Jump to content

TonyD'Amore

SMD PARTNER
  • Posts

    462
  • Joined

  • Last visited

  • Days Won

    3

Posts posted by TonyD'Amore

  1. After setting the bias controls and doing a bunch of testing, everything looked fantastic. We just got the demo car back this same night, time to remove the placeholder baseplates that we gave the shop and replace them with the real thing!

    1934609_10205448359468774_64570058405496

    1%20138_zpsauv7vdpc.jpg

    She's like 2 1/2 now

    ceeedb8c-5967-4ae4-9c96-2612e3af35c2_zps

    Now repeat the process and build another amp lol

    1%20165_zpsqolhldpy.jpg

    993875_10205489421975311_283065880037112

    10645010_10205489299612252_6910535549052

    A little orange peel on the paint (it looks worse here than in life for some reason) I blame the cold booth and the rush we were in. Production pieces will look better!

    If I find more pics I'll add them. Thanks for following along. The result was better than we hoped for, everyone who has heard these amps has been impressed. They are VERY dynamic and fast, the dynamics of the music come through producing a very "real" or "live" sound.

    Tony D

    • Like (+1 Rep) 3
  2. After we got one PCB up and running we started doing thermal testing again (beating the crap out of it). Remember a few posts back we weren't satisfied with the isolator material we were using? Yeah, this happened. You can see the tan-ish colored material. It is a phase change material that is supposed to be awesome, it was pretty awesome, just not awesome enough.

    1%20179_zpsmu7wvb1p.jpg

    Time to go home for the day now that the entire building smells like blown up MOSFETs. The search continues for a better material.

    After searching around we discovered Alumina Al203. It is a bad ass ceramic that will do both jobs excellently (electrical isolation, yet fast heat transfer). Learn about Alumina here http://www.morgantechnicalceramics.com/materials/alumina-al203

    • Like (+1 Rep) 1
  3. So it looks like the heatsink design is more than enough (even though we are not satisfied with the temperature delta between the power devices and the heatsinks yet). So at this point we called in a good friend who is a mechanical engineer to design the sheet metal top, front cover, top meter cover, and associated parts. I gave him the drawings that I had and in no time at all he sent this rendering and asked "like this?" He is the man, thanks Tom Sollars.

    38_zpsdo9hprrf.jpg

    We loved it and asked him to continue and order up some of these parts when done.

    A few weeks later we had these

    1%20175_zpsxfry9vgw.jpg

    1%20176_zpstitpmdz3.jpg

    1%20174_zpsrfkqi8ag.jpg

    1%20178_zpsik4u222w.jpg

    1%20177_zps4khdd6kg.jpg

    1%20170_zpshj4dzpuv.jpg

    • Like (+1 Rep) 4
  4. Now that we know the electricals are mostly sorted, lets check this heatsink design and see if we have enough dissipation capability. A fairly common industry test is full power sine wave into 2 ohms, and 40% power sine wave into 2 ohms. The first case stressing the power supply, the second case stressing the output section. Time to put that foam core board lid back on.

    37_zpsycaqc79v.jpg

    I don't have photos to show, but we blew it up several times. The problem was we couldn't get the heat from the power devices (power supply MOSFETs and output section BJTs) into the heatsink fast enough. Between the power devices and the heatsink you need an electrical isolator because the mounting surface of the MOSFET or BJT is connected to the center leg of the device, which has various voltages on it that you wouldn't want on the heatsink. This would cause the amplifier to not work at all. The most basic of amplifier designs use kapton tape, it is a great electrical isolator, but a terrible conductor of heat. We didn't even try kapton because I wanted this thing to perform! The next level is various thin composite materials. One of these materials turned out to be good enough for the home amplifier and if you look at some of those pics you will see a thin grey pad under the devices. We wanted better than that because at low impedance we want to be reliable. If this topic interests you check out http://www.fujipoly.com/usa/ After weeks of testing different materials we settled on some stuff from this place that is tan in color and has a high thermal conductivity. http://www.aavid.com/products You can see it in the next series of photos.

