Early Digital Volt Meters in the Physics Department: ---------------------------------------------------- Digital Voltmeters are now very common and inexpensive. Before their introduction there was no easy way to measure voltages to better than 5% (or perhaps 2% with a fancy expensive analog meter). Before digital meters the Physics Dept used a number of potentiometers and bridges for precision measurements. The best potentiometers in the Dept (one from Cambridge Instruments and a Leeds and Northrup K6) had 6 digit resolution but were slow and complicted to use at full resolution. The early digital voltmeters that I know of in the Dept are: - Data Technology Corp model DT-320 The manual for the Data Technology Corp DT-320 is from August 1967. I think the design was done in 1966. This was a Blatt-Schroeder-Forstat Rm 8 meter. This meter is 10,000 counts full-scale with 20% over-range that meets specifications and maybe 70% total over-range available. The clock for the counter is 100 kHz. The DC input attenuator is Hi Z straight through or 10 Meg Ohm divide by 100. The amplifier has gains of 1x, 10x, and 100x. The 10V full scale range is straight through Hi Z attenuator and amplifier gain of 1x. This DVM is so early that it does not have any TTL in it. It has discrete transistors and either DTL or RTL stuff in its decade counters. Its analog section uses op-amps made with discrete transistors and some ua709 ICs. The meter uses pulse transformers to communicate between the analog and digital sections. I think that the department had a couple of these meters. A couple of times I was given one of these meters to repair when I first stated working in the department in 1973. I know that one time I could not fix the meter. These meters had a digital output - obviously not IEEE-488. Data Technology Corp was in Mountain View, CA. The Data Technology Corp DT-320 meter(s) were never interfaced to the PDP-8. In the same era of mid to late 1960s, Data Technology Corp also made an early 8 bit computer called DT-1600. It had 4k to 16k bytes of memory. I do not know what happened to Data Technology Corp. - HP 3490A The date on the HP 3490A manual is May 1972. This is a 100,000 counts full-scale meter with a 20% over-range. The counter clock is 2 MHz. The input DC attenuator is straight through Hi Z or 10 Meg Ohm divide by 100. The amplifier gain is: 1x, 10x, or 100x. The 10 Volt full scale range uses Hi Z straight through input and a gain of 1x amplifier. It has a digital output and a digital control input - neither are IEEE-488. As seems common for HP they designed an ASIC for the counter/latch part of the digital section of this meter. It's hard to imagine that an ASIC was cost effective but they did get practice designing what should have been a straight forward part even in 1972. As far as I know this meter was never interfaced to a computer in the department. This was a Blatt-Schroeder Rm 23 meter. - Leeds and Northrup Precision Digital VoltMeter model 2760. The Physics Dept obtained this meter in June 1975. I believe that L&N introduced this meter in 1973. I believe that L&N intended this meter as a "standards" type of instrument vs and engineering or research type of meter. I don't know if L&N had any other digital meters. I assume that by this time their potentiometer / bridge business was failing and this may have been an attempt to save their standard business by moving to a digital meter. This is a strange complicated meter. It is 100,000 counts full scale with a 60% over-range in which it maintained full accuracy, i.e. it had full accuracy up to 159,999 counts. The high order part of this meter used a limited resolution but very high precision DCA based on fancy resistors and read relays to buck out the majority of the unknown input signal. That is, they used a classic potentiometer technique that they were well familiar with to implement this meter. For example on the 10 Volt scale this DCA bucked out all but 100 mV of the input signal. On the 10 Volt scale the output of the DCA was between 0 and 15.9 Volts and it has 159 steps of 0.1 Volt. But the precision of the DCA on this scale was better than 100 uVolt, i.e. the precision is 1000 times better than the value of its Least Significant Bit. For other ranges the output of this DCA was attenuated by 10, 100, or 1000. The actual ADC in this meter was only 2000 counts full scale. It is a dual-slope that integrated the unknown for 1000 counts and then counted down with the reference for up to 2000 counts. A full measurement took two steps: the unknown (through an attenuator or gain on ranges where this is necessary) was measured with the