I know the ESP 3a amplifier is old news in DIY but the small size possible seems to make it a good choice for tight spaces when one does not want to use a chip amp or class D.
I kept a dead AudioSource Amp One/A out of curiosity hoping to get some insight into how amps work. After getting some experience re-capping a couple of receivers I pulled it out of the crawl space to see if I could do something with it. Removing the dead amp board left me with two of the more expensive items on a DIY BOM: case and transformer. Problem is this case has only about two inches of height on the inside, to short for a lot of builds
The Project 3a amplifier as found on the ESP site seemed fairly small and simpler than the boards I’ve seen in most of the other stuff I’ve opened up.
Populating the boards from ESP was not that difficult, much easier than re-capping as nothing needed to be removed and no traces to worry about while so doing. Most of the work was fiddling with the board to make it fit into roughly the original layout of the case and some refitting.
I separated the two channels so each would have its own heat sink.
Each channel was cut to reduce their depth to allow them to be mounted in the middle of the heat sink and not over lap. They were then cut further so I could use some angle brackets to secure the boards to the heat sinks with more than the leads to the output transistors. This then necessitated some alternate wiring for the electrolytic caps C+ and C-.
Shielding between the amp and power supply was cobbled together from a piece scavenged from some other electronics.
The power supply is just the basic one outlined in the Project 3a discussion at ESP: the transformer followed by a rectifier with a capacitor between the AC connects and two 10,000uF electrolytic supply caps.
ESP suggests a 25-0-25v transformer to run the DC into the amplifier at 35+/-. The AudioSource transformer measured 36-0-36. Too high. The few pictures I’ve seen of the P3a build in a case similar to this had what looked like a small transformer in the area of the power switch. I’m guessing those were bucking transformer, but no writeup I found specifically mentioned them let alone what values were used.
An ESP 3a in a Sonos case from the ESP gallery. Similar case layout, much better workmanship than in my project. Used for guidance in my build.
Not finding a 25-0-25v transformer short enough to fit the case or finding a small enough choice for mounting a buck inside the case I went to an external bucking transformer, a 40v unit pulled from a CambridgeSoundworks MicroWorks unit. This may change once I get some time to look into other possibilities. Currently I have 33+/- coming off the power supply.
At this point I have not put in any protection or soft-start circuitry. There is plenty of real estate to add these later if I choose and I may shorten some of the wire runs when I go back in.
Finally, much of the original construction on the Amp One/A was done with sheet metal screws, including securing the output transistors to the heat sink. The holes for these were close to the diameter of M3 machine threads. I converted all these to M3 threads using a forming tap, if for no other reason I prefer smooth threads.
The amplifier sounds quite good at this stage
I kept a dead AudioSource Amp One/A out of curiosity hoping to get some insight into how amps work. After getting some experience re-capping a couple of receivers I pulled it out of the crawl space to see if I could do something with it. Removing the dead amp board left me with two of the more expensive items on a DIY BOM: case and transformer. Problem is this case has only about two inches of height on the inside, to short for a lot of builds
The Project 3a amplifier as found on the ESP site seemed fairly small and simpler than the boards I’ve seen in most of the other stuff I’ve opened up.
Populating the boards from ESP was not that difficult, much easier than re-capping as nothing needed to be removed and no traces to worry about while so doing. Most of the work was fiddling with the board to make it fit into roughly the original layout of the case and some refitting.
I separated the two channels so each would have its own heat sink.
Each channel was cut to reduce their depth to allow them to be mounted in the middle of the heat sink and not over lap. They were then cut further so I could use some angle brackets to secure the boards to the heat sinks with more than the leads to the output transistors. This then necessitated some alternate wiring for the electrolytic caps C+ and C-.
Shielding between the amp and power supply was cobbled together from a piece scavenged from some other electronics.
The power supply is just the basic one outlined in the Project 3a discussion at ESP: the transformer followed by a rectifier with a capacitor between the AC connects and two 10,000uF electrolytic supply caps.
ESP suggests a 25-0-25v transformer to run the DC into the amplifier at 35+/-. The AudioSource transformer measured 36-0-36. Too high. The few pictures I’ve seen of the P3a build in a case similar to this had what looked like a small transformer in the area of the power switch. I’m guessing those were bucking transformer, but no writeup I found specifically mentioned them let alone what values were used.
Not finding a 25-0-25v transformer short enough to fit the case or finding a small enough choice for mounting a buck inside the case I went to an external bucking transformer, a 40v unit pulled from a CambridgeSoundworks MicroWorks unit. This may change once I get some time to look into other possibilities. Currently I have 33+/- coming off the power supply.
At this point I have not put in any protection or soft-start circuitry. There is plenty of real estate to add these later if I choose and I may shorten some of the wire runs when I go back in.
Finally, much of the original construction on the Amp One/A was done with sheet metal screws, including securing the output transistors to the heat sink. The holes for these were close to the diameter of M3 machine threads. I converted all these to M3 threads using a forming tap, if for no other reason I prefer smooth threads.
The amplifier sounds quite good at this stage
Attachments
Last edited:
