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 3rd Jul 2020, 8:28 pm   #101
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Replacement for HP8640B signal generator

I have my Mi 2019A running on the bench now. All seems well. I've stood it on one inch laths to do what I can to keep it cool, as advised.

It's nice not to have a fan running all the time but in my troglodyte workshop I miss the 7-seg LEDs and illuminated meter of the 8640B. It seems a bit strange Marconi didn't bother with backlights on the LCD displays. And all this modern button pushing on the front panel instead of having an analogue knob to wind will take some getting used to.

Yes I know: I'm never satisfied.

73, Alan
Alan_G3XAQ is offline  
Old 4th Jul 2020, 12:11 am   #102
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Replacement for HP8640B signal generator

Quote:
Originally Posted by Alan_G3XAQ View Post
7-seg LEDs and illuminated meter of the 8640B.
Cheek! Them's dot matrix OBIC displays. Run half a watt apiece to help test your power supply.

I suppose I was hoping you'd have gone for the Marconi 2017 sig gen. That'a an impressive box

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 4th Jul 2020, 11:17 am   #103
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Replacement for HP8640B signal generator

Dot matrix, LED, backlighting, whatever. Never mind, I'll just keep a torch beside it. 2017s seem to go for 8 times what I paid Bob for the 2019A!

Alan
Alan_G3XAQ is offline  
Old 4th Jul 2020, 3:17 pm   #104
factory
Octode
 
Join Date: Jul 2009
Location: Cheshire, UK.
Posts: 1,706
Default Re: Replacement for HP8640B signal generator

Quote:
Originally Posted by Radio Wrangler View Post
Quote:
Originally Posted by Alan_G3XAQ View Post
7-seg LEDs and illuminated meter of the 8640B.
Cheek! Them's dot matrix OBIC displays. Run half a watt apiece to help test your power supply.

David
They look more like 7-segment LED's on mine, but it is an later one from the mid 1980's.

David
Attached Thumbnails
Click image for larger version

Name:	SAM_0662a.jpg
Views:	80
Size:	64.2 KB
ID:	210290  
factory is offline  
Old 4th Jul 2020, 4:25 pm   #105
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Replacement for HP8640B signal generator

They are indeed 7 segment ones. There was a later counter board towards the end. I don't think that many were made and I pretty much forgot about them.

Don't look as pretty but will give you two major advantages. The power demand and heat should be a good few watts better and you can find replacement 7-seg displays while OBICs are hard to find.

I wonder if Alan has one of the last units, too.

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 4th Jul 2020, 5:17 pm   #106
turretslug
Dekatron
 
turretslug's Avatar
 
Join Date: Nov 2011
Location: Surrey, UK.
Posts: 4,385
Default Re: Replacement for HP8640B signal generator

Is it the OBICs that are also used in the RA1772 and sometimes fail? If so, I'll administer myself a slight kick in the shin for not saving a load of NOS that were thrown out at work a way back!
turretslug is offline  
Old 4th Jul 2020, 5:59 pm   #107
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Replacement for HP8640B signal generator

Now your 2019A has arrived I can offer a few tips for long term use. In 'slow time' I would advise you start looking for the little 20cm long SMC (F) to SMC (M) RF cable used for servicing. It is described in the service manual.

You can either make one yourself using crimped connectors and RG174 cable or it may be possible to buy a ready made SMC-SMC cable on ebay but this will probably need a gender changing adaptor at one end. Without this cable it will be difficult to service the sig gen. Part of the inner assembly folds up like a car bonnet and this requires the existing semi rigid version of this cable to be disconnected and a flexible cable put in its place.

It will be very frustrating if you don't have this cable when you need it most.

Also, be very careful when soldering near some of the resistors on the RF board. You will see some tiny brown axial resistors that are sometimes fitted up in the air in the RF amplifier section. These resistors are very fragile and will fall apart with even the tiniest amount of sideways pressure (with the iron) on the resistor leg. Best to consider them as being made of chocolate.

If your sig gen has been used a lot in a hot environment you may have to change a few dried out electrolytic caps on the RF sections if you want to get the lowest (close to carrier) spurious performance.

