thinking of a source follower.... running a IRFP240n at around 3A.
either:
A) at 32v with another IRFP240n as a CCS. 16v center... cap coupled to speaker load.
B) at 19v or so, choke (100mH, 1ohm) loaded. 3v at top of choke. cap coupled to speaker load.
what are the advantages of both? Disadvantages?
Can I get a show of hands as to which one is "better" ?
Planning on using a BG as the cap.... maybe 5000uf or so. probably get away with a lower voltage cap in 'B'.
I know a lot of talk has been said about this...
thanks in advance.
either:
A) at 32v with another IRFP240n as a CCS. 16v center... cap coupled to speaker load.
B) at 19v or so, choke (100mH, 1ohm) loaded. 3v at top of choke. cap coupled to speaker load.
what are the advantages of both? Disadvantages?
Can I get a show of hands as to which one is "better" ?
Planning on using a BG as the cap.... maybe 5000uf or so. probably get away with a lower voltage cap in 'B'.
I know a lot of talk has been said about this...
thanks in advance.
If you have the chokes, use them!
The efficiency goes up from ~20 to ~40%, which pays off when it comes to PSU and heatsinks.
Disadvantages?
Well, biasing requires some degree of care, but the DC resistance in the choke is high enough to serve well as source resistor.
Not a problem really.
One thing that might become a problem (of not taken care of)
is slow "thumps" from the prior stage during startup.
Since there is no constant current "mechanism" in the choke
you´ll need some kind of protection during these "thumps".
A timer circuit controlling a relay that clamps the "grid leak" (tube term...) for some tens of seconds solves everything.
The efficiency goes up from ~20 to ~40%, which pays off when it comes to PSU and heatsinks.
Disadvantages?
Well, biasing requires some degree of care, but the DC resistance in the choke is high enough to serve well as source resistor.
Not a problem really.
One thing that might become a problem (of not taken care of)
is slow "thumps" from the prior stage during startup.
Since there is no constant current "mechanism" in the choke
you´ll need some kind of protection during these "thumps".
A timer circuit controlling a relay that clamps the "grid leak" (tube term...) for some tens of seconds solves everything.
yeah!.
well I am going to IT couple to the gate of the MOSFET.
maybe run a ECC99 or 6SN7.
what size choke should I use? and what should the current rating be?
thinking of 50mH - 100mH 1ohm choke. 3A bias.
would this be adequate for 8 and 4 ohms?
Z= 2(pi)fL
so the -3 for 4 ohms is around 15Hz for 50mH. should go to 100mH if I can.
what should my power supply voltage and my source voltage be to have symmetrical clipping from a choke load?????
or is this relavent as the speaker load will dominate over the choke load? (AC wise).
well I am going to IT couple to the gate of the MOSFET.
maybe run a ECC99 or 6SN7.
what size choke should I use? and what should the current rating be?
thinking of 50mH - 100mH 1ohm choke. 3A bias.
would this be adequate for 8 and 4 ohms?
Z= 2(pi)fL
so the -3 for 4 ohms is around 15Hz for 50mH. should go to 100mH if I can.
what should my power supply voltage and my source voltage be to have symmetrical clipping from a choke load?????
or is this relavent as the speaker load will dominate over the choke load? (AC wise).
Fuling said:
Hi,
Well a choke is best regarding efficiency indeed. But it raises from 10% to 20% for SE topology, not to 40%! 100 mH is a good choice. I would take the other end of the choke to ground then but then you have DC across the choke. To get rid of that you need a coupling cap. A better option IMHO would be to use a 1:1 transformer in stead of the choke. The primary and secondary can be wound bifilar then which allows for a very tight coupling between the windings. This way you also avoid switch on thumps if you let the bias current come up slowly but anyway you need some kind of servo to control the bias current.
Cheers
Hi,
For such a transformer you need basically the same core and amount of turns as for the choke. You will need a gapped core to handle the large DC current and a primary inductance of the same 100 mH. I wouldn’t go higher than 300 - 500 mT for the standing induction to keep distortion low. If you are not be able to calculate such a transformer yourself a good winding house for sure can.
