orange, around 2R5@100Hz
that's not easy
previous graph was no lesser than 5R, that's easy
Cooling
Before I go re-inventing the wheel. I'm planning to mount 4 Noctua 140mm case fans to my 4U Deluxe chassis. They will be driven by a 5.25" by PC fan controller I have which I happen to have laying around. It has thermal sensors and setpoints with a wonderful display.
Anyone already designed a good mounting solution for the fans to the heat sinks? I'm thinking a 3d printed solution to screw into would be neat.
Right now I just have a couple 140mm random case fans that were pulled from a factory install that do well enough at 7v, but could be quieter. But anything worth doing is worth overdoing right?
Before I go re-inventing the wheel. I'm planning to mount 4 Noctua 140mm case fans to my 4U Deluxe chassis. They will be driven by a 5.25" by PC fan controller I have which I happen to have laying around. It has thermal sensors and setpoints with a wonderful display.
Anyone already designed a good mounting solution for the fans to the heat sinks? I'm thinking a 3d printed solution to screw into would be neat.
Right now I just have a couple 140mm random case fans that were pulled from a factory install that do well enough at 7v, but could be quieter. But anything worth doing is worth overdoing right?
I'm finding Babysitter approach much neater, and certainly functional
if you need more air, just mount more fans ( not just one as I did)
4U, even if measly 300deep , will work as charm....... no need to ruin its looks with fans on heatsinks
as Bikini on a Hen
(and - as good engineering practice - no need for fancy controllers; plain LM317 will do, if you want to set it accordingly to heat situation; for me it works simple - fan is On in summer months, while in rest of year is simply waiting for next Summer)
edit: for Babysitter purpose, cleaned and lubed 12cm fan from dead PC PSU is equally good as any fancy expensive fan; from that stance, 33 Greenies for fan is .......... well, better invested in semis
if you need more air, just mount more fans ( not just one as I did)
4U, even if measly 300deep , will work as charm....... no need to ruin its looks with fans on heatsinks
as Bikini on a Hen
(and - as good engineering practice - no need for fancy controllers; plain LM317 will do, if you want to set it accordingly to heat situation; for me it works simple - fan is On in summer months, while in rest of year is simply waiting for next Summer)
edit: for Babysitter purpose, cleaned and lubed 12cm fan from dead PC PSU is equally good as any fancy expensive fan; from that stance, 33 Greenies for fan is .......... well, better invested in semis
Last edited:
ok thanks. But like I said, I already have the schmanzy controller that was sitting in a box for years - I almost threw it out a few years ago.
Where are you mounting the fans? I'm not getting that. Inside moving air up? I'm not terribly worried about looks. No wife factor here and I listen mostly in the dark anyhow. Yes, I went with 4U 300mm, as I'm cheap cheap cheap.... with some things...
Where are you mounting the fans? I'm not getting that. Inside moving air up? I'm not terribly worried about looks. No wife factor here and I listen mostly in the dark anyhow. Yes, I went with 4U 300mm, as I'm cheap cheap cheap.... with some things...
Hello, I finished one channel and it's running BUT after listening for a couple of hours at a moderate volume, I start to hear a distorted sound.
At first I thought that my speakers got messed up but I am suspecting it's a temperature problem in the amp, as it will run fine after cooling down.
Anybody else with a similar issue?
At first I thought that my speakers got messed up but I am suspecting it's a temperature problem in the amp, as it will run fine after cooling down.
Anybody else with a similar issue?
These are the aprox temps:
heatsink measuring close to MOSFET 42 degrees
transformer 42 degrees
diode bridge 57 degrees
body Q5 52 degrees
body of Q6 47 degrees
body of Q7 48 degrees
body of Q8 48 degrees
Didn't happen this morning running at about 350mV (.74A) at the output MOSFETs
Don't know what's going on...I suspect temps
heatsink measuring close to MOSFET 42 degrees
transformer 42 degrees
diode bridge 57 degrees
body Q5 52 degrees
body of Q6 47 degrees
body of Q7 48 degrees
body of Q8 48 degrees
Didn't happen this morning running at about 350mV (.74A) at the output MOSFETs
Don't know what's going on...I suspect temps
not the temperatures, if that what you wrote is temp. equilibrium
both channels the same behavior ?
both channels the same behavior ?
well, you lost me there
I did ask you about temperatures, and it seems that you replied .......
maybe I was clumsy - to be more precise - if those temps you claimed in post #7269, then I think it is not temperature issue .......
do you have same issue in both channels?
care to post some pics?
I did ask you about temperatures, and it seems that you replied .......
maybe I was clumsy - to be more precise - if those temps you claimed in post #7269, then I think it is not temperature issue .......
do you have same issue in both channels?
care to post some pics?
Could be a hot spot developing somewhere.... Typical places are lug connectors and screw terminal blocks and/or bad PCB solder joints, where DC rail wires leave/enter PCB's. We are talking some 3-4 A of constant current here, and that can cause a lot of heat across poor connections with high contact resistance... worth checking.
Pass DIY Addict
Joined 2000
Paid Member
Yep, I had a problem that started with some crimp connectors/spade lugs in one of my 150w Aleph-X amps. I was experimenting with different configurations of diode bridges when one of the terminal connections lit up with a bright glow due to a poor connection. It drew a ton of current and blew a trace or two on the pcb.
I suppose going with the lowest possible fuse rating on the primary side would be the only option to hopefully prevent that scenario. It happened to me as well when I was playing with shield wiring / common returns. The bare multistrand wire I was using (silly me...) touched +24V and caused the corresponding set of 0.47 ohm / 5W resistors on the power supply PCB to burn out, but no other damage. I was lucky. Then I experimented with the fuse values until I got down to 1.63A (I think...) which was reliably sustaining the main power ON surges / cap bank charging.
I typically use the fuse value that only blows every few months at switch-on ... that is my criterium for "high enough, but as low as possible" 😛
A blown fuse once in a while also assures me that I'm not over-rating the fuse 😀
A blown fuse once in a while also assures me that I'm not over-rating the fuse 😀
Does anyone have a measurement of the peak current required per channel on the Aleph J. I want to go for SMPS and need to get an idea what size to buy?
Have you found a bipolar SMPS at about 5A? Or are you going to take two that have isolated earths and stack them?
Anyway, about 1.5A/channel is a good back of the napkin current draw. How much excess you want, and remembering that the draw of a class-A amplifier is fairly constant, is up to you.
Anyway, about 1.5A/channel is a good back of the napkin current draw. How much excess you want, and remembering that the draw of a class-A amplifier is fairly constant, is up to you.
This is going to be an interesting experiment. So far we have not seen many SMPS used for the 25W Class A amplifiers, mostly for the smaller amps such as the ACA and VFET pt1 and pt2. I need to say that for the ACA and VFET pt1 (P-type), building a linear power supply has resulted in significantly improved sound every time. The improvements have been across the board, with special note to bass extension and dynamics. My larger 25W amps all use linear supplies. My Aleph J has a dual-mono CRCRC configuration.
Having said that, there are some newer SMPS supplies that are designed for audio use. Importantly, they support large enough banks of filter capacitors on the output. By 'large enough' I mean around 30 mF (30,000 uF). Preliminary reports suggest that these newer supplies may be worth trying. I recommend looking for the models that are rated for high current Class A amplifiers.
Imagine if we short the output (or high output into low impedance). The current limiter (Q3, R13, R14) sensing the voltage across R19 limits the current through Q8 to about 2.25A so the upper limit should be about 4.5A per channel.
Last edited: