READ ME! Why am I building this? Because I can and want to, mostly. Also because I don't entirely trust the DSP solution for speaker crossovers. Active analog filters are set and can't lose their programming, barring a catastrophic (and extremely rare) failure. You might say that failure for DSP is extremely rare, but it's already happened to me. After a power outage, my miniDSP 4x10 lost it's configuration settings and luckily I checked before played anything. Having the full signal at high volume going to my expensive tweeters would not be a great start to a relaxing evening of listening. The amp design is relatively simple, a typical three stage configuration with good quality parts. I tested the prototype for operational stability and thermal stability and saw no issues. Not shown in this video (but I'll do it in an upcoming one) the stability testing included adding capacitance to the output. I used a 1uf cap and the amp did not break out into oscillation. Square wave testing shows a fast enough slew rate and there will be more testing when I get the professionally made boards to do the full power tests. Note that this is an integrated amp that has 28db of gain and doesn't require a preamp. The case I built is a little rough and not even close to finished. But it is HUGE and has plenty of space for the ten channels I want to put inside. Yes, I've settled on 10 channels - 8 for the main speakers and two to spare for a possible centre channel and the other channel can run the bass shaker in my seat. I'll have to rework or remake the back panel and make a cool looking front and top panel, plus install some dividers inside. The transformers were salvaged from vintage power amps and are easily powerful enough to run all 10 channels at good listening levels with headroom to spare. Each will have its own large capacitor bank for smoothing and reserve, with a soft start circuit on each. Conveniently, the larger transformer also has a low current 22vac winding that's perfect for the active crossovers supply.
I don't hear slew rate and square wave testing mentioned much these days. Back in the day a quality amp would have these in the specs. Looking forward to this build very much.
How are you going to delay the midrange and tweeters compared to the subs? I have a huge issue with this myself. Bass is soft because the subs is delayed vs the mid&tweet that doesnt have the dsp delay the subs have.
This is my biggest fear with DSP. Like you I’m running three 4-way active speakers up front. Though they use 3” beryllium domes for the mids and 1” beryllium for the tweeters. The replacement diaphragms for the mids , if you are lucky enough to find them, are $1200-$1500 each. No way I was taking a chance, so all of the berylliums are on an active channel, but are crossed with a passive circuit so they are protected. The midbasses and woofer are fully active. But, yeah, I don’t trust computers lol. I’m glad you said you’ll make the gerbers available (might go pcbway instead of etching) as I might clone this project. I’d love to see a video on you building the active crossovers. I’d need to figure out how to make mine with their own crossover points and slopes and such. Awesome video, exactly what I was hoping for. I know some consider it boring, but I hope to get some footage of you soldering. I solder all day for a living (audio repair and restorations) so you think I’d get tired of it lol. Oh, and brilliant solution on drilling the holes! I ended up purchasing the small attachment that houses the dremel basically turning it into a tiny drill press. Your solution was a much simpler setup.
mind blown... watching you make the board, I never knew this is the way. I don't have any intelligent words, just in awe and really interested in learning, back to the video... had to pause to share. wow....
4:00 The Dremel Tool drill press is brilliant. That has a million uses that are all jumping into my brain at once. I can't believe John Heisz thought of it first instead of me. Oh well, can't win them all.
I took an old Denon 6 channel AVR and put an active 24db/octave crossover in it. Decent MOSFETs, okay Burr-Brown DACs, and a very good preamp. Dirt cheap, 75Wx3 / channel. Looks like a Denon AVR. Sounds sooo much better than a passive crossover.
Thank you for the video, I really enjoy this channel. I second the problems with breakout and prototyping boards. I am working on a sketch to run a stepper motor for my router lift and could not understand why it would not work. It turned out to be a flaky row on my breadboard. It took me hours to track it down as it was intermittent. Looking forward to the next video.