Cheers
For such a transformer you need basically the same core and amount of turns as for the choke. You will need a gapped core to handle the large DC current and a primary inductance of the same 100 mH. I wouldn’t go higher than 300 - 500 mT for the standing induction to keep distortion low. If you are not be able to calculate such a transformer yourself a good winding house for sure can.
Cheers
I don´t agree. Maximum teorethical efficiency for an CCS loaded SE amp is 25% versus 50% for a chokle loaded design.
When I had my "choked" follower running I got 30W out @ 80W power consumption.
AudioGeek: 50mH is enough for 4 ohms load and 100mH for 8 ohms. 3A bias is enough for 36W into 8R and 18W into 4R, teorethically.
A 1:1 output transformer is a very interresting idea, but there is another way to get rid of the output cap: Balanced operation.
Two low voltage/high current followers driven by a a balanced signal, speaker connected between the outputs.
That´s the concept I´m working on right now.
A cap coupled output using modern electros with a little judicious bypassing with a Black Gate is incredibly transparent, and guarantees you'll never monster a driver with DC. Believe me, this cuts the worry factor to zilch...
Offset control will require some sort of servo; this is not so bad, but it introduces complexity into a hugely elegant, simple circuit.
A +19V/-3V supply cannot be properly supported from a center tapped winding because, by definition, it's not a center tap! Such an arrangement sets up unequal flux across the winding which may cause transformer noise problems with the larger currents proposed here. A better idea is to use two separate windings, say 15Vac and 3Vac, each with a separate rectifier, then place the DC in series. This will give you a sufficient voltage margin to incorporate chokes into the supplies to assist with filtering and greatly reduce the size of the filter caps and lessen the current pulse load on the transformer; 22,000uF should be quite enough.
The DCR of the 100mH audio choke could be set to 1R, so that three amps will drop precisely three volts. If the gate is then biased for zero offset, exactly three amps will be flowing into the choke.
These SE amplifiers - significantly NOT balanced - sound very, very good. I have one, though it's efficiency is only marginally higher than a steam engine since it's loaded with a CCS. I have heard Circlotron's choke loaded SE Mosfet and it is sensational.
Cheers,
Hugh
Offset control will require some sort of servo; this is not so bad, but it introduces complexity into a hugely elegant, simple circuit.
A +19V/-3V supply cannot be properly supported from a center tapped winding because, by definition, it's not a center tap! Such an arrangement sets up unequal flux across the winding which may cause transformer noise problems with the larger currents proposed here. A better idea is to use two separate windings, say 15Vac and 3Vac, each with a separate rectifier, then place the DC in series. This will give you a sufficient voltage margin to incorporate chokes into the supplies to assist with filtering and greatly reduce the size of the filter caps and lessen the current pulse load on the transformer; 22,000uF should be quite enough.
The DCR of the 100mH audio choke could be set to 1R, so that three amps will drop precisely three volts. If the gate is then biased for zero offset, exactly three amps will be flowing into the choke.
These SE amplifiers - significantly NOT balanced - sound very, very good. I have one, though it's efficiency is only marginally higher than a steam engine since it's loaded with a CCS. I have heard Circlotron's choke loaded SE Mosfet and it is sensational.
Cheers,
Hugh
Hello Fuling,
Correct, theoretically that is the maximum for an “ideal” amp. But regarding Vg_th and you can’t drive the Fet to have zero voltage across it, losses in the choke/transformer, and the low supply voltage you may be glad to have 20 to 25% in the end at reasonable low distortion.
But you are more an expert with Zen amps I understand, so what did you actually measure in the end at say 0.5% distortion?
Regards
30/80= 37,5% efficiency. That´s what I measured in my amp.
Not a Zen expert at all, I´ve never even buit one. I just meant that alot of people seems to use them and like them even though they are cap-coupled.
AudioGeek: No, I don´t think you´ll need bipolar caps. Just ordinary lytics (ok, the best ones you can find), bypassed with something good
Let´s see, you need a pair of 100mH 3A chokes, huh?
