Hi Everyone,
This is my first post, so please bear with me.
I've just completed my first major audio project. The power amplifier comprises of two LJM L12/2 Power Amplifier modules and an LJM power supply/protection board, all built from kits from AliExpress.
The preamplifier is of my own design. It uses reed relays to select the input and as a volume control. I'm really pleased how well it works. The reed relays give a nice smooth control of volume in 1dB steps with no audible clicks (mechanically or electrically). Has anyone else tried this?
Each channel has a 3-way input switch and 6 switched attenuators of -1dB, -2dB, -4dB, -8dB, -16dB and -32dB. By selecting them in binary combinations, 64 levels are available in 1dB steps. LM4562 opamp buffers provide a constant impedace source to the next stage to stop interations between them. See attached circuit diagram.
LM317 and LM337 regulators provide +/-15v power for the opamps, +12v for the relays and DAC module, and +5v for the processor.
I'm using a Raspberry Pi Pico to receive the front pannel buttons and infrared remote control and control the relays, although an Arduino or PIC would work just as well. A 4-character alphanumeric LED display shows the volume and input source. The software is pretty simple and written in Python.
Frequency response of the preamp is 7Hz to 500KHz at -3dB. I don't have equipment to measure the distortion or noise, but expect it will be extremely low. It sounds great to me, but of course I'm biased.
I'd be interested to hear what everyone thinks, and especially ways in which it could be improved.
Gary
LJM L12/2 power amplifier:
Reed Relay Preamplifier:
This is my first post, so please bear with me.
I've just completed my first major audio project. The power amplifier comprises of two LJM L12/2 Power Amplifier modules and an LJM power supply/protection board, all built from kits from AliExpress.
The preamplifier is of my own design. It uses reed relays to select the input and as a volume control. I'm really pleased how well it works. The reed relays give a nice smooth control of volume in 1dB steps with no audible clicks (mechanically or electrically). Has anyone else tried this?
Each channel has a 3-way input switch and 6 switched attenuators of -1dB, -2dB, -4dB, -8dB, -16dB and -32dB. By selecting them in binary combinations, 64 levels are available in 1dB steps. LM4562 opamp buffers provide a constant impedace source to the next stage to stop interations between them. See attached circuit diagram.
LM317 and LM337 regulators provide +/-15v power for the opamps, +12v for the relays and DAC module, and +5v for the processor.
I'm using a Raspberry Pi Pico to receive the front pannel buttons and infrared remote control and control the relays, although an Arduino or PIC would work just as well. A 4-character alphanumeric LED display shows the volume and input source. The software is pretty simple and written in Python.
Frequency response of the preamp is 7Hz to 500KHz at -3dB. I don't have equipment to measure the distortion or noise, but expect it will be extremely low. It sounds great to me, but of course I'm biased.
I'd be interested to hear what everyone thinks, and especially ways in which it could be improved.
Gary
LJM L12/2 power amplifier:
Reed Relay Preamplifier:
Last edited:
Hi, lovely project you got there, I was wondering if you could share what kind of reed relays those are, looks perfect for my preamp aswell.
BR
Mikael
BR
Mikael
The relays are SIP-1A05 . I got mine very cheap from AliExpress. I suspect they have been removed from telecom equipment, but they work fine.
Good job on the preamp. Ideally, the 100k resistor to ground should be connected on the other side of the input switch,
to prevent pops from the coupling capacitor. Or there could be resistors to ground on both sides of the switch.
Are the reed relay switches normally open or normally closed?
to prevent pops from the coupling capacitor. Or there could be resistors to ground on both sides of the switch.
Are the reed relay switches normally open or normally closed?
I like the British flag displays.
Beware that if you're aiming for the ultimate in low distortion reed relays are not the best choice. There's a good article by Groner and Putzeys in Linear Audio Vol. 13 about that. They're usually good down to about -80 dBc (0.01%) THD. If you want lower distortion, gold point small-signal relays would be better options, but they're usually somewhat pricey ($5-6/each).
Tom
Beware that if you're aiming for the ultimate in low distortion reed relays are not the best choice. There's a good article by Groner and Putzeys in Linear Audio Vol. 13 about that. They're usually good down to about -80 dBc (0.01%) THD. If you want lower distortion, gold point small-signal relays would be better options, but they're usually somewhat pricey ($5-6/each).
Tom
Correct. Reed relays make quite a bit of distortion. Use normal signal relays with gold contacts. Your relay attenuator is built quite complicated with the many op amps. Why don't you use established designs with 1 op amp?
Thanks Rayma, that's a good point about the 100K resistors. I put them on the source side of the input switch to give a load to any capacitors which might be in the outputs of the source devices, so there should always be zero volts on the input when switching over, but I might add one to the other side of the relays as well.
I've not noticed any pops, but if I did, it would be easy enough to modify the software to set the attenuator to maximum while switching.
The relays are normally open
Thanks Tom, I did wonder about that. I doubt if I could hear 0.01%, but it's certainly worth bearing in mind. I wanted to use reed relays because they are quiet in operation, compact and long lasting, and as an experiment to see how well they would work.
Yes I do, and ramp the volume up after a short delay. The input relays are also open, so no sound gets through anyway.
The interesting part is when the relays are switching from, for example 001111 to 010000 . Reed relays are very fast, but if ones which are closed break faster than the new ones make, i.e. 001111 -> 000000 -> 010000 , there could be a short and very loud spike in volume or click. I get round this by switching in two steps, closing the new contacts 3 milliseconds before opening the old ones, i.e. 001111 -> 011111 -> 010000 . This causes the sound level to drop for a tiny fraction of a second instead, which isn't noticeable.
Yes, ideally this circuit would use normally closed relays instead, which are not available for reed types.
I did consider that, the designs I looked at all used change-over relays. If I were to use reed relays, I'd need twice as many and twice the drive circuity. Also, I was concerned about the mechanical noise of the relays and pops when switching, when neither contact was made, whereas this is silent.
You might be able to make a make-before-break action in software. I think that's what Benchmark does in their HPA-4 headphone amp.
Tom
Tom