ok i have ran into this problem before
i bought an avel toroidal transformer that is 330VA and +-35 secondaries
now i hooked up one 4700uF in the positive and negative rails
i get a whopping +-55 volts from this psu
i was wondering how you calc/approx what rails are going to be after you put caps on them
this caused problems because my amp was doing 180watts into 8 ohms and my speakers can only handle 125 watts
any comments quesions? please guide me with some useful info
i bought an avel toroidal transformer that is 330VA and +-35 secondaries
now i hooked up one 4700uF in the positive and negative rails
i get a whopping +-55 volts from this psu
i was wondering how you calc/approx what rails are going to be after you put caps on them
this caused problems because my amp was doing 180watts into 8 ohms and my speakers can only handle 125 watts
any comments quesions? please guide me with some useful info
The usual formula is you multiply the AC secondaries by 1.4 to get an approximate DC value. (Actually you take the peak values of the AC so you can also divide it by 0.70707) and then subtract 1 volt or so (or whatever your bridge diode needs to bias it on, an approximate value of 1 volt turns on most diodes) so the subtracting of 1 takes into account the voltage drop across the diode.
So 35 / 0.70707 = 49.5 -1 = 48.5 vdc.
Now your Avel (if its a Y23 model) is rated for 35 volts at 115Vac.
The usual AC you get at home is anywhere from 117 to 124 vac in most US homes. So the Actual AC on your secondaries can be anywhere from 35.6 to 37.74vac.
meaning your DC rails will be anywhere from:
( 35.6 / 0.707) -1 = 49.36 to
( 37.74 / 0.707 ) -1 = 52.37.
If you are getting, 55 volts, I am guessing your Home AC voltages are close to : 129vac or so. Using schottkeys will alter the calculation a bit but not much.
Under load, your DC will drop to approx +/- 43volts or so. More power is not a bad thing, its how you use it.
Good luck,
K-
So 35 / 0.70707 = 49.5 -1 = 48.5 vdc.
Now your Avel (if its a Y23 model) is rated for 35 volts at 115Vac.
The usual AC you get at home is anywhere from 117 to 124 vac in most US homes. So the Actual AC on your secondaries can be anywhere from 35.6 to 37.74vac.
meaning your DC rails will be anywhere from:
( 35.6 / 0.707) -1 = 49.36 to
( 37.74 / 0.707 ) -1 = 52.37.
If you are getting, 55 volts, I am guessing your Home AC voltages are close to : 129vac or so. Using schottkeys will alter the calculation a bit but not much.
Under load, your DC will drop to approx +/- 43volts or so. More power is not a bad thing, its how you use it.
Good luck,
K-
i keep frying tip142 and 147
i only have one of each in the output stage class AB
i am wondering why they keep starting on fire
its not my speaker because it wokrs fine
do i need to have one more each of the tip142 and tip147 so there is not so much current
keep in mind i am using the above rails +-55 reality -+ 50
am i exceeding the rated amperage?
i am using 8 and 4 ohm loads
the tip142 and tip147 are starting on fire with either speaker after about 30 seconds of play
i only have one of each in the output stage class AB
i am wondering why they keep starting on fire
its not my speaker because it wokrs fine
do i need to have one more each of the tip142 and tip147 so there is not so much current
keep in mind i am using the above rails +-55 reality -+ 50
am i exceeding the rated amperage?
i am using 8 and 4 ohm loads
the tip142 and tip147 are starting on fire with either speaker after about 30 seconds of play
YOU ARE EXCEEDING THE tip142/47'S PRIMARY vce RATING TO BEGIN WITH. (Sorry for caps)
They are rated at 100vce, so any rails above +/- 50 vdc will be beyond spec.
Also using one pair of those to get 180 watts into 8 ohms, you are definitely exceeding the SOAR specs too.
6-8 pairs will be more in order...
They are rated at 100vce, so any rails above +/- 50 vdc will be beyond spec.
Also using one pair of those to get 180 watts into 8 ohms, you are definitely exceeding the SOAR specs too.
6-8 pairs will be more in order...
Try any cheap alternative like the 2sd1047/ 2sb817 pair. They can be used with +/- 70 volt rails. Other characteristics are similar to the TIP142/47. 2-3 pairs will work.
Any higher then you cannot go wrong with genuine 2sa1302/2sc3281 variants or Toshiba's own 2sc5200/ 2sa1943's. Those you can use with up to +/- 100 volts if sufficiently applied i.e multiple pairs.
