UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Test Gear and Workshop Equipment

Notices

Vintage Test Gear and Workshop Equipment For discussions about vintage test gear and workshop equipment such as coil winders.

Closed Thread
 
Thread Tools
Old 24th Aug 2020, 12:49 am   #1
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Question HP 8640B displayed freq. is wrong

Intro.
As those of us who have one of these generators know, not only are they heavy but their circuitry is rather complicated. Mine has recently developed a fault, so I'm looking for some guidance to fix it.

Detail.
Briefly, the displayed freq. is wrong for all freq. range settings (detail below *). The actual output freqs. are correct (checked with a Tek. 2465 'scope and a Marconi 2435 freq. counter). When the 8640B is set to external input for the counter / display and a Marconi 2022C sig. gen. is fed to the corresponding ext. input BNC of the 8640B, the displayed freqs. on the 8640B are correct. That tells me that the counter / display section of the 8640B is working O.K. That, in turn, tells me that the input to that counter / display section is being fed with incorrect data from the rest of the 8640B - but from where and why? (See further down the page **)

* From above: "displayed freq. is wrong for all freq. range settings" - detail . .
(I have used the abbreviation 'FRSS' for 'Freq. range switch setting'.)
FRSS 0.5 → 1 MHz, display gives 3.6 to 8.5 MHz.
All FRSS 1 → 16 MHz, ditto.
All FRSS 16 → 128 MHz, display gives 58 MHz to overflow.
All FRSS 128 MHz → 1024MHz display gives 460 MHz to overflow
(N.B. The 512 → 1024 MHz option is not fitted).

Now problems with the nylon gearing for some of the front panel controls (which includes the freq. range switch) are known to be a weak spot in this generator and in mine they all appear to be O.K., but I am wondering about the peculiar printed cct. wafers that are situated and each end of the freq. range switch, but I've never heard of those particular wafers causing problems (yet ).

** My elementary knowledge of the design and architecture of the 8640B now causes me to suspect the divider / filter assy., which I believe contains many (if not all) ECL and EECL logic chips which are now unobtainable.
I do have a manual for the 8640B, but its almost as difficult to digest it as it is to dismantle the 8640B itself!

Any helpful comments / questions to aid me in fixing this sig. gen. are, of course, most welcome.

Al.
Skywave is offline  
Old 24th Aug 2020, 10:30 am   #2
factory
Octode
 
Join Date: Jul 2009
Location: Cheshire, UK.
Posts: 1,706
Default Re: HP 8640B displayed freq. is wrong

Most of these apart from the early ones have rotary switches using the PCB traces & clear plastic rotor discs, the contacts can fall off of these and this is a common problem with any HP TE that uses them. Hopefully they will be found in the case somewhere if this has happened.

I can't comment much further as mine is still in the repair queue waiting for me to fit some new brass gears.

David
factory is online now  
Old 24th Aug 2020, 5:43 pm   #3
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Question Re: HP 8640B displayed freq. is wrong

Thanks David.

A simple Q. that should help me with this . . .
I have eliminated (I think) the counter section (as per above). All the O/P freqs. and amplitudes are correct. Does that last observation eliminate the freq. divider / filter section and the PSU section as two areas where the fault lies?

Al.
Skywave is offline  
Old 25th Aug 2020, 1:06 am   #4
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Default Re: HP 8640B displayed freq. is wrong

I've had a look at the switch wipers as mentioned by David (Factory) without taking the entire switch assy. out - and the wipers look as if they are all intact. I'm prepared to remove it for a close exam, if that appears necessary, (and which is not a trivial task) but I have since had a further think.


My earlier claim that the counter / display section is working O.K. is not necessarily valid. And that is because this section contains a series of gates and dividers that must be associated with the freq. range switch. When the S/G is set to Ext. Freq. source for the counter's time-base, the freq. range switch setting will still be relevant, but with that ext. clock input, the fault remains. That, in turn, implies that the internal clock is O.K.. And that implies a logic fault in the counter / display section, which, incidentally, appears to use TTL (and not unobtainable ECL (Phew!)).

