Introduction and unboxing of the Agilent 34461A

The Agilent 34461A is a 6.5 digit multimeter designed as a replacement for the 34401A. It improves on the specifications of the 34401A and, most noticeably, includes a 4.3 inch display instead of the 34401A’s vacuum fluorescent display (VFD).

Specifications

Rather than copying the datasheet (which can be found at http://cp.literature.agilent.com/litweb/pdf/5991-1983EN.pdf), I will provide a quick summary:

• DC/AC voltage - 100 mV to 1000 V
• DC/AC current - 10 uA to 10 A
• Resistance - 100 Ω to 100 MΩ
• Frequency - 3 Hz to 300 kHz
• Temperature - 5 kΩ thermistor or PT100 RTD
• 5 V diode test

The 34461A uses a graphical 4.3 inch display rather than the more widely used character or dot matrix vacuum fluorescent. And it is this display that makes the 34461A unique among 6.5 digit multimeters. The graphical display has the advantage that it can display the readings as a bar meter, a histogram, or as a trend plot, as well as still being able to display them numerically. Whilst other multimeters have previously had histograms and trend plots, they are not as easy to read and use on a VFD, with the 34461A’s main advantage being this extra clarity.

Agilent additionally makes a 'budget' version of the 34461A in the form of the 34460A. This is roughly £100 cheaper (currently at £608, with the 34461A at £705) and to match this price cut some features are removed. These include the trend chart display, a 10x decrease in the readings memory and the LAN connectivity (which is an optional extra on the 34460A). More importantly than this, the accuracy of the 34460A is less than the 34461A. With the following table showing the 1 year DC voltage accuracy for the 34460A and the 34461A.

From a personal perspective, the differences between the 34460A and the 34461A do not justify the difference in price, especially considering the removal of the trend chart and the LAN interface. Whilst some applications might not need the extra accuracy or features, I would think that these are very limited, with the majority of people opting for the 34461A.

Additionally, both the 34460A and the 34461A do not come with GPIB as standard, with this being a user-installed optional extra. Both models also have a NISPOM sanitisation option, designed to securely clear the multimeter’s 80 MB internal storage.

Who needs a 6.5 digit multimeter?

With a resolution of 0.1 uV (on the 100 mV range) and 0.1 nA (on the 100 uA range), it is quite reasonable to ask who would need a 6.5 digit multimeter over a typical high-end 4.5 digit handheld multimeter. The first answer is, of course, those people who need to measure small values, namely scientists and those in some areas of R&D. On a more general scale, with the availability of very low powered microcontrollers, amongst other devices, it is quite easy to see the usefulness of measuring current on the order of nano amps. However, both of these are very limited applications that could use a lower resolution multimeter and a high quality amplifier.

The real advantage of a multimeter with such a high resolution is its dynamic range - its ability to measure large and small quantities on the same range setting. Whilst multimeters have a variety of ranges, it is sometimes advantageous to not change the measurement range. This could be because you want the error in the measurements to be the same or because the auto-ranging might interfere with the measurements - one example of this is when measuring current.

Although it is possible to measure DC current directly, through a current (or ampere) balance, it is not practical for a multimeter to do this. So most multimeters measure current by measuring the voltage dropped across a resistor. However, this voltage drop, referred to as the burden voltage, is passed onto the device that is being measured. As different ranges on a multimeter use different resistors, this burden voltage changes during auto-ranging. With some multimeters, this change can be around a volt - for the 34461A changing from the 10 mA range to the 100 mA range could result in the burden voltage changing by up to 0.45 V. With any device this could change the behaviour of the system, with modern microcontrollers and FPGAs this change could result in the device resetting. Having a high dynamic range, as in the case of a 6.5 digit multimeter, allows you to fix the range at say 100 mA whilst still having a 0.1 uA resolution, without having any ill effects on the device you are measuring.

There are, of course, other reasons why you might want a 6.5 digit multimeter. One in particular is accuracy - typically the error in a reading removes one or two of the least significant digits, depending on the multimeter. This can reduce a 3.5 or 4.5 digit multimeter to producing very vague measurements, and hence a higher digit count is sometimes preferable or required.

