The NPN on the neg half of the Aleph variants is a current limiter on the gain side, and it can be omitted if you like riding a motorcycle 300MPH in rain and snow.
In theory, if the source resistor hits a preset (by the voltage divider on the base) voltage drop (current draw throught fet), it shuts down the voltage to the gate, reducing the current throught the source resistor
In theory, if the source resistor hits a preset (by the voltage divider on the base) voltage drop (current draw throught fet), it shuts down the voltage to the gate, reducing the current throught the source resistor
hi
ok the lower side mpsa18 is normally off line
check me if my understanding is good ??
1-for a zen v4 with just a +v supply the aleph upper
side current source limit the max. current in both
upper and lower side transistors
2-for a aleph:
the upper side:
idle current=.7/.47=1.5 amp plus maybe 30% aleph
modulation for a max.=2 amp
the lower side:
there is no current limit from the -15v and the lower
mpsa18 is for that but ??:
for .7v on mpsa18 base you need 2.25v on .47ohm
because (2.25*100)/(221+100)=.7v
I limit =2.25/.47=4.78 amp it s very high
for a 8ohm speaker maybe 6ohm there is a natural
limit = -15/6=2.5 amp it s very below 4.78amp
thanks
ok the lower side mpsa18 is normally off line
check me if my understanding is good ??
1-for a zen v4 with just a +v supply the aleph upper
side current source limit the max. current in both
upper and lower side transistors
2-for a aleph:
the upper side:
idle current=.7/.47=1.5 amp plus maybe 30% aleph
modulation for a max.=2 amp
the lower side:
there is no current limit from the -15v and the lower
mpsa18 is for that but ??:
for .7v on mpsa18 base you need 2.25v on .47ohm
because (2.25*100)/(221+100)=.7v
I limit =2.25/.47=4.78 amp it s very high
for a 8ohm speaker maybe 6ohm there is a natural
limit = -15/6=2.5 amp it s very below 4.78amp
thanks
)
60 V secondarys, I think it's 1500 va, CRC 60,000 uf, Huge 2.5ohm 100w 55,000 uf per rail. I was thinking of regulating, but after buying tranny, decided it would be better to just burn the extra volts in a ceramic resistor then waste heatsink capacity with a reg
I've pushed them really hard, but my mind clipps before they do.
I think 38 comes out almost dead on the 45V mark, but don't sweat it. The circuit works fine anywhere from 20V to 75V.
I've pushed them really hard, but my mind clipps before they do.
I think 38 comes out almost dead on the 45V mark, but don't sweat it. The circuit works fine anywhere from 20V to 75V.
hi Brian
your big amp seem very interesting have you
a web site for details
my actual setup
www.loginnovation.com/jeapel
more questions ?
1-for the 3 220uF you use top cap like blackgate
or just junk cap ?
2-on the mini A schematic the output mosfet are irf644 is it
possible to use irfp140 or 240 more available without
change the sound ?
your big amp seem very interesting have you
a web site for details
my actual setup
www.loginnovation.com/jeapel
more questions ?
1-for the 3 220uF you use top cap like blackgate
or just junk cap ?
2-on the mini A schematic the output mosfet are irf644 is it
possible to use irfp140 or 240 more available without
change the sound ?
I did very little documentation. I did a few posts as things progressed, but many thing changed over the last few years
http://www.diyaudio.com/forums/showthread.php?s=&threadid=17365&highlight=
I used Nichicons, not boutique, but not quite junk. I plan to bypass the one in the neg feedback, but the other 2 I don't really consider critcal.
IRFP 140,240, 244 have all been used successfully. Using 4 sets of 8 matched, and a pain in the **** to install, I didn't listen to any others than the 244's I used.
A friend and I listened for a few hours, tweaking bias, current gain, and rail voltage (with a variac), and the biggest improvement was upping the rails. It seemed to open the sound and deepen the soundstage. We couldn't really tell a differece in the sound of the bias or gain, but most people seem to prefer the sound of higher bias.
To summarize, if I did it all over, I would run the rails higher, set the bias as high as the sinks can handle without running too hot, and adjust the gain so it clipps symetric. I would advise against running hotter than 150 deg. The stress of thermal cycling makes the silicon die delaminate from the copper case, then it melts down. We weren't watching, and ran mine up to 175 deg. About 40 times on and off after, the source resistors lit up like lightbulbs.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=17365&highlight=
I used Nichicons, not boutique, but not quite junk. I plan to bypass the one in the neg feedback, but the other 2 I don't really consider critcal.
IRFP 140,240, 244 have all been used successfully. Using 4 sets of 8 matched, and a pain in the **** to install, I didn't listen to any others than the 244's I used.
A friend and I listened for a few hours, tweaking bias, current gain, and rail voltage (with a variac), and the biggest improvement was upping the rails. It seemed to open the sound and deepen the soundstage. We couldn't really tell a differece in the sound of the bias or gain, but most people seem to prefer the sound of higher bias.
To summarize, if I did it all over, I would run the rails higher, set the bias as high as the sinks can handle without running too hot, and adjust the gain so it clipps symetric. I would advise against running hotter than 150 deg. The stress of thermal cycling makes the silicon die delaminate from the copper case, then it melts down. We weren't watching, and ran mine up to 175 deg. About 40 times on and off after, the source resistors lit up like lightbulbs.
