Yes, that was my idea, to use small signal transistors. They are
very fast and have high beta. (and are cheap)
How does your amp sound ?
Mike
very fast and have high beta. (and are cheap)
How does your amp sound ?
Mike
MikeB said:
Very smooth sound, and never tends to be agressive. Due the low feedback it's closer to the valve amplifiers than to the tipycal high feedback ss designs. I designed it very powerful, so it can drive 2ohm load without any quality degradation.
sajti
Sanyo's transistors are made by planar technology
Most transistos are made by expitaxi technology
Planar makes less noise than expitaxi
Most transistos are made by expitaxi technology
Planar makes less noise than expitaxi
1. The bias confuses me. In one design (a reputable amp), the VAS has 30mA bias. It uses triple darlington, the first darlington has 6mA bias, the second about 50mA bias.
Why he put 30mA bias in VAS, while the first darlington only 6mA?
2. There is a small transistor D669 and B649. What is better about these compared to BD139-BD140?
Why he put 30mA bias in VAS, while the first darlington only 6mA?
2. There is a small transistor D669 and B649. What is better about these compared to BD139-BD140?
(( I don't know the exact model but a Mark Levinson 4 series monoblock (rated at 300 8E, but actually does 500 8E, 9004E and 1.5kW 2E) uses 8 pairs of a Toshiba class TO-3PL devices.
It has 40mF per rail or 80,000uF in total with a 2372 VA toroid.
Your design seems similar to that in terms of power, I'd say as a rough estimate, you are on target. ))
Thanks K-amps ,
I do not know about the Mark Levinson but the above rating is very suitable for my purpose as my loudspeakers are 400w each , and with the above amp I can drive one two or four loudspeakers .Does the Mark Levinson have a 66-0-66v transformer too ?
EVA
These drivers were not specified in any circuit , I am planning to make a high power version of the leach amp , by increasing the supply voltage to +/- 93v and using higher voltage devices, thanks for warning me against fake devices,
OliverD
22pairs of 150w devices for a 250w amp.....22pairs means 44transistors X 150 w = 6600w how is this viable I wonder ??
For your 300w you are using 12pairs of 150w devices... and I was thinking 8pairs would be over kill for 800w and 1K5w.... for my amp , however I will be using an extralarge heatsink with a 4 inch fan , should I go for 8pairs or increase the pairs ? also which drivers should I go for.
It has 40mF per rail or 80,000uF in total with a 2372 VA toroid.
Your design seems similar to that in terms of power, I'd say as a rough estimate, you are on target. ))
Thanks K-amps ,
I do not know about the Mark Levinson but the above rating is very suitable for my purpose as my loudspeakers are 400w each , and with the above amp I can drive one two or four loudspeakers .Does the Mark Levinson have a 66-0-66v transformer too ?
EVA
These drivers were not specified in any circuit , I am planning to make a high power version of the leach amp , by increasing the supply voltage to +/- 93v and using higher voltage devices, thanks for warning me against fake devices,
OliverD
22pairs of 150w devices for a 250w amp.....22pairs means 44transistors X 150 w = 6600w how is this viable I wonder ??
For your 300w you are using 12pairs of 150w devices... and I was thinking 8pairs would be over kill for 800w and 1K5w.... for my amp , however I will be using an extralarge heatsink with a 4 inch fan , should I go for 8pairs or increase the pairs ? also which drivers should I go for.
For +/-93V I would use class G, which means less dissipation, and much more safety with same number of output devices...
sajti
sajti
Rajeev,
The Mark Levinson I was referring to is a model. 436.
It is rated 350 at 8e, 700 at 4e and 1400w at 2e.
It actually does 500at 8e, 890w at 4e and 1500w at 2e.
I do not know the trafo ratings , I am not even sure if is is a bridge balance design or regular.
But if I were to take a guess, I say the rails are around +/- 90vdc.
The Mark Levinson I was referring to is a model. 436.
It is rated 350 at 8e, 700 at 4e and 1400w at 2e.
It actually does 500at 8e, 890w at 4e and 1500w at 2e.
I do not know the trafo ratings , I am not even sure if is is a bridge balance design or regular.
But if I were to take a guess, I say the rails are around +/- 90vdc.
sam9 said:
The human ear is capable of resolving time arrival differences of less than 100,000th of a second. that is one sigle factor. then there is the dynamic linearity of transient peaking, and harmonic intergration, with respect to one another.
Add that up,and minimums of 200k with perfect linearity in all regards-- to that 200kHz, become a resonable bottom limit for a solid state device to produce good audio.
sajti said:
Hm... my belles 450, which is slightly modded, with 0.3F capacitance, runs it's rails at approximately 90 volts DC, at night due to the high AC level here in Kingston... it is typically 127AC at night. The spec wattage rating is 250 per channel, but.... it's more like, well.. you figure it out. Lots more, it is......
It uses 4 per rail of 2sj50,and 2sk135 Motorola MOSFETS. I suspect my wattage rating is a a bit higher than Dave Belles wants it to be.
If that is 1.5Kw into 4 ohms , the dissipation on the transistors can be taken as approximately 750 watts max.
With 8 pairs the dissipation will be 750/16 = 47 watts per device. If this is OK according to the derating curve for power dissipation , all should be well. You will however need to know how hot the heatsink will get at that power level. But this is for a steady state condition.
You could get away with a bit less if the maximum power level is only expected on transients and you are not considering 3KW at 2 ohms ! In that case ( reduced devices and less heat sink area) continuous sinewave testing might kill the unit!
