IM-16-12676R2 1 Abstract— AC-DC measurements have been performed on a commercial thermal transfer standard (TTS) using a pulse- driven AC Josephson voltage standard with a small cryostat enabling shorter voltage leads. A 20 mV RMS output voltage was chosen, close to the maximum value that can be obtained without low-frequency compensation (at a sigma-delta code amplitude of about 10 %) and close to the upper limit of the TTS in its 22 mV range. The measured deviation at higher frequencies, which is due to the length of the voltage leads, was an order of magnitude smaller than was observed before with a normal cryostat. The observed square dependence on the signal frequency and cable length follows exactly the expected behavior. All results agree well with the calibrations performed using a micropot reference system. Furthermore, the influence of other cables and equipment was investigated. The measurement results can be quantitatively understood both in terms of a lumped circuit description and in terms of reflected waves. Index Terms— AC-DC difference, Josephson voltage standard, measurement standards, measurement techniques, voltage measurement I. INTRODUCTION HE development of a pulse-driven AC Josephson voltage standard (ACJVS) is already going on for 20 years [1]. The focus of the developments has been on increasing the output voltage to practical levels without compromising too much on sufficiently large operating margins, i.e., the ranges over which the different bias parameters can be varied without changing the spectral purity of the output signal [2]. The output voltage can be increased to values on the order of 1 V by increasing the number of junctions in a series array and connecting a number of such arrays, each with its own bias electronics, in series [3,4]. Apart from reaching practical voltage levels, the problem of Manuscript received July 15, 2016. This work is carried out with funding by the European Union within the EMRP JRP SIB59 Q-WAVE. The EMRP is jointly funded by the EMRP participating countries within EURAMET and the European Union. H.E. van den Brom and E. Houtzager are with VSL, Dutch Metrology Institute, P.O. Box 654, 2600AR Delft, The Netherlands (email: hvdbrom@vsl.nl). O.F.O. Kieler and S. Bauer are with Physikalisch-Technische Bundesanstalt (PTB), Bundesallee 100, D-38116 Braunschweig, Germany (email: okieler@ptb.nl). reaching low uncertainties has been addressed as well. At frequencies up to 1 kHz, impressive uncertainties have been obtained already by comparisons between the pulse-driven ACJVS and a programmable Josephson voltage standard (PJVS) [5,6] and between two ACJVS systems [7,8]. At higher frequencies it is still very difficult to perform precision measurements with uncertainties lower than obtained when using conventional AC-DC thermal converters [7,9-11]. The major problem is that for high-accuracy AC-DC calibrations the device under test (DUT) and the reference converter are usually connected as close as possible to each other, such that the reference plane of the measurement is close to the input connector of both DUT and reference. In the case of a pulse- driven ACJVS, the calculable AC voltage signals are generated at low temperatures (in our case 4.2 K), whereas they are measured with equipment under test at room temperature. The voltage leads and connection cables cause the output voltage to show deviations that scale with frequency squared; deviations of almost 1 % at 1 MHz have been reported for a cryostat with voltage leads of about 1.5 m [11]. For practical applications, this limits the ACJVS output frequency to approximately 100 kHz [7,10,11]. Recent investigations revealed that the frequency dependence can be explained by reflection of waves at the high impedance of the load [12]. Calculations showed that at first order approximation the voltage at the input of the device under test is increased by a factor that scales with the cable length squared, the signal frequency squared, and the inverse propagation speed of electromagnetic waves through the cable squared, independent of other parameters. This calculation is based on the assumption that the input impedance of the DUT is high and the output impedance of the ACJVS is low compared to the characteristic impedance of the cable. The only fitting parameters are an offset length representing the apparent signal path in the DUT and the propagation speed of the waves in the cable. It was suggested that using a small cryostat would result in much smaller deviations than observed in other work [10,11]. In Ref. [13] we presented measurements obtained using a He 4 cryostat at PTB with a short cryoprobe, which enables a distance of only 70 cm between the Josephson array and the output connector at room temperature. We connected a commercial Fluke 792A AC-DC thermal transfer standard AC-DC Calibrations With a Pulse-driven AC Josephson Voltage Standard Operated in a Small Cryostat Helko E. van den Brom, Senior Member, IEEE, Oliver F.O. Kieler, Stephan Bauer, and Ernest Houtzager, Senior Member, IEEE T