10. Other Instruments
1. ELECTRONICS
I have been putting together my electronics lab for over 30 years and in that time it has collected a fair few instruments. More had to be purchased for the LHC project, both for optical measurements as well as very high speed signals and very high voltages. Due to its size, there is a separate section on [Oscilloscopes]. The instruments also needed probes, see [High Speed HV Probe], [High Speed Current probe], test jigs and specialist interfaces see [Detectors & Visualisers]. Professional probes are always best so where possible I bought what I could afford, and built what I could not.
It is normal to expect a low budget hobbyist approach to use very basic equipment but whilst this project is indeed low budget, there are no timescales to meet or customers to serve and the equipment was affordable as it was purchased mostly on eBay over many years, although for various reasons I ended up with far more than intended!
From the outset I wanted to run the project using the same quality of instrumentation I had been used to at work. I wanted the theme of exploration and discovery that runs through the project to apply to all sciences, not just optics, but such observations require accuracy of measurement.
As funds have always been tight, it was impractical to buy new test equipment except where it was affordable. Instead most instruments have been bought in used condition on eBay and the surplus market and whilst it is always best to get a fully working model, usually they are not, and something needs fixing, see [Repairs].
In 2022 I achieved one of my primary goals: the acquisition of a Time Electronics 9822 multi-calibrator that I can get professionally calibrated in order to calibrate everything else myself. The 9822 was also faulty, but I managed to fix it, see [Repairs].
PULSE GENERATORS
Under [Oscilloscopes] I explored Tek TDR based picosecond pulse generators but good as these are, they are only capable of very low output voltages and whilst excellent for reference, are of little use for high speed test circuitry.
In the case of the Pockels Q-switch driver, initially I created a fast pulse generator using a race hazard: a 74AC00 quad two I/P NAND gate with of its gates as an inverter on one input, the other input and the inverter input both driven by a rising edge, producing a short ~2ns glitch at the NAND output due to the propagation delay through the inverter. Different logic families 74xx Standard, LS, HC produce different pulse widths, but otherwise the amplitude and width are fixed. For the Pockels generator, width is less of an issue than rise and fall time, but I could not control these beyond using different logic families.
After I obtained the Tek TDR pulse generators I found this great website exploring instruments of this era and learned about the 250MHz HP 8082A, which is essentially a high speed version of the cheap Lyons Instruments PG73Ns I bought for laser timing, with the added advantage of control of rise and fall times, ideal to develop the pockels driver circuitry: http://dabbledoo.weebly.com/pulse.html
Eventually due to issues with the first, I ended up with two, each at a very reasonable £50.
GRIPES
I am particularly fond of historical Tektronix, but less so of HP whom I dislike for, in my opinion, their likeness to Apple in sticking to their own path at the expense of practicality. The HP8082A 250MHz pulse generator is a prime example. I've seen people praising its 'aircraft cockpit' layout with a huge front panel and long linear slider switches. Had they never operated one of these switches before? It is very difficult to slide across without inadvertently going too far.
Ten marks for a pretty layout but minus several hundred for impracticality. If they'd replaced those godawful sliders with rotary controls they could have cut the case width down by a third. Not only that, the inside of the beast is wildly unpopulated by vertical PCBs with huge empty spaces between them. Why! If they ruled out adding grounded metal shield plates between them on cost, they would have easily recouped that with the shorter case. Better still, replace the separate cards with one big horizontal pcb like most other manufacturers, making everything easier to debug, test and calibrate, particularly as the box top and bottom are easily removable. Just because everything in America is big does not mean instruments have to be as well, especially in a high frequency environment where you want shorter leads and instrument controls within arm's reach. A whacking big instrument in the middle of the lab bench limits what else you can stack close to the thing you're working on.
I cannot help thinking communication between departments was sadly lacking in HP at that time, see famous joke below:
Below left: the HP 8082A Pulse Generator: Below right: inside viewed from the rear:
At least when Tek made large instruments in that era, it was for a good reason - they were crammed full of electronics, although Tek also comes in for some criticism too: their TDS series of DSOs are infamous for failing electrolytics on the acquisition pcb; the U800 trigger chip on their 24xxB oscilloscopes is prone to frying (fingers crossed my 2455B won't suffer this fate), and their P6042 current probe is a pig to calibrate and stay in calibration once set, the PCB in mine having cut and strap mods on it giving me the impression the design was a bodge from inception, and don't mention their Type 109 Pulse Generator from yesteryear, see: [Projects: Tek 109 1-Shot & LF Mod] (although I don't regret having one).
https://w140.com/tekwiki/wiki/P6042
https://w140.com/tekwiki/wiki/109
Below: 109 Pulse Gen P6042 75MHz current probe
For a great front panel layout that uses the smallest possible space, look to the Lyons PG73N dual pulse generator, although sadly their own design falls short because 1/3 of the entire unit's power consumption is 8.5W to the POWER ON lamp leading to an unnecessarily hot transformer/box, and there's no fan. I may yet add one in the small compartment housing the transformer behind the IC regulators.
