OK, huge turn on thump due to active crossovers being single rail power and 10uF coupling caps. Amp does not have a muting relay. I know the obvious solution is to add a relay either to the amp output or preamp output.
But, is it a reasonable approach to reduce the issue to start with by carefully selecting the output coupling cap?
For instance, my sub plate is 20K input. So rather than the 10uF electrolytic, I could drop in a 1uF giving a 8 Hz first order filter and 1/10th the DC energy stored. Same token on my main amp, 33K in, so a .22uF gives me a 20 Hz roll off for way less DC charge. ( Plus removing electrolytics from the path which is trivial but basically good)
Somewhere; Self, Cordell, Didden, Pass, someone much smarter than I had a reason not do do so, but I can't find the reference.
But, is it a reasonable approach to reduce the issue to start with by carefully selecting the output coupling cap?
For instance, my sub plate is 20K input. So rather than the 10uF electrolytic, I could drop in a 1uF giving a 8 Hz first order filter and 1/10th the DC energy stored. Same token on my main amp, 33K in, so a .22uF gives me a 20 Hz roll off for way less DC charge. ( Plus removing electrolytics from the path which is trivial but basically good)
Somewhere; Self, Cordell, Didden, Pass, someone much smarter than I had a reason not do do so, but I can't find the reference.
The thump will always be there, by reducing the capacitor you just change it's duration. I looked into this as I ran active crossovers with tweeters direct to the amp (although my crossovers are split rail so negligible thump). In the end I came to the conclusion the only way is to add a muting relay to the amp.
It sounds like you are only running subs and fullrange though? In that case the thump may not be a big problem. You can change capacitors as you suggest but whilst that may redeuce the audibility of it, the cones will still visibly move.
It sounds like you are only running subs and fullrange though? In that case the thump may not be a big problem. You can change capacitors as you suggest but whilst that may redeuce the audibility of it, the cones will still visibly move.
Last edited:
Yes, sub and full range. Thump is a huge problem. It was not as bad with the cheap Rolls crossover, just annoying. Now it is dangerous.
Correct, looking to reduce it at this stage knowing it won't eliminate it. I already swapped the input caps to .47 as a subsonic filter so it is only the outputs with as large a charge. I guess a relay, as I am loath to add a FET or other distortion generating device.
Correct, looking to reduce it at this stage knowing it won't eliminate it. I already swapped the input caps to .47 as a subsonic filter so it is only the outputs with as large a charge. I guess a relay, as I am loath to add a FET or other distortion generating device.
It is not a case of bigger caps holding more charge therefore more thump. The thump is caused by the amount of voltage offset the caps are blocking. Smaller caps just charge faster and the thump is shorter but its severity (amplitude) is the same.
As you just want muting not DC protection your relay doesnt have to be as tough, as long as you always wait for the click before you play music.
As you just want muting not DC protection your relay doesnt have to be as tough, as long as you always wait for the click before you play music.
Ah. OK. And to boot, the caps are tiny surface mount ceramics, so a bit hard for the old eyes. Now to find some really nice dry contact low voltage relays. I remember back in the the lab days, there was a list of catch-22 issues with relay contacts a mile long! I think old data relays, bifurcated gold contacts is what is needed here.
Gold is not suitable for high currents such as amp output. It melts off. Buying a speaker disconnect relay from a PA amp manufacturer such as crown or peavey gets you a suitable one. Buying a burnt up amp of that heritage is probably cheaper. Cs800x has a nice one. If you're buying from distributors you want silver plate, but specifying contact material went the way of the dodo bird when all relay manufacturing moved to asia. Gold contacts are for currents under 50 ma. And relay contact switching on inputs to gain sections makes its own thump when the contacts are open. I tried it 40 years ago switching inputs after a second to a TI DFSII data acquisition system.
There are other methods than relays. download the schematic to PV-1.3k from eserviceinfo. Look at page 1 sector D5. Q101 along with R105 & C107 make a thump supressor that crams the input voltage up to the +voltage rail at start then gradually lets go. The additional input from U101 is a dual duty function that also crams down the input voltage if too much high frequency is detected - the DDT function. You don't need that part of it.
You can't buy J174 pfets any more (except from peavey) but you can cram the voltage down to the negative rail with a n channel jfet. With the same relaxation timer.