Comparison with 34401A

Over the 34401A, the 34461A only offers a modest improvement in specifications, with the 34461A improving on the DC and AC current ranges, with it ranging from 100 uA to 10 A, compared to the 34401A’s 10 mA to 3A. Additionally, the 34461A also includes temperature measurements and increases the diode test voltage from 1 V to 5 V.

The main noticeable difference between the 34401A and the 34461A is the change from a VFD to a graphical display. The graphical display has the advantage that it can display the readings numerically, as a bar meter, a histogram, or as a trend plot. The 34461A also removes the serial interface, in favour of USB and ethernet. Although these are quite small and iterative, the ability to use USB to communicate with the device and the ability to transfer readings to a flash disk, are very useful and reflect the change in times, with their inclusion essentially being a modern-day necessity.

Accuracy wise, the 34461A is a very close match to the 34401A, with it being an exact match, based on the 1 year figures, for resistance, continuity and DC voltage. The DC current is identical for all ranges except 3 A, where the 34461A has a lower accuracy specification than the 34401A. In this range, the 34461A’s accuracy is 0.200 + 0.020, whereas the 34401A’s is 0.120 + 0.020. However, the 10 A range of the 34461A has a better accuracy of 0.120 + 0.010, which is probably the reason why Agilent recommend using the 10 A terminal over the 3 A, when measuring currents greater than 1 A, however, by doing this you do lose precision.

The AC voltage specifications are almost identical, with the 34461A increasing the accuracy of the ADC for voltages below 100 mV and between 3 Hz and 20 kHz, from 0.04% of the range to 0.03%. Additionally, for 1 to 750 V and between 20 and 50 kHz, the ADC’s accuracy is worse, at 0.05% of the range compared to the 34401A’s 0.04% of the range. However, these are very small changes and are unlikely to affect the majority of users. The largest change in accuracy is for AC current measurements,

with the 1 A range between 3 and 5 Hz offering an order of magnitude increase in accuracy for the percentage of reading.

Overall, I find it quite surprising that the general accuracy of the 34461A is not better than the 34401A. Although the 34461A is designed as a drop-in replacement for the 34401A, an increase in the accuracy specifications would have been beneficial to all. The accuracy is around the average for 6.5 digit multimeters, looking at the 1 year DC voltage figures (in particular, the 10 V range as it is typically the most accurate) the average is 0.0035 + 0.0005, which is the 34461A’s specification. However, the Keithley 2000 (approximately £20 cheaper than the 34461A) has an accuracy of 0.0030 + 0.0005, with the best accuracy being that of the Fluke 8846A at 0.0024 + 0.0005. Although the Fluke is considerably more expensive (around £300 more), it shows that better accuracies can be achieved. Whether it was too expensive for Agilent to increase the accuracy of the 34461A or whether they were only aiming for the 34401A’s specification is something that we can only speculate on, however, it is a little disappointing that they have not improved significantly upon the 34401A’s accuracy specifications.

Comparison with other multimeters

There are quite a few different 6.5 digit multimeters that are currently available, which I will now provide an overview of. A spreadsheet containing all of the tables is available at https://docs.google.com/spreadsheet/ccc?key=0Akm-le8ckBVPdFVrWnBxS3o3enhvUnMtX19mRWR5T0E.

The table below summarises the multimeters that will be used as a comparison for the 34461A, along with how much they cost (based on prices from Farnell, which are correct as of 16/10/2013).

There is a large range in the costs of each of the multimeters, which, to a large extent, reflect their functionality.

The first comparison that will be made is that of the AC and DC voltage and current ranges for the multimeters.

From the above table, there is very little variance between the voltage specifications, with most of the multimeters ranging from 100 mV to 1000 V DC and 100 mV to 750 V AC. The exceptions to this are the Hameg, on both AC and DC, and the Flukes and Tektronixs on AC, where they have a higher top range. The current specifications have more variations, however, with the exception of the Keithleys, the DC and AC current specifications for the multimeters are the same. The variation is between both the low and high ends of the ranges, with most multimeters having a low end of 100 uA and an almost even split between the high end of 3 and 10 A - it is worth highlighting that the Hameg is an exception to this with its range topping out at 1 A. The 34461A, the Flukes and the Tektronixs all share the largest ranges for voltage and current, with the 34461A having the lowest price out of these 5.

