Hi, so I'm having a puzzling problem that I've already got some help with on this forum, and it was suggested to open a thread here so lets see if anyone might some suggestions. 
(The problems started at post #323 here: https://www.diyaudio.com/forums/mul...own-speaker-scratch-discussion-thread-33.html )
Basically, I've finished my first build, speakers, amplifier and DSP and got problems when I began to adjust the DSP and measure the system.
In order of appearance;
1. PC with upgraded soundcard, and external soundcard for measuring (XTZ microphone + REW).
2. Wired to 4x RCA-In on 2X (one for left, one for right channel) 3E 1701 DSP (I'm using the I/O to switch so I get 2x source in).
3. Balanced connection to one 3E 3255 2ch amp, and one 3E 3251 4-ch amp.
4. Amplifiers are connected to Speak-On connectors (wires run straight to drivers from there).
5. Two Connex PSU, 28V 300W (3251). 55V 600W (3255).
6. DSP is powered by the auxiliary outlet from the amplifiers.
7. All is installed in one box, on one aluminium sheet, box is of non conducting material though as I don't have any grounded sockets in my house.
8. Controls (DSP) is only Volume and toggle switch for source select.
The problem I got was that the tweeter overheated during testing, the tweeter have a cap installed on the positive wire just behind it (10-15cm).
After doing some REW measuring, one driver at the time, with a SigmaStudio project scaled down for this, I went ahead with the 3-ch program.
Now the problems started, the tweeter level (dB) dropped and I only noticed this as the whole system (one speaker) sounded terrible. Measuring in REW showed how it rolled off at higher frequencies as well.
I tried to unplug the amplifier every now and then, and it helped to bring back clarity to the sound (it really didn't sound good otherwise from Any driver/amplifier, but nothing... that would set off alarm bells, it just sounded like a 'cheap stereo').
But every time I plugged in a driver with a cap in series it started to go bad again.
I have a set of cheap 2-way coaxial drivers for car that I use when I'm not sure about what's going to come out of the amplifier (to protect my drivers, a suggestion I got from another helpful individual here).
I also tried to switch the channel on the 3251 amplifier, to see if it was one that was bad, I also changed channels (D/A) out from the DSP, but to no help.
Using the same coaxial driver and changing between midrange and tweeter line (with cap) while music played, I would say the volume was half (tried different 3251-amp channels, but DSP "tweeter out", cables both with and without cap to the driver and the cap (at least) caused reduced dB (tried both caps I had)).
Besides using cap, the only thing I've changed since I did tests with the system and programming in Sigma for the first time, is the line-in as I used the supplied RCA connections to the DSP at that time (but changing between the different line-in now doesn't help, as I've tried both DSP).
I have cats in my house and they didn't react to loud noises (above my hearing range), the dB measuring was pretty low, and I still fried a tweeter with a 'safety-cap' on it.
Does anyone have any suggestions at all?
I have tried to think of every different way of connecting all the stuff, reprogramming DSP, used both PC internal and XTZ external soundcard as source, but I might have missed something that sets of this chain of events.
AllenB suggested (among other things) oscilloscope to find potential instability, I unfortunately sold mine several years ago.
(I will try to upload pictures if needed, they are on my phone, but some can be found as well in the thread I posted a link to)
(The problems started at post #323 here: https://www.diyaudio.com/forums/mul...own-speaker-scratch-discussion-thread-33.html )
Basically, I've finished my first build, speakers, amplifier and DSP and got problems when I began to adjust the DSP and measure the system.
In order of appearance;
1. PC with upgraded soundcard, and external soundcard for measuring (XTZ microphone + REW).
2. Wired to 4x RCA-In on 2X (one for left, one for right channel) 3E 1701 DSP (I'm using the I/O to switch so I get 2x source in).