There are other methods than relays. download the schematic to PV-1.3k from eserviceinfo. Look at page 1 sector D5. Q101 along with R105 & C107 make a thump supressor that crams the input voltage up to the +voltage rail at start then gradually lets go. The additional input from U101 is a dual duty function that also crams down the input voltage if too much high frequency is detected - the DDT function. You don't need that part of it.
You can't buy J174 pfets any more (except from peavey) but you can cram the voltage down to the negative rail with a n channel jfet. With the same relaxation timer.
Last edited:
Thinking preamp (crossover) output as it is the guilty party, not the power amp output. Looking best practice, I need to add 100 Ohms on the outputs and 100K or so drain to ground. Then I can short the output.
Yea, gold does not take arcing. I forget what alloy was what we settled on back in the lab. It had to take arcing on EPO ( 90V, 45A) but low resistance at mv for sensing.
Running to a rail is an interesting concept though puts non-linear parts on the signal path. Same problem with the various FET and BJT soft mute circuits.
Yea, gold does not take arcing. I forget what alloy was what we settled on back in the lab. It had to take arcing on EPO ( 90V, 45A) but low resistance at mv for sensing.
Running to a rail is an interesting concept though puts non-linear parts on the signal path. Same problem with the various FET and BJT soft mute circuits.
Hmmm. I wonder about the "lightspeed" devices. NSL-32SR etc. Control it with a simple ramp. It would be in series but some claim it to be most excellent.
[quote Running to a rail is an interesting concept though puts non-linear parts on the signal path. Same problem with the various FET and BJT soft mute circuits./quote]
There are no active devices that are not non-linear except about 5 volts output of a 250 v plate triode (12AX7) or the curve of an EF86 tube. Bipolar junction transistors are exponential reponse devices. I suppose your plate amp is built out of EF86's which are running about $30 each these days? Mosfets are more linear than BJT's but have a huge turnon knee of about 5 v and capacitive switching whomp that must be overcome by design features so the device doesn't turn on or off during music. The non-linearity of transistors etc is overcome in most amplifiers since 1956 by negative feedback. Witness the op amp, a device with feedback approaching 100%.
The junction fet in the Peavey circuit acts as a >100 kohm resistor parallel from the input to power supply rail, when C107 is charged up. When the jfet is leaky and non-linear is the turnon period, when you don't want to hear anything, anyway. There is nothing as non-linear as a copper brass or silver relay contact covered by oxide trying to conduct a few microvolts when it first closes. Gold is nice and linear but it melts out.
There are no active devices that are not non-linear except about 5 volts output of a 250 v plate triode (12AX7) or the curve of an EF86 tube. Bipolar junction transistors are exponential reponse devices. I suppose your plate amp is built out of EF86's which are running about $30 each these days? Mosfets are more linear than BJT's but have a huge turnon knee of about 5 v and capacitive switching whomp that must be overcome by design features so the device doesn't turn on or off during music. The non-linearity of transistors etc is overcome in most amplifiers since 1956 by negative feedback. Witness the op amp, a device with feedback approaching 100%.
The junction fet in the Peavey circuit acts as a >100 kohm resistor parallel from the input to power supply rail, when C107 is charged up. When the jfet is leaky and non-linear is the turnon period, when you don't want to hear anything, anyway. There is nothing as non-linear as a copper brass or silver relay contact covered by oxide trying to conduct a few microvolts when it first closes. Gold is nice and linear but it melts out.
Last edited:
I triamp and have a little controller that powers the crossovers/preamp, waits 5 sec, powers the amps. The tweeter gets a relay that closes 5 secs after amp power on, everything else is direct coupled to the amps. Small thump but barely noticeable. These are all dual rail amps though.
Here's a relay muting circuit I built for a Korg Nutube preamp, which runs off a single 24V supply. The original concept was based on a circuit Nelson Pass posted, and then I went through a couple iterations of embellishments to try to get the timing right to completely eliminate the turn on/off thump for that particular circuit.
I've also used it on an opamp-based preamp, which had a 10uF output cap, 100K to ground, and 100 ohms in series, which I think is the same as you are proposing. In that case, my power supply had enough voltage on the unregulated side that I could add a 24V regulator for the muting relay.
The design is adaptable to other supply voltages, and you could eliminate Q2 and associated circuitry if you aren't concerned about turn-off muting. Just be prepared to spend some time tweaking component values to account for different supply voltages and timing requirements.