Comparing the accuracies of multimeters is quite difficult for the AC ranges, as each manufacturer uses different frequency ranges to specify the accuracy. However, the DC ranges can be compared easily. The table below does this using the 1 year accuracy figures for the 10 V and 100 mA ranges - both of these typically have the highest accuracy compared with the other ranges.

From this, there is very little variation in the current measurement accuracy, with the exception of the Hameg (although, comparatively the Hameg’s accuracy is so much greater that it seems likely that this figure is erroneous). There is a wider range of accuracies for voltage, with the 34461A being around the average, however, there are higher priced multimeters that have a much better accuracy (Tektronix’s DM4050 and Fluke’s 8846A). More importantly, the Keithley 2000/E achieves a better DC voltage accuracy whilst costing less. Overall, the DC voltage and current accuracies place the 34461A around the middle of the pack, reflecting its price point.

The next important comparison is the resistance and continuity, as these are likely to be used less often than the voltage and current, but more often than the other measurement modes. The table below shows the resistance and continuity ranges as well as the frequency ranges.

The continuity threshold of 1000 Ohms is common amongst all of the multimeters, with the Tektronixs adding a 1 Ohm threshold - although I am not sure of the practical uses of such a low threshold are. The resistance range of 100 Ohms to 100 MOhms is pretty much the standard for the multimeters (with the exception of the Hameg), with some of the more expensive multimeters adding an order of magnitude of range on either end, or both, of the measurement range. As most resistance measurements will be within the 100 Ohms to 100 MOhms range, the 10 Ohms and 1 GOhms are unlikely to be used particularly often, unless for specialist applications. The measurement of frequency on a multimeter is not meant to rival that of a frequency counter, with it instead being meant to measure things like mains power. The 34461A is, again, around the middle of the pack, inline with its price point. The Keithley 2000/E has a slightly higher top range, at 500 kHz, but it is the Tektronixs and Fluke that have a markedly higher top range, at 1 MHz.

As the main three features of a multimeter are voltage, current and resistance, it becomes increasingly difficult to compare the specifications of the additional features. Of the multimeters in this comparison, these additional features are capacitance and temperature measurements and a diode test functionality, which are shown in the table below.

Only the Tektronix DM4050 and the Fluke 8846A include a capacitance measurement mode, which is a shame. Unfortunately, it is not always practical to include capacitance measurements as they require a frequency source combined with some specialist processing. This is largely why it is often left to dedicated LCR meters, with the accuracy of the Tektronix and Fluke peaking at 1% +/- 0.5%. Still it would have been nice for the 34461A to include this, but it would have probably been slightly out of place. The temperature measurements use additional sensors, in particular RTDs, thermocouples and thermistors. I am personally quite used to using thermocouples, due to their wide temperature range and relatively low price, so I was surprised to see that the 34461A is not capable of using them for measurements. However, as can be seen from the table, this is not something that makes the 34461A unique, with only the Keithley 2000/E and the Hameg having the capability of thermocouple measurements. All of the multimeters, with the exception of the Hameg, are capable of testing diodes, with a varying maximum voltage level. As most diodes have a voltage drop below 1 V, the usefulness of a higher range will be largely dependant on specific applications.

All of the multimeters have some level of internal memory and the ability to be connected with other measurement devices or with a computer (using GPIB, serial, USB etc.). Whilst the majority of them are not meant as data acquisition devices, it is often useful to perform measurements over a period of time. The table below shows the reading size of the internal memory and the 4.5 digit (5.5 digit for those with *) measurement speed for the multimeters.

The 34461A compares reasonably well, with quite a high reading memory size and a reasonable 4.5 digit speed. Once again the usefulness is largely dependant on the application, if used in a measurement rack a reading rate of 1000 to 2000 readings per second is likely to be the maximum required. For normal bench use, a 1000 readings per second is unlikely to be particularly useful, with the memory depth being more useful.

The following table compares the software features and the connectivity, with Y indicating that it is included, O for optional and N for not included.

All of the multimeters are able to perform some level of statistics on their readings, this typically is the minimum, maximum, average and standard deviation. Additionally, they all can apply limits to the readings, indicating an error to the user if these are exceed. I was surprised to see that Agilent did not include the ‘mX +b’ feature which can be particularly useful. This feature allows the readings to be linearly changed through the use of a multiplier (m) and an offset (b), which can be useful for converting a measurement to a different unit. With the displaying of charts, the 34461A and the 34460A are the only multimeters that have the ability to display the readings as a bar, in a faux-analogue fashion. The 34461A, Tektronixs and Flukes can all display the readings as histograms and trend charts, with the 34461A having the clear advantage that it uses a graphical display rather than a dot matrix VFD. I am not going to talk to much about the different connectivity types as this will depend on the application, with USB being very useful for a multimeter on a lab bench.