Time to get the cover off the counter box and start scoping waveforms.
Anyway, its now early in the morning and my bed beckons (yawn) . . . there's always tomorrow.

TTFN, Al.

Last edited by Skywave; 25th Aug 2020 at 1:24 am.
Skywave is offline  
Old 26th Aug 2020, 5:53 pm   #5
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Thumbs down Re: HP 8640B displayed freq. is wrong

Very regretfully I have been forced to conclude that this HP 8640B is beyond repair: for me, a very sad day indeed. My decision has been forced by the simple fact that for this fault (and probably many others), the chosen mechanical design and layout of the various assemblies by HP makes any attempt at a repair for this fault simply impossible. So it's off for re-recycling. At present, what I will replace it with I know not - but it most certainly will not be another HP 8640!

So my thanks to David - the only one here who has responded to my cry for help.

Al.
Skywave is offline  
Old 26th Aug 2020, 7:34 pm   #6
factory
Octode
 
Join Date: Jul 2009
Location: Cheshire, UK.
Posts: 1,706
Default Re: HP 8640B displayed freq. is wrong

I'm a bit surprised at the lack of replies to help, the last one of these on here was in extremely poor condition with numerous faults, but did get repaired and that thread got to 7 pages. https://www.vintage-radio.net/forum/...d.php?t=166293

Instead of sending it for recycling, consider offering it on the forum, someone else may be interested in trying to repair it if you decide not to.

David
factory is online now  
Old 26th Aug 2020, 7:45 pm   #7
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,798
Default Re: HP 8640B displayed freq. is wrong

Sorry, I don't know my way around the counter... never had to dig into one. So can't help much.

It has to have a fast prescaler to make the external counter input work. But for counting the output frequency, it could run off of the ECL divider chain on the filter board.

There are TWO filters for each range, an upper freq range one and a lower. A pot on the tuning shaft to the cavity osc drives a comparator so it changes filter part way up each band (with hysteresis so it doesn't sit rattling at some point)

However, going too low down the divider chain would increase the needed gate time a bit too much. It would have to play with the gate time to get the right reesults.

Oh, the OBIC, dot-type displays include latches and decoders.

The later 7 seg displays will need those on the board.

As I said, its one area I haven't had to dive into.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 26th Aug 2020, 10:03 pm   #8
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Default Re: HP 8640B displayed freq. is wrong

I've owned this sig. gen. for about 20+ years and during that time it has given good service. I obtained it a good price because all the bevel edge gears were broken. (I fixed those and have repaired other 8640s since with that problem). It also came with the manual. Unfortunately, it turned out that that manual was no good for fixing the faults - there were several, at least - the manual was a far too early an edition.
The counter section was faulty. There were also indications of a fault in the R.F. divider - which uses unobtainable ECL chips. Unfortunately, the mechanical design of the counter section makes any maintenance unquestionably impossible; ditto the divider section. For the counter section, having the special riser / extender card is essential for even making a start on a repair. (Why did HP mount the top pcb upside-down?) No such item here, nor found on the 'Net. Then there are the three SMC connectors underneath which only with great difficulty are removable; re-connection is impossible on account of insufficient space to grasp and twist their threaded ends. And that lack of access space problem pervades throughout this entire instrument: HP should have used a much larger case with a sensible layout. But then the whole mechanical assy. appears to have been chosen to make maintenance / repair as difficult as possible.
For my simple hobby use, owning an 8640 is nice - but is an unnecessary luxury; a simpler, maintainable instrument is more suited to my needs. And that is what I will now search for. 'Advance Instruments' have always been favoured here, but certainly not Marconi - for same reasons as above.

Al.
Skywave is offline  
Old 26th Aug 2020, 10:05 pm   #9
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Default Re: HP 8640B displayed freq. is wrong

Quote:
Originally Posted by Radio Wrangler View Post
As I said, it's one area I haven't had to dive into.
Believe me: you don't want to have to!

