RoadTest: Agilent 34461A Digital Multimeter
Evaluation Type: Independent Products
Did you receive all parts the manufacturer stated would be included in the package?: True
What other parts do you consider comparable to this product?: Keithley 2100, 2000/E Agilent 34401A Fluke 8845A Tektronix DM4040
What were the biggest problems encountered?: None really, but the firmware could be improved.
This review also has some companion blog posts that are available at:
Overview / introduction
The 34461A is designed as a drop-in replacement for the 34401A, whilst providing additional features to modernise the multimeter. The new features are a 4.3 inch display, USB (device and host) and LAN connectivity, and the ability to measure temperature, using either a thermistor or an RTD. Additionally, some features have been improved upon, most noticeably the current measurement range has been extended to 100 uA to 10 A (from 10 mA to 3A), and the volatile readings memory has been extended from 512 to 10,000 readings.
As I have never used the 34401A, this review will look at the 34461A being a 6.5 digit multimeter and not as a pure replacement for an older model.
Review of probes
Before properly diving into the review, it is worth pointing out the probes that come with this multimeter. Agilent describes the 34461A as coming with a test lead set, which is essentially two probes with some very nice accessories. The probes are fairly standard, with nice sharp tips and silicone cables, but have a nice high quality feel to them. However, it is the accessories that really stand out for me. Included are two fine tip probes, two SMT grabbers and a single mini-grabber, with all of these fitting over the end of the main probes. The fine tip probes
are primarily meant to be used for use on PCBs, due to their length and small contact size, with them being ideal for SMDs (such as measuring the value of a surface mount resistor). Their length make them a little tricky to use at first, but you get used to them quite quickly. The SMT grabbers are what I find myself using the most.
Despite their name, they are ideal for clipping onto the ends of wires and component legs. The mini-grabber,
only comes in black and so I believe it is meant to hook onto the ground of a system. Unlike the SMT grabbers, it has a right angled contact, which doesn't always clip securely onto things, such as cases and shields.
Features and price
The 34461A has all the feature you would expect to find on a multimeter - voltage, current, resistance, continuity, frequency and diode tester - none of which can be faulted. It also includes temperature measurements using RTDs and thermistors, however, these work optimally with only two specific types (one RTD and one thermistor), both of which can be, comparatively, quite costly to find. It would have been useful if more thermistor resistances could be used and not just 5 kOhm, and it would have been a great addition if thermocouple support was included. The only feature that is missing from the 34461A, that might be expected, is capacitance measurement - this would have been a nice addition and would round off the multimeter quite nicely, however, its accuracy would be significantly worse than the other measurement modes.
Overall, the 34461A’s specification matches its price, with it being the lowest costing 6.5 digit multimeter to have a 100 uA to 10 A range and to come with LAN as standard. However, its accuracy specification has progressed very little from the 34401A. Whilst it closely matches multimeters of a similar price, it would have been the cherry on top of the cake if these had been improved on.
The multimeter has a very high build quality and feels like it can take quite a high level of abuse. The buttons on the front panel are made from rubber and have a nice feel to them. Additionally, the connectors, on both the front and back, feel like they will last a long time with frequent plug insertion and removal. Overall, I cannot fault the build quality of the 34461A.
Layout and interface
The layout of the multimeter is logical with the probe connections being in the typical orientation format (sense inputs on the right of the main inputs). The mode buttons are logically grouped, although I would have prefered a separate voltage and current button, rather than an AC and DC button.
I thought it was strange that there is a frequency button, but changing to period is done through the menu buttons underneath the screen, which seems a little inconsistent. However, as frequency is included in the accuracy specifications and period is not, I can only assume that the period mode was added quite late to the firmware as a useful addition.
The interface is quite easy to use, I especially like how holding in a button will pop up a help menu telling you what the button does, which practically means that you won’t be looking for the manual very often. I was surprised that none of the help menus include the accuracy specifications as they are often useful to have close to hand. Saving screen shots and data to a flash disk or the internal memory is hampered by the multimeter’s text entry. This is done using the four arrow buttons around the select button and is a little irritating - I found that I was often saving files with numbers appended as this was the easiest and quickest thing to do, but unfortunately not very informative. The four arrow and select buttons are only used for text and number input, and I think that a rotary selector, like those found on digital oscilloscopes, would have been a much more practical way of doing this. Inputting text would be easier, as it would allow a quick twist of the control to pass over several letters instead of pressing an arrow key for each letter, and for numbers a rotary control is more logical.
Some parts of the menu interface could be improved upon. The most obvious thing that caught me out when I first used the multimeter was that the done button, which gets you back to the previous menu, is also the clear readings button in many measurement modes - so one too many presses can delete all of your reading history (as there is no confirmation). Additionally some of the options are buried quite deeply in the menu system, with it taking 5 button presses to change the screen’s brightness or to turn the screen off. Having menus that are this deep can make it quite hard to remember where the individual settings are and can leave you playing a bit of a guessing game to find the right menu.
For all of the interface, there is a black bar that runs across the top of the menu and below the main part of the screen (where the numeric display and graphs are), which is used for when numbers are entered, for things like ranges. As this is not always needed, it would have been good if the screen filled this space by default or if the number entry was done in a pop-up window, so that a section of the screen is not constantly being wasted. In the same vein, it would also be very useful if the trend chart and histogram could have a full screen mode, where the menu disappears entirely.
I have found that general usage of the 34461A is easy to do and quite logical, however, some of the features could be improved upon.
The statistics display is a very useful addition that does not take up too much of the screen space and there is generally not that much that can be said about it. However, I think it has a bit of a strange behaviour when it overloads, as everything (maximum, average, span and standard deviation) is set to ‘overload’. This is quite noticeable in the resistance range, where if you are trying to measure a resistor, using the statistics to average out the errors, you have to essentially find a third hand to clear the readings. So it would be quite good if there was a menu option so that the statistics ignored overloads, or, even better, if you could perform statistics on a subset of the 10,000 readings memory, possibly using the trend chart to set the range.
One of the things that I was quite looking forward to testing was the probe hold mode of the 34461A, as I thought that you would be able to perform statistics on the readings. This would be incredibly useful for characterising a batch of resistors, for example, as it would save quite a bit of time. Unfortunately, the probe hold mode doesn’t work this way and is fairly limited. Its only real use is for performing measurements when you can’t see the multimeter’s screen, which is still useful, but it could have been made a lot better with the addition of statistics or the histogram view. I was also disappointed to see that the threshold level is fixed at some unknown value, however, on my Keithley 2000 the equivalent hold mode allows you to select how many readings have to be within the tolerance and the tolerance level between 0.1 and 10 %. Whilst this might seem a small gripe, it does make a practical difference as when I was testing its use on a PCB it would record different values (by 2 to 3 digits) for the same resistor, which suggests that the measurement had not stabilised. Overall, the probe hold mode is functional and I can see myself using it in the future to quickly verify if a reading is reasonable, but I don't think it is practical to use this mode when a high accuracy is required. I really think Agilent missed a trick with this and I am hoping they will improve on it in a future firmware update.
The number display on the multimeter is clear and easy to read from quite a distance, and I like how a space is inserted into the digits to logically group them. But there is relatively little that can be said about displaying a single number. The 34461A really excels with its graphical views, in particular the histogram and tend chart.
Throughout this roadtest, I have found myself switching between the number display, the histogram and the tend chart. I don't think at any point I thought that a reading would be best displayed on the bar meter. Overall I think this is because the bar meter does not add anything that cannot be displayed better on the tend chart. Many handheld multimeters include a bar meter as well as a numerical display, with the bar meter having a faster update rate. This allows you to see a precise measurement as well how stable the measurement is. I would almost guarantee that this is what Agilent wanted to have, but instead they have ended up with a bar meter that follows the numeric display exactly (that is at the same update rate), which isn't very useful. Additionally, the bar always starts on the far left hand side, which makes very little sense and can be quite confusing.