    • Like (+1 Rep) 5
  5. The PCBs have arrived, I hope they fit the mechanicals!

    20_zpsryjs3ikm.jpg

    There is a little more gap between the PCB and heatsink than I wanted. This will work for testing though!

    23_zpsinzndbkj.jpg

    26_zpsnkpwjbvt.jpg

    22_zpsfynycu0p.jpg

    19_zpslpfvn3mt.jpg

    Handed over to Magic Juan to do his thing. He built them up, tested, found a few problems and fixed them. Gave me a laundry list of changes for the next revision. That's how the ball bounces, the first revision is never perfect. There are just too many things to consider. That being said, you know Juan made it work!

    27_zpsjfm8lwsn.jpg

    29_zpsytbw4s08.jpg

    31_zpspi79hzhn.jpg

    32_zpspy1xxm9t.jpg

    30_zpsml5eqxrz.jpg

    She's ALIVE! ALIVVVVEE! The original heatsinks were two pieces bolted together as you can see here. This was because we weren't sure what the top was going to need to be, but we had the big mass of aluminum on the bottom defined. Once we figured out what we needed we made it one piece.

    33_zpsibahpfpm.jpg

    • Like (+1 Rep) 2
  6. The machine shop usually takes about a month after I give them my CAD files so I better get to designing the heatsinks and baseplate because I'm not going to be able to test the amp with foam heatsinks!

    4_zpsuhwzkla5.jpg

    Two months and revisions later we have some aluminum! 17 pounds of it for 1 amplifier!

    18-6_zpsxmjzhchz.jpg

    24_zpsaufzapp4.jpg

    11_zpsfnhdrpud.jpg

    1%2016_zpsufe3wplb.jpg

    Ok back to the power supply and current amplifier PCB design. As it turned out, I needed to use a 4 layer PCB. Like 4 circuit boards in 1, literally in layers sandwiched together. Typical PCBs in electronics have a copper thickness called 1/2 ounce. This means that for a square foot of PCB there is 1/2 ounce of copper in it to carry the current. In our design we are going to need a LOT more than that. 4x that actually so we went with 2 ounce copper, 4 layers of it.

    12_zpsumk62oev.jpg

    Each color you see here is copper on a different layer of the PCB.

    13_zpsz61pxbwi.jpg

    15_zpsozjshmfi.jpg

    17_zps0xj4lkgo.jpg

    14_zpsm8yp7m5t.png

    18-5_zpsy0lrgf3n.jpg

    Many, many late nights later I am ready to send the files off to the PCB house and order all the components to build one. Now that this PCB is done, I get to do it all over again for the right hand PCB. No there is no magic button I can press to tell my software to do the opposite of the whole board for the other side. If you try that it will reverse all of the "footprints" for the components and when you build it every part will be in backwards and some of them can't go on the board any other way. Ask me how I know this.... $600 down the drain. Eventually I had to make the right hand board from scratch like the left hand one.

    • Like (+1 Rep) 5
  7. Now that we know we have a stable amplifier design, we could change the power supply from the AC/DC supply to a DC/DC supply that can run from 12V and convert to 100-120VDC. The home amp runs at 180VDC, but it is designed to run at 8, 4 and 2 ohms. I want the mobile amp to run at 4, 2 and 1 ohm so we need less voltage to make the same power. We also need a mechanical concept. All we knew at this time is that we wanted it to be a TRUE dual mono. This means 2 amplifiers, each with their own power supplies. I'm also a fan of symmetry, I may be a bit OCD. I like to use foam core paper and foam blocks from Michael's.

    1_zpsz9n8szz9.jpg

    3_zpsnqy90u3t.jpg

    2_zpsf2n5orgx.jpg

    Hmm... I like it. But can I make a PCB that works with the design? There are many voltages on the PCB that require lots of copper to carry the current. B+, GROUND, +60V, -60V, speaker output, AC output from transformer, rectified output to the power supply secondary caps to name a few.

    11-5_zpsbqhurmmj.jpg

    Hmm, I'll have to give it more thought later. Let's design the new input stage and voltage amplifier. This amplifier needs a gain control which the home amp didn't have. I also wanted to be able to switch it into class A/B mode for those times when I'm not critically listening and those times when int 120 deg outside in PHX. Also want to add switchable unbalanced/balance input. After several weeks of design on the computer, I sent the files to my PCB house and we ordered up all the components. They both arrived in about a week and we started to build.

    7_zpsb47plvxp.jpg

    8_zpsfj3hoc0m.jpg

    9_zpslkltpqfo.jpg

    Magic Juan does it again! This little board has the input section, gain section, and ALL of the voltage amplification for the amplifier output. The signal will leave this PCB already amplified to its maximum voltage. The main PCB is just power supply and current amplifier. This will ensure the signal stays reallllly clean! Let's test it! If it's as badass as I think it will be it will be able to amplify a signal as fast as 100kHz without distorting it....and it DID!!

    10_zps6z6jniza.jpg

    Let's see a Class D amp do that!

    21_zpskbnx9fdw.png

    • Like (+1 Rep) 3
  8. It sounds amazing! Probably spent the next month just listening to it at work all day every day.

    It def needs a sexier front panel though, and thicker. We went with 1/4" aluminum, CNC'd on both sides, recessed areas for the meter PCBs and glass to sit in as well as beveled meter openings and our brand name carved into it.

    32_zpsazwarlv6.jpg

    34_zpsst3xhvsd.jpg

    33_zpsmryjlykp.jpg

    Now I send all the CAD files to the machine shop and wait.

    I also designed a top cover and bottom cover to match. 1/4 aluminum

    While waiting I ordered up some copper mesh to put over the slots to keep large things out of the amplifier and give it a unique look. As it turns out the copper mesh somehow reduced the little distortion and noise that was left even more! Not going to argue with my test equipment, I'll take it!

    36_zpsxihxbvys.jpg

    38_zpsvuftkoxw.jpg

    37_zpscy6qzqtq.jpg

    39_zpstogb7bxr.jpg

    It still gets used daily and occasionally I'll measure the temp of it. It's always between 130F - 150F depending how long it has been on and the ambient temp of my lab. Since it is Class A, it doesn't matter if it idles or being pushed hard, the temp remains nearly the same. It actually seems to cool off a little as I beat on it!

    40_zps8cspnvt9.jpg

    Now, it would be cool if I could get this kind of sound quality in my car......

    • Like (+1 Rep) 3
  9. I send my PCB design files off to the PCB house along with some $ and in a week I have some PCBs! During this waiting time we ordered up all the components for the PCB so they arrive about the same time. I choose the best capacitors I could find for the audio paths, Nichicon FG (fine gold) and Nichicon Muse (green). My partner Juan is magical with a soldering iron and debugging so it's his turn now.

    16_zpsbyyegmvp.jpg

    21_zpscwgblyju.jpg

    17_zpsq4osmgua.jpg

    Let's drill these new 'sinks for the PCB mounting holes and transistor mounting holes

    15_zpsq9t97ppd.jpg

    Proud papa Magic Juan

    20_zpsrqkpor0r.jpg

    19_zps1xc3dldk.jpg

    18_zpsrmc9tkin.jpg

    23_zpsglvwajv3.jpg

    22_zpsfonvif4e.jpg

    Testing the supply, only 180V DC

    24_zpswjiaoi13.jpg

    • Like (+1 Rep) 5
  10. Time to give props where they are due. I've learned a lot about amplifier design over the years from some of the best. Besides the books shown here I also worked with Jim Strickland and Mark Albers daily for years

    26_zpsukr4rfaw.jpg.