Al.
Skywave is offline  
Old 26th Aug 2020, 11:32 pm   #10
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,798
Default Re: HP 8640B displayed freq. is wrong

When I did muck about with the occasional 8640, my desk at work was about 100ft away from the production line where they built and debugged them, so all the test jigs were available and running. I still didn't want to delve into the counter!

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 27th Aug 2020, 2:59 pm   #11
orbanp1
Heptode
 
Join Date: Nov 2009
Location: London, Ontario, Canada
Posts: 672
Default Re: HP 8640B displayed freq. is wrong

Hi Al,

Clean the card edge connectors (even that they are gold plated), there are a number of boards "stuck together" in the counter part.
Mine had similar problems (plus the usual gear trouble, and others, it was a "for parts only" acquisition), and just by disassembling and putting that part back together did help the counter.
You might get lucky!

Regards, Peter
orbanp1 is offline  
Old 27th Aug 2020, 6:53 pm   #12
Nymrod121
Nonode
 
Nymrod121's Avatar
 
Join Date: Apr 2012
Location: Worcestershire, UK.
Posts: 2,039
Default Re: HP 8640B displayed freq. is wrong

I'd agree with Peter ... and if memory serves, Richard (trh01uk) did a very good repair job on an '8640B with frequency counter problems which involved him getting deeply involved with the ECL/TTL circuitry.

Mine is scheduled for a set of brass gears (from Paul Koretko - 'across the pond') when funds permit ...

Guy
__________________
"What a depressingly stupid machine." [Marvin: HHGTTG]
Nymrod121 is offline  
Old 1st Sep 2020, 5:49 pm   #13
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Arrow Re: HP 8640B displayed freq. is wrong

Update.
O.K., I've relented: a stay of execution. Every time I go in and out of the house I can't help but see this S.G. sitting on the floor by the front door looking all sad and forlorn. I now realise that my earlier decision to scrap it in total was simply motivated by my utter frustration, desperation and outright annoyance at HP - in my opinion - for having designed and built this item to such a very poor standard, especially mechanically. Some of the electronics is not exactly brilliant either; the manual (operator's and service) is pretty awful too. OTOH, the performance of an 8640 (when fully functional) is, of course, simply superb: that aspect is unquestionable.
So, in conclusion, yes, I lost my temper which caused me forget my three principles of any tricky repair: Patience, Perseverance and Persistence. Following a deep analysis of this S.G.'s functions, I now have a couple of strategies to implement, one of them being a 'repair by modification': in a phrase, needs must when the devil drives (or in this case, HP!)

Hopefully, more to follow - of a positive nature.

Al.
Skywave is offline  
Old 1st Sep 2020, 7:51 pm   #14
factory
Octode
 
Join Date: Jul 2009
Location: Cheshire, UK.
Posts: 1,706
Default Re: HP 8640B displayed freq. is wrong

I hope it goes well, I found using the pdf manuals very frustrating with the last repair on the forum.
I've got a dead tree format manual on the way from the US now, hopefully it is the correct version for mine, "Artek" have a list of prefix's it covers, I checked it matched before buying.

David
factory is online now  
Old 1st Sep 2020, 9:05 pm   #15
gingpeakin
Hexode
 
Join Date: Mar 2009
Location: Worthing, Sussex, UK.
Posts: 274
Default Re: HP 8640B displayed freq. is wrong

I can also sympathise with you - there sometimes seem to be so many variations of the same model with enough differences to make it a real headache when fault finding. Not just HP but Tektronix are as guilty.
On another note I've repaired an HP5326B counter that had a faulty ECL chip by using a combination of other ECL glue logic chips stuck together. The parts are usually fast enough to get away with that sort of thing.
Ging
gingpeakin is online now  
Old 1st Sep 2020, 9:53 pm   #16
factory
Octode
 
Join Date: Jul 2009
Location: Cheshire, UK.
Posts: 1,706
Default Re: HP 8640B displayed freq. is wrong

The HP 5326 series, that's another with major design changes, I have earlier one in the repair queue, the later ones had the motherboard redesigned to allow space for the 10544 ovenized oscillator.