For your application 2 pairs will be more than adequate, heck even one pair will work but its better to be conservative.
Any higher then you cannot go wrong with genuine 2sa1302/2sc3281 variants or Toshiba's own 2sc5200/ 2sa1943's. Those you can use with up to +/- 100 volts if sufficiently applied i.e multiple pairs.
For your application 2 pairs will be more than adequate, heck even one pair will work but its better to be conservative.
how does this work when reading the SOAR
so reading the tip142 i see that vce at like 1 volt can hadle about 10 amps
so that would be more the fine for 150 watts at 8 ohms
with 55 volt rails correct?
the only problem is that it cant handle more the 100vce which would happen becuase the total swing is 55*2 =110 right
so vce is to high? correct?
i am trying to figure this out becuase i keep frying tip142
i have great heatsinking so that is not the problem
i am wondering if i am reading my specs sheet wrong
heres my scenario
i play music ts fine i turn the amp up higher and its loud and playing fine then all of a sudden the tip142 and 147 fry
i am using class AB
so when i turn it up the signal is swinging bigger and exceeds my vce for the transistors right? then fries?
so if i use -+45 or 50 rails i should be fine
because max amps would be 5.625 or 6.25 into 8 ohms
and the tips can handle 10 amps
tell me if my thinking is correct and what is not correct
thanks
so reading the tip142 i see that vce at like 1 volt can hadle about 10 amps
so that would be more the fine for 150 watts at 8 ohms
with 55 volt rails correct?
the only problem is that it cant handle more the 100vce which would happen becuase the total swing is 55*2 =110 right
so vce is to high? correct?
i am trying to figure this out becuase i keep frying tip142
i have great heatsinking so that is not the problem
i am wondering if i am reading my specs sheet wrong
heres my scenario
i play music ts fine i turn the amp up higher and its loud and playing fine then all of a sudden the tip142 and 147 fry
i am using class AB
so when i turn it up the signal is swinging bigger and exceeds my vce for the transistors right? then fries?
so if i use -+45 or 50 rails i should be fine
because max amps would be 5.625 or 6.25 into 8 ohms
and the tips can handle 10 amps
tell me if my thinking is correct and what is not correct
thanks
roofingboom said:
I wouldn't expect them to run them at more than ~45V / 50W a piece and that's a bit high. The rule of thumb to get a device in the SOAR is dividing max disipation power and voltage (125W/100V in your case by at least 2). Try to add at least two more devices per polarity with 45V rails! and you'll be fine, or use the devices K-amps suggested.
roofingboom said:
You need to read the SOAR curves correctly. At 1 volt maybe it CAN do 10 amps but at 110 volts (which your rails are) the devices will only do 0.3 amps or less. (Look at the curve where the X axis says 100volts then read on the Y axis the amps it will deliver.
Based on this, you know that you easily surpass the secondary breakdown of the device with the rails you are running it.
As rule of thumb, choose a device with VCE twice of your rails. So if your rails are 110, choose a device with VCe over 220v. The most popular choice as I said before is a 1302/3281 variant.
The 1302/3281 pair can do 7 amps at 20 volts. Or 5 amps at 28 volts without reaching the secondary breakdown voltage..
1 pair of those will be fine for a nominal 8 ohm load, but I'd suggest going to 2 pairs.
Your TIPS will NOT work in this scenario (unless your have 10 pairs or so...) The devices may short out even before they get warm...
K-
PS: Even though you are only applying 55 volts to each NPN/PNP device, as the two switch back and forth, whilst one is conducting fully near rail potentials, (Lets assume a emitted follower where the collector of the PNP is -55v and the collector of the NPN is +55v) it will dump all of its voltage to the load and source the same voltage to the other device. So lets say the PNP device is fully conducting at a given portion of the sinewave, it will source -55v to the emitter of the NPN device (which already has +55v on it's collector) and thus the device will be at a potential of 110volts.