I needed a 16+ channel amplifier for a bespoke project. There are some available on market - they were either too expensive or poorly spec'ed. So I purchased a rack case, a couple of big switch mode supplies and a pile of amplifier modules (TPA3116) and have been amazed at the performance. Plus it gets all the custom features I wanted - analogue current meter, external solar power option and phoenix input harness...
Great video. I've done toner transfer pcbs before but I used ferric chloride. Muriatic acid looks sooo much cleaner to work with. Building a 12 channel, scratch built amp is a worthy and bonkers project. I hope that it continues to be worth filming for you. The anecdote about the glitchy dsp is super pertinent since I am in the early planning stages for my own bonkers audio project. Thanks again.
How many watts is the amplifier? Why do the runaway diodes turn each way on the diagram? Isn't it DC that runs through the bias circuit? And what good do D1, D2 and R3 do? Don't they just create a false frame point? Thanks for many great videos!
In the end is this for a 2.1 setup? For some reason I thought you were building a surround sound amp. That's something I'd like to do, but I can't figure out how to get all the different Dolby modes. Got any suggestions?
1:49 If you throw the PCB into water close to boiling point and wait for 5 minutes, the paper will separate on its own. Then you won't have to worry about cleaning up paper residue. The method I mentioned will not work if the water is not hot enough.
They say a 3D printing can make a laser-iron technology outdated. They use a one layer printing, a proper nozzle, and some other settings. A result is remarkable. If you have a 3D printer you could try, sir.
I've thought about it and also using the CNC, even just to draw the traces. But I don't make enough boards to justify the time it would take to perfect the process, and using this method I can have a board ready to populate in less than an hour.
@@IBuildIt Have tried milling my boards : first build a small CNC myself ... bought a small CNC ... converted my old 3D printer ... Well, the good old iron/etching beats that easily. I am probably gonna give the the 3D LCD printer a try which looks very promising, but might be a bit of a hassle for double sided boards. Then again I only use that for prototyping, for the final board I use JLCPCB (gonna try PCBWAY for my next project) because it's even cheaper per board to get them there compared to the cost of a copper clad board. By the way, even for prototyping-boards it's cheaper to get those made by JLCPCB (yup, they will drill those thousand holes for no extra cost)
@@IBuildIt Yep. It was about $20 minimal few years agom if I'm not mistaken. Plus a quite high fee for a number of different designs on a same PCB (and still is quite high) Now I see it $5, and this is like "why a h.. not to order?!"
Great work, the output inductor is connected in parallel with a res , maybe r16? i can;t see it on your pcb , and in the schematic there is not any inductor, it's good to have the inductor in parallel with a 5-10 ohm resistor, it makes the inductor " behave nicer " , land ike one thing that I recommend, ( But you might want to test it before, it might be good, if your amp is stable , temperature wise, and won't go into thermal runaway or the quissance current will drop to much when it gets hot. instead of the two diodes and the pot to set the Bias , I would use a " VBE multiplier ", its more stable controlling the quiss current ( bias ) vs temp and all that stuff. Works fine with diodes to, depends on the diodes. You could use a " VBE multiplier " and leave the diodes in place as well ( only the two diodes without the pot) in case the transistor fails ( very unlikely, ) it serves as a protection ,and the output stage won't go into smoke.
Is this by any chance heading for a Linkwitz type speaker project ? When one hears mention of miniDSP and 8 to 10 amplifiers the mind heads that way. Love your stuff....Take Care.
Hi John, I really enjoy your audio stuff videos, since building speakers was one of my entries to woodworking. With this video I wonder why you etch your PCBs instead of carving them on your CNC router.
It would take a lot of trial and error to get the CNC to even come close to the quality that I can get very quickly by etching. I can make a board ready to populate in less than an hour.
Nice project John. Are you going to do on board supply and regulation? If you are running all those amps off of one transformer it might be an option to think about. Mark
Thanks Mark. There's not much to be gained by either regulating or putting the supplies right on the amp boards. It needlessly complicates an already very complex project and while i have a big case, it's still going to be a tight fit getting everything I need inside. The supply for the active crossovers will be regulated, of course.
