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 10th Sep 2020, 1:31 pm   #21
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Question Re: HP 8640B displayed freq. is wrong

Yes David: that thought had crossed my mind.
Question is, though, where to start looking for the cause? Having successfully implemented some repairs, this display problem was not there. Return to this S.G. a day later, switch it on - and then this problem is manifest!

Al.
Skywave is offline  
Old 10th Sep 2020, 2:38 pm   #22
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: HP 8640B displayed freq. is wrong

Sounds like the usual suspects, dry joints, maybe a bad connection inside a package... those counters run awfully warm. They're rather cramped and airflow is somewhat limited. I think those OBIC displays (if it has that variety ) dissipate 0.5 to 1 Watt per digit. They did move to 7 seg displays later that ran cooler, but by then we had all the 8640s we needed in the plant and were working on the 8657. There was more push to get the frequency coverage up and manufacturing cost down than there was to keep the phase noise performance. Cellphones were starting to take over the world.

Some shotgun resoldering maybe? selective freezer on chips is difficult because of wanting it working while you spray.

Still, if you'd gone down the Marconi road, you'd be fighting those ****** griplets...

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, 8:17 pm   #23
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Default Re: HP 8640B displayed freq. is wrong

The displays are 7-segment. The S/N of this generator is 2428/A25521 - if that's any help.

As for general fault finding - freezer spray or otherwise - very difficult if not impossible with the counter / TB assy. on account of the 'sandwich' assy. and not having the appropriate extender card.

I've spent many, many days trying to fix this S.G., but will try a bit more. Not that I achieve 100% success with any repair, but I just hate admitting defeat when faced with a problem of which I like to think "Ah-Ha! I'm familiar with this - yes, I can fix it!"

I don't know what it is about me and R.F. signal generators, but on recollection, over the many years I've had more trouble with S.G.s that any other item of T & M kit - hobby-wise or in employment - especially Marconi's stuff.
And then this HP S.G. threw me a challenge

Al.
Skywave is offline  
Old 10th Sep 2020, 9:03 pm   #24
factory
Octode
 
Join Date: Jul 2009
Location: Cheshire, UK.
Posts: 1,706
Default Re: HP 8640B displayed freq. is wrong

Another late model with ordinary 7 segment LED displays, mine is prefix 2350A and the correct manual arrived a couple of days ago (has lots of error/change sheets too).

David
factory is offline  
Old 18th Sep 2020, 12:07 am   #25
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 brief, Sept. 14th.

I am making some headway, but the going is really difficult and is very, very time-consuming.
Hopefully more to come and on a more positive note.

TTFN,
Al.
Skywave is offline  
Old 18th Sep 2020, 5:04 am   #26
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: HP 8640B displayed freq. is wrong

The serial number says it was built in summer 1984, in the US. I think the 8640 was originally developed at Stanford Park, but may have moved over the production lifetime. South Queensferry made the ones sold in the EU, though a few were swapped across continents to balance production against demand.

Do you think you may have done something nasty to a sig gen in a previous life?

You've reminded me that my Marconi 2008 needs some therapy, looks like a PSU rail down at first guess.

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 24th Sep 2020, 10:28 pm   #27
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Default Re: HP 8640B displayed freq. is wrong

The problem I reported in my OP has now been cleared. Actually, I completed that fix a couple of weeks ago, but since then have been trying to fix another - no FM: that is currently W-I-P and is being a right to fix. (Nowhere have I read any reports claiming that fixing a faulty HP 8640B is a trivial affair )

This is a long, detailed report. Its main aim is to hopefully assist others who may experience this described fault with their HP 8640B.

