Solartron LM 1420 Digital Voltmeter (mid-sixties)

[Kat Manton]

Hi,

I just couldn't resist this, and that was without really knowing anything about it beyond it probably being fairly old as it uses nixie tubes. Interesting, definitely. It looks like I'll have to admit that I collect old T&M equipment at some point...

I found three Solartron price lists (1961, 1962 and 1965); it's in the 1965 price list but not the 1962 one. So that dates it to "older than I thought"; I was originally assuming early seventies. That just makes it more interesting. Mid-sixties? Digital? Advanced technology!

It's in a bench case and has an extra bit attached to the back identified as 'EX 1476'. The basic instrument, 'LM. 1420 DC Integrating Digital Voltmeter (0.05%)', was £375; with the bench case and EX. 1476 it was £425. Sticking that into an on-line inflation calculator, that's over £6,000 in today's money. I think that places it firmly in 'precision laboratory instrument' territory, not mere 'test gear'!

Google Books provided a bit more information:

Nuclear Engineering International: Vol. 9, 1964

"Twice the accuracy, and ten times the sensitivity, of any other instrument in its price range is claimed for the new Solartron LM 1420 digital voltmeter, which is also designed so that it may be used as a versatile counter. There is said to be an inherent rejection of supply frequencies or noise and common mode interference can be ignored. The instrument has a 4-digit display, employing cold cathode number tubes; with this exception, the circuitry is all solid-state. Six ranges give nominal full-scale readings from 20 mV to 1000 V. The normal sensitivity is 10 uV per digit; switching can change this to 5 or 2.5 uV per digit. By the use of an external resistor, currents as small as 1 nanoamp may be measured. For data logging, the displayed voltage is presented in BCD code. To realize its full potential, a completely new data logging system has been designed, based on this instrument. An additional sub-unit will soon be available to enable the LM 1420 to measure ac voltages."

Canadian electronics engineering: Volume 9 , 1965

"It's the new Solartron LM 1420. Frankly, it's impossible to find a unit with the features of this new integrating digital voltmeter. Unless you're willing to pay $1000 more. This is because the 1420 inherently rejects 60 cycle noise, and its method of operation reduces the effect of transient noise to insignificant terms."

Other results suggest that, for the time, it was quite an impressive, well-regarded and popular instrument and it undercut the competition by some margin. I wonder how many survive?

Searching Google Scholar for 'Solartron 1420', I get over 200 results in various papers. It looks like it was quite a common feature in research laboratories from its introduction right through the seventies.

All this makes me feel like it deserves a full restoration. It needs it; it looks like it's been neglected and stored somewhere damp.

I'll close this post with a few photographs, work is already under way so watch this space for the next instalment

Kat

[Brian R Pateman]

Oh, that takes me back!

We still had a few of these when I was at BAC in the early 1970s. Absolutely lovely instrument and head and shoulders above pretty much anything else around at the time.

These and their successor were the standard for use in the production test rigs I dealt with. They hardly ever required much attention when calibration was due.

Perhaps I should look out for one to keep my 7150 company.

[Kat Manton]

Thanks, Brian; you're confirming my suspicion that it's a good 'un. I already like it a lot and it still isn't quite working properly yet.

Originally Posted by Brian R Pateman

Perhaps I should look out for one to keep my 7150 company.

Funny you should mention that; I took a couple of photos of it with my 7150. The LM 1420 is deceptively large, around the size of a typical 'scope. I don't think I'll leave them together for long, the 7150 might feel small and insignificant and decide to sulk.

While I'm posting photos, here are a couple more of the rear panel; I can only guess the purpose of the 'EX 1476 Fan Out Unit' (and I'm guessing that it buffers the BCD outputs for connection to several somethings.) Attached to that is something else, I'm not sure what it is but it isn't wired up to the EX 1476.

Kat

[Brian R Pateman]

As far as I remember the EX 1476 Fan Out" was a precursor to GPIB found on the 7150 among others.

I have a very dim recollection that we used this interface to provide inputs to a rudimentary data logger to collect information when we were developing the test fixtures.

I think that the other unit might have been a sample / hold facility.

It was a very long time ago and I don't have any information on the interfaces now. It is possible that WME_bill might have something.

[John_BS]

Kat: Nice find: Hours of fun ahead!

And when it's working, you'll have a voltmeter with 5Gohm input impedance on the 2v range!