Overall, the 34461A is the lowest costing 6.5 digit multimeter with a 100 uA to 10 A current range, with the 34460A being the lowest priced with a 100 uA to 3 A current range. Additionally, as morgaine pointed out in the comments, the 34461A is the lowest priced multimeter that includes LAN as standard. The 34461A’s accuracy specifications are inline with its price, although its DC voltage specification could be improved upon. The use of a graphical display, rather than a VFD, makes the 34461A and 34460A unique amongst 6.5 digit multimeters. However, it would have been nice if the 34461A supported capacitance measurements and temperature measurements using a thermocouple, as well as including an ‘mX +b’ mode.

Unboxing

When I received the multimeter it was slightly daunting to see a bright orange sticker on the side saying 'Sell by September 1, 2013'. As with any high precision device, the calibration is key to its accuracy, so I was a little worried that Agilent might sell a device with only a few weeks of calibration left. Thankfully, this bright orange sticker refers to nearly 4 and a half months after calibration, so it is likely this sticker is to ensure the multimeter arrives with at least 6 months left of its 1 year calibration.

The multimeter was pretty well packaged and felt like it could withstand quite a few bumps during transport. The box had some very light damage, with the edges being slightly crushed, but the box feels like it could withstand a lot more.

When the box is opened, the first thing you see is the accessories box.

This contains the pack of probes, the power lead, the software discs and a USB device cable.

The box does not contain a printed manual or a getting started guide, with this available on the CDs and on the Agilent website. I have mixed feelings about the lack of printed material. On the one hand, large manuals, often in many different languages, are cumbersome and won't be used very often. But on the other hand, they can be very useful during the first weeks of ownership, providing a quick look-up reference. Considering most lab setups have computers in them, I can completely understand why Agilent has chosen to do this, but I still miss having something to idly flick through.

After removing the accessory box, you are left with the multimeter sat in the bottom of the box with foam around its front and back, with the calibration certificate, in a nice bright yellow envelope, sat on top.

Additionally, the box I received contained two additional power cables placed on either side of the multimeter. These weren't fastened in anyway and there was no place for them in the end pieces of form. It seems to me that it is highly unlikely that Agilent, or any other company, would do this, considering there was already a power cable in the accessory box. And so, the only conclusion I could come to is this was something that was done particularly for this unit. This multimeter was shipped to the UK from the US, with the power lead in the accessory box being American. The power leads placed next to the multimeter were for the UK and Europe. I therefore think it is likely that element14/Newark added these extra cables to the box prior to shipping. Supporting this, the box looked as if it had been opened and the re-taped prior to shipping. The following image shows all of the accessories included, as well as the 3 different power leads.

The following picture shows the multimeter with the end foam pieces still on.

These pieces of foam are nothing special, with them being the same that is normally used for general equipment packaging. They are quite thick and hold the multimeter really well - when the multimeter is in the bottom of the box, these pieces of foam hold it snugly, so that it doesn't move at all. So hopefully this should survive being shipped by even the roughest of carriers.

Removing the foam gives us the first proper look at the multimeter.

Initial impressions

Build

The first thing I noticed about this multimeter, after unboxing, was its weight. Although it is not a particularly heavy piece of test equipment, at 3.76 kg (8.3 lbs), its small size gives it the perception that it is a lot heavier. This isn't of any real concern as, being a bench multimeter, it is not designed to be portable, and whilst it is heavy, it is not so heavy to make moving it difficult. It is intriguing why it is so heavy, it is around 900 g heavier than a Keithley 2000, for example. The majority of the weight seems to be due to the transformer, which is positioned almost halfway back on the unit's left-hand side (looking from the front).

However, this is the only real large component (see the image above), which leaves the rest of the weight in either the PCB or in the structure, which is more likely.