The image above shows this, with a quick glance at the bar meter giving the impression that the reading is below -100 mV. This, to me, makes the bar meter useless, as at no point should any display make it possible to mis-read the measurement and I have no idea how this feature managed to pass the initial usage tests.
The histogram view is incredibly useful and I like how it is not dependant on the reading memory, with it instead using 400 bins internally. This means that it is not limited to 10,000 readings and instead can go significantly higher (although I don’t know the exact limit). This is particularly useful if you want to see how a measurement varies over time, as you could leave the device on for many days collecting data. The histogram works very well and I have three issues with it. The first can be seen in the following image.
When there is a sharp peak in the histogram, the rest of the histogram tends to zero and so becomes very difficult to see. To help in these cases, it would be nice if a log scale button was introduced, so the histogram could be redrawn with a logarithmic frequency. The second issue is that you can not change the number of bins used without it clearing the data. Whilst I understand that changing the range of the histogram requires it to be cleared, choosing between using 400 or a 100 bins shouldn’t, especially when the data is stored as 400 bins internally. This can be seen through using SCPI commands, as ‘CALC:TRAN:HIST:ALL?’ returns 400 bins despite how many bins are specified on the multimeter or the number of measurements. And the third issue is that you cannot get the histogram data from the multimeter without using USB, LAN or GPIB. Which seems, to me, to be a very odd omission as you can use the front panel to save the reading memory to a flash disk, but not the histogram data.
There is currently a bug with the histogram view, that hopefully Agilent will fix quickly, where the histogram does not update when the multimeter is not triggering. Unfortunately for me, I found this out whilst characterising a pack of resistors (see this blog article Practical use of the Agilent 34461A). Whilst the data had been correctly placed into the bins, the display instead showed the following.
The histogram incorrectly displays the data from a previous set of current measurements, whereas the numerical display and the statistics are for the resistance measurements. This was a little disappointing, as I am surprised that no one had come across this before, however, Agilent did confirm that this was a bug in the firmware within a day, or so, of me reporting it to them. So I can only hope that a firmware update will be available in the near future.
Update : Agilent have given a date for the next firmware release. As this isn't a critical bug and the development teams are busy with new products, they have said that a new firmware version will be released in December 2014. Whilst I appreciate the work involved in fixing the bug and testing the firmware, I am a little disappointed that this date is so far off (the rep informed me of this date in February) and that fixing a bug takes almost a year. After the initial quick response and acknowledgement of the issue, I was expecting a new firmware version to be released considerably faster.
The trend chart is probably one of the best uses of the 34461A’s 4.3 inch screen and I think any user of the this multimeter would be using it frequently. Whilst the histogram is good at showing the span of readings and what ranges are the most common, the trend chart shows how the readings change over time. Like the histogram, it too is independent of the reading memory, however, unlike the histogram, you can not access the data the multimeter uses for the displaying of the graph. I am not entirely sure how useful this data would be, as it is likely that it is downsampled so that it can be kept in memory, however, it would be nice to be able to plot it on a computer with a higher resolution than what the 34461A’s screen can show. Although, for any practical use, the remote control options would be used instead. The only issue I have with the trend chart is something that one of the other roadtesters has pointed out, where it does not autorange by default. The default scale is the current measurement range of the mode, which is an odd scale to choose as a default. There is an option to change the scaling to auto, but it seems like it should be the factory default. This can be mitigated completely with the use of the store and recall settings feature. This allows you to save all the current settings of the multimeter and recall them when needed, or, what is particularly useful, set a state to use on power on. By doing this, my 34461A uses autoscale as the default for the trend chart - but I do feel that it is a few extra steps that an end user has to perform for what should be the default.
