Since I'm using two different transformers, I need two separate power supplies. These supplies have a soft start circuit built in and in this video I go over how that works. In a nut shell, the thermistor is all that's needed for the soft start to work. But there are a couple of good reasons why it's better if you bypass it after it has done its job. First is that if you bypass it, it will cool down and last longer. Second is that when the amplifiers are not drawing enough current, the thermistor will cool down anyway, and as it does its resistance goes up, which lowers the supply voltage. When you short it with a relay, the relay carries all of the current without heating up. But there has to be a short delay before the relay switches on so that the thermistor has time to work. I used the switching circuit from Rod Elliott's website to do that, and he gives a more detailed description of how it works: sound-au.com/project39.htm To power the soft start bypass I dropped the 40VAC from the power transformer down to less than 10VAC with a resistor and a pair of 1uf capacitors. That lower AC voltage is then rectified to power the circuit and switch the 12VDC relay. This has the advantage of not needing a small transformer to run the soft start. Also it doesn't heat up like resistors would to drop the rail DC down to a usable voltage. On the board as well is what's known as a ground lift. It consists of a 10 ohm power resistor and two diodes wired back to back. The lift "separates" the supply ground from the chassis (earth) ground, but still provides a physical connection between the two. This can help greatly with ground loop problems that cause hum. In the video I also show the power distribution boards that feed the individual amp boards. These reduce the wire runs inside the case and also put the supply capacitors very close to the amp boards, which should improve (very, very slightly) transient response. They are also a convenient place for the rail fuses for each amp board. Since I have two supplies and 10 amplifiers, one supply will power 6 amp boards (these will drive the tweeters, midranges and midwoofers in my 4-way speakers) and the second will power the other 4. That other 4 will drive the woofers in my 4-ways, plus the bass shaker in my seat and have one channel left over for a possible centre channel speaker.
Love seeing the journey…. Thanks. Very interesting about the NTC thermistor bypass as I have two identical guitar amps one had a problem which also involved replacing a NTC thermistor, but before I did that I bridged (NTC) to see if it was the problem and I thought it sound ‘better’ then after the replacement.
I'm putting in a soft start on a 2 kw transformer, a resistor, that's switched out after 100 ms, or so, with diode-RC-relay coil-and reverse diode to the input rectifier, to discharge the RC capacitor quickly on power-down...the 2 kw transformer takes out a 16 Amp breaker, unless you get lucky and the plug makes contact near the AC line zero-crossing!
I had never considered reducing wire length by just etching them onto a board. I build guitar pedals from time to time and stuffing 4 pots and 3 switches worth of wiring into a hammond box is double plus not fun. I would much rather struggle with designing a board or two instead. Thanks.
Nice to eliminate the "loose wires" with the distribution boards...you want pretty good connectors. Non-glamorous stuff can be important to performance and longevity! I've been designing and building for industrial and high-volume products, and details matter!
The bridge rectifier is nice and big, but check datasheet to see how much current it can handle continuously without extra heatsinks. I would have used the GBU / GBJ style rectifiers that sit vertically and placed to the edge of the board and could be mounted easily to a heatsink. As they are now, the bridge rectifier will radiate heat through the copper on the board into the 85c electrolytic capacitors. I also noticed using 1n4002 ... i'd just standarize on 1n4007.. Maybe it would have made more sense to bring in 12v or 5v from a stand-by power supply instead of sourcing it from the transformer... and use a RC timer or something to turn on a mosfet or transistor which turns the relay after some period. Maybe add some high value resistors across the big capacitors to slowly discharge them when not in use, just in case someone will want to service it and not get zapped?
The rectifiers don't even get close to hot; my AC dropping supply on board is genius; there is an RC timer (and mosfet) there already; and the amplifiers draw bias current even when you shut the power supply down which bleeds down the voltage. Hello and welcome to the real world.
In the bad old days, every television set had a "Global" resistor, basically a big old NTC power resistor, to kill the inrush current of tube designs, with all the heater current and big old transformers... Now, I have to DIY the same old solution...