The freq. display fault. There was more that one defect; each, which in turn, masked the others.
1. The is an ECL programmable counter in the pre-scaler section that functionally sits between the main UHF oscillator and the time-base / counter boards. That had failed, but I was fortunate to find a spare. There was also a total absence of the -5 v. power rail - dealt with in 3 below.
2. On the time-base panel I found a counter i.c. whose TTL O/P was rather low for a logic 1. It was also temperature sensitive. A replacement i.c. effected a cure.
3.The most difficult and annoying fault to locate was an intermittent - which, in my opinion, is caused by a really poor design. What follows is very difficult to explain in a few words, but I consider that it is worthwhile going into some detail since this fault took many hours to identify and fix - and for reason stated in my opening paragraph.
The time-base panel has an edge connector, into which are fed +5v. for the TTL on the two aforesaid boards, the pre-scaler, and -5 v. for that pre-scaler (ECL), plus the coded signals to produce the appropriate readings on the freq. display. Those signals and psu rails are delivered to the time-base panel on an additional accompanying edge connector. The two edge connectors provide the transfer of those signals and psu rails by means of a square cross-sectional block of insulating material to which a number of springy wires are wrapped around its periphery. So there are two sets of connections: input pcb to block; block to time-base panel. Those springy wires are each slightly raised on opposite parallel planes at one point each in an attempt to ensure a good clean connection for each signal & power line. That whole sub-assy., board, block and edge connectors, rely on appropriate pressure on those connectors only provided by adequate mechanical tightness of the time-base / counter screening cover screws - and nothing else. The aforesaid block relies on a couple of pins to obtain correct alignment, for which there is considerable tolerance. To me, overall, a singularly inadequate method of obtaining a reliable, secure set of connections. However, once all the electrical connections had been thoroughly cleaned and polished, plus getting the counter screening case adequately aligned (there are considerable tolerances in the fixing holes) and getting the torque on the screening case screws just right, the freq. display then gave the correct numbers. Phew!

However, the fun doesn't end there. There is now the non-trivial matter of re-assembling the upper part of the front panel. Like everything else mechanical in the 8640B, you soon realise that this section (and its many parts) was not designed with maintenance in mind. And as is the case with so many entries in the manual, what is provided in it for re-ass'y. is confusing, poorly arranged and difficult to understand. Eventually, I gave up with the manual and used the time-honoured approach of "This will only go back together again correctly one way!". A few hours later, with persistence wearing thin, I was still persevering and thinking same; however, I won in the end.

But the battle is not yet totally won: there is no FM. This now W-I-P and will be dealt with in another post or thread, q.v.

Al.

Last edited by Skywave; 24th Sep 2020 at 10:35 pm.
Skywave is offline  
Old 24th Sep 2020, 11:00 pm   #28
Julesomega
Nonode
 
Julesomega's Avatar
 
Join Date: Dec 2012
Location: Stockport, Greater Manchester, UK.
Posts: 2,085
Default Re: HP 8640B displayed freq. is wrong

Thanks for posting all your findings Al. I have subscribed to this thread and will probably print it out to add to the manual when you have succeeded with yours, just in case...
__________________
- Julian

It's good here
Julesomega is offline  
Old 25th Sep 2020, 12:11 am   #29
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Default Re: HP 8640B displayed freq. is wrong

In order to reduce that post's length, I did omit details of bits of 'one-off' test items that I fabricated; how to use them & interpretation of the results. Also hints of how best to dismantle and re-assemble the A9 module - which is far from easy . I have stored that data for my own use (hopefully never needed ). When I have finished all repairs (perhaps that should read "if" ), I'll collect all such stuff and accordingly present it.

As is universally recognised, the HP 8640B is a quality S.G. - in terms of performance and utility. As for its mechanical construction, etc. - I think I've made my viewpoint of that aspect abundantly clear.

TTFN,

Al.
Skywave is offline  
Old 25th Sep 2020, 7:55 am   #30
trh01uk
Octode
 
trh01uk's Avatar
 
Join Date: Nov 2010
Location: Ross-on-Wye, Herefordshire, UK.
Posts: 1,648
Default Re: HP 8640B displayed freq. is wrong

Quote:
Originally Posted by Skywave View Post
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.

Al,

I can't follow the details of your problem, but I can say that the counter section of my 8640B failed a few years ago. I took it apart and fixed it. Yes, I needed extender cards, so I just went and bought some. Not sure why you are saying it is impossible to fix. Sure the ECL chips may be hard to find - but not impossible. For a start there are other scrap 8640Bs out there that will have them in......Ebay regularly has parts of this generator on offer.

