Following on from my previous blogs about Connecting Loudspeakers to Amplifiers and Passive Dummy Loads for Amplifiers here is a project for a modest passive dummy load with a twist – I have included a differential ‘scope probe to facilitate monitoring of the signal on an oscilloscope.
I managed to score a new-old-stock Grayhill switch from Ebay. This is a 2-pole 3-way switch with 90° indexing and a 5A current switching capability. I decided to implement this design:
However, I wanted the option of open-circuiting the load which I could do more elegantly with a 4-way switch, but a 3-way switch is what I have so I added a second switch to isolate the load. I chose to use a key-operated switch because it has a good current-carrying capacity, 90° indexing to match the Grayhill and reminds me of the importance of not inadvertently disconnecting the load from the amplifier. This gives the following schematic:
But what about the differential ‘scope probe? Well the problem with connecting an oscilloscope to almost anything is that (unless the ‘scope is a portable battery-powered design) the chassis of the scope (and thus the earthy side of the ‘scope probe) will be connected to mains earth. We may not want to ground the output of our amplifier at all – and, even if one side is grounded, we don’t want to accidentally connect our oscilloscope ‘back-to-front’ and short the live output of the amplifier to ground.
We could buy an isolated differential ‘scope probe to solve this problem but they are silly-money expensive. So instead we will cheat a little and use a battery-powered op-amp to sort-of-isolate the dummy load from ground. Here’s a simplified schematic:
The LT1360 from Linear Technologies is a high speed, very high slew rate (800V/us) operational amplifier with a gain-bandwidth product of 50Mhz.
In the above circuit, the input is connected across the load resistor bank (see below for full schematic) and the output is connected via a 50Ohm BNC cable to our oscilloscope. The input impedance of the op-amp is a relatively low 5 MOhm, which is comparable to the 2 MOhm of input resistance. Without getting into the math, suffice to say that the R3/PR1 combination is about 9700 Ohms for a X10 probe-effect at 1 MOhm ‘scope input impedance (or a X20 probe-effect at 50 Ohms ‘scope input impedance). So, assuming fresh batteries, the input range is about 0-53V rms (representing 350W into an 8 Ohm load). Thus we have good headroom considering the load is rated for 200W at 8 Ohms (and will get hot quite quickly with a load of 100W).
The big question is whether we should a.c. or d.c. couple the input. If we d.c. couple then any d.c. offset on the load will show on the oscilloscope. On the other hand a.c. coupling allows for a better a.c. dynamic range. Given that we are going to allow for the connection of a multimeter as well as a ‘scope a.c. coupling is the way to go.
Next we need to add all the extra bits to make the circuit complete. We chose to use one 9V battery with a TC1044S charge-pump to generate the -9V. Also we implemented a low-battery-voltage LED (illuminating at about 8.1V) using an MCP111-450E. The circuit uses about 9mA so the battery should last a while. Here’s the full schematic (click for a larger version):
C1 and C2 set the low frequency roll-off at less than 1Hz. If you reduce C1 and C2 to 10nF the low frequency -3dB point will increase to 16Hz or so. C3, C4 and C5 roll-off high-frequencies. With the values shown, the upper -3dB point is at 42kHz. If you reduce c3 to 47pF and c5 to 10pF, the upper 3dB point will increase to about 150kHz. A middling compromise would be C3 = 220pF, C4 = 47pF and C5 = 10pF giving an upper -3dB at 85kHz.
The project is built into a Hammond 1590R1 diecast enclosure. I got the heatsink from Ebay – it’s 120 x 100 x 20mm cut down to 100 x 100mm. They key-operated switch is a Burgess P2-3FBC-209 and SW3 and SW4 are available from CPC or Farnell as 1MD1T2B4M7RE and 1MS1T2B4M7RE respectively.
The intended use for this project is for test and repair of valve (vacuum tube) amplifiers with modest output power so we are not too worried about the lack of forced cooling. It’s worth noting that the unit is effectively a radiant heater, so 100W continuous power input will get the case too warm to touch in a few minutes and even a few Watts will get it hot eventually.