ADC to determine the value that is loaded into the DCA; then with the DCA bucking out the bulk of the unknown and with 100x the gain at the input to the ADC, the ADC was used to measured the difference between the unknown and the precision DCA output. If the unknown did not change significantly then subsequent measurements used only the 2nd step in this process. The Dept may have obtained this meter because it has a 10 mV Full Scale range which is a factor of 10 better than most other digital meters of the time. The 10 mV Full Scale range has a 100 nV least significant digit and would have been useful with magnet current shunts and low temperature transport samples. The analog section used a some integrated circuit op-amps, e.g. ua725, HA2-2535-5, LM307, LM201 (often fronted with dual discrete FETs) and modular op-amps, e.g. 4017 and 4087. The input attenuator section used only straight through Hi Z or divide by 1000 at 10 Meg Ohm. The 2000 count ADC itself is a very interesting minimalist design. Coupling between the analog and digital sections used 8 optical couplers and a one pulse transformer for the ADC clock counts. The analog section was on about 10 cards counting the: current, AC, Ohm, and 10 mV full scale options. The 10 mV full scale option used a separate low level op-amp of the carrier type with a discrete cmos and transformer full wave input chopper. The digital section is a complicated mix of DTL and early TTL on about 7 cards. This meter must have been expensive to manufacture. It has 20 circuit boards (including all options and 2 mother cards), uses 42 read relays, and weighs 42 lb. It must have been complicated to match up the first step of the measurement with the 2nd step as the load on the unknown would have changed. The 1000 count ramp up time of the ADC would not have provided any normal mode rejection on the DC ranges so a number of other RC filters were required. As far as I know this meter ran well and never needed any repair. This was a Blatt-Schroeder Rm 23 meter. This meter was definitely interfaced to the PDP-8. - DANA model 5900 Dana Laboratories, Inc. Irvine, CA. The date on the DANA 5900 manual is February 1976 but earlier versions of the manual go back to at least June 1975. This meter is 100,000 counts full-scale and could handle 60% over-range, i.e. it can count to 160,000, i.e. 5 1/2 digits. It has a digital output and a digital control input - neither are IEEE-488. This DVM was definitely interfaced to the PDP-8. The ramp up integrate was for 100k counts at 6 MHz so the ramp up time was 16.666 msec. The ranges were the standard 0.1V, 1V, 10V, 100V, 1000V. The input attenuator was Hi Z straight through or divide by 10 or divide by 100. The input amplifier gain was 1x, 10x, 100x. 10V full scale is straight through attenuator and gain 1x. This was a Blatt-Schroeder meter. Note that DANA, Dana Laboratories, at some point became Racal-Dana. At some point during this series of meters the 5900 had a little brother, the 5000, and a big brother, the 6900. They all looked the same. The 6900 was 6 1/2 digits vs the 5 1/2 digit 5900. - There were various setups to interface DVMs to the PDP-8 computer for experiments and data logging. I'm certain that 2 of the departments good DVMs were interfaced to the PDP-8. The final version of this interface was quite rational and was built on "Vector cards" that plugged into a small homemade card file and serviced multiple different DVMs. The engineering of this final setup was good. This was all before the days of common buses back when interfacing things to a computers was its own field of activity. - At some point all of these meters faded away because: - For high precision stuff we needed to build a Current-Comparator in a bridge setup with SQUID null-detector. This was built starting in 1978. - For general work IEEE-488 meters were purchased. This started with quite a few Fluke model 45 meters which worked well and are still used in some setups. I currently have the manuals for the: Data Technology Corp model DT-320, HP 3490A, DANA model 5900, and the Leeds and Northrup model 2760. As far as I know only the DANA still exists and it was pulled out of a recycle pile. - At some point the HEP group got a Fluke 8505A digital meter. It's nice meter to have around because it has 4 wire Ohms. I think that in its averanging mode it is a 6 1/2 digit meter. At some point it became a Fluke-Philips item but I think that it started out as a 100% Fluke designed meter. I think that it started in 1983 and the one that HEP has is from sometime after 1986. It is a "modern" meter with IEEE-488, a cpu, and software calibration. It has