Thats theory, in reality, the emitter resistors drop some voltage as well but know that each device could be subjected to full rail voltages from both sides.
i understand that you could in fact have both rails on the collect and emmiter of the PNP and same for the NPN
but one point im not clear on is this
so lets says that the sine wave is at peak on the positive half
so the NPN is ON and the PNP is off
the PNP is getting close to haveing -55 on its collector and +55 on its emmiter but it is bisaed off so there is no flowing current
however the NPN is ON and since its really close to the peak Vce of the NPN is really small so it should be able to source around 10 amps
am i thinking correctly and reading the SOAR right
that would mean that i could get by with 1 pair of transistors if i didnt exceed the VCE requirment
correct or not correct?
basically i am saying that i dont need like 8 or 10 pairs
i think i am correct tell me what you think
but one point im not clear on is this
so lets says that the sine wave is at peak on the positive half
so the NPN is ON and the PNP is off
the PNP is getting close to haveing -55 on its collector and +55 on its emmiter but it is bisaed off so there is no flowing current
however the NPN is ON and since its really close to the peak Vce of the NPN is really small so it should be able to source around 10 amps
am i thinking correctly and reading the SOAR right
that would mean that i could get by with 1 pair of transistors if i didnt exceed the VCE requirment
correct or not correct?
basically i am saying that i dont need like 8 or 10 pairs
i think i am correct tell me what you think
Yes you should not need 8-10 pairs. 
With those Devices I would not cross +/- 40 volts the way they are blowing up.
I have used some MOT MJ15003/4 with Rails up to +/- 90vdc !!! With no ill effects. (even though it is not rated above =+/- 70vdc.)
So it also depends on how conservative the manufacturer is.
With those Devices I would not cross +/- 40 volts the way they are blowing up.I have used some MOT MJ15003/4 with Rails up to +/- 90vdc !!! With no ill effects. (even though it is not rated above =+/- 70vdc.)
So it also depends on how conservative the manufacturer is.
"I have used some MOT MJ15003/4 with Rails up to +/- 90vdc !!! With no ill effects. (even though it is not rated above =+/- 70vdc.)"
I can think of a few reasons why you might "get away" with this.
A- The true max on actual devices off the line falls into a bell curve distribution. The stated spec is at the far low end tail so that 99.9% of all production passes. This means that a buyer can count on the spec and in fact virtually all devices will exceed spec. However, you have no way of knowing by how much. A DIYer can take advantage of this but if you were turning out 1000 amplifiers per day this would be a prescription for financial disaster.
B- With 90vdc rails you may well rarely if ever approach them (that's one gnarly amp!) thus the devices never see the full swing possible.
C- SOAR charts usually specify a time domain - a V/I condition that is fatal if continuous my be quit innocent of if it exists only on peaks.
Going back to the earlier posts there is a situation I'm unclear on. We usually speak of the rail under load but virtually all transformers yield higher rails at idle than the simple spec states. At idle the Vce will be the difference between rail and gound - no problem. But I wonder if an initial voltage or pulse can occur quicker than the rails drop. Could this result in briefifly exceeding the Vce rating even though the normallyn loaded rails would not? Similarly is there a danger when driving a signal into an amp with no load or only ascope such that the rails don't drop enough? What I'm getting at is say I have +/-40V rails loaded but +/-50V unloaded, -- are devices rated 80Vce max still risk under some circumstances?
I can think of a few reasons why you might "get away" with this.
A- The true max on actual devices off the line falls into a bell curve distribution. The stated spec is at the far low end tail so that 99.9% of all production passes. This means that a buyer can count on the spec and in fact virtually all devices will exceed spec. However, you have no way of knowing by how much. A DIYer can take advantage of this but if you were turning out 1000 amplifiers per day this would be a prescription for financial disaster.
B- With 90vdc rails you may well rarely if ever approach them (that's one gnarly amp!) thus the devices never see the full swing possible.
C- SOAR charts usually specify a time domain - a V/I condition that is fatal if continuous my be quit innocent of if it exists only on peaks.
Going back to the earlier posts there is a situation I'm unclear on. We usually speak of the rail under load but virtually all transformers yield higher rails at idle than the simple spec states. At idle the Vce will be the difference between rail and gound - no problem. But I wonder if an initial voltage or pulse can occur quicker than the rails drop. Could this result in briefifly exceeding the Vce rating even though the normallyn loaded rails would not? Similarly is there a danger when driving a signal into an amp with no load or only ascope such that the rails don't drop enough? What I'm getting at is say I have +/-40V rails loaded but +/-50V unloaded, -- are devices rated 80Vce max still risk under some circumstances?
"I have used some MOT MJ15003/4 with Rails up to +/- 90vdc !!! With no ill effects. (even though it is not rated above =+/- 70vdc.)"