The second thing that jumped out at me was the stickers on the top of the multimeter. I can understand, to an extent, putting a sticker on the top advertising the demo functionality, even though a message saying this pops up when the multimeter is switched on. But I can't understand the 'win a prize' sticker. Sure everyone likes to win stuff, but when your product is being bought largely by companies, and not individuals, you have to wonder how many people will be bothered to even read this, never mind go to the website, before they rip it off. To me, these stickers seemed more fitting to a toy than a piece of test equipment, and I would have preferred this information to have been included on something like an A5 piece of paper or card. However, this is really quite a minor point.

Overall, the 34461A has a very high build quality and feels very sturdy. Agilent, on their website, say that it can withstand being dropped off a bench, which I can believe, although I won't be testing this. The rubber edging on the front and back is quite hard (I thought that it would be a soft rubber to absorb impacts) and takes a bit of wrestling to remove - which in my book is a bonus, as you don't want these to easily slide off. The rubber is held in place by its shape, not by screws like some other test equipment. The tilting bail is quite sturdy and the one I received can be a little difficult to use - as it has to be pulled away from where it joins the instrument. I currently find that I can easily pull one of the sides out to rotate it, but pulling the second side is quite difficult. Hopefully this is something that will become easier with time, but even if it doesn't, it isn't a major concern as you are not likely to be changing the arm's position that often. The buttons on the front of the multimeter are of a good quality and feel like they will stand many years of use without any issue. Additionally, all the connectors, on the front and back, are of a good quality and feel quite solid - so I don't foresee any problems there.

Usage

After plugging in the multimeter, I was disappointed to see my suspicions were true - the 34461A has a soft-off power button rather than a hard-off. For certain test equipment it is  beneficial for them to always be partly powered, for example frequency counters, signal generators and spectrum analysers that ideally need to keep their crystal oscillators at a set temperature to guarantee a certain level of accuracy. However, a multimeter, much like an oscilloscope or power supply, does not have these requirements and so has no real reason to always be consuming power when the device is off. Most bench multimeters use a front power button that isolates the device from the mains, hence hard-off, and it is rather unfortunate that the 34461A does not do this. Agilent quote that the accuracy specifications are valid 30 minutes after device power on, if the use of a soft-off reduced this time, I could see its benefit, however this does not appear to be the case.

The multimeter takes approximately 30 seconds to turn on, which is the longest I have seen for a multimeter. Part of this time is used to perform self-tests, but the vast majority is spent booting up the operating system. The 34461A uses the Windows Embedded CE 6 operating system, which is Microsoft’s real-time operating system (RTOS). Windows CE is found in many embedded devices, including cash machines (ATMs), point of sale terminals (i.e. tills) and in industrial control. It is most often used in embedded applications where graphics have to be shown to an end user, such as animations and videos. Whilst the 34461A does have a graphical display, the use of Windows CE seems to be Agilent using a sledgehammer to crack a nut, with other RTOSs being available which have a smaller footprint (read faster boot times) and similar features. However, I would bet that Agilent went with Windows CE so that they could spend more time concentrating on the multimeter’s accuracy and features, rather than having to debug issues with the operating system. To be fair to Agilent, it is unnoticeable that the 34461A runs Windows CE, other than the relatively long boot time.

The 34461A’s connection layout is logical, with the probe connections following the de-facto standard (sense inputs to the left and current below) and the interface connections, on the back, leaving a reasonable amount of room between each other so inserting and removing cables isn’t difficult. The front panel buttons are generally well layed out, with the six soft buttons, below the screen, being just the right amount, considering the screen size. The mode buttons are logically grouped, but I personally would have prefered a slightly different grouping. The 34461A has two buttons for voltage and current, which are

• DC voltage / DC current
• AC voltage / AC current

as I rarely use AC, I would personally prefer

• DC voltage / AC voltage
• DC current / AC current

however, I completely understand that it is probably more logical to group them the way Agilent has done.

Overall, my first impression of the multimeter is that it is easy to use, with it being simple to change between both measurement modes and display modes. I particularly like that holding in the buttons brings up a little help description of what the button does, with snippets from the manual describing how connections should be made. Although this is only likely to be used infrequently, it provides a quick reference which essentially makes up for the lack of a printed manual. This first impression is based on the basic usage of the multimeter, with the review and later blog posts providing a more in-depth review of the multimeter’s usability.

Edits

1 (18/11/13) - added link to multimeter comparison spreadsheet and that 34461A is cheapest with LAN (as pointed out by morgaine).