Not checked yet, P-A, to my discredit.
One thing we should add is a 10R series resistor at the output with about 10nF to ground at the load end. This will ensure doubly that RF is not an issue, and will certainly reduce power somewhat for 24R cans, but for the majority will have no effect and will ameliorate any thumps.
I had hoped by now that someone might have built and tested one, but can't complain since I did not do it myself.
Hugh
One thing we should add is a 10R series resistor at the output with about 10nF to ground at the load end. This will ensure doubly that RF is not an issue, and will certainly reduce power somewhat for 24R cans, but for the majority will have no effect and will ameliorate any thumps.
I had hoped by now that someone might have built and tested one, but can't complain since I did not do it myself.
Hugh
What about a protection/delay board? Placed at the output, providing a optional impedance switch, 1 second delay, optional zoble network, and even a push button on. Does that appeal? It fixes all the issues without having to redesign the entire amp.
Corrections made to schematic, thanks again aparatusonitus
Corrections made to schematic, thanks again aparatusonitus
I think the prototype should first be built in ratsnest to see if there are switch on thump issues. Strictly that should be done before the board is completed. If there is a problem, then a protection board is a good idea, otherwise, it almost doubles complexity.
A second option, worth considering, is to redesign for single rail and cap couple the output. This offers magnificent protection, thought there may still be a small thump as switch on. Kurt, do you have a relay in mind for protection? A good one should use gold plated contacts; audio is extremely touchy about relays.
Hugh
A second option, worth considering, is to redesign for single rail and cap couple the output. This offers magnificent protection, thought there may still be a small thump as switch on. Kurt, do you have a relay in mind for protection? A good one should use gold plated contacts; audio is extremely touchy about relays.
Hugh
Take your time, Kurt, there are no deadlines, maintain the interest but don't wear yourself out. Your board is a very good one, but these operational issues should first be resolved with a breadboard. I breadboard every circuit first; otherwise, you finish up with lots of pcbs you can't use, which represents an appalling waste of resources.
The relay idea is not often used to my knowledge with HPA. I think it should be possible to design out the turn on thump.
Hugh
The relay idea is not often used to my knowledge with HPA. I think it should be possible to design out the turn on thump.
Hugh
I managed to create a abomination of the amps output stage, with a 1.95volt LED and no DC adjust. I managed to almost burn myself on Q4 with my small little heatsink, does anyone know how much heat Q4 is theoretically puting out? The heatsinks that I used on the board layout might be too small. Please note I didn't use the spected heatsinks for the test, they were a much smaller slid on heatsink.
I'll hopefully have time to refine the railspliter the simple two resistor virtual ground circuit next week as well as install a proper DC adjust
I'll hopefully have time to refine the railspliter the simple two resistor virtual ground circuit next week as well as install a proper DC adjust
Kurt,
Each output, Q3 and Q4, dissipates (0.12A x 15V) = 1.8W of heat. If the heatsink is 16C/W, you could expect it to run at around 29C above ambient. If ambient is 25C, that's 54C, which is pretty damn hot. At 65C you cannot keep you fingers on a heatsink for more than about three seconds, depending a bit on your tolerance for pain.
You can consider reducing the quiescent current, depending on what headphone impedance you will be using. For 24R cans, then 120mA is preferred, but for 75R cans or higher, you could reduce the output quiescent to around 50mA (R11 at 22R), which would greatly lower dissipation in the heatsinks.
These changes should be acceptable to the design without additional modifications.
Cheers,
Hugh
Each output, Q3 and Q4, dissipates (0.12A x 15V) = 1.8W of heat. If the heatsink is 16C/W, you could expect it to run at around 29C above ambient. If ambient is 25C, that's 54C, which is pretty damn hot. At 65C you cannot keep you fingers on a heatsink for more than about three seconds, depending a bit on your tolerance for pain.
You can consider reducing the quiescent current, depending on what headphone impedance you will be using. For 24R cans, then 120mA is preferred, but for 75R cans or higher, you could reduce the output quiescent to around 50mA (R11 at 22R), which would greatly lower dissipation in the heatsinks.
These changes should be acceptable to the design without additional modifications.
Cheers,
Hugh
Member
Joined 2009
Paid Member
KnoghtOfAwesome,
I think your design is looking very nice indeed.
If heat is a worry - I think you'll have more flexibility for assigning heatsinks if you don't mount them on the pcb but allow the power devices to hang off the pcb and be mounted to a larger heatsink, or even the amplifier casework.
I think your design is looking very nice indeed.
If heat is a worry - I think you'll have more flexibility for assigning heatsinks if you don't mount them on the pcb but allow the power devices to hang off the pcb and be mounted to a larger heatsink, or even the amplifier casework.
Last edited:
I complete my very crude breadboard of the ASPEN, I hadn't installed C1,C2,C3 or C4... None the less, but it was stable! I took the off/on offest measurements and found no massive offsets. Was there a reason as to why there was suspicion of such a large offset? My rather simplistic multimeter measured none.
Good news?
Good news?
Due to time constraints dealing with things other then breadboarding, R6 and C4 are the only parts not installed at the moment.
Although this massive initial offset is no where to be seen, I've seen maybe ~200mV from not turning the two PSU's on at exactly the same time. It's also possible that this is from adding the 100uF capacitor, mostly turn on offset starts at ~30mV and moves down to a rather "normal" offset value.
When I get a chance I'll add R6 and C4
Although this massive initial offset is no where to be seen, I've seen maybe ~200mV from not turning the two PSU's on at exactly the same time. It's also possible that this is from adding the 100uF capacitor, mostly turn on offset starts at ~30mV and moves down to a rather "normal" offset value.
When I get a chance I'll add R6 and C4