I’m building a switch box to do A/B comparisons of two amplifiers as well as two preamps. I’m including a diagram of the relay contact assignments of the relays that handle the audio signals that I am routing through the device. The two selector relays (3 and 4) are turned on/off remotely from two different remote control units. At this point I have the preamp relays (4 and 5) wired up and switching properly from its remote unit. The signal path is correct and I am able to switch between two preamps, but there is a loud pop each time signal relay (5) turns on or off. Using a scope (without a preamp attached), I’ve measured the spike going into the RCA plugs of my device at the moment of the relay switching and it ranges from 50mv to just under 200mv and lasts 250-300ms. If these spikes indeed account for the popping noise they shouldn’t be enough to harm a speaker, but they are quite unpleasant.
I have 1N4148 flyback diodes across the 12v voltage input pins of the relay, but I’ve actually tested for the surge without the diodes and it is about the same. My best guess is that voltage from the switching pins is able to jump to the signal pins. I would have thought that they would be well isolated. I’ve tried installing .5uF caps across all the connections (power and signal) but hasn’t helped.
Is there any way to fix this problem?
Is the problem that 12v relays are too high powered for my application? I haven’t found any 5v 4PDT relays in the format that I’m using now. Low power relays signal relays could potentially work, but that would completely change the design and scope of the project. Or is there something else that I could do to supress the pop?
I’m attaching a diagram showing how I configured the relays and a couple of photos of my build so far. I’m really a beginner with this stuff, so please please excuse any shortcomings. I welcome any advice you might have regarding anything in this design since I dreamed this up on my own. Thanks, in advance, for your help!
--Jonathan
I have 1N4148 flyback diodes across the 12v voltage input pins of the relay, but I’ve actually tested for the surge without the diodes and it is about the same. My best guess is that voltage from the switching pins is able to jump to the signal pins. I would have thought that they would be well isolated. I’ve tried installing .5uF caps across all the connections (power and signal) but hasn’t helped.
Is there any way to fix this problem?
Is the problem that 12v relays are too high powered for my application? I haven’t found any 5v 4PDT relays in the format that I’m using now. Low power relays signal relays could potentially work, but that would completely change the design and scope of the project. Or is there something else that I could do to supress the pop?
I’m attaching a diagram showing how I configured the relays and a couple of photos of my build so far. I’m really a beginner with this stuff, so please please excuse any shortcomings. I welcome any advice you might have regarding anything in this design since I dreamed this up on my own. Thanks, in advance, for your help!
--Jonathan
Attachments
Last edited:
It's not easy to figure all that out at a distance tbh but what I would suggest is you disconnect the preamp feeds and add shorting links across where the preamps feed in and just see if the preamp relays then switch silently.
Loud clicks are often a result of switching a change in DC conditions so also make sure all signal inputs to the relays are at 0.000 volts and also that the power amp inputs are also at 0.000 volts when not connected to anything.
Loud clicks are often a result of switching a change in DC conditions so also make sure all signal inputs to the relays are at 0.000 volts and also that the power amp inputs are also at 0.000 volts when not connected to anything.
Hi Mooly. Thanks for your quick response! I don't detect any DC at all from any of the preamps or between any of the connection. But when I shorted the inputs as you suggested the pop softened quite considerably. Hardly an issue at that level. How might you suggest getting rid of this when the preamps are connected?
Look for a make before break switches. It’s often the brief open circuit created by the switch that causes the pop.
Before I started this project using relays I built a simple box with some dpdt mechanical switches. The switches were not On-Off-On but rather simple On-On toggles. The funny thing is that selecting between the preamps did not cause a pop like it does with the relay setup. Shorting the preamp cables (as Mooly suggested) quieted the problem, so I'm hoping that there might be an electronic solution that might function like a "soft start" when the relay fires. I don't know if this is possible or where I might start for that. If the pop is caused by the brief open circuit I suppose that's at least some good news for the amp switching section of the project since it sounds like the relays I have are break before make switches. Thanks for your idea.
You have to look at all possibilities and its not easy 🙁
Is there any possibility the relay coil current is flowing in the signal grounds and appearing there as a change in level? The relay coil switching should be isolated from the audio.
I had this problem switching hydrophone inputs to a TI DFSII in the seventies. Adding a relay to short the DFS input to ground during the switch over did not help. Just made 2 pops instead of one.
Radio stations and studios use CDS cell optocouplers to do this. CDS cells turn on & off slowly, thus no pop. They used to use incandescent lamps to drive the CDS cell, (Wurlitzer 4500 organ) but now they use LEDs. Newark sells them, about $3 apiece, Optocouplers Silonix NSL-32SR3 is the datasheet I have. You have to use 2 of them for double throw, and use somethng like an 7474 flip flop to drive them at opposite times. Don't forget the resistor to keep the TTL from sinking more than 10 ma driving a diode.
Radio stations and studios use CDS cell optocouplers to do this. CDS cells turn on & off slowly, thus no pop. They used to use incandescent lamps to drive the CDS cell, (Wurlitzer 4500 organ) but now they use LEDs. Newark sells them, about $3 apiece, Optocouplers Silonix NSL-32SR3 is the datasheet I have. You have to use 2 of them for double throw, and use somethng like an 7474 flip flop to drive them at opposite times. Don't forget the resistor to keep the TTL from sinking more than 10 ma driving a diode.