I recommend that you offer yours for sale as "for parts or restoration" - and given you say it generates RF at the right level and frequency, it seems to me that the counter part is a nice to have extra, but its not essential. While mine was broken, I just hooked up my stand alone frequency counter to it to get the frequency to what I wanted. Sure that's a bit mandraulic, but its do-able.

My reading of the thread about a replacement for the 8640B told me that there really isn't anything else that performs as well (phase noise?) and is affordable for most of us. So you will be doing a service if you recycle yours by offering the parts (one by one if you like!) on here, or some similar sales point.


Richard
trh01uk is offline  
Old 25th Sep 2020, 8:00 am   #31
trh01uk
Octode
 
trh01uk's Avatar
 
Join Date: Nov 2010
Location: Ross-on-Wye, Herefordshire, UK.
Posts: 1,648
Default Re: HP 8640B displayed freq. is wrong

And I would further add as a general comment to my post above to everyone interested in keeping our 8640Bs going, that it may be time for some serious re-engineering. Already we have brass gears being made by a guy in India I gather from another helpful thread on here. We have extensive brain power on this forum, judging by what I see in many posts - it would be worth getting some collective wisdom on how to replace whole blocks on this generator - or maybe some external solution - which will keep these valued bits of kit going for another generation.


Richard
trh01uk is offline  
Old 25th Sep 2020, 8:14 am   #32
trh01uk
Octode
 
trh01uk's Avatar
 
Join Date: Nov 2010
Location: Ross-on-Wye, Herefordshire, UK.
Posts: 1,648
Default Re: HP 8640B displayed freq. is wrong

I know this is a bit late now - but to bolster my point about re-engineering around unobtainable parts, I enclose a paper I picked up from somewhere when I was doing my own repairs around the counter area. It agrees the original HP EECL parts were unobtainable in 1999, and found a modern replacement in the Motorola MC100EL31D, a chip I note is still available on Ebay for £17.49. It maybe that the job could be re-evaluated now and a better replacement found some 20 years later.

Richard
Attached Files
File Type: pdf Repair of ECL dividers.pdf (107.1 KB, 60 views)
trh01uk is offline  
Old 25th Sep 2020, 2:04 pm   #33
factory
Octode
 
Join Date: Jul 2009
Location: Cheshire, UK.
Posts: 1,706
Default Re: HP 8640B displayed freq. is wrong

Quote:
Originally Posted by Skywave View Post
3.The most difficult and annoying fault to locate was an intermittent - which, in my opinion, is caused by a really poor design. What follows is very difficult to explain in a few words, but I consider that it is worthwhile going into some detail since this fault took many hours to identify and fix - and for reason stated in my opening paragraph.
The time-base panel has an edge connector, into which are fed +5v. for the TTL on the two aforesaid boards, the pre-scaler, and -5 v. for that pre-scaler (ECL), plus the coded signals to produce the appropriate readings on the freq. display. Those signals and psu rails are delivered to the time-base panel on an additional accompanying edge connector. The two edge connectors provide the transfer of those signals and psu rails by means of a square cross-sectional block of insulating material to which a number of springy wires are wrapped around its periphery. So there are two sets of connections: input pcb to block; block to time-base panel. Those springy wires are each slightly raised on opposite parallel planes at one point each in an attempt to ensure a good clean connection for each signal & power line. That whole sub-assy., board, block and edge connectors, rely on appropriate pressure on those connectors only provided by adequate mechanical tightness of the time-base / counter screening cover screws - and nothing else. The aforesaid block relies on a couple of pins to obtain correct alignment, for which there is considerable tolerance. To me, overall, a singularly inadequate method of obtaining a reliable, secure set of connections. However, once all the electrical connections had been thoroughly cleaned and polished, plus getting the counter screening case adequately aligned (there are considerable tolerances in the fixing holes) and getting the torque on the screening case screws just right, the freq. display then gave the correct numbers. Phew!
Al.
Is that awful connector your describing like these, which are used in a few 1970's era HP counters?
Click image for larger version

Name:	SAM_1675a.jpg
Views:	133
Size:	110.7 KB
ID:	216395
It's a shame you didn't post some pictures before fault finding, as if we had seen those connectors we could have mentioned they were unreliable.
And yes removing that board, cleaning & refitting cured it's fault too.

