HP8405A Vector Voltmeter 1000MHz

[G0HZU_JMR]

I have a slightly unplanned Christmas project (as of today)

I bought an old 1GHz HP8405A Vector Voltmeter to use as test gear for a research project and it arrived today. But sadly, it doesn't appear to be working. It was advertised as fully working but it turns out the last time it was used was 4 years ago and it was working fine then.

So I'm hoping it will be a simple repair, eg a loose connection or a tired cap or two somewhere.

I've started this thread in the hope there is a common failure mode for this old meter that I can look for. I really hadn't planned on having to repair it as I was hoping to use it this weekend. So I'm looking for a super quick fix if possible.

However, I did have a nice surprise when I lifted the top cover. It is spotlessly clean inside and looks almost new despite it being about 45years old. Lots of clean shiny metal and clean PCBs and not a speck of dust or dirt anywhere. Amazing. I'm going to go over all the caps with an ESR meter and then test the PSU section first.

At the moment the fault is that the voltmeter for both A and B positions fly over to a wobbly FSD on all ranges with or without any input and the unlocked neon is lit even if I try input frequencies on all frequency ranges on the selector dial. I've tried wiggling the switches lots of times to clean them but this hasn't helped.

So has anyone been here before with one of these and know a few common issues? I need to get it working as per its spec as I want to measure phase and amplitude very accurately up at VHF with light loading of the circuit under test.

Thanks in advance for any info about these old meters...

[GMB]

What state are the probes in? And do you have the x10 dividers on them?

They are critical to its operation and rather easily blown up, especially if used without the dividers.

If it doesn't get a lock and if both inputs are not present then it can give rather random-looking readings.

[G0HZU_JMR]

I haven't managed to check the probes yet. Sadly there are no x10 tips but I was planning to make a crude set of these for basic measurements at HF/VHF. I've only had the meter delivered this evening so I've not done anything other than lift the top cover.

I was kind of hoping the probes would be OK because the seller seemed convinced the meter was working the last time it was used. I tested it using a Marconi 2019 set to 0dBm at 2MHz and fed this to the probes via an RF splitter. But no joy.

I have taken out each PCB in turn and it looks to be completely unmolested inside with no evidence of solder rework on any boards. I guess I'll have to look at the probes next as they are easy to damage.

Have you any experience of repairing the probes? They look fairly basic with a sampling (quad diode) mixer and a FET inside and hopefully they will come apart easily.

It really does look to be in lovely condition inside and definitely worth repairing.

[GMB]

Well start with checking all the power supplies.

I have had very little trouble with mine, but I did once think it was broken only to realise that the random noise readings were down to operator error!

[G0HZU_JMR]

I've just checked all the large caps with my lashup ESR/impedance meter and found two 40uF caps in the PSU that are extremely high Z at 100kHz. The impedance is off the scale so they must both be dead.

All the others looks to give typical readings for their cap value.

[G0HZU_JMR]

I did a bit of surfing and found this useful probe repair guide By Richard Carey W6RLC

http://www.peakbagging.com/Electroni...A%20repair.pdf

Looking at the schematic for the PSU I doubt the failed caps will cause my symptoms so I will try and take the A channel probe apart tomorrow.

It looks like I have the later 'C' probe type (same as Richard) and this needs an 0.05" allen key to help take it apart.

[Radio Wrangler]

Probe diodes were said to be marked HP1517, which is an abbreviation that expands to 1901-0517, I expect, and that's the in-house stock number. Sold outside, they became 5082-5414.

They will have been put in as matched quads, matched for threshold voltage and capacitance.

You can get a reasonable match by selecting parts for forward voltage using the diode test function on a DVM if you have a bundle of schottky diodes to go at.

Power supplies first
Then check the VCO runs on all ranges
Then check probe A
Then move a sig gen around and try to get a beat on probe A output against the internal VCO, Find out why it doesn't capture and lock

I fixed a vector voltmeter last week. THe boss had bought it from the grand sell-off at the Agilent plant (Now bulldozed and scheduled for housing) a couple of yars ago. All that was wrong was a loose meter glass trapping a pointer.

David

[G0HZU_JMR]

That's lots of useful info David, thanks

I'll nip into town tomorrow to see if I can get the right Allen key for the probes. However, as both channels are behaving the same I doubt the probes are faulty. I suppose it is still possible the owner blew up both probes (4 years ago) and forgot why he stopped using it

Normally, I would enjoy fixing something like this but I need to use it asap for a research task. I have a VNA here that could probably be used instead but I wanted the versatility of the VVM especially with the 20kHz IF and the recorder outputs at the back as these would have been ideal for the task.

[G0HZU_JMR]

I put the PSU on the extender card and had a prod around and soon spotted that the -20V rail was stuck at -26V.

I quickly traced the fault to Q4 having gone very leaky across the CE junction. Replaced it with a 2N3053 for now and the -20V rail is -20.0V with no need to adjust it.
The VVM has now stopped going wobbly FSD on channel A and B but now appears to be a bit deaf on both channels and it is still unlocked. But at least it appears to measure 'something' on the voltmeter on both A and B. With a bit of luck I'll make some more progress tomorrow

[Radio Wrangler]

The thread title's suffered a typo, I think. The vector voltmeters were 8405A (moving pointer meters) and the later 8508A (LED displays).

I thought I'd posted something last night, but I must have forgotten to click the post button when I packed in.

The VCO drives a pulse generator which drives a step recovery diode to create sampling impulses a few hundred picoseconds long, with much faster rise and fall times. Finding a scope fast enough to see these is a problem, so looking at the drive pulses to theSRD and then looking for an IF passing through 20kHz as you sweep an input signal to a probe past any harmonic of the vco frequency is about the best you can do with the usual sort of test gear. Note that sampling efficiency falls as the sampling pulse width becomes significant compared to the period of the input and there is a scheme for compensation.