@John Heisz - Speakers and Audio Projects You going with two big capacitor banks and leterrip! Your smaller demand amps could do with some capacitance on board, save any voltage modulation if there's current demand from the woofer amps.
The way to look at this is that each of these transformers came from power amps that would very competently drive two 1970's era speakers to absurd volume levels. Now, each will be powering a number of separate amps, but only two speakers in total. In other words, same but different. Each board has decoupling caps on the rails already and I don't anticipate any serious power supply sag.
I'm currently using 14 channels, but spread out over 4 different amps, so this takes the place of two of those, plus the DSP plus a snakes nest of cables.
Use on the market cheap and good tpa3255 chips or better yet just buy a bunch a fosi v3 amps and jerry rig them together with plans for ppl to replicate.
Use acetone, not lacquer thinner. I'm a professional painter, and different lacquer thinners can have oily residue. They are a combination of many petroleum solvents, and acetone is better because it is just 1, and is water and oil soluable!!!
READ ME!
Why am I building this?
Because I can and want to, mostly. Also because I don't entirely trust the DSP solution for speaker crossovers. Active analog filters are set and can't lose their programming, barring a catastrophic (and extremely rare) failure. You might say that failure for DSP is extremely rare, but it's already happened to me. After a power outage, my miniDSP 4x10 lost it's configuration settings and luckily I checked before played anything. Having the full signal at high volume going to my expensive tweeters would not be a great start to a relaxing evening of listening.
The amp design is relatively simple, a typical three stage configuration with good quality parts. I tested the prototype for operational stability and thermal stability and saw no issues. Not shown in this video (but I'll do it in an upcoming one) the stability testing included adding capacitance to the output. I used a 1uf cap and the amp did not break out into oscillation.
Square wave testing shows a fast enough slew rate and there will be more testing when I get the professionally made boards to do the full power tests.
Note that this is an integrated amp that has 28db of gain and doesn't require a preamp.
The case I built is a little rough and not even close to finished. But it is HUGE and has plenty of space for the ten channels I want to put inside. Yes, I've settled on 10 channels - 8 for the main speakers and two to spare for a possible centre channel and the other channel can run the bass shaker in my seat. I'll have to rework or remake the back panel and make a cool looking front and top panel, plus install some dividers inside.
The transformers were salvaged from vintage power amps and are easily powerful enough to run all 10 channels at good listening levels with headroom to spare. Each will have its own large capacitor bank for smoothing and reserve, with a soft start circuit on each.
Conveniently, the larger transformer also has a low current 22vac winding that's perfect for the active crossovers supply.
I don't hear slew rate and square wave testing mentioned much these days. Back in the day a quality amp would have these in the specs. Looking forward to this build very much.
How are you going to delay the midrange and tweeters compared to the subs? I have a huge issue with this myself. Bass is soft because the subs is delayed vs the mid&tweet that doesnt have the dsp delay the subs have.
I stopped using delay on the DSP after making a listening comparison. In my opinion it's not needed in a small room.
@@impuls60 You can physically delay the components.
Try building tda 7294 bridge amplifier. Very simple to build. Thanks for video
I won't be building an amp of any kind but I enjoy seeing you tackle projects. You go John!
This is my biggest fear with DSP. Like you I’m running three 4-way active speakers up front. Though they use 3” beryllium domes for the mids and 1” beryllium for the tweeters. The replacement diaphragms for the mids , if you are lucky enough to find them, are $1200-$1500 each. No way I was taking a chance, so all of the berylliums are on an active channel, but are crossed with a passive circuit so they are protected. The midbasses and woofer are fully active. But, yeah, I don’t trust computers lol. I’m glad you said you’ll make the gerbers available (might go pcbway instead of etching) as I might clone this project.