3. Balanced connection to one 3E 3255 2ch amp, and one 3E 3251 4-ch amp.
4. Amplifiers are connected to Speak-On connectors (wires run straight to drivers from there).
5. Two Connex PSU, 28V 300W (3251). 55V 600W (3255).
6. DSP is powered by the auxiliary outlet from the amplifiers.
7. All is installed in one box, on one aluminium sheet, box is of non conducting material though as I don't have any grounded sockets in my house.
8. Controls (DSP) is only Volume and toggle switch for source select.
The problem I got was that the tweeter overheated during testing, the tweeter have a cap installed on the positive wire just behind it (10-15cm).
After doing some REW measuring, one driver at the time, with a SigmaStudio project scaled down for this, I went ahead with the 3-ch program.
Now the problems started, the tweeter level (dB) dropped and I only noticed this as the whole system (one speaker) sounded terrible. Measuring in REW showed how it rolled off at higher frequencies as well.
I tried to unplug the amplifier every now and then, and it helped to bring back clarity to the sound (it really didn't sound good otherwise from Any driver/amplifier, but nothing... that would set off alarm bells, it just sounded like a 'cheap stereo').
But every time I plugged in a driver with a cap in series it started to go bad again.
I have a set of cheap 2-way coaxial drivers for car that I use when I'm not sure about what's going to come out of the amplifier (to protect my drivers, a suggestion I got from another helpful individual here).
I also tried to switch the channel on the 3251 amplifier, to see if it was one that was bad, I also changed channels (D/A) out from the DSP, but to no help.
Using the same coaxial driver and changing between midrange and tweeter line (with cap) while music played, I would say the volume was half (tried different 3251-amp channels, but DSP "tweeter out", cables both with and without cap to the driver and the cap (at least) caused reduced dB (tried both caps I had)).
Besides using cap, the only thing I've changed since I did tests with the system and programming in Sigma for the first time, is the line-in as I used the supplied RCA connections to the DSP at that time (but changing between the different line-in now doesn't help, as I've tried both DSP).
I have cats in my house and they didn't react to loud noises (above my hearing range), the dB measuring was pretty low, and I still fried a tweeter with a 'safety-cap' on it.
Does anyone have any suggestions at all?
I have tried to think of every different way of connecting all the stuff, reprogramming DSP, used both PC internal and XTZ external soundcard as source, but I might have missed something that sets of this chain of events.
AllenB suggested (among other things) oscilloscope to find potential instability, I unfortunately sold mine several years ago.
(I will try to upload pictures if needed, they are on my phone, but some can be found as well in the thread I posted a link to)
Amplifier oscillating above 85 kHz, outside the feline auditory range? It has to be a pretty big oscillation, so it should be possible to measure that with a diode and a DC voltage meter. If you have a dirt cheap digital multimeter, it may use just a diode to rectify AC voltages when you set it to an AC voltage range.
Is there anything in the line that has a frequency in that range? The amplifier itself perhaps?
I do have a cheap multimeter, that's pretty much what I got right now. lol
And I actually did try to measure voltage yesterday, to see if there was a difference and how much I actually was "cranking", but something got in the way so I never finished. I was everywhere at that point, I knew I had to try as many things as possible so I could eliminate things. lol
I could dust off my old CD player, to remove my computer, as it's connected in a lot of different ways at the same time.
I'll put my cats in the trade section, maybe someone has a bat they no longer need.
I do have a cheap multimeter, that's pretty much what I got right now. lol
And I actually did try to measure voltage yesterday, to see if there was a difference and how much I actually was "cranking", but something got in the way so I never finished. I was everywhere at that point, I knew I had to try as many things as possible so I could eliminate things. lol
I could dust off my old CD player, to remove my computer, as it's connected in a lot of different ways at the same time.
I'll put my cats in the trade section, maybe someone has a bat they no longer need.
Yes, at least if it's not to interfere with the actual crossover in the DSP.
I'm using a 0.39uF Jantzen cap.
I'm going to try and be a bit more methodical in my testing here, I feel like I rushed things, but maybe I didn't.
I will use external audio source for listening, and PC for measuring only.
I'm thinking there's an instability when programming. There's an overload of the amplifier. And the tweeter that's bad might set bad things in motion as well.
So I need to sort out what's what, when I can't measure with oscilloscope.
I'm assuming I made one or several mistakes, but maybe I got a dud of an amp as well.
I'm still open for suggestions though!
I'm using a 0.39uF Jantzen cap.
I'm going to try and be a bit more methodical in my testing here, I feel like I rushed things, but maybe I didn't.
I will use external audio source for listening, and PC for measuring only.
I'm thinking there's an instability when programming. There's an overload of the amplifier. And the tweeter that's bad might set bad things in motion as well.
So I need to sort out what's what, when I can't measure with oscilloscope.
I'm assuming I made one or several mistakes, but maybe I got a dud of an amp as well.
I'm still open for suggestions though!
If you incorporate its effect in to your calculations, there are wide ranging possibilities.
Yes it's just for safety, I'm crossing the tweeter at around 2500hz in the crossover and 24dB filter, so it shouldn't have an impact on the response.
How many others are using one?
I don't have any safety futures otherwise, well a DC... what's it called... well, a blocking thing in the DSP. lol
I need to do some coherent work so I start memorizing things, it's been drawn-out for too long with too many other things on my mind.
How many others are using one?
I don't have any safety futures otherwise, well a DC... what's it called... well, a blocking thing in the DSP. lol
I need to do some coherent work so I start memorizing things, it's been drawn-out for too long with too many other things on my mind.
I thought I had sent the midrange signal to the tweeter, so I changed, but I was right at first. lol
There is something happening once in a while, high pitched sound that starts when you change settings in the DSP on the fly. As soon as you reboot it's back to normal for a while.
I also noticed some feedback between the microphone and speakers a couple of times.
And the 3251 has some noise, not much, but it wouldn't work for any speakers that's just in front of you.
To add insult to injury I noticed now that I hadn't stabilized the wood as good as I had thought. So I might have to correct the paint in the future as well.
It's a bit much right now... lol
There is something happening once in a while, high pitched sound that starts when you change settings in the DSP on the fly. As soon as you reboot it's back to normal for a while.
I also noticed some feedback between the microphone and speakers a couple of times.
And the 3251 has some noise, not much, but it wouldn't work for any speakers that's just in front of you.
To add insult to injury I noticed now that I hadn't stabilized the wood as good as I had thought. So I might have to correct the paint in the future as well.
It's a bit much right now... lol
I have noticed things like that in other electronics as well, and is sort of part of the game to reboot, as long as you don't fly a commercial airplane.
As there's so many things connected in several planes, if there's anything that is sensitive I can see how it could happen.
That's why I want to separate things, to eliminate both connections and signal source.
It could even be a problem with the larger amplifier, as the DSP I mainly use is powered by that one.
But it's not any small amount of power that the tweeter got, it has melted on the front plate! lol
I'm second guessing a lot of my work right now, but it should be solvable, I hope. Maybe start saving for two Hypex plates instead, but it's a lot of money.
As there's so many things connected in several planes, if there's anything that is sensitive I can see how it could happen.
That's why I want to separate things, to eliminate both connections and signal source.
It could even be a problem with the larger amplifier, as the DSP I mainly use is powered by that one.
But it's not any small amount of power that the tweeter got, it has melted on the front plate! lol
I'm second guessing a lot of my work right now, but it should be solvable, I hope. Maybe start saving for two Hypex plates instead, but it's a lot of money.