Then there's the HP 8175A Digital Signal Generator - it has 4 BNCs on the front panel but they're not connected in the purely digital version I have (I may yet see if I can hook them up internally to one of the pods). Instead you have to use its output pods sitting on the end of 5 permanently attached 4 foot cables. Mine didn't come with the pouch that normally holds them on the top but it's such a huge beast and my lab is so small, I have other equipment on top of it that the pouch would disallow. Not only do I have to find somewhere extra to put those huge cables (they look far less irritating in the datasheet photo below), but those FP BNCs would also have been very useful for quick debug. Instead HP chose to make life difficult for owners shelling out no less than $11,000.
Below: the HP 8175A Digital Signal Generator and those cables. At least mine works perfectly.
Another example of poor HP forthought is their 4952 protocol analyser, outwardly appearing as a neat design incorporating the interface pod into its case front. That's where the usefulness ends: it never occurred to me to buy one with the Option 002 extra memory board fitted, without which it is impossible to use the 4952A's built-in 3.5" disc drive to copy (i.e. back up) its master boot disc which of course is formatted in HP's proprietary format, totally incompatible with IBM PC floppy drives. I only discovered this after I got it and after copying didn't work, I read the manual. The additional memory is of course a non-trivial affair involving the addition of a new, unobtanium PCB, nothing as practical as filling some vacant memory chip sockets. What a good job it is nowadays pretty much all I need to perform a similar task is a freely available PC terminal emulator app, and Microchip's USB PICkit ICP (In Circuit Programmer) that also doubles up as either a mini Logic Analyser (LA) or Serial Analyser.
Below left: the HP 4952 Protocol Analyser Below right: its HP 18179 RS232-C /V24 I/F
There are some HP instruments I do like. For absolute reference, I rely on equipment I have had NIST calibrated in the past that I know is renowned for maintaining its accuracy over many years. One such example is my 5.5 digit HP3468A DMM (I also have an HP3478A with its universally hated, obscurely labelled front panel disable switch). 5.5 digits is a price/performance compromise, but fine for me.
ENVIRONMENTAL
TEMPERATURE
Magnavox MVX48 / IR-210S monochrome 48x48 thermal imager
(no temperature but great at finding overheating components)
3-5µm, -22°F to +1500: ±0.18°F / -30°C to +800°C: ±0.1°C
Chinese Cason CA380 IR thermometer (°C/°F), -32°C to +360°C
(an expensive Calex IR thermometer only lasted a couple
of years)
Cheap eBay China LCD thermometers
Pico TC-08 RS-232 8-channel thermocouple data logger
Tagem 871A dual 0.25% thermocouple meter
Lab-grade alcohol thermometer that is safer than mercury
but less accurate.
HUMIDITY
RS pocket thermo-hygrometer
2-98%RH, 0-50°C / 32-122°F
Precision Gold N21FR DMM with sound/light/temperature/humidity
35-100dB ±0.1dB, 0.1-20k Lux ±5%, 0-50°C ±3%, 25-95%RH ±6%
Cheap eBay China LCD thermo-hygrometers
SOUND
Tenma 720860 35dB-130dB sound meter
LIGHT
Lutron LX-101 0-50k Lux meter
MAGNETIC FIELD
Alpha Lab dc 0.1-20k Gaussmeter
ELECTROSTATIC FIELD
Alpha Lab 1" Surface 0-20kV dc Voltmeter
RADIATION
Ukraine Ecotest Terra-P beta gamma Radiation Detector
Alpha Lab 100XE RF/EMF Trifield meter:
0-100mGauss, 0-1kV/metre, 0-1mW RF power
PRECISION VOLTAGE & CURRENT SOURCES
Long ago when I worked for a large aerospace company that had taken the decision to dispose of a large amount of its ageing test equipment, I found myself trying to qualify an analogue multiplier design with what was left: fine adjustment pots on variable lab supplies, clearly not up to the job. I decided I would not let this happen to me in my own lab, and researched eBay for an affordable precision voltage source, eventually settling on General Resistance Dial A Source units as a cost/performance compromise: they weren't especially cheap but they were affordable and were specified for 5 bits of accuracy. It is one decision I now regret as they have been very problematic, see [Repairs: General Repairs].
To begin with they were accurate to 4 places and found a lot of use in my own designs. Unfortunately they only began to fail once I'd bought three of them, by which time I felt I had no choice but seek to repair them, which has proved very difficult with them still not fixed in 2022. This was also the initial inspiration behind my eventual acquisition of a Time Electronics 9822 multi-calibrator.
I bought two DAS-45s and a DAS-46:
PROGRAMMABLE POWER SUPPLY
I have an Amrel PPS-1202 dual channel 0-16V 0-4A programmable power supply with a resolution of 5mV.
When it arrived I found one channel stuck in constant current mode regardless of setting, but the eBay seller gave me a 75% refund so it only cost €50. See [General Repairs].