If it hasn't been used for a while it might be worth pressing [DELTA] [RF LEVEL] 10 [dB] and then use the UP/DOWN keys to exercise the RF attenuator across +10dBm down to -127dBm a few times. My 2019A never needed this but my 2019 does need this if it hasn't been used for many months (or longer).
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 4th Jul 2020, 6:23 pm   #108
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Replacement for HP8640B signal generator

Quote:
2017s seem to go for 8 times what I paid Bob for the 2019A!
They do seem to be rare beasts these days. In my first job there were several of them in the RF lab and nobody was allowed to try and pick one up on their own because they were so heavy. The close in phase noise was supposed to be slightly better than the HP8640B and these sig gens had a motorised VFO. This meant that if you had the sig gen tuned to about 300MHz and dialled in 500MHz on the keypad then the sig gen would turn its own VFO control for the internal oscillator. So it looked like it was being tuned by the invisible man!

We also had one on loan where I currently work and this was for a specific task that required very low close in phase noise across the HF bands. We ended up having it for many years as the customer didn't want it back. Sadly, it went for disposal in the end. Even I didn't want it because of the huge weight of the thing.

At work I do have a fairly comprehensive library of phase noise plots of most of our sig gens and I do remember measuring the 2017 on one of our E5052A signal source analysers. I usually do phase noise plots at 10, 14, 20, 28, 50, 70, 100, 144, 200, 300, 432, 512, 750, 1000MHz and upwards if the sig gen supports it.

Our 2017 was a bit tired and it had lots of close to carrier spurious terms on it. This may have been due to a few dried out caps. I doubt the sig gen would have been like that when it was new. It was slightly better than the 8640B for close in phase noise and I think Marconi designed it like this.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 4th Jul 2020, 7:00 pm   #109
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Replacement for HP8640B signal generator

All three of my 8640Bs are the LED dot matrix type of display. Yes, they make the 5V regulator board turn a toasty brown.

I can probably get the necessary SMA lead made by one of the eBay vendors. Or maybe it is time to buy myself the crimp tool and go DIY. I'll exercise the attenuator once I've typed this. After warming up both my TF2073 and the 2019A for two hours yesterday evening I confirmed both agreed within a gnats about a signal level of -10dBm. Not bad seeing as neither have been calibrated in aeons.

Even the 2019A made me grunt and my eyes bulge as I lifted it at almost arm's length onto its shelf in the corner of the bench. I can well believe a 2017 would have broken this camel's back. Now I have the convenience of the 2019A my current game plan is to keep one 8640B for low noise work, sell one, and keep the scrapper for spares. I've already sold my HP 11710B down converter.

I was musing on the rsgb-workshop group about how best to filter a 300MHz signal from mixing 280MHz and 24.1MHz (answer: a SAW filter for preference, or saw up some brass stock and make an interdigital filter). Inter alia a chap not ten miles from me said he has for sale an 8444A tracking generator with the OPT059 internal 500MHz option! So making a proper TG would be "easy" were it sadly not the case he wants too much money. I'll wait until my 8568B arrives from Cyprus (on a pallet!) and maybe see if his asking price has come down. If he sells it on eBay for loadsa money I'll probably be on here looking for sympathy in five years time.

73, Alan G3XAQ
Alan_G3XAQ is offline  
Old 4th Jul 2020, 8:10 pm   #110
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Replacement for HP8640B signal generator

Yes, the OBIC displays are used in the RA1772. I've fancied a 1772 in the collection and have been keeping an eye open for some time, so I have kept a bunch of OBICs for such a contingency.

They have latches, decoders and LED drivers in them.

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 4th Jul 2020, 10:26 pm   #111
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Replacement for HP8640B signal generator

Note that the service cable uses the smaller SMC connector series and not SMA. It looks like this cable is available on ebay for about a fiver with £6 postage from China. However, I think there will be cheaper and better quality solutions closer to home.

I warmed up the 2019A this evening and ran it with some calibration checking software I wrote a few years ago. This automatically sweeps the generator across its full range in small steps and measures the power level using a lab power meter. The power meter corrections are included in the software including interpolation between points and this gives very good results with some of my newer sig gens. Across 10-3000MHz the newer sig gens are usually flat within +/- 0.05dB and within +/- 0.15dB up to 4000MHz.