Fluke's "recirculating remainder" aka R2 ADC. This ADC generates binary weighted serial output bits. The input attenuator is either Hi Z straight through or divide by 64. The basic full scale input range of the ADC itself is 20 Volts so it really is a full 6 1/2 digit meter. I think that it uses the cpu to average readings to simulate this high resolution mode. - Physics Lecture-Demo had a Systron Donner model 7004 DMM which has a kind of simple design that looks like they were working hard to control cost. It is 10,000 count full scale meter with 30% over range and has built in: current, AC, and Ohm functions all on 2 relatively small simple circuit boards. The counter clock frequency is 120 kHz. The ramp up integrate is for 10,000 counts of 120 kHz or for 1/12th of a second or for 5 cycles of 60 Hz line frequency. The input attenuator is Hi Z straight through, divide by 10, and divide by 100. The input amplifier has gains of: 1x, 10x, and 100x. This meter has nixi tube readout and LM301, LM311, ua723 age analog parts with the usual TTL digital suspects. It is from 1972. I don't think that it ever worked very well. I think there were lots of problems with its push-button function / range switches not making good low resistance contact (and its analog design depended on low resistance contact to achieve a stable readout value). - Sometime before any of the digital meters listed above some one in the Dept must have had an HP / Dymec voltage to frequency type "digital" meter. I do not know who had this. Parts of such a meter were found in the hallway in about 2001 when we moved out of the old PA building. This was *probably* a Dymec model 2211 voltage to frequency converter or else something close to that. What remains are just the George Philbrick USA-3 op-amp and some parts from the precision pulse generators, e.g. the saturating blocking oscilator transformers HP part no 5080-1414. I assume that this instrument was from the early to mid 1960s. -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- HP 3440A Digital Volt Meter ---------------------------- This meter came from the Lansing radio show in July 2012. It has a model 3443A plug-in: auto ranging high gain module. The 3440A's Serial Number 637-07838 The 3443A's Serial Number ??? As far as I know, the Physics Department never had a HP 3440A meter. Designed of the HP 3440A started about 1961 and HP sold it until about 197?. HP built and shipped more than 10,000 HP 3440A digital voltmeters in only five years. The electronics of the 3440A are designed by Chuck Near and Dave Cochran. Chuck Near was fresh from MSU Engineering school when he worked on the 3440A design. The HP 3440A Digital Volt Meter has a full scale of 10,000 counts with a 5% over-range. It has a digital output and a digital control input - neither are IEEE-488. The HP 3440A was HP's 2nd DVM. The first HP DVM (outside of Dymec) was the model 405 which was all tube. Inside the HP 3440A: Decade counter and display: 5212L-4A Series 401 part number stamp Rev. A pcb 26-May-1967 date stamp there are 4 of these modules 8x 1850-0062 transistors 27th week of 1966 date code A 1850-0062 transistor is a tested 2N404A CR1:8 diodes are 1901-0025 which is a 1N645 Transfer Hold Brn Blk Red Grn CR9:13 diodes are 1910-0016 which is a ? Germanium stear Blk Blk Brn Blu Both types of diodes have a "T" symbol on them The nixi tube is a 1970-0009 National Electronics Inc. date code 6719 400 kHz Gated Synchronous Oscillator: This is part of the power supply module A9 Q4 is the Oscillator Gate transistor 1850-0062 tested 2N404A RCA CR15 is one direction of the L circuit Germanium Diode 1910-0016 Q5 is the oscillator 2N1304 TI Q6 is the emitter follower 2N3638 Fairchild Q7 is 1st part of the Schmitt Trigger 2N3645 ?? Mot Q8 is 2nd part of the Schmitt Trigger 2N2190 TI Ramp (very interesting and hard to understand at first): Ramp Gate Q11 1850-0062 tested 2N404A RCA Ramp Gate Diode CR1 1901-0025 junction diode 1N645 Q15 1853-0001 PNP Si Fairchild Q16 1854-0003 NPN Si Mot Q17 1854-0022 NPN Si Fairchild Ramp Reference QCR1 1820-0001 transistor&diode Mot The Ramp RC is and there is Volts Plus-Minus Sign Display module and the Count Gate and Overflow FF Module: A4 Module: 03440-60401 Rev. C Display tube is 1970-0012 National Electronics Inc. 2x 2N404A TI 3x 1850-0062 TI a tested 2N404A 2x 2N3638 Fairchild 1x 2N1304 TI Diodes: 5x Brn, Blk, Red, Grn ?? 1x Blk, Blk, Brn, Grn ?? on a white body Power Transformer: With 123 VAC RMS as read on the Fluke meter I see the following open circuit secondary voltages: Red to Orange 291 VAC 145.5 VAC either to CT Green to Yellow 89 VAC 44.5 VAC either to CT