rate perhaps I'll see it by Tuesday.  <fingers crossed>

-Pat

Great to hear it is on the move again. Is it a bit odd how on the 100k mode the 3 digits worked fine when i took a picture but not later? Possible clue.

[neo]

I have come up with a possible way to resolve the error without checking every single part, this idea is of course just a possibility and not intent. I could always get a newer one that isn't in great shape and the main boards ^{except display itself} from it, because as near as i can there would be no difference except age. _{Don't worry about me killing one ^yet to do that i would have to be able to buy another or find one inordinately cheap.}
As i said, short of just checking every little thing that it might be i am out of ideas. I get lost quickly when trying to understand the schematics.

[Cubdriver]

Mine is still sitting on the bench waiting for me to continue looking at it - I've been helping a friend do some painting and that has sucked up all my free time.  Hopefully later this week I will have time to dig into the schematics and try figure out how the blanking works.

This issue with yours still seems to me to be some sort of bad connection since it seems to work and then not work depending on its mood and the moon phase.  That's a  problem to try to trace.

-Pat

[neo]

Mine is still sitting on the bench waiting for me to continue looking at it - I've been helping a friend do some painting and that has sucked up all my free time.  Hopefully later this week I will have time to dig into the schematics and try figure out how the blanking works.

This issue with yours still seems to me to be some sort of bad connection since it seems to work and then not work depending on its mood and the moon phase.  That's a  problem to try to trace.

-Pat

 Yea i have definitely felt  , might help if i was smart enough to understand the schematics but then again it might help if the schematics were on paper so i could actually work with them.
The two screws on the display board that hold it in are missing so i take that to mean someone has tried to fix it before, which would ironically mean this problem is why it was so cheap AND fully functional except from the display

Short excerpt from my head.
"So it could be as simple as say, a single bad solder joint?
Simple fix but it could take hours to find especially if it doesn't look obvious.
Well are the card's loose?
No the card's are not loose, i checked each one.
Got any ideas?
No do you?
Let's try wiggling a card again.
OK."

Yes, i talk to myself. Don't all mad scientists?

[neo]

Well something interesting happened, i tried turning it on just to see the 3 digits work and it didn't do anything. Then later it was back to its normal state of semi functioning.

[neo]

A couple relevant bits, HP 1820-0986 is in fact directly replaceable with DM86L75N because they are one and the same.

Took me an hour to find that out, and to find the datasheet. http://www.ralphselectronics.com/productimages/SEMI-DM86L75N.PDF

Directly relevant to my repair, IC#2 was replaced sometime post '93.

[Tharkun]

Hi it’s been a while now but did you end up solving this problem?  I have exactly the same counter and it started doing exactly the same thing.  Just wondered if you had solved it...

Tim