Hi ChocoHolic ! (BTW. are you really addicted to chocolate ?)
The blockingcaps you are talking about, they really make sense,
but what exactly happens if they are left out ?
in the ~100uF Range they are discharged very fast, so they can
only have effect in the upper fruquencyrange. Do they help
stabilizing ?
Michael
The blockingcaps you are talking about, they really make sense,
but what exactly happens if they are left out ?
in the ~100uF Range they are discharged very fast, so they can
only have effect in the upper fruquencyrange. Do they help
stabilizing ?
Michael
Hi Mike!
*OFFTOPIC*
Yes, I love chocolate.
Especially swiss chocolate...!!!!
But it's pretty powerfull stuff: 600kcal/100g ....
So this allows a max. average consumption of 300g per day,
otherwise.....
*ONTOPIC*
The blocking caps help to stabilze and prevent oscillation, without slowing down the amp.
If you do not place them, then fast current changes of the output transistors can cause HF-signals on the rails (you can never avoid some inductance in the layout and wiring of the rails). In worst case this HF-pollution can cause self oscilation.
Sometimes it only causes poor signal response if a load is connected and a rectangular signal is applied. But I never had a amp, where I could simply skip these caps without negative influence on the result.
But there is one observation, which I did not fully understand.
It always turned out that the blocking from the positive rail to negative rail is much more important than the center tap connection to ground. So the length of the ground connection to the center tap of both caps never caused issues, if the loop from positive rail through both caps towards the negative rail was short.
I sometimes already got perfect results, just with one cap from positive rail to negative rail. This indicates an effect, which seems to be dominated by the interaction of the currents in
the NPN and PNP output transistors.... still a lot things, which I have not investigated in detail....
Bye
Markus
*OFFTOPIC*
Yes, I love chocolate.
Especially swiss chocolate...!!!!
But it's pretty powerfull stuff: 600kcal/100g ....
So this allows a max. average consumption of 300g per day,
otherwise.....
*ONTOPIC*
The blocking caps help to stabilze and prevent oscillation, without slowing down the amp.
If you do not place them, then fast current changes of the output transistors can cause HF-signals on the rails (you can never avoid some inductance in the layout and wiring of the rails). In worst case this HF-pollution can cause self oscilation.
Sometimes it only causes poor signal response if a load is connected and a rectangular signal is applied. But I never had a amp, where I could simply skip these caps without negative influence on the result.
But there is one observation, which I did not fully understand.
It always turned out that the blocking from the positive rail to negative rail is much more important than the center tap connection to ground. So the length of the ground connection to the center tap of both caps never caused issues, if the loop from positive rail through both caps towards the negative rail was short.
I sometimes already got perfect results, just with one cap from positive rail to negative rail. This indicates an effect, which seems to be dominated by the interaction of the currents in
the NPN and PNP output transistors.... still a lot things, which I have not investigated in detail....
Bye
Markus
Take a look in "1500 Watts in 2 ohms"
This amplifier can hold your supply, 90 volts i suppose.
But the one constructed by hurricane, a friend, was no good, maybe because the supply, i think so, voltage was going down in a big amount.
Read entire thread, because some transistors replacements will be more safe, the output i know you have, as drivers also, this way only input, and those are cheap.
It could not reach 600 watts, because supply goes to 64 volts or something alike, i measure it working and driving hard punches... this way, no good supply condition, if your supply have adequate power, it will go over 1 Kw without problems.... i suppose can reach 1.5 Kw, because simulations.
This power, real world was impossible, no supply, no charge, and not enough speakers to be melted.
Good Luck......poor family.....my God!
Are you sure your house is enough strong?
regards,
Carlos
This amplifier can hold your supply, 90 volts i suppose.
But the one constructed by hurricane, a friend, was no good, maybe because the supply, i think so, voltage was going down in a big amount.
Read entire thread, because some transistors replacements will be more safe, the output i know you have, as drivers also, this way only input, and those are cheap.
It could not reach 600 watts, because supply goes to 64 volts or something alike, i measure it working and driving hard punches... this way, no good supply condition, if your supply have adequate power, it will go over 1 Kw without problems.... i suppose can reach 1.5 Kw, because simulations.
This power, real world was impossible, no supply, no charge, and not enough speakers to be melted.
Good Luck......poor family.....my God!
Are you sure your house is enough strong?
regards,
Carlos
wonder_b said:
wonder,
What ever happened to the idea of using an emitter follower as a current amp?? It seems to me, if your outputs aren't darlington packages(would be unstable temp. coefficient), if you want stability use an emmitter follower type connection....(like the wonder.jpg picture) It is harder to drive a lower Z with a collector output.(more distortion)
Try using larger driver transistors. maybe TO-220 size. maybe even put them on a small heat sink. if they don't change temp., then the temp. coefficient of the drivers doesn't change. This is important if you use the drivers to bias the output devices. You may want to use a servo transistor to bias outputs, or thermal runnaway may be of concern... unless your heat sink is huge & temp doesn't get to HOT!!!
rule of thumb...don't parralell BJT's unless necessery
Choco is right in order to have current sharing RE is required, but this will also reduce the gain and efficiency. (consult the AC model or Thevinen circuit for a BJT). If 'little' re was the same in every transistor, it would be a perfect world.
Choco is right in order to have current sharing RE is required, but this will also reduce the gain and efficiency. (consult the AC model or Thevinen circuit for a BJT). If 'little' re was the same in every transistor, it would be a perfect world.
I will make a new schematic with the caps and post it after work I think they mite be worth looking at also the output transistors will be on there own pcb on the heatsink 600mm long x 2 (water cooled pic to come) I take it the caps would go on there and these are nonbipolar electrolytic's are they
cunningham could you post a pic to show me what you mean and what transistors would use for this mod
cunningham could you post a pic to show me what you mean and what transistors would use for this mod
For this high power (especially for the 12 pairs) I recommend to use single driver, but same type, as the output devices. The predriver should be MJE15032/33, this transistors like a bit higher bias, to sound great. In my amplifier I use them with 80mA, but with this high rails it's quite risky....
Thermal runaway another problem. For the predriver, and driver use separated heatsinks, and keep them as cool as possible. Use darlinton to set the bias (Vbe multiplier), and mount it to the top of one of the output devices, preferably the hottest one (which is far away from the air/water intake).
sajti
Thermal runaway another problem. For the predriver, and driver use separated heatsinks, and keep them as cool as possible. Use darlinton to set the bias (Vbe multiplier), and mount it to the top of one of the output devices, preferably the hottest one (which is far away from the air/water intake).
sajti
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.