Both the histogram and the trend chart do not have any cursors, so it is not possible to scan back through the trend chart to find the value of a certain point or find the height and value of a bin in the histogram. This is a little disappointing, as it is something that could quite easily save time. Instead, the data has to be exported to a computer, where it has to be plotted and looked at. As I have said above, the arrow buttons are only used for text and number input, so they could be easily used for moving a cursor in the graphs.
The Agilent Windows software for controlling the multimeter is generally quite good and easy to use, although there are a couple of annoying inconsistencies (see the blog post). I particularly like how it allows you to use cursors on the generated graphs, whilst still being quite intuitive. The histogram display, however, is not as well done as graph, with it having some issues with the autoscale; and the limits functionality is largely useless, as the software doesn’t notify the user when the limits have been exceeded. On the whole, the software is fairly useful, but I don’t see myself using it particularly much in the future.
I found the iPad application unusable, which might have been partly due to the age of the iPad I was using, however, from seeing the available features of the application, there seemed to be no real advantage of using it over the 34461A’s built-in interface. As I said in the blog post, the iPad application seems largely a gimmick than a functional feature - something that might be brilliant at showing the multimeter to executives at trade shows, but not something you are likely to use in a real environment.
Using the multimeter remotely is fairly straightforward and works as you would expect it to. Although I did feel that the manual lacked some more basic level documentation. For example, the command snippets in the command descriptions are very brief, which is likely to confuse those who are new to remote control and automation. Which could easily be remedied with one or two more in-depth examples or a simple tutorial. Additionally to this, some basic documentation on the sockets interface in the manual would be appreciated, as the port number is only given in a fairly old (2007) application note.
There is one particular gotcha when using the remote capability of the 34461A which is annoying. When the multimeter is removed from remote mode, back to local mode (i.e. the multimeter’s front panel) any data in the volatile buffer is cleared. It doesn’t matter how local mode is entered, either remotely or using the front panel, the buffer is always cleared. This can be quite a large annoyance, as sometimes you want to look at the data on the multimeter that was captured remotely, but, unfortunately, it is not possible to do.
I was quite surprised to see that the 34461A had a soft power switch, as there is nothing in a multimeter that needs to be powered when off, unlike, for example, equipment that contains temperature controlled oscillators. So I was quite intrigued to see how much power it consumed.
I measured the power using a fairly cheap, household plugin power meter, so these measurements are not going to be incredibly accurate, but should provide a useful estimate.
On the day I measured the power consumption, the mains voltage was running a little high at 249 V, which might have affected the figures a little as well. Whilst in standby, the power consumption was 1.4 W and when powered on it was around 9.8 W, with a peak of 14.9 W during boot.
A power consumption of over a Watt when powered off is quite excessive. I mentioned in my first blog post on this multimeter that it didn't have a true power switch, and morgaine commented that it might be because the device can be remotely powered on and off. As the manual and datasheet do not mention a remote power on capability, I decided to find out myself.
When in standby with an Ethernet cable plugged in, the lights on the Ethernet port on both the 34461A and the connected router remain off with wake-on-LAN packets being ignored by the multimeter. As I don't have any GPIB interfaces to communicate with the multimeter, I was not able to test if any power on capabilities existed.
As I wasn't particularly satisfied with this and I wanted to see what was powered, I decided to open the multimeter and look at the test points that are given in the manual. Out of all of these, only the first test point has the required voltage on it. This test point has a 25% margin on it and is essentially there to check that the transformer is working and the device is plugged in. However, from looking around inside a lot more can be learned.