I've also used it on an opamp-based preamp, which had a 10uF output cap, 100K to ground, and 100 ohms in series, which I think is the same as you are proposing. In that case, my power supply had enough voltage on the unregulated side that I could add a 24V regulator for the muting relay.
The design is adaptable to other supply voltages, and you could eliminate Q2 and associated circuitry if you aren't concerned about turn-off muting. Just be prepared to spend some time tweaking component values to account for different supply voltages and timing requirements.
Those LED+cds cell "linear optocouplers" are quite effective on high impedance circuits (>600 ohms) in series with a signal. The ramp has to be slow enough to not be audible. Same RC chargeup ramp as the Peavey jfet circuit. The jfet is cheaper.
All solid state. Remember, silver oxide has the same conductivity as silver which is why we use it for RF connectors. Copper/nickel/tin/lead etc. all have issues unless sufficient wiping action. Some have suggested mercury wetted. It was 30 years ago, Rhodium maybe? Darn, I used to have some old T-Bar relays. You could use 20 or so contacts in parallel.
With a PAS2, I would think you would have a pair of MK IV's! ( Or MV50, I preferred) For the price of a car, a pair of Cary OTL were about as good as tubes get.
Besides the high failure rate of FET muting, due to the RCA connectors more than the FET, a great many designers whom I resect hold them in contempt.
Found more specs on the "lightspeed" Maybe better than a Radio Shack POT, but not exactly low distortion in the range I am talking. .00-something cumulative end to end.
More research to follow. One of those subjects that does not seem to have an SOP.
With a PAS2, I would think you would have a pair of MK IV's! ( Or MV50, I preferred) For the price of a car, a pair of Cary OTL were about as good as tubes get.
Besides the high failure rate of FET muting, due to the RCA connectors more than the FET, a great many designers whom I resect hold them in contempt.
Found more specs on the "lightspeed" Maybe better than a Radio Shack POT, but not exactly low distortion in the range I am talking. .00-something cumulative end to end.
More research to follow. One of those subjects that does not seem to have an SOP.
Looks workable. Goal is slow turn on, fast turn off. As I have a SMPS feeding a cap bank, I could isolate the bank with a diode, so the 24V supply would drop quicker than the linear half.
If I am just shorting the signal to ground ( adding 100 Ohms or so to protect the Op Amp) then the quality of the relay contacts does not matter much as they are not in the signal path.
Two ideas:
1. Use one of those rest-room-blower-delays set to minimum time to power the amp after the preamp/crossover. Hidden in a wall-wart-like housing, those with plug and socket on opposite sides, so it does not look improvised.
2. Soft-start the crossover so slowly that the dc blocker comes up slower than the amp can follow. (It gets corrected by the feedback loop like a DC drift.) This may well take 10-20 seconds though, and is not easy if the crossover is not your own design.
1. Use one of those rest-room-blower-delays set to minimum time to power the amp after the preamp/crossover. Hidden in a wall-wart-like housing, those with plug and socket on opposite sides, so it does not look improvised.
2. Soft-start the crossover so slowly that the dc blocker comes up slower than the amp can follow. (It gets corrected by the feedback loop like a DC drift.) This may well take 10-20 seconds though, and is not easy if the crossover is not your own design.
As your preamp/crossover is single rail you could do the shorting with a mosfet. This would be easy to implement by a simple CR network on the gate.
Last edited:
Worked up from parts in the scrap bin. Spice says probably OK. Slow on, fast off.
So I just add a 100 Ohm on the crossover outputs, add a 100K bleeder to ground.
I found a 24V 4PDT "communications" relay.
I have some 2n5551's left over, and a 5.1V zener. Breadboard time.
Time resisters will be a 50K pot.
So I just add a 100 Ohm on the crossover outputs, add a 100K bleeder to ground.
I found a 24V 4PDT "communications" relay.
I have some 2n5551's left over, and a 5.1V zener. Breadboard time.
Time resisters will be a 50K pot.
Attachments
If a "telephone" relay, it may very well have rhodium plate contacts.
I had 240 of 24 v 4pdt telephone relays from electronicsurplus.com to program stops for a hammond organ, that were stolen last month by the burglar. Probably already in the scrap copper pile somewhere. I have now zero parts & tools just in time for electronics (rainy) season.
I had 240 of 24 v 4pdt telephone relays from electronicsurplus.com to program stops for a hammond organ, that were stolen last month by the burglar. Probably already in the scrap copper pile somewhere. I have now zero parts & tools just in time for electronics (rainy) season.