I have done a bit along these lines. Since I began with tubes, higher voltages are normal to me. The idea that a high rail draws out and flattens the center of the S-shaped gain curve is no mystery to tube folks.
Solid state people tend to think in terms of current. (You'll note that there is an inherent prejudice here--tubes do voltage easily, so tube people tend to think voltage is everything. Solid state people have current to spare so they preach the gospel of high current and ignore voltage.) Current, particularly if you leave it on all the time (class A) is a nifty thing to have.
So who's right? Both are. When putting a circuit together--or voicing it afterwards--you've got three parameters to play with: Voltage, current, and the amount of heat the device can safely dissipate before self-destructing. It's a good idea to leave some elbow room on all three, but once you establish a region where your circuit will operate reliably, you're free to adjust voltage against current and see how you like the results.
I put an Aleph-X on the bench and ran a relatively small signal through it--something like 1W. Then I ran the rail voltage from something pretty lean, perhaps 12V, on up to 20 or 25V. The distortion dropped dramatically. It's not hard to see why. If you look at the spec sheet for a MOSFET, you'll see that there are a number of things that change with voltage. As an example, the capacitance of the device drops, which makes it easier to drive, which in turn means that the front end has better control over what's going on. And, of course, the front end is behaving better itself, so it's a win-win proposition. The downside manifests itself in price (bigger transformer, caps, and heatsinks), and more heat (potential reliability problems and your air conditioner hates you). Just keep your wits about you and you'll do fine.
Grey
Solid state people tend to think in terms of current. (You'll note that there is an inherent prejudice here--tubes do voltage easily, so tube people tend to think voltage is everything. Solid state people have current to spare so they preach the gospel of high current and ignore voltage.) Current, particularly if you leave it on all the time (class A) is a nifty thing to have.
So who's right? Both are. When putting a circuit together--or voicing it afterwards--you've got three parameters to play with: Voltage, current, and the amount of heat the device can safely dissipate before self-destructing. It's a good idea to leave some elbow room on all three, but once you establish a region where your circuit will operate reliably, you're free to adjust voltage against current and see how you like the results.
I put an Aleph-X on the bench and ran a relatively small signal through it--something like 1W. Then I ran the rail voltage from something pretty lean, perhaps 12V, on up to 20 or 25V. The distortion dropped dramatically. It's not hard to see why. If you look at the spec sheet for a MOSFET, you'll see that there are a number of things that change with voltage. As an example, the capacitance of the device drops, which makes it easier to drive, which in turn means that the front end has better control over what's going on. And, of course, the front end is behaving better itself, so it's a win-win proposition. The downside manifests itself in price (bigger transformer, caps, and heatsinks), and more heat (potential reliability problems and your air conditioner hates you). Just keep your wits about you and you'll do fine.
Grey
It's funny I never thought about it, but voltage is the one parameter everyone considers static. Everyone is obsessed with hitting the exact rail voltage. I just didn't want to be under and loose potential output wattage.
When I first built my Alephs, I used some super meaty 60V ct 30A trannys (or something outrageous like that) that I found in the local surplus shop and fed them with a 22A variac. You know me, if there's a knob, I'll play with it. I also used pots for R21 and R19, and scaled up the output current sense resitors because my speakers are 4 ohm.
But I think Grey's right. Up the voltage and you decrease distortion and add headroom. (or you end up building a 250W amp instead of 100w and only use the first 100W)
This thinking makes me lead toward building my AlephX with bigger rails and slightly underbiased for its voltage, or maybe rely more on the Aleph ac current gain.
Oh, and thanks again Nelson.
When I first built my Alephs, I used some super meaty 60V ct 30A trannys (or something outrageous like that) that I found in the local surplus shop and fed them with a 22A variac. You know me, if there's a knob, I'll play with it. I also used pots for R21 and R19, and scaled up the output current sense resitors because my speakers are 4 ohm.
But I think Grey's right. Up the voltage and you decrease distortion and add headroom. (or you end up building a 250W amp instead of 100w and only use the first 100W)
This thinking makes me lead toward building my AlephX with bigger rails and slightly underbiased for its voltage, or maybe rely more on the Aleph ac current gain.
Oh, and thanks again Nelson.
I'm not saying that high bias isn't effective. It is. The trick is not to allow yourself to be blinded to the 'other' parameter by group-think lemmings following a narrow-minded vision of what's important. As much of both as you can stand is the ticket. Just don't burn up your output devices.
Big heatsinks are a really good idea.
Or water cooling.
Grey
Big heatsinks are a really good idea.
Or water cooling.
Grey
Water cooling and larger heatsinks is what I need for my car project.
I have been wondering this- what does bias effect the most sonically? Bass, highs, width? I'm toying with the idea of upping the bias on my Mini-A after I get more heatsink under it...but it would probably only go from 1.1A to 1.7A, and a ton more heat.
I would like to do a current source amp to play with on my horns in the car at some point.
I have been wondering this- what does bias effect the most sonically? Bass, highs, width? I'm toying with the idea of upping the bias on my Mini-A after I get more heatsink under it...but it would probably only go from 1.1A to 1.7A, and a ton more heat.
I would like to do a current source amp to play with on my horns in the car at some point.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.