Last edited:
Hi Indianajo. This sounds exactly like what I need: a soft start circuit for the preamp relays! This is foreign territory for me as I haven't work with microcontrollers like the 7474 (quite ironic given that I was an IT guy for over 20 years). So, I must apologize for needing to ask more questions.
I have a conceptual understanding of how the optocoupler works and what the flip-flop does and get that I'll need a resistor if the current driving the 7474's diode is too high. I have a dedicated 4v rail powering the LEDs on the front of the case, so powering the circuit shouldn't be too difficult to implement. Would I want to use 7474d?
Can you provide a schematic or maybe just describe in a bit more detail how I would wire up the circuit you recommend? Reviewing the pinouts from the 7474 datasheet I'm wondering how I would control the input to the "data" and "clear" pins. Is an Arduino or similar device required? Wishful thinking, but is there any way to design this without the need for a flip-flop (i.e. diodes)?
Many Thanks! --Jonathan
Last edited:
The NSL-32SR3 has on resistance 160 ohms at 5 ma in current and 60 ohms at 20 ma. If you use toggle switch to drive the optocoupler, you do not need a flip flop to remember the momentary button push you used to select one input or the other. The LED drops 2.2 v whatever current is driving it, so if you have a 4 v power supply you need a 180 ohm series resistor to limit current to 10 ma. The toggle switch pole contact would switch 1 of two optocoupler LEDs to ground.
If you find a mechanical interlock 1 of 5 switch at a surplus house, you do not need flip flops. Just tie each button contact to one optocoupler and the common pole to ground. These tend to be for motors with copper contacts, so oxidation may be a problem.
OTOH if you want momentary push buttons to select which input to be on, you need something to remember what button was last pushed. 7474 have a set and reset inputs which when pulled to ground, pull the output to that state. Set makes Q high (off). 7474 output pulls to 0.4 v at 16 ma, so you need 140 ohm resistor to limit current on the LED to 10 ma. When using 7474 as RS flop you need to tie the clock input to ground. You need to pull up each input with a switch to ground with a 1000 ohm resistor to + voltage. 7474 need 5 v power supply.
If you have more than one input, you use 1/2 a 7474 for each input and a momentary switch for each reset. A momentary switch to all the sets tied together turns all the flops off.
If you cannot buy 7474, 74ls74 sinks less current in the off or low state. On states of TTL not useful for driving LED, source like 0.4 ma or less.
I think a 74HC74 sinks 1 ma off, which is not useful. But different power supply voltages can be used.
Datasheets for IC's can be downloaded from datasheetcatalog.com.
I find berylium contact DIP sockets oxidize slower than brass contact ones. Oxidized contacts on DIP legs stop the current. I buy 0.1 inch hole project boards for DIP from Newark. They have one that connects each DIP pin to 5 other holes so you do not have to try to make a wire stick to the socket contact. I use 30 ga wirewrap wire in the other holes, or 24 ga solid wire if there is significant current.
Arduinos may be very handy, but I have never been able to buy a matched set that might actually work. You have to buy a CPU board, also input and output boards, and they have to be same generation and maybe manufacturer. My microcontroller experiment sat there and did nothing. I was able to program 8080's on intel SBC boards in 1976-78 but the SBC boards were $1000 each, the card rack was $800, the assembler software was $1300 and the programming test box was $6600. Not something you do at home.
If you find a mechanical interlock 1 of 5 switch at a surplus house, you do not need flip flops. Just tie each button contact to one optocoupler and the common pole to ground. These tend to be for motors with copper contacts, so oxidation may be a problem.
OTOH if you want momentary push buttons to select which input to be on, you need something to remember what button was last pushed. 7474 have a set and reset inputs which when pulled to ground, pull the output to that state. Set makes Q high (off). 7474 output pulls to 0.4 v at 16 ma, so you need 140 ohm resistor to limit current on the LED to 10 ma. When using 7474 as RS flop you need to tie the clock input to ground. You need to pull up each input with a switch to ground with a 1000 ohm resistor to + voltage. 7474 need 5 v power supply.
If you have more than one input, you use 1/2 a 7474 for each input and a momentary switch for each reset. A momentary switch to all the sets tied together turns all the flops off.
If you cannot buy 7474, 74ls74 sinks less current in the off or low state. On states of TTL not useful for driving LED, source like 0.4 ma or less.
I think a 74HC74 sinks 1 ma off, which is not useful. But different power supply voltages can be used.
Datasheets for IC's can be downloaded from datasheetcatalog.com.
I find berylium contact DIP sockets oxidize slower than brass contact ones. Oxidized contacts on DIP legs stop the current. I buy 0.1 inch hole project boards for DIP from Newark. They have one that connects each DIP pin to 5 other holes so you do not have to try to make a wire stick to the socket contact. I use 30 ga wirewrap wire in the other holes, or 24 ga solid wire if there is significant current.