David
factory is online now  
Old 1st Sep 2020, 11:03 pm   #17
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Angry Re: HP 8640B displayed freq. is wrong

Thanks for your responses: all read with interest.


In past employments, I frequently designed and built items of test gear for specialized purposes and wrote the operator's and maintenance manuals. To produce a quality manual is not an easy matter. It takes time: to get the right balance of detail and verbosity is a real challenge, and, most importantly, it must be vetted by someone who knows nothing about the item that the manual is for. (It's far too easy for the author to assume 'knowledge' and think 'but that's obvious!')

Here is an extract from the operator's and maintenance manual for the HP 8640B.

Internal AM and FM is provided by the modulation oscillator.
The oscillator drives either:
the AM mod. circuits and an AM Output port
or
the FM mod. circuits and (the) FM Output port
or
all four.

"All four" ? The above gives two options! Seems to me that the author should have used 'or' instead of 'and'. But even with that change, things are still not very clear. E.g. if "all four" are driven, does that result in simultaneous AM and FM of the final O/P signal? etc., etc.

This is just a small but representative example of sloppy writing & documentation that seems so prolific in the 8640 manual.

As is well-documented, a weak spot in the design is the freq. range and FM deviation controls, particularly the infamous plastic gears. Of all materials, why was plastic chosen? And why, in the manual, are there no circuit diagrams and detailed descriptions of that complicated 'gearbox' ass'y.? It is most certainly needed. Why wasn't that very intricate mechanical inter-switching not implemented in logic gates? Using traditional logic would have provided a much more reliable method than a mechanical solution.

Why did the author consider it necessary to devote a significant section of the manual giving elementary tuition on logic gates, etc? That is generic info. - which was obtainable elsewhere at the time of the design and build of the 8640 - and is not specific to this S.G.

I could give many more examples to justify my viewpoints, but I think the above representative samples are sufficient to illustrate my various criticisms.

Al.
Skywave is offline  
Old 10th Sep 2020, 12:17 am   #18
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Question Re: HP 8640B displayed freq. is wrong

Seem to be making some headway with this generator. Have cleared a couple of faults, but now the freq. display is flashing (in all places) at about a rate of 2 - 3 times per second. I suspect that that is due to something fundamentally wrong, rather than another failed i.c.

Any ideas, please?

Al.
Skywave is offline  
Old 10th Sep 2020, 9:41 am   #19
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,798
Default Re: HP 8640B displayed freq. is wrong

Quote:
Originally Posted by Skywave View Post
Why did the author consider it necessary to devote a significant section of the manual giving elementary tuition on logic gates, etc? That is generic info. - which was obtainable elsewhere at the time of the design and build of the 8640 - and is not specific to this S.G.
Indeed it is. But fashions change in management, and things like that are decreed from above to be inserted into all manuals.... and then some time later someone has a different bee in their bonnet.

Also remember that the military had some input, particularly in terms of the skill level manuals had to be written for.

As a profit-making corporation with shareholders and stock market analysts to pander to, the cost of the extra paper would be calculated and had to be justified. HP gear sold to a very wide range of customers, and as you illustrate, you can't keep them all happy at once.

HP was mostly competent, at most things, and has a rather good track record of moving along the sate of the art. We had fun doing it as well. There are a few blunders, and a couple of spectacular blunders, but taken in proportion to the size of the organisation, it's not a bad record. You also have to look at each product in terms of the era it was developed and introduced in. These instruments were doing things you just couldn't do with jellybean parts. It was also an era when most things were not as reliable as we are now used to.

There were ECL divide-by-2 chips on the open market in the sixties. But the 8640 needed something which wouldn't be the prime limitation on its phase noise. That would have been disappointing and would have wasted the effort which went into the cavity oscillator.

I know they're all a long way beyond their expected lifetime, but the strengths of the 8640 are so good that it's very disappointing that they are being let down by plastic slide switches, plastic gears.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 10th Sep 2020, 9:43 am   #20
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,798
Default Re: HP 8640B displayed freq. is wrong

I think the flashing display is used to indicate overrange.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 11:34 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.