LCR testers
I have several of these and I rate the DER-5000 the best. The only downside is you need to run its (very simple) calibration sequence first to get the highest accuracy, but it takes a whole minute to complete:
https://www.deree.com.tw/de-5000-lcr-meter.html
A while back I bought a Hartley Measurements 'Model 350' on eBay UK out of curiosity as the only information I could find on the web was the following patent describing it as producing 'a 30 kV pulse with a 1μs rise time at a frequency between 1-100Hz': https://patents.google.com/patent/US5221561A/en I thought this might be useful for TEA laser experiments but sadly when I came to inspect it I found it did the 1-100Hz bit, but my Tek P5015A showed only a 750V 6ms pulse, of no use to me.
2. OPTICS
Unlike electronics instruments of the past, it is rare to find optics instruments of any era with free service manuals, let alone schematics (Oriel Instruments is one of the exceptions), so I chose these instruments even more carefully.
SPECTROMETERS
The first optics instrument I bought was a research grade PC parallel port driven Stellarnet EPP2000SRC 200-1080nm, 25µm slit (1.5nm resolution) 12-bit spectrometer. Like Tek instruments of yesteryear, its actual working range range is impressively wider at 190nm to 1085nm, spanning all project wavelengths.
When I got the LIBS prototype working I was able to use its optical trigger to capture 4ns wide single shot MK367 Q-switched pulses much to the surprise of Stellarnet. It occurred to me it would be a good idea to get its external electrical trigger option fitted before I built the LHC as I could then trigger it independent of the light pulse.
The trigger upgrade was only available with an expensive 16-bit upgrade which could only be done in the USA, but it also upgraded the parallel port to USB 2.0. After much deliberation I decided it was worth the equally expensive two-way insured shipping. It went away in a white box and came back in a black box, the closest version on their website (2017) being the Black Comet C-SR variant, so I'm guessing they upgraded it to one of those: http://www.stellarnet.us/spectrometers/black-comet-sr
I have been very impressed by this spectrometer and never regretted its expense. It has proven to be a reliable workhorse and like my Rigol DSO, the go to instrument for numerous applications.
The next optics purchase was of course the PCM401. The Spectrapro Sp-275 monochromator that came with it has a working range of 185nm to 1400nm [I16, page 5]; the PCM401 only uses 200nm - 1000nm.
Years later I finally acquired the much faster PCM403, sadly without the software installation disks I had seen offered with the original $7k one I could not afford, but a calamity in shipping from the Philippines destroyed its PC. The seller offered me a 50% refund without return (shipping was prohibitive) and I accepted this as the PCM401 came with prototype PCM403 software installed on its PC which thankfully worked with the new spectrometer and I was able to largely resurrect the system.
POWER & ENERGY METERS
Early on I realised I would need to be able to measure laser power and energy, especially to assist in the setup of NLOs, but also to be sure I could safely attenuate beam energy for observation on the cameras and the PCM401 spectrometer; also be good to know what I was hitting the target with.
After quite a bit of research I settled on Laser Precision's range of energy and power meters and sensors. My choice was a combination of an Industry leader with a proven track record, low price on eBay USA, and reliability and repeatability of measurement with minimal need for calibration due to their pyroelectric and thermopile probe technologies, which also offer a wide bandwidth from UV to FIR. I was also pleased to discover the company was approachable and emailed me all pdf service manuals.
What I don't like is they had organised their historical product range in the same manner as Tek by creating a huge range of models each doing a slightly different thing. In particular, the older models were either energy meters or power meters, and not every model supported every probe. Despite researching in quite some depth, lack of information in their scant spec sheets meant this only became apparent when they arrived. Consequently as well as an excess of Tek oscilloscopes, I also have an excess of optical meters! However as with the Tek, I budgeted well and paid little for some nice equipment, the only repairs being LCD electro-luminescent light replacements and a shorted tantalum capacitor in one probe, see [Repairs].
The instruments I bought are:
LP Rk-5710 Power Radiometer
LP RkP-575 pyroelectric shutter power probe 10W max
LP RkT-30 thermopile power probe 30W max
LP RJ-7620 Dual Energy Radiometer
LP RjP-735 1J pyroelectric probe x 2
Scientech combined 20W/20J meter
Scientech 380401 150J 90GW/cm² 100mm calorimeter (for the target pulse)
Here the Rk-5710 power radiometer is measuring a 300mW laser diode using the shuttered Rkp-575 probe.
The DMM confirms the LD supply voltage.
My 4'x 2' breadboard is custom built. Since I cannot regulate
lab temperature, if I have issues with this I shall replace it with one made from delrin (plastic).
2022 update: I finally found an affordable optical breadboard!
Sequoia-Turner spectrophotometer 340 in very nice condition including the full set of 4 filters covering the standard range from 300nm to 999nm. An optional DUV accessory extends it down to 198nm.
Unfortunately it's faulty and the US company I saw offering a service manual for it and that partly influenced my decision to buy it, revealed afterwards they'd lost it in a fire.
Fortunately it was only $60, albeit with $40 shipping to the UK.
How to make reflectance measurements:
https://www.youtube.com/watch?v=3YHYKFyPdvA