David

[G0HZU_JMR]

Yes, sorry about the typo in the title. I'm so used to HP gear that starts with HP85xx. I got the number correct in the first post but not the title.

I couldn't find anywhere with the correct Allen key so I filed a 1.5mm key down and this enabled me to take the probes apart. Both probes look original and unmolested inside and the diodes test out OK on a meter.

But I tried swapping the guts across from A to B and looked at the VTO frequency. In the original state the VTO could lock down to 1mV at 2MHz input. But if I swap the tiny PCBs inside the probes then the VTO can't lock even with a big signal at A.

So it looks like probe B could be damaged in some way. This assumes that it is OK to swap the probe guts from A to B?

I'm not too happy with this... The seller said the unit was fine but I'm finding more and more problems with it. How can anyone sell something like this and claim it is working if they haven't used it in 4 years?

Can a forum mod correct the number in the title to HP8405A please?

[G0HZU_JMR]

I'll try and trace the signal path on channel A next as it must surely be a good sign that the VTO locks and tracks with input levels down to 1mV on Channel A. It also loses tracking if I deliberately mistune the preselector control away from the input frequency.

When fitted in the correct probe holder each FET on the little probe PCB seems to be biased up as per the manual and the diodes appear healthy so maybe you can't swap the tiny probe PCBs? I'm still clinging to the hope that the probes are OK.

[G0HZU_JMR]

Nearly there... (?)

I've replaced a couple more dead transistors in the sampler and channel B now seems to read correct on the voltmeter when fed from the external two way splitter. It agrees spot on with my Racal 9300 true rms meter at 2MHz.

But channel A is not yet correct.

What must have happened is that the two caps in the regulation loop in the PSU went high ESR and this must have upset the -20V PSU and caused it to fail in the high voltage state. (Q4 dead)

This appears to have set off a series of component failures in the meter.
I've had enough for tonight but hopefully I'll have it running OK this weekend

[Sean Williams]

Thread title changed as requested.

[G0HZU_JMR]

Thanks!

The good news is that I think I've just got it working The probes were OK after all. However, once I got channel A working the phase meter section wasn't working

This turned out to have several dead components.

What is really odd is that the areas where this VVM failed in the sampler and the phase meter section were the areas where it has been redesigned and upgraded wrt the original circuits in the manual. Mine has S/N starting with 742- so it has nearly all the changes in the appendix in the manual.

The strangest fault is to do with the upgraded 'voltage limit' adjuster section in the current source sections of the phase meter.

Basically, some fixed resistors are replaced with trimmers to allow fine setting of this limit. But on both limiters the (physically big) 1k ohm trimmer was totally open circuit across all three pins! Also a zener diode had failed. I've put it back to the original fixed resistor scheme for now and the current sources are working properly and phase meter now works.

That was hard work... I have a box of dead parts but luckily I've been able to fit reasonable equivalents.

At some point in the new year I'll see if it's possible to get original replacements because I'd like to keep it all as original as possible. But for now I think I have a fully working HP8405A VVM

[Radio Wrangler]

Glad you got there.

With a reasonable signal generator and a test fixture, The VVM is the absolute bee's wotsits for measuring quartz crystal parameters. Look up the Cathodeon 'Pi network' crystal fixture. They haven't been made for decades, but they haven't been bettered either. Any still left going are priceless, but you can build one. All it is is a ninimum loss pad matching 50 Ohms down to about 12 ohms and a switched 30pf capacitor.

David

[Herald1360]

Congratulations! I haven't used one for 30 odd years, but I remember it as seriously useful for cold measurements on ATUs and the like.

[G0HZU_JMR]

Yes, it should prove to be a very versatile tool

So far I've checked a 10.7MHz LPF with it for phase response against a decent VNA and it agreed within about 1 degree at 2MHz, 3MHz, 5MHz and 10MHz so I think it is now basically working OK.

I'm now where I was hoping to be when I unboxed it. I planned to spend a few hours evaluating it and possibly calibrating it for my particular needs. I'm hoping to use it to look at negative impedances. eg to evaluate a negative capacitor circuit and also plain old negative resistance circuits.

I need to poke around various parts of the circuit and the VVM should be ideal for this with a Hi Z x10 tip. I plan to (crudely) make these tips myself and I'm hoping they will perform OK for my needs.

The original owner forgot to include the mains lead and manual and is sending these in the post so I'm hoping to have a decent manual as it's the original HP manual that came with the meter when new. At the moment I'm using a mains lead from a HP431C power meter as it's the same type.

[G0HZU_JMR]

Here's a link to the HP Journal from May 1966 and it shows a few of the things you can use this meter for.

http://www.hpl.hp.com/hpjournal/pdfs...Fs/1966-05.pdf

The precision frequency comparator application on page 7 looks interesting as it can be used to measure the error between two frequency references.
I can't quite read the fuzzy text but is seems to say the precision will be 1x10^-13 over a few minutes.

I'd be tempted to datalog the (rear) recorder output of phase over the test time to see if it can produce good results here

[Radio Wrangler]

Now if you look at that HPJ article, you'll find the project team group photo. Dead centre in the shot is Seigfried Linkwitz who is a very interesting character.

Look him up and you'll find the author of a number of loudspeaker articles in Wireless World, and he's still develping speakers for fun. Along the way he did something a bit bigger. He's the originator of a certain cinema surround system that competes with Dolby.

His speakers are very open dipoles and are very suited to DIY, but need large rooms for space around them.

David