I can think of a few reasons why you might "get away" with this.
A- The true max on actual devices off the line falls into a bell curve distribution. The stated spec is at the far low end tail so that 99.9% of all production passes. This means that a buyer can count on the spec and in fact virtually all devices will exceed spec. However, you have no way of knowing by how much. A DIYer can take advantage of this but if you were turning out 1000 amplifiers per day this would be a prescription for financial disaster.
K-Amps: True but being a "Tim the toolman" kind of DIYer Iam, half the fun is in exploring new territories. I also believe, what i had (marked as MJ15003) were probably MJ15022/23/24/25's or similar production run. As long as the substrate meets the specs of both devices, why not save costs by large production runs and just re-labelling.
B- With 90vdc rails you may well rarely if ever approach them (that's one gnarly amp!) thus the devices never see the full swing possible.
K-amps: Approached them till clipping. Loaded with an 8 ohm Set of Subs. The OP stage had 16 devices/8 pairs of the MJ15003/4's. Fully loaded rails were +/- 78vdc for a 8 ohm load
C- SOAR charts usually specify a time domain - a V/I condition that is fatal if continuous my be quit innocent of if it exists only on peaks.
K-Amps: Correct yet depending on manufacturers, the tolerances will vary.
Going back to the earlier posts there is a situation I'm unclear on. We usually speak of the rail under load but virtually all transformers yield higher rails at idle than the simple spec states. At idle the Vce will be the difference between rail and gound - no problem. But I wonder if an initial voltage or pulse can occur quicker than the rails drop. Could this result in briefifly exceeding the Vce rating even though the normallyn loaded rails would not? Similarly is there a danger when driving a signal into an amp with no load or only ascope such that the rails don't drop enough? What I'm getting at is say I have +/-40V rails loaded but +/-50V unloaded, -- are devices rated 80Vce max still risk under some circumstances?
K-Amps: I am in the same boat as you. Unless our friend can describe the rails when the devices blow, we'd be guessing.
I can think of a few reasons why you might "get away" with this.
A- The true max on actual devices off the line falls into a bell curve distribution. The stated spec is at the far low end tail so that 99.9% of all production passes. This means that a buyer can count on the spec and in fact virtually all devices will exceed spec. However, you have no way of knowing by how much. A DIYer can take advantage of this but if you were turning out 1000 amplifiers per day this would be a prescription for financial disaster.
K-Amps: True but being a "Tim the toolman" kind of DIYer Iam, half the fun is in exploring new territories.
I also believe, what i had (marked as MJ15003) were probably MJ15022/23/24/25's or similar production run. As long as the substrate meets the specs of both devices, why not save costs by large production runs and just re-labelling. B- With 90vdc rails you may well rarely if ever approach them (that's one gnarly amp!) thus the devices never see the full swing possible.
K-amps: Approached them till clipping. Loaded with an 8 ohm Set of Subs. The OP stage had 16 devices/8 pairs of the MJ15003/4's. Fully loaded rails were +/- 78vdc for a 8 ohm load
C- SOAR charts usually specify a time domain - a V/I condition that is fatal if continuous my be quit innocent of if it exists only on peaks.
K-Amps: Correct yet depending on manufacturers, the tolerances will vary.
Going back to the earlier posts there is a situation I'm unclear on. We usually speak of the rail under load but virtually all transformers yield higher rails at idle than the simple spec states. At idle the Vce will be the difference between rail and gound - no problem. But I wonder if an initial voltage or pulse can occur quicker than the rails drop. Could this result in briefifly exceeding the Vce rating even though the normallyn loaded rails would not? Similarly is there a danger when driving a signal into an amp with no load or only ascope such that the rails don't drop enough? What I'm getting at is say I have +/-40V rails loaded but +/-50V unloaded, -- are devices rated 80Vce max still risk under some circumstances?
K-Amps: I am in the same boat as you. Unless our friend can describe the rails when the devices blow, we'd be guessing.
I'd say go with 2 pairs now that these are blowing at 10% overspec on VCE. But if you are very tight on space/ budget then one pair could work for an 8 ohm load, knowing that it will not be uber-reliable.
If you believe in "hunches" I'd guess sooner or later you will blow the one pair with 50 volt rails too. ;-)
If you believe in "hunches" I'd guess sooner or later you will blow the one pair with 50 volt rails too. ;-)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.