John

[AVO_VCM163]

The 4 page technical Sales Data Sheet is HERE (3.6MB pdf).

The fan-out buffers the high Z, unloadable sources for data recorders etc.
It can also be used as a counter/timer since it used a V to f conversion system.

Excellent reliable instruments as were most Solartron instruments at that time.
R

[gingpeakin]

Looks lovely - I have a crumbly Solartron 1010. It's been in a loft for 20 years at least. It's got one of those displays with the layers of perspex with a pattern of holes in the shape of the numbers - one per character with a bulb for each. Ican't remember what they were called. A pair of ME1400 electrometer pentodes give input resistance of >1000M ohms. The weston cell inside has given up the ghost and the whole thing is quiet corroded but it looks like it's all there. I'm saving this one for when I retire!

Ging

[SteveSmit]

It does make me smile to see mention of the Solartron 7150 here on a vintage electronics forum. Where I work we stilll have two of them kept in cal and used daily, one connected by IEEE on a computer controlled calibration system (which is equally as old) as well as as one of the older Solartron 7051s, with the LED display.

[Brian R Pateman]

Originally Posted by SteveSmit

It does make me smile to see mention of the Solartron 7150 here on a vintage electronics forum. Where I work we stilll have two of them kept in cal and used daily,

I'm glad to hear it! At least there are some places where proper instruments are in use.

[Kat Manton]

Hi again,

Okay, now to try bring this thread up to date with where I'd got to before I started the thread, so I can stop writing about it and get back to fiddling with it...

Armed with little information and a vague guess as to how it worked, I removed it from the bench case, looked at it, wondered what I'd let myself in for this time, decided I needed a diagram and put it back together. At least it seemed to be 'all there' with no obvious damage, missing bits, signs of fire etc.

Half an hour later I gave in to temptation. As I knew it'd already been powered before I acquired it, I reasoned that it'd be unlikely to do any more harm to apply mains and see what happened.

This is what happened:



That was accompanied by a high-pitched whining noise and the sound of a relay chattering frantically. I fiddled around with the controls; the first digit remained resolutely stuck on both '1' and '2' and the second on '8'. The last two digits did, at least, change. No smoke or unpleasant smells, either. So there was some life in there!

The sensible thing to do would've been to leave it at that until a manual had been sourced. But where's the fun in being sensible? All I needed to do is get some idea of what was in there; after all, any complicated thing is just a lot of simple things connected together...

(Bear in mind this all took place before I started the thread. I really should write these things up as I go along; I'd have had more information to work with if I'd done that.)

Back inside, it was fairly easy to divide it into functional blocks. I was already armed with the word "integrating" from the price list so I suspected that it comprised an Integrating ADC.

Power supplies are pretty easy to identify. The boards with the nixie tubes attached looked like they contained BCD-to-decimal converters and drivers. The boards without nixie tubes attached looked like they made up various bits of counter, so the screened box on the RHS would be providing... something to count..?



There aren't many connections to the screened box bit. A couple of screened wires run from the PSU bits on the LHS to the rear of the box and underneath there are three feed-through insulated pins with four wires connected (red, white and two black.) The wires from the input socket disappear through a grommet and there's a section of the range switch outside the box which is wired to the decimal point neons.

So I stood it on one end, turned it on and looked at the signals emerging from those three pins with a 'scope. I found what looked like clock and gate pulses and the gate pulse period changed when I fiddled with the range switch and the 'volts zero' preset.

Great, it looked like some of it was working; I surmised that the counter bit was being provided with something to count but wasn't counting it.

I figured I'd first look for obvious faults on the boards comprising the counter. Everything looked pretty much okay, there was some evidence of previous repairs but nothing stood out. However, I noticed a few of the diodes (Hughes HG1005) had been replaced (with something other than Hughes HG1005) on several boards... Hmm... maybe they're a bit unreliable..?

I systematically worked through all the boards, checking each diode with the 'diode check' range on my trusty old Beckman hand-held DMM. Three were open-circuit. I replaced them with NOS CV7130 (OA91 in uniform) after determining they were a close-enough match.

With that done, I applied power again. Now all the display digits responded to twiddling the 'set/zero' presets, I could get the display to read zero and the relay wasn't chattering; it looked like it might actually be working to some extent!