I really enjoy your videos, especially your exploring the validity of the so called "accepted facts" on topics related to amateur speaker design and construction. I would say that I possess a moderate competence regarding speaker building, but I am woefully lacking in the area of electronics. You obvious are skilled in this area. How did you get your start and gain your expertise. I have wanted to get into this hobby, but the beginner stuff has no application to me. Is there a resource that allows a new enthusiast to build interesting things from the beginning? What was your path and how did you get to the place you are today?
Question about C3 and C4. What is the purpose of wiring a pair of 1uF caps in parallel? Why not use a single 2.2uF 100v cap? Just what you had on hand?
Redundancy. It will work with just one 1uf cap, but (apparently) this AC dropping method can be tough on the cap, so I put in a second one in case one blows.
You could have had wider copper traces for + - & gnd. except at lug tap outs , instead of etching away all that usable & useful copper.... Also place some 3-10 uF caps ( ac snubber types preferably ) @ the other extreme end of bus layout..
I could have, but didn't, and I shouldn't have to point out that these are already made and will work perfectly fine as they are. I actually like boards with just the right amount of copper, rather than poured planes. As for bypass caps, they are on the amp boards these supply, close to the output transistors where they are supposed to be for best operation.
All in, probably less than $1000 in materials, but some of those are used and cost very little (the transformers, for instance). And that doesn't include the time it takes to design and build everything.
Since I'm using two different transformers, I need two separate power supplies. These supplies have a soft start circuit built in and in this video I go over how that works.
In a nut shell, the thermistor is all that's needed for the soft start to work. But there are a couple of good reasons why it's better if you bypass it after it has done its job.
First is that if you bypass it, it will cool down and last longer.
Second is that when the amplifiers are not drawing enough current, the thermistor will cool down anyway, and as it does its resistance goes up, which lowers the supply voltage.
When you short it with a relay, the relay carries all of the current without heating up. But there has to be a short delay before the relay switches on so that the thermistor has time to work. I used the switching circuit from Rod Elliott's website to do that, and he gives a more detailed description of how it works: sound-au.com/project39.htm
To power the soft start bypass I dropped the 40VAC from the power transformer down to less than 10VAC with a resistor and a pair of 1uf capacitors. That lower AC voltage is then rectified to power the circuit and switch the 12VDC relay.
This has the advantage of not needing a small transformer to run the soft start. Also it doesn't heat up like resistors would to drop the rail DC down to a usable voltage.
On the board as well is what's known as a ground lift. It consists of a 10 ohm power resistor and two diodes wired back to back. The lift "separates" the supply ground from the chassis (earth) ground, but still provides a physical connection between the two. This can help greatly with ground loop problems that cause hum.
In the video I also show the power distribution boards that feed the individual amp boards. These reduce the wire runs inside the case and also put the supply capacitors very close to the amp boards, which should improve (very, very slightly) transient response. They are also a convenient place for the rail fuses for each amp board.
Since I have two supplies and 10 amplifiers, one supply will power 6 amp boards (these will drive the tweeters, midranges and midwoofers in my 4-way speakers) and the second will power the other 4. That other 4 will drive the woofers in my 4-ways, plus the bass shaker in my seat and have one channel left over for a possible centre channel speaker.
Love seeing the journey…. Thanks. Very interesting about the NTC thermistor bypass as I have two identical guitar amps one had a problem which also involved replacing a NTC thermistor, but before I did that I bridged (NTC) to see if it was the problem and I thought it sound ‘better’ then after the replacement.
I like the way you are breaking it down into components.
The process of making them ready for video also helps me, keeping me on track and not jumping ahead and doing the more interesting parts :)
On one of my little ham radio projects long ago I installed a soft start. I learned to flip the switch really slow. 👍🏻😂
Leaving it on all the time works, too :)
I'm enjoying this series.
Very few engineers, like yourself, would show actual schematics. Nicely done.
I'm putting in a soft start on a 2 kw transformer, a resistor, that's switched out after 100 ms, or so, with diode-RC-relay coil-and reverse diode to the input rectifier, to discharge the RC capacitor quickly on power-down...the 2 kw transformer takes out a 16 Amp breaker, unless you get lucky and the plug makes contact near the AC line zero-crossing!