Are EI transformer kits available in your area?
If so, get a pair of >250VA kits, remove the primary windings and fill the bobbin with 1,4mm wire.
The core must be assembled with an airgap, talk to Circlotron about how to adjust that.
OR: Have a look at hammondmfg.com.
Following pieces might be of interrest:
195T5
193V
159ZE (three or four in series)
Re: yeah!.
Make sure you use a 220 ohm or so gate resistor on the mosfet - on mine I have the gate lead and the resistor lead each cut to about 4mm before being overlapped and soldered.
Once you have it going you will wonder how you put up with anything else before. You will =not= be disappointed.
3 amps will let you swing to -24v into 8 ohms The supply rail will need to be +24v plus a couple of volts extra so the mosfet when swinging positive doesn't have to turn on too hard, going into it's nonlinear region, but don't go nuts. I use 27.7 volts @ 3.5 amps, 80mH.AudioGeek said:
Make sure you use a 220 ohm or so gate resistor on the mosfet - on mine I have the gate lead and the resistor lead each cut to about 4mm before being overlapped and soldered.
Once you have it going you will wonder how you put up with anything else before. You will =not= be disappointed.
chokes found.
Digikey has those BIG *** 195T5 chokes (100mH 0.67dcr 5A) for 57.00 each.
would these work???
So... 3x0.67 is 2v across the choke.
hmmm since these are not enclosed, would they emit tons of interference???? or should I really be looking for enclosed chokes?
Thanks!!!!
p.s. these weigh 14 lbs each.
Digikey has those BIG *** 195T5 chokes (100mH 0.67dcr 5A) for 57.00 each.
would these work???
So... 3x0.67 is 2v across the choke.
hmmm since these are not enclosed, would they emit tons of interference???? or should I really be looking for enclosed chokes?
Thanks!!!!
p.s. these weigh 14 lbs each.
195T5 should work fine.
Don´t worry about them not being enclosed, most audio chokes and transformers aren´t anyway.
Since they can handle 5A you have the opportunity to crank up the bias as much as the heatsinks and PSU can handle, not being locked to 3 A. No such thing as too much bias, soundwise.
Heat is an issue, though!
No such thing as too big heatsinks either!
2V across the choke is nothing. I dropped 4,8V @ 3A when my amp was up and running. Just calculate for it when you design the PSU.
14 lbs, that´s nice
Don´t worry about them not being enclosed, most audio chokes and transformers aren´t anyway.
Since they can handle 5A you have the opportunity to crank up the bias as much as the heatsinks and PSU can handle, not being locked to 3 A. No such thing as too much bias, soundwise.
Heat is an issue, though!
No such thing as too big heatsinks either!
2V across the choke is nothing. I dropped 4,8V @ 3A when my amp was up and running. Just calculate for it when you design the PSU.
14 lbs, that´s nice
bias at choke current handling
is there any issue with choke saturation running the bias so close to the choke rated current?
If my bias is say 4A... and the choke is rated 5A... do I have only a +1A to -1A available output swing??
lemme see.... 3A x 28v = 84W.... toasty.
I have a heatsing that was running 2x 12v x 3.25A = 78W so they should work fine. (Yes they were HOT to the touch, but mosfets worked fine and lasted).
What coupling cap should I use??? 4700uf good enough?
next got to work out a bias scheme.... probably run from the ground of the IT coupling to the gate.
thinking around 6v bias (2v across choke + 4v for 3.25A at 28v).
make it adj. of course.
is there any issue with choke saturation running the bias so close to the choke rated current?
If my bias is say 4A... and the choke is rated 5A... do I have only a +1A to -1A available output swing??
lemme see.... 3A x 28v = 84W.... toasty.
I have a heatsing that was running 2x 12v x 3.25A = 78W so they should work fine. (Yes they were HOT to the touch, but mosfets worked fine and lasted).
What coupling cap should I use??? 4700uf good enough?
next got to work out a bias scheme.... probably run from the ground of the IT coupling to the gate.
thinking around 6v bias (2v across choke + 4v for 3.25A at 28v).
make it adj. of course.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.