There were old methods developed before there were good oscilloscopes. As suggested earlier in the thread, if there is an oscillation then a signal diode might be used to rectify it into a signal that could be read by a voltmeter. Once there is some way to measure, even if sort of crude, at least its something. Then, for a oscillation to occur in the first place there has to be some coupling from something like an output back to an earlier point in the signal flow, with some gain from input to output in the circuit between those two points. The necessary coupling can happen due to layout problems or sometimes it can be coupled through the power supply (maybe including ground). The idea is to introduce changes into the possible coupling mechanisms and see if the measured output changes. Sometimes capacitance from a finger touching a part of the circuit may show a change in the measured output. The idea is the find a point in the circuit where the output is very sensitive to small changes in capacitance (or some other perturbation). That might be very close to were a trial fix could be applied.
I thank you for taking the time to explain this to me, I have to say it might be a little bit more than what I have knowledge for at this point.
I was interested as a kid with building circuits, but I gave up as my hands are anything but steady. lol
I decided to put everything together in hope of avoiding grounding problems, but maybe I made it worse.
I will look for some better equipment for measuring, even my multimeter is a bit well used to be honest, it's usually on garage duty.
I was interested as a kid with building circuits, but I gave up as my hands are anything but steady. lol
I decided to put everything together in hope of avoiding grounding problems, but maybe I made it worse.
I will look for some better equipment for measuring, even my multimeter is a bit well used to be honest, it's usually on garage duty.
I finally got some time over today to get back to it all, today I used an old CD player instead of PC for music.
I tried my cheap coaxial speakers on all 4 channels on the 3251 with both regular cable and a cable with the cap, it almost completely mutes the signal instantly.
I changed the programming to one without crossover and a single D/A out, just to play for a while. The channel on the amp that destroyed the tweeter have been playing music for an hour or two now through the coaxial driver, no problem.
I had a 50k pot for volume control previously, I soldered a new 10k to have the same as from the factory. I have changed between those two and the factory one, sometimes there's a change in volume (factory goes lower).
And honestly it's pretty loud when it's turned down all the way, so do people usually add internal volume control as well? Maybe one on each channel to adjust for sensitivity?
No bad noises when programming during playback either, I just need to reconnect it once in a while ("connect download").
So I guess it's safe to say that playing music from the PC while programming is one problem.
I really don't like it when problems goes away too easy, and it feels even worse to remove the cap in an attempt of Not destroying the other tweeter...
I tried my cheap coaxial speakers on all 4 channels on the 3251 with both regular cable and a cable with the cap, it almost completely mutes the signal instantly.
I changed the programming to one without crossover and a single D/A out, just to play for a while. The channel on the amp that destroyed the tweeter have been playing music for an hour or two now through the coaxial driver, no problem.
I had a 50k pot for volume control previously, I soldered a new 10k to have the same as from the factory. I have changed between those two and the factory one, sometimes there's a change in volume (factory goes lower).
And honestly it's pretty loud when it's turned down all the way, so do people usually add internal volume control as well? Maybe one on each channel to adjust for sensitivity?
No bad noises when programming during playback either, I just need to reconnect it once in a while ("connect download").
So I guess it's safe to say that playing music from the PC while programming is one problem.
I really don't like it when problems goes away too easy, and it feels even worse to remove the cap in an attempt of Not destroying the other tweeter...
I played some more music (CD), using the SB tweeter and midrange just laying on the floor and it sounded pretty good.
I removed the cap in the cabinet for the tweeter and re-installed the drivers yesterday evening. But it didn't sound good anymore, hooked up the coaxial drivers instead so I wouldn't hurt them.
It really sounds like bad reception on a radio, the volume drops slightly, the sound gets muddy and crunchy, the odd sparkle of statics.
Yesterday I could hear that the 3255 amp driving the woofer sounded the same, so it's not isolated to the 3251 amp.
I switched to the second DSP for maybe 10 seconds (don't want to push my luck even with the coaxials) and it didn't sound that good either.
That sort of eliminated the two direct powering paths; PSU-Amp-DSP (6 affected channels from 2 amplifiers and 2 DSP units, independently).
So, can someone point me in the direction of what diod to buy for use with the multimeter?