    The next step here is to design some heatsinks, make them, dump heat into them and see how they react. I don't have pics from that time but this is the result, the tunnels increase the surface area of the heatsinks big time! Surface area is what you need to move the heat to the atmosphere. The holes are 0.750" diameter. So the surface area lost by having the holes is (.750 / 2) ^2 * pi x 2 (top and bottom) = about 0.88 square inches x 39 holes = 34 square inches. BUT the amount of surface area gained is (.750 * pi * 7 inches deep = 16.5 square inches x 39 holes = 643 square inches! So each hole effectively gave us 20x the surface area. Not to mention a great chimney effect that starts to flow air without a fan.9_zpsevu0rh9n.jpg

    Lets make a set of mechanical and have them anodized black

    10_zpscb9h6c7v.jpg

    Now we need to loose all the prototype stuff and design the actual PCB that will contain the voltage stage, current stage and sit nicely on our new heatsinks

    11_zps107qvbsm.jpg

    13_zps84wdfcqu.jpg

    14_zpsnq0tivfl.jpg

    • Like (+1 Rep) 5
  11. Since the unveiling of our A series amplifiers and before, I get the occasional email that reads "hey, I know you're busy, but could you design a few amps for me real quick" It usually makes me giggle a little and then smh. It takes me back to 1999 right before I started with RF. I also had no idea what was involved designing an amplifier from the ground up.

    No disrespect intended if you are reading this and have asked me to make you some amps before. How would you know what is involved if you haven't done it before? This has prompted me to go through literally thousands of photos on my phone and pull off any amplifier design pictures that have to do with the new amplifiers we just introduced. I will make my best attempt at posting them here in chronological order, and make some brief comments here and there. I also threw in a few pictures of my daughter taken during those times as a measure of time.

    The A series amplifiers are a direct COPY of the home audio amplifier I designed for my own use (changed power supply to 12VDC, added fans, remote turn on, balanced/unbalanced input and ability to put it into Class A/B mode (home amp is always class A)) so I have included the photos of designing that as well because all of the time I spent developing the circuits went directly into the mobile amplifiers. Ok enough of me talking..

    On with the show!

    Before I build any circuit I like to try to simulate it in PSPICE, here you see a nice clean output and the predicted distortion

    6_zpshyn7rhn4.jpg

    7_zpsej06a81y.jpg

    My daughter was born during this time

    5_zpsdsp0klj4.jpg

    Here you see the power supply for the home amp. AC to DC supplies like this are very easy to design. The small circuit board is the input stage and voltage amplifier stage. This is where a lot of the magic is and many months of reading and testing were done to get it to work. Unfortunately I couldn't find any pictures during that time. We went through at least 3 different voltage amplifier designs, PCBs and testing before we came up with what we use today.

    1_zpstidb7f7l.jpg

    Moved tested input stage and voltage amplifier stage onto main PCB instead of it having it's own PCB. (In the car amp it still has its own PCB for the least amount of noise possible) Each tiny resistor and transistor has been studied for voltage, current, and power dissipation.

    12_zpsra3340hq.jpg

    Prototyping the current amplifier stage

    2_zps03rj0pku.jpg

    It works! Now lets build 2 voltage amplifiers, 2 current amplifiers, connect them together and add a power supply! We also added a prototype SMD VU-DIN meter for S&G.

    3_zpsmel4kemv.jpg

    Ok let's put this contraption in some kind of case so we can move it around without breaking anything. A SMD AD-1 Amp Dyno prototype case worked well

    4_zpsqs1vyhou.jpg

    Keeping the same case, I designed a simple face plate for it and had my machine shop CNC it.

    We put it all together with some prototype meters. Shown here on the distortion analyzer making a sweet ass sine wave with only 0.0068% THD! WHOA! Cleanest amp i've ever tested, home, pro, or car.

    8_zpslmqi9tnt.jpg

    • Like (+1 Rep) 10
×
×
  • Create New...