Okay, that's a long enough post; I hope I'm not boring anyone with these lengthy accounts. There's more to come, stay tuned for the next thrilling instalment!

Kat

[mhennessy]

I hauled something like this out of a skip some 20 years back - it was very similar, but in a 4U rack case. The internal construction was much the same, though the layout a little different because of the increased width. Reading through this thread is making me regret not keeping it. Indeed, even the nixie tubes would have been worth saving. I think it's my Mum's fault; she rather inconsiderately moved, which meant a quick and ruthless sort-out in her attic was required

Looking forward to more updates...

[Kat Manton]

Hi again,

I found that I could now zero the display by alternately switching between 'volts zero' and 'current zero' and adjusting the associated preset. I'm not sure what the correct procedure should be, but these two adjustments seemed to be interdependant; this method seemed to work.

Switching to 'set 1.019' produced a low reading, below 100mV, which could be wound up and down with the associated preset but wouldn't go anywhere near 1.019. I recognised '1.019' when I first examined the instrument as the voltage of a Weston cell, so that meant there was one inside and either it'd died or I still had faults to find.

Lacking an input connector (and not knowing what it was at the time) I stood it on one side, unsoldered the wires to the socket and lashed up a bench PSU with croc-clip leads to the orange and red wires (having determined they were the inputs and the remaining black wire connected to the screened box via a resistor.) I also connected my 7150 in parallel.

It worked! On the 2V range, I could set the PSU to somewhere around 1V, then adjust the 'set 1.019' preset until the 1420 and 7150 agreed, then wind the PSU up and down. The readings broadly matched; not too bad considering I'd only roughly adjusted the volts/current zero presets, didn't know how I was supposed to adjust them and it seemed to be drifting a bit.

Now if I'm remembering correctly, this was with negative readings. When I swapped the polarity round it all went a bit odd. Switching through the ranges with the same applied voltage, I got dramatically different readings depending on the selected range. 10.00V on the 20V range would become 0035V on the 1000V range. I swapped the connections back and it behaved again. I tried connecting the black wire (I assumed, correctly, that this was the 'guard' connection) to one or other input; it made some difference but readings still varied wildly, but only for one polarity of input. Very odd.

I'd noticed that when swapping input polarity, I heard the (previously chattering) relay operate. The relay position changed depending on input polarity. Without knowing what the relay was doing, it seemed like as good a guess as any to suspect that the relay's contacts might need cleaning.

I just needed to get at it; I could see it (or some of it) attached to the screened box. After disconnecting the wires to the relay coil (whereupon one of the tags broke off) and the wires for the D.P. neons, removing knobs, nuts and screws I found I could slide the sub-assembly to the rear, swing the rear end out sideways (the PSU and output wires are long enough) and finally withdraw it.

The relay is a custom-made item using bits of standard relay; the coil is outside the screened box and the contacts are inside.

Interesting; they're taking both screening and isolation very seriously! (I'll throw in a couple of photos of this sub-assembly with the covers removed and the boards hinged out, too.)



The relay contact cover is glued on and I didn't want to disturb it unless I had to. So I noted down the connections and disconnected it. My Anatek 'Blue' ESR meter doubles as a handy digital low-ohms meter; checking the contacts with this (using a cocktail stick to wedge it in the energised position) I found it was fine. Oh well...

Incidentally, it's a DPCO relay wired to swap two signals when energised, reversing polarity at some point in the circuit. On one of the counter boards, there's a transistor marked 'relay drive'; it appears that if counter reads below zero, the relay is energised if it wasn't or de-energised if it was.

Ah well, back to the drawing board. Just to eliminate the relay from suspicion, I connected the two pairs of wires first one way, then the other and found I still had the same odd behaviour; working fairly well with one input polarity but wildly inaccurate with the other.

Then I had a thought. You remember that Sun SS1000 computer with the bizarre PSU fault which turned out to be caused by electrical leakage across a PCB? Leakage... this instrument has a high input resistance and the input circuitry is supposedly well isolated. What if it wasn't? I was using a mains-powered bench PSU as a voltage source; while the outputs are supposedly isolated from mains earth, I bet they're not that well isolated.

I lashed up another arrangement of croc-clip leads with one output of the bench PSU connected to the meter input and the other output connected to 'jack-up' (or down) the first output relative to mains earth.