Arduinos may be very handy, but I have never been able to buy a matched set that might actually work. You have to buy a CPU board, also input and output boards, and they have to be same generation and maybe manufacturer. My microcontroller experiment sat there and did nothing. I was able to program 8080's on intel SBC boards in 1976-78 but the SBC boards were $1000 each, the card rack was $800, the assembler software was $1300 and the programming test box was $6600. Not something you do at home.
Last edited:
More wishful thinking: Can I use the same remote controlled (latching on/off) device (that I'm already using to supply 12v to trigger the relay) to activate/deactivate the momentary switch? Or can that same 12V on/off itself serve as the trigger for the 7474?
Thanks -- Jonathan
If you have a latching on off device, you can use that to control the various optocouplers. Just pick a resistor suitable to make the LED draw 10 or 20 ma. 20 ma gets lower resistance, 10 ma makes the LED last longer. Does the latching device have its own input buttons? No momentary switches required. If controlling ma level currents with switches, gold contact switches have less trouble with oxidation block over the years compared to brass, copper, or silver. I like cherry switches. (PS -2.2)/.01=R
Op amps tend to have microamp input currents, so 150 ohms series that current of the optocoupler is trivial.
Op amps tend to have microamp input currents, so 150 ohms series that current of the optocoupler is trivial.
I'm happy to report that the "pop" is gone. I capacitor coupled (47uF MKP) the output from both channels of the preamps (between Relay 5 Pins 9/10 and 11/12) and added a resistor (4.7k) to the signal ground of the RC "preamp out" on the back of the case. That quieted things a bit. The thing that made the big difference was getting the signal ground out of the relay path. In other words, the signal ground runs from the input terminals to the output terminals directly and the relay only switches the signal. I should mention that the signal ground is not tied to the chassis ground. When I tried tying it to the star ground near the rectifier board it picked up RF noise.
I tested my preamp switch using several different amps and preamps and it worked really nicely as I had hoped it might.
Thanks everyone for all of your input and suggestions. I truly appreciate it! I know that I'll soon have some questions about the amp switching module. I know I'm going to have to be really careful to ascertain that I don't run the risk of two amplifiers connecting simultaneously to one set of speakers. Rather a high-stakes proposition, right?
Thanks again--Jonathan
I tested my preamp switch using several different amps and preamps and it worked really nicely as I had hoped it might.
Thanks everyone for all of your input and suggestions. I truly appreciate it! I know that I'll soon have some questions about the amp switching module. I know I'm going to have to be really careful to ascertain that I don't run the risk of two amplifiers connecting simultaneously to one set of speakers. Rather a high-stakes proposition, right?
Thanks again--Jonathan
Thanks so much Mooly! Here's what I have so far. I described the capacitor coupling wrong in my last post and can't seem to edit it. Basically, the signal goes through the capacitor (not the relay) with a resistor to signal ground on the output side of the cap. --Jonathan
Attachments
May I know what are those gold power resistors used for, dummy loads for tube amplifiers?
What relay are you using to switch the audio signals?
What relay are you using to switch the audio signals?
Hi Chris. Yes, exactly! I have two sets of wire wound resistors, one for each amp under test. Without any voltage on the relays each amp is connected to its own set of 8 Ohm resistors. Both amps receive a signal continuously. When the control relay fires, only one amp at a time can be switched to the speaker outputs.
Is it true that SS amps don't need a load? I have been operating under the assumption that they do. Still, I have 3 tube amps so I need to have the dummy load either way.
About the relays: They are inexpensive Van Aimes relays from China, but from what I an tell they are made to a high standard. Silver alloy contacts and I don't think they are making relays for the hobby market. I hope I'm not wrong. BTW they are just a few bucks a piece.
Thanks -- Jonathan
Is it true that SS amps don't need a load? I have been operating under the assumption that they do. Still, I have 3 tube amps so I need to have the dummy load either way.
About the relays: They are inexpensive Van Aimes relays from China, but from what I an tell they are made to a high standard. Silver alloy contacts and I don't think they are making relays for the hobby market. I hope I'm not wrong. BTW they are just a few bucks a piece.
Thanks -- Jonathan
Thanks, what is the resistance of those power resistors?
And what is the 4pdt relay contact rating?
And what is the 4pdt relay contact rating?
That has to be blocking DC in which case it would have a desired effect.
They don't need a load and should be unconditionally stable without. This covers 99.99% of designs.
Valve amps are different because of the transformer coupling to the load (so the same would/could apply to solid state if it was an unusual transformer coupled design). With no load such a transformer could generate very high voltage transients if signal (and also power on and off transients) was present and that can cause breakdown of the insulation on the transformer windings.
@JKDenver Don't forget when you charge the coil in the relay and then open it up, it will create a high flyback voltage that will disrupt your ground. So, typically you put a capacitor and resistor between the coil lines of the relay to isolate and suppress this voltage locally before affecting the ground of the nearby components. If your solution placed the RC circuit on the pre-amp line, its very possible it would be even more effective to move that to the relay itself. Of course I don't know your circuit, but just a thought I had on the source of your pops.