My next build is a preamp. Having a dedicated head amp and speaker amp I am happy with this is the gap in my chain. I set some simple parameters to begin with, namely it needed to have multiple inputs and outputs, should use active amplification and be single ended. Looking at the options available to me in kit form I decided to go with one of the other three of the trifecta of early 2000s top diy designs in the form of the AMB “M Cubed”. It takes its name from the topology which is defined by using an active ground amplification. From the website:
In addition to the left and right channels, in this amplifier the “ground” wire of the headphone is actively driven by a third channel of the same topology. The ground channel amplifier sources or sinks the return current from the transducers, which would otherwise have been dumped into signal ground or power supply ground. This shifts responsibility for the high current reactive load of the headphones from signal ground to the tightly regulated power supply rails, thus removing the primary source of signal ground contamination. The headphone transducer “sees” symmetrical output buffers with equal impedance and transfer characteristics on both sides, rather than an output buffer on one side and a capacitor bank of the power supply ground on the other.
The intent then is to create a clean, low impedance output with a neutral sound signature.
For the power supply I chose to use the Sigma 11, also from AMB in a 24v configuration. I ordered both boards with any specified matched transistors from AMB.
Having made the basic decisions I have spent the last few months designing, redesigning, scrapping and starting over, the signal path around the amplifier stage. I’ve changed the number of inputs and outputs, added and then removed gain staging, both passive and active versions. Toyed with using relay switching instead of mechanical switches and tried out front panel layouts endlessly.
While I did this I built out the PSU board - a relatively simple pcb - which is ready for the final parts now.
Just an FYI because it’s not obvious, if you intend to use that single big heatsink, on what I assume are regulators at the top of the board in I assume TO 220 packaging, the metal tab on them will be the output, not ground as you may assume, so you need to ensure it’s isolated from the heatsink if more than one shares the heatsink or you have a short.
The little plastic gromets and silicon pads they SOMETIMES include with heatsinks do that, but always check it with a meter.
Next stage. The main board is stuffed with the smaller components, resistors, diodes, mosfets, capacitors. I enjoy this bit the most.
With this board there are choices to be made as it optionally has a second pot as either a gain or a bass boost. I’m choosing to have neither here as I don’t feel a need for it to have more gain and I am not keen on an EQ that only allows for one frequency area to be adjusted. Interestingly this board also allows for a wide variety of sizes for the larger components without bending the pins. You can see the additional pads around the capacitors.
Still to fit are the larger rectifier fets with their heat sinks and larger resistors at the outputs.
Higher voltage electrolytic come in wildly varying sizes, so it’s nice to have.
Though that looks like you can choose to use less larger devices to make up the total capacitance.
I had some time to work on the amplifier this weekend. So first up, the completed power supply. As @Polygonhell pointed out the heat sink absolutely must have isolation mountings to avoid a short. The specifications recommend them anyway in lower height cases. The reason I have chosen this larger, single, heat sink is to do with my case layout.
Below is the toroidal transformer that will be paired with the above pcb. I do really like these things as objects. There is something wonderfully real about them compared to many modern electrical components.
Finally the now completed main board. You can easily see the two channels plus ground opamp stages to the right of the board and then the three matched pairs of rectifiers to the left.
Next up will be preparing the VU meter header board and the opamp mounts. The recommended opamps are only available in a SOIC (very small, surface mount size) format this days. I may change them out in the future but first I want to hear the amp as it was designed. Additionally it is reputedly one of the more picky amplifiers when it comes to opamp swapping. Indeed the specifications have a whole table of minor changes to be made depending on the voltage rating of a chosen opamp. Additionally it only accepts fet input opamps which also limits the available choices.
It’s just really dumb as well, I’ve been trying to reduce the heat from one of the rectifiers in my tube amp project, and there is a roughly 20C difference for the part itself using thermal paste vs the silicon pad. The manufacturer even quotes it’s example heatsink requirements using thermal paste
But of course you can’t isolate the heatsink if you use thermal paste.
I would love to know why they decided the body of the component should be Vout, and not Gnd or better yet disconnected.
Definitely redo 1, while you’re at it, might as well redo 2. But 1 is definitely problematic… or will eventually be. It looks like they’re the feet of the heatsink so not a real issue, it’s just OCD.
Yes. Yes they are. You just picked out the two that are better on the opposite side I’ve already re-flowed them using flux once. I may do so again if I can’t ignore your ocd. Hah!
I have a 1/4 inch chisel tip I use only for Heatsink supports, the copper busbar I decides to use to ground my amp and de-soldering, if your iron lets you swap tips easily, it makes soldering stuff like that much easier.
I have constructed a mezzanine from sheet aluminium to act as a shield and as a platform to carry the amplifier section. However because of the very small tolerances both the sections only have vertical clearance of a couple of millimetres. For this reason all the heat sinks have isolation fittings and I have used an M6 nylon bolt to secure the transformer to be sure of not creating a winding short.
The case has been hard to work with. Literally! My first time working with a steel case and it was not easy. The inside is pretty scratched up, due to needing to sand/dremel the boring holes flat. The front and back panel were very slightly warped too but I think no one else will notice that as they won’t have built it. The mezzanine worked well and I am pretty pleased with it for a first attempt at metal work. I can see where the shortcomings of my tools has limited what I can do but on the whole I am pleased with the result.
One more warning/tip. If those wire might touch the edge of the aluminum plate, my experience tells me it will eventually cut through the insulation. So I’d recommend putting some tape around the edge of the plate or better fasten the wires in a way they can’t touch the edge of the sheet.
Good pointer, thanks. I have a rubber grommeted hole for any wires that need to pass up to the top deck. Hmm. I think I might tale up the sides anyway as a precaution.
@Polygonhell@Gothique you guys were talking about spare parts scattered in your bins. Saw this from Nelson Pass and reminded me of your comments. This is a lot of spare parts, and it’s a joy to see a guy whose hobby and his work are the same passion.
Starting the final assembly now. Adding audio cable and front panel controls. It’s slow going as I am taking care with cable location to try and minimise emi exposure and heat sinks.
Note that while it’s considered “good” to ground inputs to the case, it can introduce ground problems, the same issue exists with running ground across the inputs. In your case there is no guarantee ground is common across all connected devices, and ideally you want to pull all those grounds together at a single point, because the material between them acts as very small resistors, and so they end up slightly floating in relation to each other, and power ground. This is why star grounds are considered the best solution, but they aren’t always feasible.
If you don’t get a ground hum when you listen ignore this advice. But I spent hours massaging ground paths on my tube amp to eliminate miniscule hum’s, and I’m convinced it’s all total voodoo at this point. I ended up removing the case ground on the RCA inputs which resolved most of it, removed the two remote grounds I had put in to simplify wiring, and in the end I routed everything individually back to the central bus bar, though a lot of it was just to be sure.
I’ve already re-flowed them using flux once. I may do so again if I can’t ignore your ocd. Hah!