Lo and behold, I could now obtain near-correct readings with either polarity, or wildly inaccurate readings with either polarity, depending on both the polarity and voltage of the offset relative to mains earth of the input voltage. How inaccurate? How about '0238' displayed on the 1000V range for 10V input! Interesting and definitely a clue...

Now... I'll leave you in suspense to await the next instalment (which I'll try to post later today.)

Kat

[Top Cap]

Hi Kat, Just seen this thread and yes! Those diodes were a little bit dicey. I repaired quite a few bits of Solartron equipment, without a service manual, by just checking the diodes first! The Orange coloured ones, if you come across any, were even more prone to failure. When I worked at Whitehall communications centre (hope I am allowed to say that, don't want MI5 knocking on the door) they had a Solartron multimeter that everyone over a considerable time had tried to get going. First thing I did was to check the Orange diodes, replaced four of them and the multimeter worked! The guys in the centre could not believe it, just 15 minutes to fix, lol your post has brought that happy memory back. Good luck with the restoration. Les

[dragonser]

Hi,
when I was studying at Chelsea College back in the mid 1980's they were selling off some of the equipment and I bought a Lm1420.
I seem to remember that if you put it on one of the high impedance voltage ranges without it being connected to anything, the relay chatters and the display get lots of different readings.
I "think" I got a copy of the service info when I bought the Lmn1420, but finding it may not be straightforward !
useful to know about the diodes being unreliable !
regards Peter
P.S. there is something nice about Nixie tubes and the other display using perspex slides with numbers etched onto them. But I can't remember what they are called !

[Kat Manton]

Originally Posted by Kat Manton

(which I'll try to post later today.)

(Which I'll fail completely to do; sorry, I haven't been firing on all cylinders for the last few days.)

Right, where were we? Ah yes. Leakage.

I found electrical leakage where I wouldn't have expected any. I wasn't entirely sure what to expect as I wasn't armed with the information I now have, but, for a start, somewhere around 4 megohm between the chassis of the screened box and both inputs joined together didn't seem right. The chassis isn't used as a common or 0V connection. There didn't seem to be anything connected to the chassis except...

One 4700 pf disc ceramic capacitor. Not knowing what voltage rating it was supposed to be I replaced it with a 3 kV one; I think that'll suffice.

The original one is visible just above the centre in this photo:



While I was in there, I noticed a group of five ropey-looking WIMA capacitors, lifted one end of one of them to measure it and found it was leaky. Tracing the wiring, one is used for input filtering with the filter switched off and all five with it on. I didn't think that'd do much for the input resistance, so removed the lot pending replacement.

The outer insulation looked crazed, probably allowing moisture to get in.



Still, after removing these, I was getting readings (taken with Megger BM8/2) of tens of megohm where I was expecting much higher.

It also looked like something has been spilled inside this bit. Hopefully, something has been spilled inside this bit; if not, the Weston cell has leaked. Something has contaminated various areas of the central board carrying the precision resistors, Weston cell and the previously-mentioned filter capacitors. Whatever the 'something' is, it seemed to be conductive. Wiping areas of the board around the edges with a damp cotton bud then drying the area with the other end improved the leakage readings a little.

Great. Just what I need, a high-resistance conductive coating all over some rather sensitive bits of circuitry...

As I'm well aware of what's inside a Weston cell (and know more than a little about heavy-metal poisoning) I carefully removed it and sealed it in a resealable bag for the time being, pending further investigation. Then I cleaned as much of the board as was now accessible, bringing about further improvements.

While I was doing that, wires were dropping off 'Board 12' far too easily. The exposed conductors had corroded. I could've just stripped the ends and re-soldered them but, assuming liquid contamination, it would've seeped inside the insulation. There was only one thing to do... replace all the corroded wiring. That involved lots of notes, sketches, coffee, photographs, patience and coffee...

As I ended up with 'Board 12' completely disconnected and noticed it too was contaminated, I washed it (after removing the few components which weren't sealed.)

I'm not sure how well it shows up after scaling down, but it may be possible to make out traces of contamination on the PCB and some rather grotty-looking wiring. It was worse in real-life; the camera, it lies.



To my amazement, after everything had plenty of time to dry out and I'd wired it all back up again, it still worked!

Not only that, but measurements of either polarity on any range were now a lot closer to reality!