David
factory is offline  
Old 25th Sep 2020, 3:12 pm   #34
trh01uk
Octode
 
trh01uk's Avatar
 
Join Date: Nov 2010
Location: Ross-on-Wye, Herefordshire, UK.
Posts: 1,648
Default Re: HP 8640B displayed freq. is wrong

Looking at the number of ways on those connectors, and the avoidance of exact board alignment, no doubt the original designer would respond to all this criticism with a "seemed like a good idea at the time" type of comment!

Richard
trh01uk is offline  
Old 25th Sep 2020, 5:44 pm   #35
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Arrow Re: HP 8640B displayed freq. is wrong

Quote:
Originally Posted by factory View Post
Is that awful connector your describing like these, which are used in a few 1970's era HP counters?
Attachment 216395
Yes.

Al.

Quote:
Originally Posted by factory View Post
It's a shame you didn't post some pictures before fault finding . .
But that's being wise after the event.
It was only during the process of fault-funding that that block became suspect. Hence, it was only after cleaning it, etc., and then the whole ass'y. reinstated and thus finding that the fault had been cleared, that substantial suspicion fell upon that block. But if I had known in advance what I now know, then yes, a photo. would have been a good idea.

Al.
Skywave is offline  
Old 25th Sep 2020, 7:21 pm   #36
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: HP 8640B displayed freq. is wrong

Quote:
Originally Posted by trh01uk View Post
Looking at the number of ways on those connectors, and the avoidance of exact board alignment, no doubt the original designer would respond to all this criticism with a "seemed like a good idea at the time" type of comment!

Richard
T'wasn't me!

I have an alibi!

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 Sep 2020, 12:32 pm   #37
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Arrow Re: HP 8640B displayed freq. is wrong

Attached (as a downloadable pdf file) are the design details in depth of the HP 8640B. Plus, on p. 13, you also get a photo of the engineers involved. Note: I don't see David there!

Al.
Attached Files
File Type: pdf The HP8640B design in detail.pdf (2.77 MB, 77 views)
Skywave is offline  
Old 26th Sep 2020, 12:54 pm   #38
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: HP 8640B displayed freq. is wrong

Whenever a new month's HPJ flopped into our in-trays, all work stopped as we immediately perused the mini biographies of the people responsible. Primarily we were interested in the hobbies. Making stained-glass windows was big. Coaching youth soccer teams was seen as a ploy to look like a better civic person. Churchy things, ditto. Rebuilding a home, a motorbike or mustang seemed more normal. Then we'd get round to looking at the instruments and what went into them.... we were looking for things to nick, of course.

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 Sep 2020, 1:10 pm   #39
trh01uk
Octode
 
trh01uk's Avatar
 
Join Date: Nov 2010
Location: Ross-on-Wye, Herefordshire, UK.
Posts: 1,648
Default Re: HP 8640B displayed freq. is wrong

I always aspired to work for HP, and actually managed to get a job at South Queensfeery around 1984/5 I think (can't be sure of the year as memory has failed me). Then before I could start, HP suddenly announced there was a recession on, and they were not recruiting anyone, and managers were going to take a 10% pay cut. Sounds like a great firm to work for - and I regretted losing that opportunity.

Richard
trh01uk is offline  
Old 28th Sep 2020, 12:06 am   #40
Skywave
Rest in Peace
 
Join Date: Jun 2006
Location: Chard, South Somerset, UK.
Posts: 7,457
Arrow Re: HP 8640B displayed freq. is wrong

Quote:
Originally Posted by trh01uk View Post
{snip} . . . . to everyone interested in keeping our 8640Bs going, that it may be time for some serious re-engineering.
W.r.t. my 'no FM' problem, based on recent detective work, that is the approach I am now pursuing.

Al.
Skywave is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 2:38 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.