When the device is turned on, there are a few seconds when the screen remains white before the boot loader kicks and the first grayscale progress bar appears. During this time, the two LEDs on the measurement board, marked as FPGA and uController are off, and it is only when the first image appears on the screen that these turn on (the FPGA remains solid, but the uController flashes until the point where the self-test begins), which is accompanied by a quiet click, possibly that of a relay. This gives the impression that the measurement board is unpowered until the first image appears. From measuring voltages on this board this seems to be the case, in particular there is a Texas Instruments TPS54040 (a 42 V 0.5 A step-down converter - U1005 on the board) located near one of the transformer connections, that has its enable line low until this moment. Additionally, there seems to be no power lines going to the IO connections at the back of the unit that are powered when off. Whilst this doesn't rule out a remote power control, it seems very unlikely that one is currently present in the 34461A. Which really begs the question as to how a soft power switch, an LED and an RTC can consume over a Watt of power. Personally, I would have preferred a true power switch as I can see myself regularly switching this multimeter off at the wall.
The 34461A is a very good multimeter and has a very reasonable price, that is consistent with its competitors whilst offering some slight improvements. The measurements it takes cannot be faulted, with the hardware being almost perfect, with the only drawbacks being the soft-power switch and the text entry / arrow buttons. However, the firmware, although being functional, could be improved in quite a few places.
This is quite lucky for Agilent, as they could easily create a firmware update to add cursors into the trend chart and histogram, and to improve the statistics and probe hold modes, but whether they will do this or whether they will keep these in mind for their next multimeter is something we can only guess on.
Some of the documentation, in terms of remotely programming the multimeter, is lacking a little in areas, but this is something that could be easily improved.
Overall, if you need a 6.5 digit multimeter, this is most likely the one you will want. Whilst the firmware could be improved, its display and graph modes will make life easier when you are measuring on your bench, as they are considerably easier to read than a VFD.
Update after 6 months
I have been using the 34461A fairly regularly over the last 6 months (since I received it), for a range of different measurements and it has become the multimeter that I always seem to go to. Even considering the 30 second start-up time, I prefer to use the 34461A over both my Keithley 2000 and my handheld multimeter, for even the simplest of measurements that only require a couple of digits of accuracy. I think the main reasons for this is that the multimeter is so nice to use and has all the extra features (in particular the trend plot and statistics) that are lying ready for when you need to investigate your readings a little more.
The vast majority of the time, I have been using the DC voltage and DC current, and to a lesser extent, resistance modes of the multimeter. With the AC voltage, AC current and temperature modes remaining fairly unused. I did, however, find myself using the frequency mode to calculate the energy cost of a function on a microcontroller. Which surprised me, as it was not a function that I expected I would ever use.
I have, so far, not used the bar graph display once and, from the type of measurements I have been taking, I think it is unlikely I ever will. I was hoping that I might find a use for it, and be less negative about it than I was in my original review, but, unfortunately, my opinion of it has not changed and I still see it as an awful way of displaying the data.
I mentioned, when I first wrote this review, about the fairly large power consumption of the multimeter when it is switched off, and over the last 6 months I have found myself turning the 34461A off at the wall rather than leaving it in standby. This is partly due to the power consumption and partly due to the front panel standby LED. Whilst I wish that Agilent had gone with a true power switch, it occurred to me when looking at other Agilent test gear that it is likely that the front panel is a generic module they use across many devices, including those that require devices to be powered in standby. So it was probably too complicated, or would potentially affect the reliability of the multimeter, if they had to redesign the front-panel for a true power switch. Which, at least, makes the design decision understandable.
Overall, I am very happy with the 34461A as it is a very high quality instrument, that is easy to use. Whilst I am disappointed with Agilent that they are going to take until December to release a new firmware version (which fixes the histogram bug), the multimeter is very well built, with the firmware being very reliable. There are a few things in the firmware that I would have prefered to have been different - such as defaults and extra features - but these are very minor issues, which shouldn’t deter anyone from buying what is a fairly perfect multimeter.
1 (01/12/13) - added link to the third blog post.
2 (15/05/14) - added update for firmware bug fix, remote use, extended the conclusion to include positives and negatives, added the 6 month update and added link to the fourth blog post.