I’d love to see a video on you building the active crossovers. I’d need to figure out how to make mine with their own crossover points and slopes and such. Awesome video, exactly what I was hoping for. I know some consider it boring, but I hope to get some footage of you soldering. I solder all day for a living (audio repair and restorations) so you think I’d get tired of it lol. Oh, and brilliant solution on drilling the holes! I ended up purchasing the small attachment that houses the dremel basically turning it into a tiny drill press. Your solution was a much simpler setup.
mind blown... watching you make the board, I never knew this is the way. I don't have any intelligent words, just in awe and really interested in learning, back to the video... had to pause to share. wow....
“I love it when a plan comes together” -John “Hannibal” Smith
4:00 The Dremel Tool drill press is brilliant. That has a million uses that are all jumping into my brain at once. I can't believe John Heisz thought of it first instead of me. Oh well, can't win them all.
I took an old Denon 6 channel AVR and put an active 24db/octave crossover in it. Decent MOSFETs, okay Burr-Brown DACs, and a very good preamp. Dirt cheap, 75Wx3 / channel. Looks like a Denon AVR. Sounds sooo much better than a passive crossover.
Great video. Thanks John. Greetings from the UK!
Great video. I look forward to seeing more of this project.
Thank you for the video, I really enjoy this channel. I second the problems with breakout and prototyping boards. I am working on a sketch to run a stepper motor for my router lift and could not understand why it would not work. It turned out to be a flaky row on my breadboard. It took me hours to track it down as it was intermittent. Looking forward to the next video.
Thanks Paul :)
I needed a 16+ channel amplifier for a bespoke project. There are some available on market - they were either too expensive or poorly spec'ed. So I purchased a rack case, a couple of big switch mode supplies and a pile of amplifier modules (TPA3116) and have been amazed at the performance. Plus it gets all the custom features I wanted - analogue current meter, external solar power option and phoenix input harness...
Great video. I've done toner transfer pcbs before but I used ferric chloride. Muriatic acid looks sooo much cleaner to work with. Building a 12 channel, scratch built amp is a worthy and bonkers project. I hope that it continues to be worth filming for you. The anecdote about the glitchy dsp is super pertinent since I am in the early planning stages for my own bonkers audio project. Thanks again.
How many watts is the amplifier?
Why do the runaway diodes turn each way on the diagram? Isn't it DC that runs through the bias circuit?
And what good do D1, D2 and R3 do? Don't they just create a false frame point?
Thanks for many great videos!
Can't imagine doing this but the drill press was genius!
Great post. Thanks for sharing. Reminds me of my days in high school shop making PCBs.
That's awesome... Keep it up brother
Great tutorial for using normal things for high tech use! Thanks
I love your drill press, can't wait to see how many people ask for plans!! LOL
In the end is this for a 2.1 setup? For some reason I thought you were building a surround sound amp. That's something I'd like to do, but I can't figure out how to get all the different Dolby modes. Got any suggestions?
1:49 If you throw the PCB into water close to boiling point and wait for 5 minutes, the paper will separate on its own. Then you won't have to worry about cleaning up paper residue. The method I mentioned will not work if the water is not hot enough.
Genius! I'll never be as smart as John.
Holy smokes, that is going to be awesome.
I'm disappointed you used a store bought oscilloscope... Cop out! 😂 You do amazing things. Keep it up
They say a 3D printing can make a laser-iron technology outdated. They use a one layer printing, a proper nozzle, and some other settings. A result is remarkable. If you have a 3D printer you could try, sir.
I've thought about it and also using the CNC, even just to draw the traces. But I don't make enough boards to justify the time it would take to perfect the process, and using this method I can have a board ready to populate in less than an hour.
@@IBuildIt Have tried milling my boards : first build a small CNC myself ... bought a small CNC ... converted my old 3D printer ...
Well, the good old iron/etching beats that easily.
I am probably gonna give the the 3D LCD printer a try which looks very promising, but might be a bit of a hassle for double sided boards.
Then again I only use that for prototyping, for the final board I use JLCPCB (gonna try PCBWAY for my next project) because it's even cheaper per board to get them there compared to the cost of a copper clad board.
By the way, even for prototyping-boards it's cheaper to get those made by JLCPCB (yup, they will drill those thousand holes for no extra cost)
Yes, I was just setting up to order the ones I need and it's actually insane how low the price is. They must be producing it for next to nothing.
@@IBuildIt Yep. It was about $20 minimal few years agom if I'm not mistaken. Plus a quite high fee for a number of different designs on a same PCB (and still is quite high) Now I see it $5, and this is like "why a h.. not to order?!"
Great work, the output inductor is connected in parallel with a res , maybe r16? i can;t see it on your pcb , and in the schematic there is not any inductor, it's good to have the inductor in parallel with a 5-10 ohm resistor, it makes the inductor " behave nicer " , land ike one thing that I recommend, ( But you might want to test it before, it might be good, if your amp is stable , temperature wise, and won't go into thermal runaway or the quissance current will drop to much when it gets hot. instead of the two diodes and the pot to set the Bias , I would use a " VBE multiplier ", its more stable controlling the quiss current ( bias ) vs temp and all that stuff. Works fine with diodes to, depends on the diodes. You could use a " VBE multiplier " and leave the diodes in place as well ( only the two diodes without the pot) in case the transistor fails ( very unlikely, ) it serves as a protection ,and the output stage won't go into smoke.
I'm using Onsemi thermal track output transistors - the diodes are pins 4 and 5 on those devices, not discrete ones.
@@IBuildIt Wow I look dumb now, I didn't really look what output transistors you used. Great then !.
Do you know that MC12 can play a clip of WAV music through your design so you could listen to it. Few knows this trick
Is this by any chance heading for a Linkwitz type speaker project ? When one hears mention of miniDSP and 8 to 10 amplifiers the mind heads that way.
Love your stuff....Take Care.
Hi John, I really enjoy your audio stuff videos, since building speakers was one of my entries to woodworking. With this video I wonder why you etch your PCBs instead of carving them on your CNC router.
It would take a lot of trial and error to get the CNC to even come close to the quality that I can get very quickly by etching. I can make a board ready to populate in less than an hour.
But can it drive a 1Ω load at 2Hz like an Audio Research monoblock? ;) After all you never know when you might want to arc weld with your power amp.
Nice project John. Are you going to do on board supply and regulation? If you are running all those amps off of one transformer it might be an option to think about.
Mark
Thanks Mark.
There's not much to be gained by either regulating or putting the supplies right on the amp boards. It needlessly complicates an already very complex project and while i have a big case, it's still going to be a tight fit getting everything I need inside.
The supply for the active crossovers will be regulated, of course.
@John Heisz - Speakers and Audio Projects You going with two big capacitor banks and leterrip! Your smaller demand amps could do with some capacitance on board, save any voltage modulation if there's current demand from the woofer amps.
The way to look at this is that each of these transformers came from power amps that would very competently drive two 1970's era speakers to absurd volume levels.
Now, each will be powering a number of separate amps, but only two speakers in total. In other words, same but different.
Each board has decoupling caps on the rails already and I don't anticipate any serious power supply sag.
Nice chassis.
John why would you want a 10/12 channel amplifier?
I'm currently using 14 channels, but spread out over 4 different amps, so this takes the place of two of those, plus the DSP plus a snakes nest of cables.
I would be worried about earth loops / humm....
When you are done put done the toilet. It will clean toilet and kill roots in drain at same time.
большой проект. успехов!
Normally want to sand your work after you clean it, or you can sand muck into the pease
Use on the market cheap and good tpa3255 chips or better yet just buy a bunch a fosi v3 amps and jerry rig them together with plans for ppl to replicate.
Use acetone, not lacquer thinner. I'm a professional painter, and different lacquer thinners can have oily residue. They are a combination of many petroleum solvents, and acetone is better because it is just 1, and is water and oil soluable!!!