Don't forget inductive loading and the SOA limits. Better to play safe by having more transistors and heatsinking to ensure reliability of the circuit.
High power circuitry has an impressive way of melting down !
Cheers,
Ashok.
sajti
Thanks , I opted for class AB because the Leach amp design is was
available on the net and the same is tried and tested for years . Further I
may be wrong BUT I feel that Amps with Higher voltage supply sound
better. Lastly I do not have any design of a good class G amp.
KBK
Your Amp must be delivering 500W at 8E per channel with +/-90v and
that high value of caps the supply voltage would not drop much on audio
load.You have also mentioned that the mains supply voltage at night is
127volts.
ASHOK
It is 800w at 4E and 1500w at 2E , but I will be using the amp mostly at
4E .
How did you calculate 47watts per device . You should have divided
1500 by 16 or 750 by 8 which equals to 93.75 w per device also the
amp is not 100% efficent . IF we overrate 20% it should be 112.5w per
device and now I feel this is too much for the device , but as I said I will
be using the amp at 4E , here the power will be max 1000w , hence the
number of devices should be just sufficent but not over kill as I thought
eirlear . Do you want me to increase the output devices to 12 pairs 0r
even more ??
Thanks , I opted for class AB because the Leach amp design is was
available on the net and the same is tried and tested for years . Further I
may be wrong BUT I feel that Amps with Higher voltage supply sound
better. Lastly I do not have any design of a good class G amp.
KBK
Your Amp must be delivering 500W at 8E per channel with +/-90v and
that high value of caps the supply voltage would not drop much on audio
load.You have also mentioned that the mains supply voltage at night is
127volts.
ASHOK
It is 800w at 4E and 1500w at 2E , but I will be using the amp mostly at
4E .
How did you calculate 47watts per device . You should have divided
1500 by 16 or 750 by 8 which equals to 93.75 w per device also the
amp is not 100% efficent . IF we overrate 20% it should be 112.5w per
device and now I feel this is too much for the device , but as I said I will
be using the amp at 4E , here the power will be max 1000w , hence the
number of devices should be just sufficent but not over kill as I thought
eirlear . Do you want me to increase the output devices to 12 pairs 0r
even more ??
rajeev luthra said:
Hi,
my experience is to use the transistors up to the 75% of their voltage limit. With high voltage, not the dissipation is the most important thing, but the second breakdown. So, for the 250V rated Motorola devices are useful to +/-93V. It looks OK, but please keep in Your mind that the main voltage can be 10% higher in the night hours, so it would be nice to reduce the voltage down with 10%. So 85V is acceptable by my point of view. This results still high output power, - say 350-400W/8ohm.
sajti
If You want to calculate the maximum dissipation of Your amplifier, here is the formula:
Pdmax=(U^2/Rload)*0.2027. In Your case Pmax=(93^2/4)*0.2027=438W With 8 pairs this means 27Ws for each output device. Please keep in Your mind, that this is true only for 4ohm load! If You use 2ohms the dissipation is double!!!!
sajti
Pdmax=(U^2/Rload)*0.2027. In Your case Pmax=(93^2/4)*0.2027=438W With 8 pairs this means 27Ws for each output device. Please keep in Your mind, that this is true only for 4ohm load! If You use 2ohms the dissipation is double!!!!
sajti
rajeev luthra said:
IMHO higher rails might mean more power but tey also translate as harder/ harsher mids and mid-trebles. I have gotten warmer sound by simply reducing rails... This ofcourse is not a rule as several variables affect sonics, however this is a correlation I found in some cases.
Thanks sajti ,
2SA1943/2SC5200 are 230v devices you recomend 75% as useful
hence I go for a transformer of 60-0-60v so that the DC is +/- 85volt , Ok I
will do that , now the best thing to do is to increase the no of filter caps ,
so that the voltage drop on load is minimum , how much mfd per rail
would you recommend if I want max power possible at 4E
Motorolla MJ15024/25 are also avilible but I think the
2SA1943/2SC5200 pair will be better
How did you calculate 27w per device at 4E , but 27w per device will be
at 8E , for 8pairs .
K-amps
The rails have been reduced from +/- 93v to +/- 85v I hope the sound will
be warmer . Should I continue with 8 pairs or go for 12 pairs , now the
supply and power will be less , what drivers do you recommend .
2SA1943/2SC5200 are 230v devices you recomend 75% as useful
hence I go for a transformer of 60-0-60v so that the DC is +/- 85volt , Ok I
will do that , now the best thing to do is to increase the no of filter caps ,
so that the voltage drop on load is minimum , how much mfd per rail
would you recommend if I want max power possible at 4E
Motorolla MJ15024/25 are also avilible but I think the
2SA1943/2SC5200 pair will be better
How did you calculate 27w per device at 4E , but 27w per device will be
at 8E , for 8pairs .
K-amps
The rails have been reduced from +/- 93v to +/- 85v I hope the sound will
be warmer . Should I continue with 8 pairs or go for 12 pairs , now the
supply and power will be less , what drivers do you recommend .
rajeev luthra said:
Hi,
the MJ15024/25 are better for dissipation (TO-3 package better for high power applications), but 2SC5200/2SA1493 better in sound quality....
For this power I would use 30000uF per rails at least, but more is better. Practically 6-8 pcs. of 4700uF/100V low esr caps looks nice for each rails....
The dissipation is: Pd= (85^2/4)*0,2027=366W for 16 output devices. This means about 23W for each. It's very safe, and 2ohms is also possible with very large heatsink, or with fan cooling.
sajti
- Status
- Not open for further replies.