I had never considered reducing wire length by just etching them onto a board. I build guitar pedals from time to time and stuffing 4 pots and 3 switches worth of wiring into a hammond box is double plus not fun. I would much rather struggle with designing a board or two instead. Thanks.
Nice to eliminate the "loose wires" with the distribution boards...you want pretty good connectors. Non-glamorous stuff can be important to performance and longevity!
I've been designing and building for industrial and high-volume products, and details matter!
Wow. Your level of knowledge is super impressive.
This thing is going to look and be amazing.
The bridge rectifier is nice and big, but check datasheet to see how much current it can handle continuously without extra heatsinks. I would have used the GBU / GBJ style rectifiers that sit vertically and placed to the edge of the board and could be mounted easily to a heatsink. As they are now, the bridge rectifier will radiate heat through the copper on the board into the 85c electrolytic capacitors. I also noticed using 1n4002 ... i'd just standarize on 1n4007..
Maybe it would have made more sense to bring in 12v or 5v from a stand-by power supply instead of sourcing it from the transformer... and use a RC timer or something to turn on a mosfet or transistor which turns the relay after some period. Maybe add some high value resistors across the big capacitors to slowly discharge them when not in use, just in case someone will want to service it and not get zapped?
The rectifiers don't even get close to hot; my AC dropping supply on board is genius; there is an RC timer (and mosfet) there already; and the amplifiers draw bias current even when you shut the power supply down which bleeds down the voltage.
Hello and welcome to the real world.
In the bad old days, every television set had a "Global" resistor, basically a big old NTC power resistor, to kill the inrush current of tube designs, with all the heater current and big old transformers... Now, I have to DIY the same old solution...
Nice going John… enjoying this.
Awesome vid! Very inspiring!!!
I really enjoy your videos, especially your exploring the validity of the so called "accepted facts" on topics related to amateur speaker design and construction. I would say that I possess a moderate competence regarding speaker building, but I am woefully lacking in the area of electronics. You obvious are skilled in this area. How did you get your start and gain your expertise. I have wanted to get into this hobby, but the beginner stuff has no application to me. Is there a resource that allows a new enthusiast to build interesting things from the beginning? What was your path and how did you get to the place you are today?
That would make a good topic for a video, actually. In a nut shell, I learned by doing and mostly failing, and have no formal education.
@@IBuildIt Thanks for the reply. I definitely would be interested in a video addressing this subject. Thanks for all the great videos!!!
Do you mean negative temperature coefficient?
By back to back diodes do mean inverse parallel diodes?
Question about C3 and C4. What is the purpose of wiring a pair of 1uF caps in parallel? Why not use a single 2.2uF 100v cap? Just what you had on hand?
Redundancy. It will work with just one 1uf cap, but (apparently) this AC dropping method can be tough on the cap, so I put in a second one in case one blows.
@@IBuildIt okay, makes perfect sense.
I might have missed it or just forgotten it if you already showed this, but do you tin or laquer or otherwise protect the copper lines from oxidation?
On a board I'll keep I usually tin the traces. These ones I sprayed with clear finish, mainly because I'm low on solder for tinning.
make a video on how to make custom amplifer
You could have had wider copper traces for + - & gnd. except at lug tap outs , instead of etching away all that usable & useful copper....
Also place some 3-10 uF caps ( ac snubber types preferably ) @ the other extreme end of bus layout..
I could have, but didn't, and I shouldn't have to point out that these are already made and will work perfectly fine as they are. I actually like boards with just the right amount of copper, rather than poured planes.
As for bypass caps, they are on the amp boards these supply, close to the output transistors where they are supposed to be for best operation.
They say if you have to ask the price, you can't afford it. I'm sure I can't. But I'm very curious to know the cost of the DIY project like this.
All in, probably less than $1000 in materials, but some of those are used and cost very little (the transformers, for instance). And that doesn't include the time it takes to design and build everything.
@@IBuildIt but it will be custom built for a specific purpose so you save tons of $ doing it yourself.
@@IBuildIt Oo less than I thought. Thanks! You could sell plans for it...
Coefficient.... not constant. Sorry :-)