Anything to consider regarding the multimeter, like minimum specifications?
Is there any reason I could suspect something like EMI as a problem or part of it?
TV, TV box, my regular stereo, various cables are within a meter from it when I'm testing. Most of it is connected through the same socket in the wall as well.
Any suggestions is helpful right now, even a bad one. lol
I removed the cap in the cabinet for the tweeter and re-installed the drivers yesterday evening. But it didn't sound good anymore, hooked up the coaxial drivers instead so I wouldn't hurt them.
It really sounds like bad reception on a radio, the volume drops slightly, the sound gets muddy and crunchy, the odd sparkle of statics.
Yesterday I could hear that the 3255 amp driving the woofer sounded the same, so it's not isolated to the 3251 amp.
I switched to the second DSP for maybe 10 seconds (don't want to push my luck even with the coaxials) and it didn't sound that good either.
That sort of eliminated the two direct powering paths; PSU-Amp-DSP (6 affected channels from 2 amplifiers and 2 DSP units, independently).
So, can someone point me in the direction of what diod to buy for use with the multimeter?
Anything to consider regarding the multimeter, like minimum specifications?
Is there any reason I could suspect something like EMI as a problem or part of it?
TV, TV box, my regular stereo, various cables are within a meter from it when I'm testing. Most of it is connected through the same socket in the wall as well.
Any suggestions is helpful right now, even a bad one. lol
Some common signal diodes are 1n4148, 1n4448, 1n916, which are all very close to the same thing if not exactly the same (depending on manufacturer). You might put a small cap from the meter side of the diode to ground. Maybe .1uf or .01uf.
The meter may also show evidence of an audio signal. If a true oscillation is occurring, it may keep happening if you disconnect or turn off incoming audio. That may help to distinguish if the meter is showing audio or evidence oscillation.
Other things to look for might be odd changes in node voltages and or changes in temperature of components as the distortion symptom changes.
The meter may also show evidence of an audio signal. If a true oscillation is occurring, it may keep happening if you disconnect or turn off incoming audio. That may help to distinguish if the meter is showing audio or evidence oscillation.
Other things to look for might be odd changes in node voltages and or changes in temperature of components as the distortion symptom changes.
Thank you, a lot of good suggestions!
I can confirm that the 3251 doesn't get any way near the heat now compared to back when I did microphone measuring and fried the tweeter.
So energy was clearly not coming from evil spirits in the room.
I have noticed interference when I moved around cables a couple of times, but I had placed everything very poorly on the floor, low level in, USB programmer, 230V, speaker cables, pretty much everything crossing over each other.
So it's not idiot-proof. lol
And when I connect the programmer to the DSP it can send out a crack to the speakers loud enough to make you check your trousers.
I do need to play music to hear anything (now, as I don't connect the PC with anything but USB programmer, it clearly didn't like that I use it for playback at the same time).
I should probably get some shielded cables for the input, or possibly shield them with something.
I tried to go over the use of a cap, but I must admit that I'm lost. lol
Is it on the rectified side to... chassi ground..?
I need a nap, followed by some strong coffee. lol
I can confirm that the 3251 doesn't get any way near the heat now compared to back when I did microphone measuring and fried the tweeter.
So energy was clearly not coming from evil spirits in the room.
I have noticed interference when I moved around cables a couple of times, but I had placed everything very poorly on the floor, low level in, USB programmer, 230V, speaker cables, pretty much everything crossing over each other.
So it's not idiot-proof. lol
And when I connect the programmer to the DSP it can send out a crack to the speakers loud enough to make you check your trousers.
I do need to play music to hear anything (now, as I don't connect the PC with anything but USB programmer, it clearly didn't like that I use it for playback at the same time).
I should probably get some shielded cables for the input, or possibly shield them with something.
I tried to go over the use of a cap, but I must admit that I'm lost. lol
Is it on the rectified side to... chassi ground..?
I need a nap, followed by some strong coffee. lol