See below for the result with my Marconi 2019A. This generator hasn't been calibrated in over 25 years but the result is still quite good up to about 500MHz. Above 750MHz the ALC system is struggling a bit and so there is about +/- 0.35dB ripple. However, a quick glance inside at the basic technology used for the output amplifier and ALC detection will show that this amount of ripple is to be expected at higher frequencies.

The Marconi 2019 is best viewed as a 520MHz signal generator (the Marconi 2018) with the bonus of 512-1040MHz via a doubler but at reduced noise and levelling performance across the doubled range.
Attached Thumbnails
Click image for larger version

Name:	Mi2019A_level.jpg
Views:	59
Size:	58.9 KB
ID:	210336  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 4th Jul 2020, 10:57 pm   #112
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Replacement for HP8640B signal generator

I have several modern spectrum analysers here with a digital IF and these offer very good log accuracy so I could have a look at how accurate the 1dB steps are across most of the attenuation range of the 2019A. I won't be able to do this until tomorrow though. I haven't tested my 2019(A) generators like this for many years so maybe it is worth doing it to see if they are both working OK.

There will be some uncertainty introduced by the spectrum analyser but the best one has typical level accuracy and log scale fidelity as per the images below. I'm not sure I believe the analyser really is this good but Agilent seemed to think so. Obviously, the first plot will be taken with a signal source with very good VSWR to minimise mismatch uncertainty. In the real world, this level of performance is a bit artificial because mismatch uncertainty will degrade the flatness.

It was supposed to be the best spectrum analyser on the planet 18 years ago and the performance charts below are worlds apart from anything that came before it. It's probably best to do this with an automated routine using GPIB and I'll try and lash something together tomorrow. It should work on my other sig gens too so will be a neat way to check them over each year.

I'm expecting the 2019s to perform very well in this test when used up to 520MHz.
Attached Thumbnails
Click image for larger version

Name:	PSA_response_band0.jpg
Views:	59
Size:	54.0 KB
ID:	210342   Click image for larger version

Name:	PSA_scaleF.jpg
Views:	63
Size:	49.6 KB
ID:	210343  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 5th Jul 2020, 6:16 pm   #113
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Replacement for HP8640B signal generator

I ended up doing this manually as it requires quite a lot of user input. I tested the 2019A at 28MHz from 0dBm down to -127dBm.

I normalised out the slight error at 0dBm on the spectrum analyser and then stepped down in 1dB steps to -127dBm. The results were very good. From 0dBm down to about -90dBm each step size was typically within 0.05dB. By -90dBm I saw the odd 0.1dB error but this could be due to noise.

By -127dBm I had to use the internal preamp and renormalize at -80dBm and then step down again in 1dB steps. The biggest error was about 0.5dB by -127dBm but I think some of this was noise lifting the level slightly on the analyser.

Back in the early 1990s at work we used to measure the noise figure of our receiver and downconverter designs using a 2019 and the HP8568B using the gain method. This has higher uncertainty compared to the Y factor method but if you have a good (accurate) sig gen with good source VSWR and a good coax cable and a decent spectrum analyser it is possible to get reasonably reliable results using this method.

The HP8568B has a decent noise marker function and it was our gold standard for measuring the spectral density of noise back in those days. We also had a very good high level noise source and some BPFs that were calibrated for noise bandwidth at ( I think) the NPL. These filters were looked after very carefully to preserve the noise bandwidth figure. This allowed us to use a HP431C + HP478A thermistor power meter to cross check results and prove that the HP8568B noise marker function was reliable enough for the stuff we were doing back in those days. We were typically measuring noise figures in the 6dB to 15dB range and I think the overall uncertainty wasn't much more than +/-1dB. I doubt it was as bad as +/- 2dB.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 5th Jul 2020, 8:50 pm   #114
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Replacement for HP8640B signal generator

These numbers suggest the 2019A is fancier than I will ever make use of. At what frequency were those NF measurements?

Alan
Alan_G3XAQ is offline  
Old 5th Jul 2020, 8:58 pm   #115
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Replacement for HP8640B signal generator

It was a long time ago but most of our noise figure measurements in the early days were at 70MHz or 140MHz.

