The author’s DIY Hummer, made from an AC/AC wall‑wart PSU.
If you want to identify studio equipment that’s susceptible to grounding problems, a simple DIY project based on a wall‑wart PSU can help.
My Ground Control article in SOS August 2023 (https://sosm.ag/ground-control) raised the topic of good and bad audio equipment design practices in respect of correctly managing interference and noise currents. In poor designs, unwanted noise currents from ground loops or RFI (radio‑frequency interference) introduced via the shields of cables connected to other equipment are allowed to intermingle with the audio signal currents within the amplifier circuitry, adding hums and buzzes to the wanted audio. In contrast, good designs ensure such noise currents are kept well away from the audio circuitry, leaving the device blissfully quiet and trouble‑free.
The Hummer
The Windt Hummer test was first proposed by John Windt in an AES paper in 1995.Unfortunately, there’s no way of telling whether a piece of equipment is designed well or badly in this respect just by looking from the outside — but there is a simple engineering test to find out, and the necessary equipment can be constructed with moderate electrical DIY skills and a spare wall‑wart power supply. This device, called a Hummer, was invented by John Windt and first described in the AES Journal (June 1995, Volume 43, No. 6). It was created in response to another article published in the same issue by Neil Muncy, who discussed at length equipment susceptibility to noise currents injected via external audio cables — something he christened the ‘pin‑1 problem’.
Conceptually, the Windt Hummer is nothing more than a reliable source of an audible noise current that can be injected (harmlessly) into the ground paths of any audio device. If a hum or buzz becomes audible (or the noise floor changes noticeably) at the output of the device under test, it clearly has a ‘pin‑1 problem’ and will be highly susceptible to ground‑loop or RFI issues when installed in an audio system. If all remains quiet under test, it is a well‑designed device that handles noise currents properly. It really is that simple!
So, the source of noise current in the Windt Hummer is a standard (double‑insulated) wall‑wart power supply with an AC output. Any AC PSU with an output current rating of more than 30mA and an AC output voltage of between 6 and 24 Volts is acceptable — my own tester was built using a redundant 1.9A 9V AC PSU from an old phone charger that I found in the back of a drawer. The wall‑wart transformer provides a safe, electrically isolated (ie. floating) and low‑AC signal at 50 or 60 Hz (depending on the local mains supply). The actual voltage isn’t particularly important, although I’d recommend 6‑12 V if possible. The current being injected into the device under test is restricted using a current‑limiting resistor. Windt’s original design passed 100mA and that seems excessive to me, so the incarnation described here provides around 50mA of current.
If a hum or buzz becomes audible at the output of the device under test, it clearly has a ‘pin‑1 problem’ and will be highly susceptible to ground‑loop or RFI issues.
Hearing a pure 50Hz hum might not be very easy in some listening conditions; a signal containing higher harmonics would be far more practical. Of course, the local mains supply may already have plenty of harmonic distortion, but a couple of diodes across the output to rectify the AC signal absolutely guarantees it! Moreover, if one of those diodes is an LED it will provide helpful visual confirmation that current is really flowing into the device under test.
The Hummer is designed to inject its noise current via an audio cable’s shield connection, returning via the device’s chassis (and/or the protective earth in the mains supply). A croc‑clip can be attached temporarily to the chassis (or just held against it), so that’s easy. Windt’s original design used another croc‑clip to access the audio connector shield connection (XLR pin 1) too, but I think that presents too great a risk of accidentally touching an audio input terminal (pins 2 and 3): a 12V AC signal is equivalent to +24dBu which, if accidentally touched to an audio input terminal, will max out a line input and probably destroy a mic input! So, to negate the risk of sending the Hummer signal anywhere other than the ground path, I strongly recommend using an XLR plug as the second connector, with the Hummer signal wired to pin 1 and nowhere else. If the device under test doesn’t have XLR inputs, you can substitute a TRS plug (Hummer signal to the sleeve terminal) or stick with the XLR and use standard adaptor plugs or cables to convert from the XLR to any other format — that’s what I do.
So, the complete parts list is: low‑voltage AC wall‑wart, red LED (ensure it has a peak current rating of more than 50mA), IN4001 diode, resistor (value from table shown), croc‑clip, male XLR‑3 connector, and heat‑shrink sleeving. You’ll also need a soldering iron and the ability not to burn your fingers or set the house on fire during the construction!
Adapting an AC PSU to create a Windt Hummer test device is a very simple DIY job.
Depending on the construction of the redundant wall‑wart power supply, it might be possible to build some of the components into the wall‑wart itself, which I did. Alternatively, it might be more practical to build them into the XLR plug, with a small hole drilled into the side of the shell to accommodate the LED (with copious heat‑shrink sleeving to prevent short‑circuits in the wiring).
The current‑limiting resistor’s value and power rating depend on the output voltage of the wall‑wart power supply. If you like sums and algebra, the calculations are from Ohm’s Law in the form: R=V/I. We need 50mA (0.05A) of current, so with a 12V AC wall‑wart the calculation is 12/0.05 = 240Ω. In practice, some of the AC voltage will be lost across the diodes, which means the actual current will be slightly less than 50mA, but that will help with power dissipation.
Talking of which, 50mA passing through the current‑limiting resistor will make it warm, so it needs to be rated to handle that power dissipation. The formula required to check this is: Power = Voltage squared, divided by the resistance. In this example, that’s 144/240=0.6W, and a 1W resistor will do the required job — it’s not as if this Hummer is going to be left connected all day long, as current is normally only passing briefly during testing. If you can’t find a calculator, the table gives the required E24‑series resistor values (5% tolerance is fine) and suitable power ratings for common wall‑wart AC output voltages.
As I mentioned earlier, given a choice I’d recommend using a 6, 9, or 12 V wall‑wart, since the current‑limiting resistor will be physically smaller and easier to install inside the wall‑wart or XLR. If there isn’t space there, you can fit it to the croc‑clip instead, supported and protected by some heat‑shrink sleeving.
Testing
To test an audio device, disconnect all inputs and outputs except for the output being used to monitor for nasty noises. Turn the monitoring volume right down, then power up the unit and advance the volume control, listening for any hums or buzzes (there shouldn’t be any at this stage!). If possible, advance the volume sufficiently to assess the level and nature of the noise floor (which is hopefully just normal Gaussian hiss). Make a note of this volume setting for later before turning it right down again.
Next, plug the Hummer’s XLR into one of the unit’s audio inputs and attach the croc‑clip onto a suitable part of the chassis. The Hummer’s LED should illuminate to indicate a noise current is being injected. If it doesn’t, try the croc‑clip somewhere else such as a fixing screw or the shield terminal of another connector. With the LED illuminated, carefully advance the monitoring system’s volume while listening for hums and buzzes. If a buzz or hum is evident disconnect the croc‑clip (the LED will extinguish) to confirm it is genuinely coming from the Hummer’s noise current.
If the equipment is well‑designed there shouldn’t be any buzz or hum at all, but it’s also worth checking the noise floor to see if the noise current has degraded it. Set the volume control to the noted position from earlier and check the noise floor remains at the same level and sound character as previously. If possible, check all of the device’s inputs and outputs (via suitable adaptors) as some may exhibit problems while others won’t, depending on the internal construction and wiring.
If the tested unit exhibits hum and buzz problems when using the Windt Hummer, it is highly likely to suffer from ground‑loop noises and/or RFI issues when installed in an audio system using standard audio cabling. As a result you will almost certainly need to employ one‑end‑only cable techniques (see the aforementioned Ground Control article) or transformer isolation boxes to sidestep such problems.