Following on from my previous article on matching amplifiers and loudspeakers, this article offers some ideas for making passive dummy loads.
Active dummy loads are fine, but if you need to test an expensive amplifier at full load, you do need to be pretty confident that you are connecting a load which is going to be dependable and reliable. Also, quality active-load instruments are expensive.
So let’s say you’re going to build a dummy load using conventional aluminium-clad power resistors. These are available on eBay for a good price. The sweet spot for price at the time of writing is to buy 100W 4 Ohm resistors in pairs.
Note that the green no-brand Chinese resistor is supposedly four times the power of the Welwyn WH25 part but it is barely twice the size and twice the mass. This bring us to an important consideration: The specified power rating of this type of resistor is not the free-air rating, but assumes that the resistor is mounted on a heatsink. The Welwyn resistor is assumed to be mounted on a 30cm x 30 cm x 1.5mm (one foot square by 1/16″) aluminium heatsink.
It is also worth noting that if we compare the dimensions of the Chinese resistor with the Welwyn datasheet, the supposedly 100W Chinese resistor is about the same size as a 50W Welwyn part. This indicates that the Chinese manufacturer is basing his 100W rating on the assumption that either you are going to mount it on a substantially better heatsink than Welwyn specify or you are going to cool the resistor with a fan, or both.
Note that I don’t have any problem with the construction quality of the Chinese resistors (or, for that matter the tolerance – which is within 1% on the four samples I purchased).
The actual resistance of our dummy load is not critical. As we noted in our previous article, the impedance curve vs frequency of a typical full-range driver is anything but flat and the nominal 8 Ohms value of a driver generally represents the minimum impedance the driver exhibits. Also, the d.c. resistance of an 8 Ohm driver is likely to be as low as 6.5 Ohms. So, if we use a 3.9 Ohm resistor in lieu of 4 Ohm, an 8.2 Ohm resistor in lieu of 8 Ohm or a 15 Ohm resistor in lieu of 16 Ohms, we are not going to have a problem with the amplifier. However, we can get 4 Ohm, 8 Ohm and 16 Ohm resistors from eBay, so that’s OK.
We could buy a 4 Ohm resistor, an 8 Ohm resistor and a 16 Ohm resistor and connect them individually to speaker cables and use them separately, but that would be too easy. If we are going to have an integrated design, the basic configuration for our dummy load is thus:
However, there is a disadvantage with this arrangement. For any given voltage, we know that the power dissipation:
P = V²/R
So, the smaller the resistance, the greater the power. We have 300W-worth of resistors, but when we want to get 4 Ohms, we are wasting 200W of power dissipation.
If we take four 4 Ohm resistors and connect them in series-parallel like this the net result is 4 Ohms at 400W power dissipation:
So, let’s see what we can do with four 4 Ohm 100W resistors. Firstly, if we have an SPST switch to hand, we can do this:
With four binding posts, we could get stereo 8 Ohms @ 200w/channel, 4 Ohms @400w, or 16 Ohms at 400W.
Alternatively, if we have a couple of switches and only two binding posts, we could do this:
One DPDT toggle switch and one other (SPST, SPDT, DPST, or DPDT) and two binding posts are required to give 4 or 16 Ohms @ 400W or 8 Ohms @ 200W. Be sure to buy heavy-duty mains-rated types.
Here’s a general arrangement of the wiring for the above schematic (188SV) assuming you’ve bought two identical DPDT switches:
With both switches up the load is 8 Ohms @ 200w, with SW1 down, the load is 4 Ohms @400w. With SW2 down (or both SW1 and SW2 down), the load is 16 Ohms @400W.
To make our design a little more foolproof, we can use a rotary switch. High-current rotary-wafer switches are relatively expensive and you might be tempted to use a signal-level switch like a Lorlin CK-series or an Alpha SR25 or SR26 series, but they don’t have adequate current-switching capacity and they only have 30° indexing, so I don’t recommend you use them. If you deal with Mouser or Digikey, you could consider the NKK HS16-2, TS2N or PS2, or the Grayhill 44 range (but with the Grayhill 44, get at least 45° indexing to ensure 5A current rating) . Or maybe you have managed to extract something from an old piece of test gear, or catering equipment. Anyway, here’s the schematic:
This gives 4 Ohms or 16 Ohms into 400W or 8 Ohms into 200W.
If you can’t get hold of a wafer-switch, the alternative is a cam switch. Some cam switches operate the same way as wafer switches and can be wired as above. However, some have a different operating principle to wafer switches. Isolated-type cam switches have a rotating cam which closes one or more pairs of SPST contacts at each actuator position. The canonical design for the cam switch is the Kraus & Naimer CA10 series. The Chinese have started ripping off the Kraus & Naimer design and the sellers on Ebay and Alibaba have no idea of the specification they are selling, so the actual switch you get may not match the description. They are good quality copies though. Here’s a picture of a copy:
Bear in mind that with genuine Kraus & Naimer, you have to buy the switch body and the actuator separately. There are many models you could press into service, but if you’re going to order a new one, I suggest you go for one of the following options:
- Wafer-style: CA10-A250-600
- Isolated Style: CA10-A750-600
Here is the schematic for the isolated cam style:
Modifications to the basic design
You may wish to add additional binding posts so you can connect to a multimeter or oscilloscope at the input to the load. Also, you may want to be able to measure the off-load voltage of your amplifier in which case you might to have a four-position switch with an off position.
In the next part of this series we will present our own dummy load project.