I did also do some noise spectral density comparisons between the HP8568B and two modern analysers this evening. I set a sig gen to generate a 2MHz wide noise pedestal at 70MHz and measured the centre of the noise with all three analysers to get a reading in dBm/Hz. The two modern analysers both agreed within 0.1dB after both were allowed to run an internal calibration just before the measurement. The modern analysers really do a great job of measuring noise.

With the HP8568B it isn't quite so fast and easy but after running its internal calibration routines it was only in error by about 0.2dB when measuring the noise using the noise marker function. This isn't bad for an analogue analyser that is over 30 years old and it also hasn't been to HP/Agilent/Keysight for a formal calibration in about 15 years.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 5th Jul 2020, 9:42 pm   #116
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Replacement for HP8640B signal generator

I spent a bit more time playing with the HP8568B and I think it does need a bit of a service. The log accuracy was only just in spec with a 1dB error by -70dB and it should be much better than this. I think the fine RBW filters need adjusting slightly too. It is probably just in spec in these areas but there should be a greater margin I think. I have only used this analyser a handful of times in recent years and it would definitely benefit from a good service.

One thing to note with the 2019A is that the internal modulation oscillator isn't referenced to the main 10MHz OCXO. The frequency accuracy of the internal modulation oscillator isn't that great. This means you have to take into account any slight error in mod frequency if you wanted to check the deviation using Bessel nulls for example. Otherwise it should be fine.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 5th Jul 2020, 11:04 pm   #117
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Replacement for HP8640B signal generator

I can see I'll be bugging you for recalibrating my 8568B when it lands at my house in a month!

In the meantime I'm still half-heartedly wondering about modifying a tracking generator. It has been suggested a note to Golledge might get them to supply me with a 300MHz SAW filter for experiments on creating the LO3+21.4MHz signal. I wonder: how much does the LO3 frequency move around? That will determine which of their flters is best suited for the task.

73, Alan
Alan_G3XAQ is offline  
Old 5th Jul 2020, 11:22 pm   #118
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Replacement for HP8640B signal generator

I really am a bit rusty on this analyser these days. I don't think LO3 is swept at all. It is a bit like a better behaved LO2 in that it sits inside the whole PLL loop and any errors at LO3 get cancelled out by the PLL at LO1. I'll have a better look tomorrow.

What will move will be the 'sweep' oscillator at the end of the LO chain on narrow spans of <=1MHz. This oscillator deliberately puts a swept error into the system that gets corrected out by sweeping LO1 to counter it.

My advice would be to either not bother with a tracking generator or to use a basic 8444A system and live with the drift on narrow spans. I'm not sure about taking out various LO signals from within the HP8568B. You do increase the risk of producing higher internal birdies within the analyser if the LO signals are extracted out of the analyser.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 6th Jul 2020, 12:27 am   #119
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Replacement for HP8640B signal generator

Cables are two way things. People only tend to think of the direction they want signals to run, but sproggies generated in the tracker can go backwards into the analyser and corrupt its LO signals, giving the analyser sproggies.

Soma analysers have LO outputs specifically for external trackers, others have LO outputs for external mixers. They are often provided with a little terminating resistor load in the appropriate connector, kept nearby with a little chain.

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 6th Jul 2020, 11:03 am   #120
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Replacement for HP8640B signal generator

The guy on the HP forum who has successfully modified an 8568B+8444A to use all 3 LOs and keep the TG transparent down to the SA noise floor has gone to some lengths to keep the reverse isolation high. He has a chain of isolators on the LO1 and LO2 signal feeds and a lot of bandpass filtering on LO3.

I was being tempted by all this because I need to replace the broken LSI output stage on the 8884A that is coming as part of the Cyprus package. One route might have been to buy a second TG just for its output stage and then I'd have most of that one left doing nothing. In the event the local chap with a TG for sale wants more than I can thoil so I'm going cold on the idea anyway. You two telling me it it's a bad idea is another nail in the coffin.

It's a shame that I'll be ending up with a system that is in some ways less functional than my old TF2370. I don't have space to keep both. Maybe I need to become more used to using my NanoVNA and the PC screen grabber although cluttering up the bench with a laptop whenever I want to use it feels clunky to me. Of course, if David has one of those HP VNAs he mentioned for sale for a lot less than the prices on eBay...

73, Alan G3XAQ
Alan_G3XAQ is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 5:35 pm.


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.