That brings this thread up to date, it's still in bits but it's now behaving more like a volt-meter, less like a random-number generator

I'm still not sure I've fully dealt with the leakage issues; I'll have another look at the data sheet and see if actual leakage measurements resemble the specification in any way. I have a feeling they don't and I may have to dismantle things a lot further and clean it all more thoroughly.

But... I think it's worth the effort. Now I've seen the original specification I seem to have a potentially very good and useful instrument, if I can get it working to that original specification...

Kat

[Kat Manton]

Hi again,

I hope nobody's getting bored of this, it's turning into quite a saga!

Anyway, from the data sheet linked to in post #6:

• Low input to chassis: > 10⁵ MΩ
• Screening box to chassis: > 10⁴ MΩ

Using my Megger BM8/2 on the 500V range, I measured:

• 'Red' input to chassis: 1500 MΩ
• 'Orange' input to chassis: 1400 MΩ
• 'Guard' input (connected to screening box via a low-value resistor) to chassis: 1000 MΩ

That's definitely not as good as it should be.

As I previously mentioned, I'm not entirely sure how to calibrate it, but by twiddling the 'volts zero' and 'current zero' adjustments with the range switch set appropriately, I can (eventually) get it to read zero.

But, if I apply a voltage to the input, I get a reading in the 'volts zero' and 'current zero' positions. I'd have expected the input to be disconnected in these positions. Leakage across range-switch wafers..?

(With 10V applied, the reading is '+0008' in the 'volts zero' position and '+0038' in the 'current zero' position.)

The Weston cell is knackered, after leaving it the right way up for a week it measures 0.068V on my 7150. It doesn't appear to be used as a reference for measurement, just as a calibration voltage source, so I haven't bothered to refit it. (I think I'll devise something using a buried-zener precision voltage reference IC to replace it.)

I've attempted to calibrate it against my 7150 though. Applying around 1V on the 2V range, I adjusted the 'set 1.019' preset to obtain the same reading as the 7150.

If I then switch to the 20V range, it's now off by a bit (and not exactly linear either):

• 10.00V input, reading 10.03V
• 20.00V input, reading 20.10V

(It's similarly 'out' on the 200V and 1000V ranges, using the Megger BM8/2 as a voltage source.)

It's also drifting a lot even though it's been left on for over an hour.

So there's plenty more work to do. I'm think it's contaminated with something, causing electrical leakage all over the place. These photos show the more obvious signs:



I may end up stripping-down and rebuilding the contents of the screened box and may have to replace the range switch. That'll be... er... fun...

I'm going to set it to one side for now, pending arrival of the service manual. Before I do anything else I need to study the diagrams and it'd be nice to know how to calibrate it!

Still, I've made some progress. This is all I've had to replace so far:



And it is now measuring voltage, just not as accurately as it should:



Not bad considering its age and the state it was in. I'm determined to restore it to full working order though, I don't give up easily

Kat

[Alistair D]

"So there's plenty more work to do. I'm think it's contaminated with something, causing electrical leakage all over the place. These photos show the more obvious signs"

If you suspect contamination obtain some pure IPA (not the beer) and do some gentle cleaning. Last week I recovered a mobile phone that went to the toilet accidently. So far so good.

Al

[Kat Manton]

Hi Al,

I have some IPA but it's a 1-litre tin, my aerosol of the stuff ran out a while ago and I'd forgotten how useful it was for this sort of thing. I've already thought about getting another one so I can use it to give the range switch and a few other bits a wash. Unless I do a lot of dismantling it'd be easier to spray the stuff on until it's dripping out and maybe rinse out some of the contamination.

That's if it's soluble in IPA; I could be dealing with the results of storage somewhere damp, something spilled inside it and leakage from the Weston cell. Your guess is as good as mine as to what would dissolve it!

I'm awaiting the arrival of some Russian military surplus capacitors to replace those WIMA ones; I'll try a bit more cleaning when those turn up and I've got it in bits again.

Kat

[dave cox]

Kat,
I had quite some fun and games fixing my 7065 here which had some leaky/noisy FETs and general 'leakage' problems due to accumulated detritus (and the PSU decide to go AWOL once I had screwed it all back together). I found it impossible to meaningfully measure anything in the high impedance circuits and had to proceed by trial and error when cleaning. It was worth persevering though, as eventually I estimated input impedance at 50G ohm on 10V range.

What sort of reference does this meter use ? BTW, The 7065 uses an ovenised zener.